Behavioral Finance JAFFW2008

Volume 18, No. 2Fall/Winter 2008

Editors:Betty J. SimkinsRamesh P. Rao

Charles W. Smithson

Academic Assoc. Editors:Edward I. Altman

Kenneth A. BorokhovichChristine A. BrownJennifer S. Conrad

Javier EstradaMark J. Flannery

Gerald D. GayStuart I. Greenbaum

Allaudeen S. HameedAndrea J. Heuson

Takato HirakiBrian M. Lucey

Sotiris K. StaikourasLaura T. Starks

David A. WalkerRalph A. WalklingSamuel C. Weaver

Lawrence W. LiconMartin R. Young

Practitioner Assoc. Editors:Niso Abuaf

Donald ChewMike Edleson

John FraserGene Guill

Andrew J. KalotayIra G. KawallerJoseph V. Rizzi

D. Sykes Wilford

Academic Contributions

Behavioral Finance: Quo Vadis? Werner De Bondt, Gulnur Muradoglu, Hersh Shefrin, and Sotiris K. Staikouras

The Effects of Institutional Risk Control on Trader BehaviorRyan Garvey and Fei Wu

Why Do People Trade?Anne Dorn, Daniel Dorn, and Paul Sengmueller

The Long-Term Value of Trade InformativenessMichel Rakotomavo

Shareholder Theory-How Opponents and Proponents both Get It WrongMorris G. Danielson, Jean L. Heck, and David R. Shaffer

Student Managed Investment Funds: An International PerspectiveEdward C. Lawrence

Practitioner ContributionBehavioral Basis of the Financial Crisis

Joseph V. Rizzi

Roundtable

University of Rochester Roundtable on Bankruptcy and Bailouts: The Case of the US AutoIndustry

Panelists: Thomas Jackson, Charles Hughes, James Brickley, Joel Tabas, and Clifford SmithModerator: Mark Zupan

InterviewPioneers in Finance: Vernon Smith Interview

Terrance Odean and Betty J. Simkins

Case StudyThe 2008 Federal Intervention to Stabilize Fannie Mae and Freddie Mac

W. Scott Frame

Book ReviewsBook Review: Ending the Management Illusion: How to Drive Business Results Using thePrinciples of Behavioral Finance By Hersh Shefrin

Andrea Heuson

Book Review: The Venturesome Economy by Amar BhidéColby Wright

Financial PuzzlesStewart C. Myers

THE FINANCIAL MANAGEMENT ASSOCIATION INTERNATIONALOFFICERS—DIRECTORS—EDITORS

Journal of Applied Finance (ISSN 1534-6668) is published by the Financial Management Association International, an affiliateof the Finman Corporation. It is published semi-annually. The Editors and the Association assume no responsibility for theviews expressed by the authors.

Membership dues in the Association include a one-year subscription to the journal. Membership fees: New Professional $100,Renewal Professional $70, New Sustaining $125, and Renewal Sustaining $95. An application form is available inside this issue.JAF subscriptions for libraries are available. Contact Financial Management Association International, University of SouthFlorida, College of Business Administration, Suite 3331, Tampa, FL 33620-5500, Telephone: (813) 974-2084 for further information.

Memberships, Subscriptions and Address Changes: Write Financial Management Association International, University ofSouth Florida, College of Business Administration, Suite 3331, Tampa, FL 33620-5500.

Manuscripts: Electronically submit your submission form and a PDF file at www.fma.org. A submission fee is required forevaluation of each manuscript: $200 for non-FMA members, $130 for doctoral students who are not FMA members, and $100for FMA members (U.S. dollars). The non-member submission fees include an FMA membership for the submitting author. Styleinformation for manuscripts is located on the inside back cover of this journal.

Permission to Quote or Republish: Blanket permission is granted to any individual wishing to use articles appearing in Journalof Applied Finance for educational (university classroom) purposes. Written permission from the Financial ManagementAssociation International or the Editor is not required. To make any other requests for permission to quote or republish, write toFinancial Management Association International, University of South Florida, College of Business Administration, Suite 3331,Tampa FL 33620-5500. Telephone: (813) 974-2084; Fax: (813) 974-3318; Email: [email protected]; Website: http://www.fma.org

Copyright © 2008 Financial Management Association International, an affiliate of the Finman Corporation. Printed by DartmouthPrinting Company, Hanover, NH. Printed in the U.S.A.

PresidentDouglas R. Emery

University of Miami2008-2009

Secretary/TreasurerAjay Patel

Wake Forest University2002-2012

Chairman, Finman CorporationJennifer Conrad

UNC-Chapel Hill2008-2011

Vice President-ProgramG. Andrew Karolyi

The Ohio State University2009-Reno, Nevada

Vice President-Financial EducationRobert Parrino

University of Texas-Austin2008-2010

Vice President-Global ServicesAlexander J. Triantis

University of Maryland2007-2009

Vice President-Practitioner ServicesO. Rawley ThomasLifeCycle Returns

2008-2011

Editors, Survey and Synthesis SeriesJohn Martin

Baylor UniversityJames Schallheim

University of Utah2004-2010

Editor, Financial ManagementWilliam G. Christie

Vanderbilt University2006-2011

Editors, FMA OnlineExecutive EditorBetty J. Simkins

Oklahoma State University2005-2009

EditorsJohn Finnerty

Fordham UniversityMark Flannery

University of FloridaSheridan Titman

University of Texas at Austin2005-2009

Editors, Journal of Applied FinanceBetty J. Simkins & Ramesh P. Rao

Oklahoma State UniversityCharles W. Smithson

Rutter Associates2007-2010

Journal of Applied FinanceVolume 18 Number 2 Fall/Winter 2008

EDITORSBetty J. Simkins

Oklahoma State University

Charles SmithsonRutter Associates

ASSISTANT EDITORHeidi Carter

Oklahoma State University

Ramesh P. RaoOklahoma State University

ASSOCIATE EDITORSAcademicEdward I. AltmanNew York UniversityKenneth A. BorokhovichCleveland State UniversityChristine A. BrownUniversity of Melbourne, AustraliaJennifer S. ConradUniversity of North CarolinaJavier EstradaIESE Business School Barcelona,SpainMark J. FlanneryUniversity of FloridaGerald D. GayGeorgia State UniversityStuart I. GreenbaumWashington University, St. LouisAllaudeen S. HameedNational University of Singapore

Takato HirakiKwansei Gakuin University, JapanBrian M. LuceyTrinity College, DublinSotiris K. StaikourasCass Business School, LondonLaura T. StarksUniversity of Texas at AustinDavid A. WalkerGeorgetown UniversityRalph A. WalklingDrexel UniversitySamuel C. WeaverLehigh UniversityLawrence W. LiconArizona State UniversityMartin R. YoungMassey University, New Zealand

PractitionerNiso AbuafIndependent ConsultantDonald ChewMorgan StanleyMike EdlesonMorgan StanleyJohn FraserHydro OneGene GuillDeutsche BankAndrew J. KalotayAndrew Kalotay Associates, Inc.Ira G. KawallerKawaller & Co.Joseph RizziCapGen FinancialD. Sykes WilfordEAQ Partners; The Citadel

SPONSORSOklahoma State UniversityUniversity of South Florida

Andrea J. HeusonUniversity of Miami

JAF REFEREES

JAF would like to thank all the referees who have reviewed manuscripts since the last issue. We appreciate the efforts ofour reviewers for responding as soon as possible and for providing constructive comments for the authors.

Tom AaboAarhus School of Business

James J. AngelGeorgetown University

Thomas M. ArnoldUniversity of Richmond

Chenchu BathalaCleveland State University

TK BhattacharyaCameron University

Kenneth A. BorokhovichCleveland State University

Helen M. BowersUniversity of Delaware

Christine A. BrownUniversity of Melbourne

Kelly R. BrunarskiMiami University

Antonio CamaraOklahoma State University

David A. CarterOklahoma State University

Don ChanceLouisiana State University

Donald H. Chew, Jr.Morgan Stanley

Jennifer S. ConradUniversity of North Carolina

Arnald R. CowanIowa State University

Frank D’SouzaLoyola College Maryland

Michael EdlesonMorgan Stanley

Javier EstradaIESE Business School

Michael G. FerriGeorge Mason University

John R.S. FraserHydro One, Inc.

Gabriele GalatiDe Nederlandsche Bank

Jacqueline L. GarnerDrexel University

Gerald D. GayGeorgia State University

Stuart L. GillanTexas Tech University

Radha GopalanWashington University

Stuart I. GreenbaumWashington University

Yilmaz GuneyUniversity of Hull

Benton E. GupUniversity of Alabama

Allaudeen HameedNational University of Singapore

Joel T. HarperOklahoma State University

Scott E. HeinTexas Tech University

Andrea J. HeusonUniversity of Miami

Jonathan M. KarpoffUniversity of Washington

Eric KelleyUniversity of Arizona

Sivarama KrishnanUniversity of Central Oklahoma

David R. LangeAuburn University Montgomery

K.C. MaStetson University

Cathy NidenLECG, Inc.

Thomas J. O’BrienUniversity of Connecticut

Tim OplerTorreya Partners

Christos PantzalisUniversity of South Florida

Janet PayneTexas State University San Marcos

Ivilina PopovaTexas State University San Marcos

Jack S. RaderFinancial Management Association

International

Daniel A. RogersPortland State University

Kasper RoszbachRiks Bank

W. Gary SimpsonOklahoma State University

Charles SmithsonRutter Associates

Sotiris StaikorosCass Business School, London

Mathijs van DijkEramus University

David A. WalkerGeorgetown University

Larry WallFederal Reserve Bank of Atlanta

Samuel C. WeaverLehigh University

Melissa A. WilliamsUniversity of Houston — Clear Lake

Wendell L. LiconArizona State University

John D. MartinBaylor University

Journal of Applied FinanceVolume 18 Number 2 Fall/Winter 2008

Academic Contributions

7 Behavioral Finance: Quo Vadis? Werner De Bondt, Gulnur Muradoglu,Hersh Shefrin, and Sotiris K. Staikouras

22 The Effects of Institutional Risk Control on Trader Behavior Ryan Garvey and Fei Wu

37 Why Do People Trade? Anne Dorn, Daniel Dorn, and Paul Sengmueller

51 The Long-Term Value of Trade Informativeness Michel Rakotomavo

62 Shareholder Theory-How Opponents andProponents both Get It Wrong Morris G. Danielson, Jean L. Heck, and David R. Shaffer

67 Student Managed Investment Funds: An InternationalPerspective Edward C. Lawrence

Practitioner Contribution

84 Behavioral Basis of the Financial Crisis Joseph V. Rizzi

Roundtable

97 University of Rochester Roundtable on Bankruptcy and Bailouts: The Case of the US Auto Industry Panelists: Thomas Jackson, Charles Hughes,

James Brickley, Joel Tabas, and Clifford SmithModerator: Mark Zupan

Interview116 Pioneers in Finance: Vernon Smith Interview Terrance Odean and Betty J. Simkins

Case Study

124 The 2008 Federal Intervention to Stabilize Fannie Mae and Freddie Mac W. Scott Frame

Book Reviews

137 Book Review: Ending the Management Illusion: How to DriveBusiness Results Using the Principles of Behavioral Finance By Hersh Shefrin Andrea Heuson

139 Book Review: The Venturesome Economy by Amar Bhidé Colby Wright

Financial Puzzles142 Stewart C. Myers

4

4

Letter from the EditorsThis is our second issue as editors of the Journal of Applied Finance (JAF). We were delighted to note that the first issue

was received very well by our readers and we wish to thank all who sent us their feedback. We received positive feedback onthe new layout of JAF with sections on academic and practitioner contributions, roundtable discussions, case and clinicalstudies, interviews, surveys, and book reviews. We will strive to keep these same features consistent across issues. Ourcontinued success though depends on our ability to attract submissions in each of these categories, especially roundtables,surveys, case and clinical studies, and interviews. We would like to encourage our readers to be actively engaged in thesetypes of submissions. As your editors, we are more than happy to work with you to develop submissions in these categories.We would also like to appeal to our practitioner community to consider writing for JAF and are willing to do what we can toidentify academics that they can partner with.

In This Issue

As we mentioned in our first issue, one of our goals is to have one or two themes for each issue. In this issue our focus ison behavioral finance. Our lead article (Behavioral Finance: Quo Vadis?) provides the reader an overarching view of behavioralfinance from its inception to the current state and beyond. It is based on a panel discussion held at the FMA-Europe meetingsin Prague, 2008. The panelists included Werner De Bondt, Gulnur Muradoglu, Hersh Shefrin, and Sotiris Staikouras, whoalso authored the piece. Our academic contributions also include several articles on trading behavior: “The Long-TermValue of Trade Informativeness” by Michel Rakotomavo, “The Effects of Institutional Risk Control on Trader Behavior” byRyan Garvey and Fei Wu; and “Why Do People Trade?” by Anne Dorn, Daniel Dorn, and Paul Sengmueller. There is also aprovocative article on shareholder theory by Morris G. Danielson, Jean L. Heck, and David R. Shaffer (“Shareholder Theory-How Opponents and Proponents both Get It Wrong”). In addition, our academic and practitioner readers will also find thearticle on student management investment funds by Ed Lawrence to be of interest (Student Managed Investment Funds: AnInternational Perspective)

The behavioral finance theme is continued in our practitioner contribution by Joseph Rizzi titled “Behavioral Basis of theFinancial Crisis.” This article also continues a theme from our first issue on the subprime crisis.

The behavioral finance theme continues in our “Pioneers of Finance” interview feature. Professor Vernon Smith, recipientof the 2002 Nobel Prize in Economics. In this interview, Professor Smith shares his perception of how experimental economicsand behavioral economics are related. He goes on to provide some insights from his research on speculative bubbles inexperimental markets that help us understand the recent bubble is US residential real estate, The interview was conducted byTerry Odean and Betty Simkins. We thank Terry for his contribution.

Our readers will also enjoy our case/clinical study contribution by Scott Frame titled “The 2008 Federal Intervention toStabilize Fannie Mae and Freddie Mac”. As a Federal Reserve insider, Scott Frame provides insights into the Fannie andFreddie debacles that only an insider can provide. The case study also continues the subprime crisis theme from our firstissue.

Our roundtable features a time topic: Bankruptcy and Bailouts: The Case of the US Auto Industry. The panelists featureseveral prominent auto industry executives and faculty members from the University of Rochester. We thank the Universityof Rochester for sponsoring the roundtable, Mark Zupan for moderating, and Don Chew for editing it.

Our book review section features two books. One is titled Ending the Management Illusion: How to Drive BusinessResults Using the Principles of Behavioral Finance written by Hersh Shefrin, one of the pioneers in behavioral finance. Thereview was written by Andrea Heuson. The second is by Amar Bhidé, another prominent author who is an authority onstrategy. The book is titled The Venturesome Economy and is reviewed by Colby Wright.

This issue concludes with the solution to the Financial Puzzle that appeared in the Fall/Winter 2007 issue, along with twonew puzzles by Stu Myers.

5

In Closing

In closing, we would like to highlight themes/topics we are considering for future issues. These include valuation, corporaterestructuring, and dividends and share buybacks. We welcome your contributions and any suggestions you may have forJAF.

Sincerely,

Ramesh P. Rao Betty J. Simkins Charles SmithsonPaul C. Wise Chair Williams Cos. Professor of Business Founding PartnerOklahoma State University Oklahoma State University Rutter Associates LLCEmail: [email protected] [email protected] [email protected]

6

Behavioral Finance: Quo Vadis?

Werner De Bondt, Gulnur Muradoglu, Hersh Shefrin, and Sotiris K. Staikouras

7

Behavioral finance endeavors to bridge the gap betweenfinance and psychology. Now an established field,behavioral finance studies investor decision processeswhich in turn shed light on anomalies, i.e., departuresfrom neoclassical finance theory. This paper is thesummary of a panel discussion. It begins by reviewingthe foundations of finance and it ends with a discussionof the future of behavioral finance and a self-critique.We describe the move from the standard view thatfinancial decision making is rational to a behavioralapproach based on judgmental heuristics, biases, mentalframes, and new theories of choice under risk. A newclass of asset pricing models, which adds behavioralelements to the standard framework, is proposed.

Werner De Bondt is a Professor of Finance at DePaul University in Chi-cago, IL. Gulnur Muradoglu is a Professor of Finance at Cass BusinessSchool in London, UK. Hersh Shefrin is a Professor of Finance at SantaClara University in Santa Clara, CA. Sotiris K. Staikouras is a Senior Lec-turer in Finance at Cass Business School in London, UK.

Proponents of behavioral finance argue that poorlyinformed and unsophisticated investors might lead financialmarkets to be inefficient. The debate between neoclassicaland behavioral finance is wide ranging, and sometimesexplains differences in policy recommendations on suchissues as financial regulation, corporate governance, or theprivatization of social security. It had immediate impactworldwide including emerging markets (Muradoglu, 1989).

Behavioral finance emerged as a field in the early 1980swith contributions by, among others, David Dreman, RobertShiller, Hersh Shefrin, Meir Statman, Werner De Bondt and

Richard Thaler. Soon, this small group of financial economistswas meeting regularly with psychologists — including PaulAndreassen, Daniel Kahneman, and Amos Tversky — at theRussell Sage Foundation in New York. Five or six yearslater, the National Bureau of Economic Research beganorganizing semi-annual meetings. From its beginnings as afringe movement, behavioral finance moved to a middle-of-the-road movement, with spillover effects on marketing,management, experimental economics, game theory, politicalscience and law. Now behavioral finance is poised to replaceneoclassical finance as the dominant paradigm of thediscipline.

Traditionally, economists model behavior in terms ofrational individual decision-makers who make optimal useof all available information. There is ample evidence that therationality assumption is unrealistic. The path-breaking workof Herbert Simon, Tversky and Kahneman, Lola Lopes, andothers on bounded rationality, judgmental heuristics, biases,mental frames, prospect theory, and SP/A theory has providednew foundations for financial economics. Behavioral financestudies the nature and quality of financial judgments andchoices made by individual economic agents, and examineswhat the consequences are for financial markets andinstitutions. Investment portfolios are frequently distorted,with consequent excess volatility in stock and bond prices.Examples include the stock market crash of 1987, the bubblein Japan during the 1980s, the demise of Long-Term CapitalManagement, the Asian crisis of 1997, the dot-com bubble,and the financial crisis of 2008. Most everyone agrees that itis problematical to discuss these dramatic episodes withoutreference to investor psychology.

The term “behavioral finance” has a variety of meanings.Our paper aims to provide an over-arching view of the field.It is a summary of a panel discussion. The paper is writtenfor a wide spectrum of readers, including financialpractitioners. It begins by examining the current state offinance, reviews some fundamental questions, and then

8 JOURNAL OF APPLIED FINANCE — FALL/WINTER 2008

introduces behavioral concepts. “Behavioral finance: QuoVadis?” Sections I and II review modern and behavioralfinance, respectively. Section III briefly delves into theefficient markets literature. Section IV discusses key buildingblocks of the behavioral approach. Section V explores somenew ideas in behavioral asset pricing and behavioral corporatefinance. Section VI provides a self-critique. Section VIIconcludes.

I. What Is Finance?

Let us start by defining finance. Even though the realeconomy and finance are linked, we usually make a distinctionbetween the two. The real economy is where goods andservices are produced and consumed, and where wealth iscreated. The world of finance is mostly seen as a sideshow.Even so, finance serves important functions such as thepayment system, the pooling and transferring of funds, savingand investing, contract design, organizational architecture,and risk management. Anyone who contemplates the functionsof finance, and the financial institutions involved in them(e.g., the banking system; insurance companies; moneymanagement firms; pension funds; rating agencies, and soon), soon realizes that the central unifying concept is assetvaluation. Certainly, the theory of value, and comparisons ofprice and value, is what much of finance is about. Of course,valuation also impacts the decisions investors make aboutthe composition of their portfolios and the decisions whichmanagers make about the sources and uses of funds in theirfirms.

Modern (or neoclassical) finance is the paradigm that hasgoverned thinking in academic finance since the late 1950s.It flows from a philosophical tradition (the 18th centuryEnlightenment) that aims to reconstruct society withindividual rational action as its centerpiece. Modern financeis built on two pillars. The first pillar is the concept of“beautiful people”, defined as logical, autonomous agentscharacterized by expected utility maximization (over time),risk aversion, Bayesian updating, and rational expectations.The second pillar is the concept of “beautiful markets” i.e.depending on the problem-at-hand, perfect, liquid,competitive, complete markets. Based on these two conceptsas well as the mutual adjustment of demand and supply (plusan assortment of auxiliary assumptions), various asset pricingtheorems are derived. In equilibrium, all agents reach theiroptimum. Investment portfolios are mean-variance efficient.Only systematic non-diversifiable risk is priced. There areno opportunities left for rational arbitrage. Conditional onwhat is known about the future, price equals value.

What is the role of institutional factors such as marketorganization, regulatory framework, tax systems etc. inneoclassical finance? To a first approximation, there is none.

Rational agents work around institutional frictions and therebyrender them immaterial to market outcomes. Of course, theprocess may take time. Merton Miller made this type ofinstitutional arbitrage a favorite lecture theme. He spokeabout institutions as potential distortions, though ultimatelyneutral mutations. Miller’s comments were often formulatedin the context of regulatory barriers to financial innovation,but the link with the Miller-Modigliani theorems and the workof Ronald Coase is obvious. Robert Merton’s views aresimilar. His writings say that the basic functions of financeare the same, always and everywhere. What does change isthe technological and regulatory environment. That is whybanking in 2008 is different from banking in 1908, and whybanking in Switzerland is different from banking in Egypt.

How do modern finance theorists plead their case? Theymostly reason in a logically deductive way starting fromaxioms that have a priori normative appeal.1 In the past,modern finance theorists rarely administered surveys(Muragdoglu, 1989) and they did not run experiments,although this is starting to change(Muragdoglu, Salih, andMercan, 2005). Still, many financial economists believe thatthe swaying power of data cannot match the power of logic.

II. What Is Behavioral Finance?

Behavioral finance does not assume rational agents orfrictionless markets. It suggests that the institutionalenvironment is vitally important. The starting point is boundedrationality. Paul Slovic (1972) writes that “a fullunderstanding of human limitations will ultimately benefitthe decision-maker more than will naive faith in theinfallibility of his intellect.” That economic and financialintuition is fragile may clash with our aspirations for mankind,but it looks more plausible than the opposite view thatinvestors and advisors (as well as bankers and corporatemanagers) know perfectly well what to do.

Behavioral finance is the study of how psychology impactsfinancial decisions in households, markets and organizations.The main question is: What do people do and how do theydo it? The research methods are mostly (but not exclusively)inductive. Behavioral researchers collect “facts” aboutindividual behavior (based on experiments, surveys, fieldstudies, etc.) and organize them into a number of “super-facts.” The psychology of decision-making can be exploredin various ways. A quarter-century ago, most effort went intocognition. Consider, for instance, the heuristics and biasesliterature pioneered by Tversky and Kahneman (1974) andKahneman and Tversky (1979). Their main focus was on

1The normative approach asks how decision-makers logically should actwhile the positive approach looks at how decisions are truly made.

9DE BONDT, MURADOGLU, SHEFRIN, & STAIKOURAS — BEHAVIORAL FINANCE: QUO VADIS?

questions such as: How do people think? How do they decide?Current work continues to draw on cognitive research. Inaddition, it studies emotion (mood; affect) and socialpsychology (especially herding behavior).

What has been learned? The central insights of behavioralfinance are described in Barberis and Thaler (2003), Danielet al. (2002), De Bondt (2002, 2005, 2008a), Dreman (1995),Shefrin (2001a, 2002) and Thaler (1993).2 There are threeclasses of findings. First, thereis a catalog of biases, i.e.,predictable mistakes such asoverconfidence in judgment,wishful thinking,procrastination, myopia, etc.Intuition is fragile. Note that itis not alleged that financialintuition is broken, only that itcan break. Specific errorsdepend on context, but are systematic nonetheless. Theresearch examines psychological mechanisms whichilluminate how the human mind works. It also explains whyfinancial judgment is fallible.

The second class of findings relates to the speculativedynamics of asset prices in global financial markets. Here,the main insight is that the systematic errors of unsophisticatedinvestors (“noise traders”) create profit opportunities forexperts, even if noise traders create a great deal of risk.Investor sentiment matters. Widely-shared misconceptions(that may be self-reinforcing) cause transient price bubbles,large and small. Certainly, rational arbitrage matters too but,since most people’s investment horizons are short, arbitragedoes not wipe out inefficiencies.

The third class of findings has to do with how decisionprocesses shape decision outcomes.3 Here too, the study offiascoes is informative, since it guides us to decision processvariables that are critical. Numerous specific applications ofthis finding appear in Nudge, a book authored by RichardThaler and Cass Sunstein (2008). One striking example hasto do with organ donation (Johnson and Goldstein, 2003).The U.K. participation rate in organ donation is approximately15% whereas in Belgium it is over 95%. What explains this

difference? For an answer, we look to the decision processplus the well-known fact that people tend to stick with thestatus-quo. In case of a fatal car accident in the U.K., the lawassumes –-unless the driver signs his license to the contrary-– that his bodily organs will not be donated. In Belgium, thedefault solution is the opposite, i.e. the driver’s organs aredonated. Note that in either country all it takes to modify thedefault is a signature.4

Why is behavioral researchoften so convincing? Onereason is that “good” behavioralresearch depends on supportfrom multiple sources. Forinstance, laboratory researchpermits any reader who doubtsthe results to replicate theexperiment “at home.” Further,many studies rely on surveys or

observe individual behavior (e.g., trading records) in a naturalenvironment (e.g., Odean, 1998, 1999). Lastly, behavioralresearchers also make use of conventional market-level priceand volume data. This “one-two-three punch,” we believe,provides a discipline to behavioral theorizing that is farsuperior to what is typical for research in modern finance.Decision anomalies (in the laboratory), matched withanomalies in the behavior of individual agents (in a naturalenvironment), matched with market anomalies (when socialinteraction allows fine-tuning) produce a powerful body ofevidence. Take, for example, investor overreaction. Certainly,experiments teach us that subjects do not update beliefs inBayesian fashion (De Bondt, 1993, Muradoglu, 2002).Second, when asked, investors tell us that they like to buypast winner stocks but that they stay away from past losers.Regardless of what investors say, their trading records confirmthe bias.5 Third, at the market level, we find predictablereversals in share prices (De Bondt and Thaler, 1985). Thelaboratory, financial behavior, and market results appear tobe connected.

III. Price and Value

Milton Friedman (1953) and Eugene Fama (1965) arguethat, even though naive investors may push security pricesaway from intrinsic values, more sophisticated traders will

Behavioral finance is based onthree main building blocks,namely sentiment, behavioralpreferences, and limits toarbitrage.

2These works lay emphasis on investment and asset pricing. However,Shefrin (2005) focuses on behavioral corporate finance. Apart from agencyand asymmetric information problems, there are behavioral costs thatobstruct the corporate value maximization process.

3This type of research is especially relevant to the study of organizations.Everyday we learn more about committee decision-making (e.g. boards),the role of top managers in the creation of corporate wealth, and the prosand cons of bureaucratic formalities and red tape. As president of theAmerican Finance Association, Michael Jensen asked that we break openthe black box called the firm. Behavioral finance is contributing to thateffort.

4Our economist friends emphasize incentives. We ask them: What incentivescheme may achieve the same outcome (95% participation) that aseemingly minor adjustment in the decision process produces effortlessly?

5Ironically, investors are more likely to hang on to losers than to winners ifthe changes in value occurred while the stocks were part of their portfolio(Shefrin and Statman, 1985).


find it worthwhile to correct any mispricing. In other words,competitive rational arbitrage guarantees that, at all times,the market valuation of any security reflects what is —andwhat can be— known about its future cash flows and theopportunity cost of capital. Based on market efficiency,finance academics have made two main assumptions aboutsecurity valuation. First, securities have an intrinsic valuebased on their fundamentals; and second, their prices are notpredictable on the basis of publicly available information.Among others, Fama (1965) argued that the competitiveactivity of arbitrageurs will bring security prices into linewith fundamentals. Thus, the arbitrage activity of rationaltraders will prevail (over irrationals) as long as securities haveclose substitutes. Over the decades, this perspective, theefficient markets hypothesis, has been examined by manyscholars.

Behavioral finance has provided evidence whichcontradicts the notion of efficient markets. An example isthe case of “Siamese twins” stocks (Rosenthal and Young,1990; Froot and Dabora, 1999). Consider the share pricemovement of Royal Dutch/Shell Group, where Royal Dutchstock trades in the US/Netherlands and Shell stock trades inthe U.K. The two companies’ original merging interests wereon a 60:40 basis for Royal Dutch and Shell respectively. Thusa ratio of 1.5 (price of Royal Dutch relative to Shell) shouldhave been achieved in order for the prices to reflectfundamentals. Froot and Dabora (1999) and Lamont andThaler (2003) find that the relative price ratio ranges from15% overvalued to 35% undervalued. This contradicts “thelaw of one price.” In relation to these stocks, there is alsoevidence that noise trader risk is a significant impediment toarbitrage (Scruggs, 2007).

The efficiency of security prices has also been challengedby Graham (1949), Nicholson (1968), Basu (1977), Dreman(1977, 1980), and many others who believe that stocks withlow price-to-earnings (PE) ratios are undervalued and stockswith high PE ratios are overvalued. Investors, these authorssuggest, are overly pessimistic about the prospects of lowPE stocks. Since the crowd avoids them, investing in low PEstocks is a profitable contrarian strategy.6 De Bondt and Thaler(1985) extend this idea with their analysis of investoroverreaction and with the finding of predictable pricereversals for long-term winner and loser stocks. Poterba andSummers (1988) obtain analogous reversals for national stockprice indexes.

There are other widely documented phenomena which aredifficult to reconcile with efficient markets. Consider thefollowing examples:

• Price volatility that is not linked to news: Cutler et al.(1991) show that during periods with “no” major newsannouncements equity prices experience some of their largestone-day moves. A vivid example was the 22.6% drop in theDow Jones Industrial Average on October 22, 1987. Roll(1984, 1988) offers systematic evidence of market volatility,not associated with information arrival.

• Excess volatility: Keynes (1936, pp. 153-4) observes how“day-to-day fluctuations in the profits of existing investments…tend to have an altogether excessive, and even absurd,influence on the market.” This comment anticipates Shiller’s(1981, 1993) work on equity volatility. There, it is suggestedthat fluctuations in economic fundamentals alone (e.g.,dividends) cannot possibly account for the observed aggregateprice movements.

• Earnings momentum: Stock prices “underreact” to annualand quarterly announcements of corporate earnings causinga post-announcement drift in returns, markedly for firms withlow institutional shareholdings (Bartov et al., 2000). Bernardand Thomas (1989, 1990) were among the first to establishthis effect, but the research goes back to Ball and Brown(1968).7

• Price momentum: For holding periods up to one year,Jegadeesh and Titman (1993, 2001) and others show trendsin share prices of individual stocks, i.e., past winner stocksremain winners, and past losers remain losers.8 Yet, beyondone year, momentum is often followed by reversals. Europeanand emerging markets exhibit similar patterns (Rouwenhorst,1998, 1999; Muradoglu, 2000). Small firms feature moremomentum than large firms (Jegadeesh and Titman, 1993;Grinblatt and Moskowitz, 1999; Lee and Swaminathan,2000). Price momentum may be due to positive feedbacktrading. That is, when large increases in stock prices pull innew investors, the inflow of funds causes prices to rise further.It is probable that the phenomenon is also partly explainedby earnings momentum, investor underreaction, and thegradual dissemination of news. Grinblatt and Han (2005) andFrazzini (2006) suggest that momentum can be explained bythe disposition effect, a concept introduced by Shefrin andStatman (1985) whereby investors sell winners too early andhold losers for too long.

• Equity premium puzzle: Historically, the spread betweenthe return on equities and fixed income US government

6Other price-scaled ratios, e.g., the book-to-price ratio, also forecast stockreturns. See, e.g., De Bondt and Thaler (1987) and Fama and French (1992).

7Corporate news that is not directly related to earnings also predicts returns.See, e.g., Michaely et al. (1995) on dividends or Ikenberry et al. (1995) onshare price repurchases. For a critique of these findings, see Fama (1998).

8Trends are also visible in stock indexes of US industries and investmentstyles, and in stock indexes of foreign equity markets. See Chen and DeBondt (2004) and De Bondt (2008b) for details.


securities has exceeded 6%. It is difficult to reconcile themagnitude of this premium with modern asset pricing theory(Mehra and Prescott, 1985) since it implies that therepresentative investor is exceedingly risk-averse.

• Size and calendar effects: Small firms earn anomaloushigh returns. There is also ample literature on calendar effects.For example, there are curious patterns in equity returnsrelated to weekends, the turn of the month, and the turn ofthe year (Siegel, 1998; Keim, 1983, 1986; Reinganum, 1983;Roll, 1983).

The main point of the above examples is that businessfundamentals alone do not explain the structure and dynamicsof asset prices. Behavioral finance offers promising, plausiblealternative explanations for some of these phenomena. In thenext section, we describe some of the key psychologicalbuilding blocks of the behavioral framework.

IV. Key Building Blocks

Behavioral finance is based on three main building blocks,namely sentiment, behavioral preferences, and limits toarbitrage. By sentiment is meant investor error. Errorsoriginate at the level of the individual but can manifestthemselves at the level of the market. Behavioral preferencescapture attitudes about risk and return which do not conformwith the principles of expected utility theory. In neoclassicalfinance, rational information traders exploit the behavioralinconsistencies of irrational noise traders, and in so doinglead prices to be efficient. Proponents of behavioral financesuggest that there are limits to the process of arbitrage, andas a result prices need not be efficient. We next describeeach of these building blocks in greater detail.

Psychology shows that people’s beliefs are oftenpredictably in error. In many cases, the source of the problemis cognitive. That is, the problem is a function of how peoplethink. Some psychological mechanisms have been modeledas heuristic rules of thumb. By and large, heuristics performwell but, sometimes, they lead to systematic error. A few biasesin beliefs are described below.

• Anchoring is a form of bias where beliefs rely heavily onone piece of information, perhaps because it is was availablefirst, and are not sufficiently adjusted afterward. For instance,investor forecasts may anchor on the price at which theybought a security (De Bondt, 1993; Muradoglu and Onkal,1994). “Conservatism” is closely related. Investors may placeexcessive weight on past information relative to newinformation, i.e., they underreact.

• Representativeness is overreliance on stereotypes.Investors who regard recent time-series trends asrepresentative of an underlying process are vulnerable toextrapolation bias. The “law of small numbers” is a related

bias whereby people behave as if the statistical properties ofsmall samples must conform to the properties of largesamples. Investor overreaction is partly rooted inrepresentativeness. The “gambler’s fallacy” is also connectedto representativeness but leads investors to make unwarrantedpredictions of reversal.

• Availability bias means that investors overweighinformation that is easily accessible, e.g., that is easily recalledfrom memory or that corresponds to a future scenario that iseasy to imagine. People are likely to remember events thatreceive a lot of attention by the media and this influencestheir behavior (see, e.g., Barber and Odean, forthcoming).

• Overconfidence implies that individuals overvalue theirknowledge or abilities. It has many consequences. Forinstance, overconfidence may lead investors to underestimaterisk or to overestimate their ability to beat the market.Overconfidence bias may also cause excessive trading. Danielet al. (1998, 2001) suggest that investors suffer from acombination of overconfidence and self-attribution bias, i.e.,people attribute success to their own skills, but blame failureon bad luck.

Investor preferences constitute the second key element offinancial models. In this regard, there are several behaviorally-based preference frameworks. The best known is prospecttheory, developed by Kahneman and Tversky (1979) todescribe the manner in which people systematically violatethe axioms of expected utility theory. Prospect theory differsfrom expected utility theory in that probabilities aresubstituted by decision weights, and the value function isdefined over gains and losses, not final wealth.9 Otherbehavioral preference frameworks include SP/A theory,change of process theory, regret theory, affect theory, andself-control theory.

The following list describes some of the most importantfeatures of behavioral preferences:

• Loss aversion portrays investors’ reluctance to realizelosses. Tversky and Kahneman (1992) argue that peopleweight losses twice as much as gains of a similar magnitude.Unlike what is assumed in neoclassical finance, loss averseinvestors may be inconsistent towards risk. People may prefer

9Fellner (1961) introduces the concept of decision weight to explainambiguity aversion. Kahneman and Tversky (1979) state: “In prospecttheory, the value of each outcome is multiplied by a decision weight.Decision weights are inferred from choices between prospects much assubjective probabilities are inferred from preferences in the Ramsey-Savageapproach. However, decision weights are not probabilities: they do notobey the probability axioms and they should not be interpreted as measuresof degree or belief.”


to avoid risk in order to protect existing wealth, yet mayassume risk in order to avoid sure losses.10

• Mental accounting refers to how people categorize andevaluate financial outcomes (Henderson and Peterson, 1992).Shefrin and Thaler (1988) assume that people categorizewealth in three mental accounts: current income, currentwealth, and future income. Itis furthermore assumed thatthe propensity to consume isgreatest from the currentincome account and smallestfrom the future-incomeaccount. One consequence isthe tendency to treat a newrisk separately from existingrisks, usually called narrowframing.11 Narrow framingposes dangers. Investorsmay act as if they are risk averse in some of their choices butrisk seeking in other choices. Shefrin and Statman (2000)develop behavioral portfolio theory in single and multiplemental account versions (SMA and MMA). In the SMAversion, investors integrate their portfolios into a single mentalaccount; in the MMA version, investors prefer securities withnon-normal, asymmetric distributions that combine downsideprotection (in the form of a floor) with upside potential.

• Myopic loss aversion combines time horizon-basedframing and loss aversion. Investors are more averse to riskwhen their time horizon is short than when it is long (Haighand List, 2005). Benartzi and Thaler (1995) argue that thesize of the equity premium suggests that investors weighlosses twice as much as gains, and that they evaluate theirportfolios on an annual basis.

• Self-control refers to the degree to which people cancontrol their impulses. Thaler and Shefrin (1981) analyze howpeople exhibit self-control with respect to saving behavior.Shefrin and Statman (1984) develop a theory of dividendsbased on this idea, where mainly elderly investors have apreference for dividends. Shefrin and Statman (1985) referto self-control when they explain how investors deal with theimpulse to hold onto losing investments for too long (seeLease et al., 1976, for empirical evidence).

• Regret aversion stipulates that investors may wish to avoidlosses for which they can easily imagine having made a

superior decision (ex post). Regret helps to explain thedividend puzzle if, ex ante, investors want to avoid the regretof having sold shares that later went up in price. Such regretsmay also encourage investors to hold on to loser stocks(Shefrin and Statman, 1985). Koening (1999) argues thatinvestors will bet on good assets, in order to avoid regret,

which in turn could possiblytrigger some sort of herdingbehavior.

Finally, limited arbitrageplays a crucial role inbehavioral asset pricing. Torepeat, a basic tenet of modernfinance is that arbitrageursforce prices to converge totheir true fundamental values.Yet, research has uncovered aseries of financial market

phenomena that do not conform to the notion that full arbitrageis always carried out. For this reason, behavioral asset pricingmodels focus on the limits that arbitrageurs face in attemptingto exploit mispricing. Markets are not frictionless because oftransaction costs, taxes, margin payments, etc. Therefore, theactions of noise traders (i.e., traders with biased beliefs, notbased on fundamental information) may cause prices to beinefficient. As a result, arbitrage can be risky (Shleifer, 2000).Mispricing has been the focus of many studies, e.g., Cornelland Liu (2001), Schill and Zhou (2001), or Mitchell et al.(2002).

V. Behavioral Analogues to NeoclassicalAPV and SDF-based Pricing

Asset pricing theory and corporate finance are in theprocess of becoming behavioralized. At the moment, thebehavioral approach is somewhat piecemeal, whereas theneoclassical approach is more coherent and integrated.Shefrin (2005, 2008a,b) argues that in the future finance willcombine the best of neoclassical and behavioral elements,thereby presenting a coherent, integrated framework fordescribing how markets are impacted by psychologicalphenomena.

Behavioral asset pricing emphasizes that asset prices reflectinvestor sentiment, broadly understood as erroneous beliefsabout future cash flows and risks (Baker and Wurgler, 2007).Sentiment impacts the prices of all assets, and drives thedifference between what behavioral and neoclassical financetell us about the relationship between risk and return. In thisregard, consider the global financial crisis that began in 2008.Academics, media, and policy makers have all contributedto the question of what caused the crisis. In a New York Times

Sentiment impacts the prices of allassets, and drives the differencebetween what behavioral andneoclassical finance tell us about therelationship between risk andreturn.

11In the traditional approach, investors judge a new gamble via itscontribution to total wealth.

10In The Theory of Moral Sentiments, Adam Smith (1759) says that “wesuffer more when we fall from a better to a worse situation than we everenjoy when we rise from a worse to a better.” Smith’s observation capturesthe modern notion of loss aversion.


article, Lohr (2008) discussed the failure of financialengineering to incorporate the human element. Notably, thebehavioral stochastic discount factor (SDF) approachdeveloped by Shefrin incorporates the human factor intofinancial engineering. Lohr says that Wall Street analysts diduse risk models that correctly predicted how the market forsubprime mortgage backed securities would be impacted bya decline in real estate prices. However, analysts attachedtoo low a probability of a major decline in real estate prices.This type of situation is typical of events that take place in abehavioral SDF model, where investors collectively commiterrors in their judgments of probabilities, thereby leadingsome derivatives and their underlying assets to be mispriced.

One of the most important points made in behavioralcorporate finance is that although the principles taught intraditional corporate finance have great value, psychologicalobstacles may prevent organizations from putting them intopractice (Shefrin, 2005). Many normative aspects oftraditional corporate finance remain intact. Yet, they need tobe augmented so that there is a narrowing in the gap betweenwhat academics preach and what managers do. Tomorrow’smanagers should understand why people, includingthemselves, make mistakes, and how as managers they shoulddeal with market inefficiencies.12 The new approach shouldbe specific, not general, and focus on how to make decisionsabout capital budgeting, capital structure, mergers andacquisitions, payout policy, and corporate governance. In thisregard, Shefrin (2008a,b) introduces the concept of“behavioral adjusted present value.” He begins withtraditional adjusted present value (which combines net presentvalue and financing side effects) but adds a component tocapture the effects of inefficient prices.

Shefrin (2008a, b) suggests that an appropriate startingpoint for discussing the asset pricing paradigm transition isthe book written by John Cochrane (2005). Cochrane’sexcellent work is built around the concept of a stochasticdiscount factor. His approach offers a unified treatment. Inparticular, the capital asset pricing model (CAPM), Fama-French multifactor model, and models for the yield curveand option prices all appear as special cases of a generalSDF framework. For example, the CAPM corresponds tothe special case when the SDF is a linear function of thegrowth rate of aggregate consumption in the economy. Theweakness of the neoclassical SDF approach is that itsunderlying assumptions are behaviorally unrealistic.

Although an extensive discussion is beyond the scope ofthe present study, a point worth addressing is whether

behavioral assumptions alter the basic neoclassicalrelationship between the SDF and mean-variance frontier.They do not. What they do is alter the shape of the SDF andthe ingredients of mean-variance portfolios. In neoclassicaltheory, the SDF is monotone declining. However, Aït-Sahaliaand Lo (2000) and Rosenberg and Engle (2002) find that,during the first half of the 1990s, the SDF features anoscillating shape that supports the predictions based onbehavioral assumptions. Moreover, using survey expectationsdata, Shefrin (2005, 2008) predicted that the shape of theSDF would change during 2001-2004, with a decline in theleft portion displayed in Figure 1. Notably, Barrone-Adesi etal. (2008) report that during 2002-2004 the left portion ofthe SDF does indeed feature a flat shape.13

Mean-variance analysis is very useful for bringing out theimplications of behavioral phenomena for the pricing of allassets. To see how different behavioral and neoclassicalmean-variance portfolios can be, consider figure 1. Thisfigure contrasts the equilibrium returns to two mean-varianceportfolios, one neoclassical and the other behavioral, asfunctions of aggregate consumption growth in the economy.The return to a neoclassical mean-variance portfolio isessentially linear, and corresponds to the return fromcombining the risk-free security and the market portfolios.In contrast, the return to a behavioral mean-variance portfoliooscillates with economic growth, reflecting the impact ofinvestor sentiment. The construction of efficient portfoliosunder the neoclassical paradigm is done by combining aninvestment in the risk-free asset and the market portfolio.The theoretical outcome of such combination is known asthe two-fund separation theorem (Tobin, 1985).14 Behavioralmean-variance portfolios satisfy the two-fund separationtheorem. However, the risky asset used to construct behavioralmean-variance portfolios features the use of derivatives.

It is a well-established fact that investors requirecompensation to assume risk. Risk can take any form infinancial markets but, in broad terms, the neoclassicalframework focuses on fundamental risk. The behavioralapproach adds sentiment risk. Therefore, behavioral riskpremiums serve as compensation for bearing both sentimentand fundamental risks. Behavioral risk premiums, like theirneoclassical counterparts, will be associated with betas andfactor pricing models. To illustrate this point, consider figure2. This figure displays a mean-variance return pattern whoseshape is that of an inverse U. Notably, such a shape is implied

12Behavioral corporate finance emphasizes organizational heuristics andbiases. Such heuristics and biases were endemic to financial firms involvedin the global financial crisis that began in 2008.

13If investors underestimate the probability of extreme negative events,which is part of the “black swan” phenomenon emphasized by Taleb (2006),then the SDF will typically be upward sloping in its left tail.

14In the case of leveraged portfolios, the theorem still holds but a negativeposition with respect to the risk-free asset is held.


FIGURE 1Contrasting the return to a neoclassical mean-variance portfolio and the return to a behavioral mean-variance portfolio, as functions ofaggregate consumption growth in the economy.

FIGURE 2Special case of Figure 1, when behavioral mean-variance return function has the shape of an inverse-U. This figure also shows that theneoclassical mean-variance return is approximately linear. In the CAPM, the mean-variance function is exactly linear.

75%

80%

85%

90%

95%

100%

105%

110%

96%

97

%

99

%

10

1%

103

%

104

%

10

6%

Consumption Growth Rate g (Gross)

Mean-Variance Return

(Gross)

Behavioral MV Portfolio Return

Neoclassical Efficient MV Portfolio Return

Gross Return to Mean-variance Portfolio:

Behavioral Mean-Variance Return vs Efficient Market Mean-Variance Return

0.95

0.96

0.97

0.98

0.99

1

1.01

1.02

1.03

95

.82

%

96

.64

%

97

.48

%

98

.31

%

99

.16

%

10

0.0

1%

10

0.8

7%

10

1.7

4%

10

2.6

1%

10

3.4

9%

10

4.3

8%

10

5.2

8%

10

6.1

9%

Consumption Growth Rate g (Gross)

Mea

n-v

aria

nce

Ret

urn

Neoclassical MV Portfolio Return

Behavioral MV Portfolio Return

Return to a Combination of the Market Portfolio and

Risk-free Security


by the work of Dittmar (2002). When the inverse U shape isquadratic, the risk premium for any security can be expressedas a function of two factors, the return to the market portfolioand the squared return to the market portfolio. Modelsinvolving squared returns to the market portfolio areassociated with the analysis of coskewness. The work ofBarone-Adesi and Talwar (1983), Harvey and Siddique(2000), and Barone-Adesi et al. (2004) indicates thatcoskewness is important in the determination of riskpremiums. Much of theexplanatory power of size,book-to-market equity andmomentum plausibly derivesfrom coskewness.

VI. Strengths andWeaknesses

Behavioral research hasfour major strengths. First, ithas proven itself to beproductive. For example, ithas led to a series of newempirical findings. Examplesinclude over- and underreaction in share prices, the new issuesand stock repurchase puzzles, and the role the stock spliteffect. Second, with its focus on the impediments to optimaldecision-making, behavioral finance brings a pragmaticapproach to the study of financial decisions. For instance,insights from behavioral finance help our understanding ofhow to structure the relationship between a firm’s investorsand its managers. Certainly, the behavioral approach suitsthe professional business school which aims to educatemanagers and to improve their expertise. Third, behavioralfinance potentially brings a new type of discipline to socialscience research. Discipline fundamentally impliestriangulation i.e. the synthesis of data from multiple sources.(“Finance you can believe in” requires more thanmathematical proof.) The final strength of behavioral financeis simply that it is a stimulating field of scholarship. Peopleand money: What can fascinate more? Perhaps the appeal ofbehavioral finance is that it is social science, but with strongemphasis on both the social and the science.

Behavioral finance also has weaknesses. As mentioned inthe previous section, it lacks the unified theoretical core ofneoclassical finance, and can be lacking in discipline. Forexample, there is no single preference framework toaccommodate the features in prospect theory, SP/A theory,

regret theory, self-control theory, and affect theory.15 Indeed,a series of recent works has identified the limitations ofprospect theory in explaining the behavior of real worldinvestors.16 There are multiple behavioral explanations formomentum, not all mutually consistent.17 In this regard, manybehavioral asset pricing models are eclectic and ad hoc. Somemodels rely on the assumption that prices are set by arepresentative behavioral investor, even though aggregationtheory indicates that such an assumption is unwarranted. As

for the winner-loser effect,there is no clear explanation asto why reversals only appearto occur in January.

To be sure, behavioralfinance is a work in progress.It is unfinished. Indeed, at thepresent time, many researchersrefer to “behavioral finance” todescribe their work18 but thereis no common accepteddefinition of what it is.Perhaps, this is not an issue inthe long-term. After all, themain goal of behavioral

finance is to behavioralize finance, not to create a separatefield of scientific study.

The second weakness can be described by analogy. Just asa study of the economic function of payments and settlementscannot tell us much about the practical organization of thepayment system (cash vs. credit cards etc.), in the same way,undivided focus on psychological mechanisms (e.g. impulsesand predispositions, or psychophysics) does not allow anadequate interpretation of economic and financial events. Anindividual is much more than a biological organism; (s)he isalso a person, a social-historical creation. Reality is sociallyconstructed. Philosophers often compare man’s conduct to

One of the most important pointsmade in behavioral corporatefinance is that although theprinciples taught in traditionalcorporate finance have great value,psychological obstacles mayprevent organizations from puttingthem into practice.

15This list is hardly exhaustive. Investors also have preferences whichinclude issues that go beyond returns, an example of which is sociallyresponsible investing. See Statman (2008).

16See Hens and Vlcek (2005), Barberis and Xiong (forthcoming), andShefrin (2008)

17There are at least four separate theories to explain why markets exhibitshort-term momentum but long-term reversals. Some psychologicalexplanations, such as Barberis, Shleifer, and Vishny (1998) emphasizeunderreaction. Other psychological explanations, such as Daniel,Hirshleifer, and Subrahmanyam (2001) emphasize overreaction. Grinblattand Han (2005) emphasize the disposition effect.

18Hong and Stein (1998) develop a behavioral model in which someinvestors rely on fundamental analysis and other investors rely on technicalanalysis. However, there are no specific psychological elements in theirmodel.


that of a stage actor. People enact roles. Their motives, outlookand self-image are shaped by what is expected from them insociety. Hence, research in behavioral finance shouldexamine the tangible content of people’s thought processes.19

Evidently, this issue cannot be resolved without reference tosocial, cultural and historical factors. We need to look moreinto the content, structure and style of intuitive economicstories. For example, how do Swiss citizens (who in majorityrent) think about home ownership, and likewise how doAmericans? In general, what sorts of economic arguments(true or false) sound plausible to investors, persuade themand motivate their actions? Investment bankers, clientrelationship and financial marketing managers, among others,would be interested in answers to these questions. Yet, so far,behavioral finance has little to say.

Third, behavioral finance must move beyond the narrowmicro-level study of typical “mistakes.” If not, too muchbehavior remains unintelligible. Yes, US data suggest thatCEOs, entrepreneurs and investors tend towards unrealisticoptimism – an error with perilous consequences. But, onemay ask, what causes over-optimism? Is it context-specific?Does it stem from past personal success? Or is it anincontestable part of the American character? A morefundamental critique is to pose the related question: What isa mistake? Economists take a hard line. Error, they say, isstrictly about the contrast between actions that are taken andactions that rationally should be taken in accordance with anindividual agent’s costs and benefits. Economists’ chiefconcern is efficiency. However, the concept of error is elastic.James March and Chip Heath (1994) draw a useful distinctionbetween the economic “logic of consequences” and the morebroadly applicable “logic of appropriateness.” Consider, forinstance, someone who breaks the rules of etiquette. His normviolations may be embarrassing, perhaps inexcusable, but mayalso make little practical difference. Still, collective beliefsand norms often make all the difference. For example, asidefrom efficiency, there are other criteria of economic andfinancial organization such as sustainable development orequity and fairness. These may be “protected values,” i.e.,people reject all trade-offs for money.

Finally, there is a disconnect between the emphasis inbehavioral research on human frailties and the reality that inmany corners of the globe people lead a pretty good life.Why are we collectively so strong, yet as individuals so weak?Why does societal rationality transcend individual rationality?

Orthodox economic theory places the pinnacle of rationalityin the brains of individual people whose self-interest drivesmarket prices.20 It blames social evils on dysfunctionalincentives and disarray, mainly in corporate bureaucraciesand government. The truth may be nearly the opposite.Rationality and well-being derive from organization,spontaneous or deliberate. Why are institutions so cruciallyimportant? The reason is that everyone in society dependson everyone else. We sell 99% of what we produce, we buy99% of what we consume, and we lead better lives for it.Incessant technological progress, product and servicestandardization, and economic organization are central. Thesecret is encapsulated by the motto of the 1933 ChicagoWorld’s Fair: “Science finds, industry applies, manconforms.”

Technological artifacts make us smart for several reasons(Norman, 1993). First, technology greatly extends man’scognitive capabilities. Because we forget, we use a notepador we access the Internet. Second, technology is coupled withlabor specialization. For example, experts make decisions(e.g. in relation to the nation’s supply of electricity) that tensof thousands are incapable of making for themselves. Third,technology embodies knowledge. Few of us know exactlyhow the watches on our wrists function. Fortunately, we donot need to know. It is enough that we are able to read thetime.21 Finally, technological artifacts often allow cheapreplication. So, good products or ideas spread quickly.However, people and machines have to work together..Technology can be easy or difficult to use. Similarly,administrative organization can be effective or ineffective.Smart technology and organization are human-centered(Reason, 1990). In the short run, this is a matter of design,i.e., of pragmatic behavioral research. Over longer periods,it is the outcome of an evolutionary process. To ask about the“logic” of American corporate law or the dashboard of anautomobile is a bit like asking who designed the Frenchlanguage, to what purpose and under which specifications.

That in modern society the balance between individual andinstitutional forces has shifted often gets on our nerves. Welament that man must “conform,” that personal freedom islost when either law regulates what we do or largecorporations – e.g. because of network externalities – controlour choice options. Yet, man is limited by his brainpower,habits, and conception of purpose. Organization produces

21Occasionally, however, society forgets why some systems or technologieswere designed the way they were, and this can be very costly. Recall theY2K problem.

19It is difficult to interpret human action without knowing first how peoplethink about a problem. An extravagant illustration, far removed fromfinance, has to do with the September 2001 attacks in New York. Thequestions that we would ask in relation to these evil acts are as follows:How did the perpetrators comprehend the world, and how did theyunderstand their self-interest so that they wanted to be suicide-pilots?

20 Austrian, institutional and evolutionary economics do not. Theseeconomists espouse the private enterprise system but call attention to thefact that its assumed virtues (innovativeness, responsiveness, administrativeparsimony) have no solid basis in microeconomic theory.


predictability. This is fundamental. Rules and regulationscoordinate society while reducing the individual’s need tothink.22 Of course, financial technology is often customer-friendly and performs brilliantly. Take, for instance, the ATM-machine. Still, it is easy to come up with counter-examples.Retirement saving plans and asset allocation tools can be mademore effective. The US mortgage debt crisis of 2007-2008 isa gigantic drama from which, one can only hope, the industrywill learn. The global wave of financial deregulation thatallowed unparalleled growth in the use of complex derivativesmay produce even more spectacular failures since quantitativerisk models disregard rare events and try to model whatarguably cannot be modeled. In every instance, the solutionof these problems starts with the recognition that people arehuman. What is required is “financial ergonomics,” adiscipline that engineers financial products and servicesaccording to human needs and that optimizes well-being andoverall system performance. Behavioral finance holds thepotential to create much value for society but it also has agreat deal of work to do.

VII. Conclusion

Over the last few decades, our understanding of financehas increased a great deal, yet there are countless questions

begging for answers. On the whole, financial decision makingprocesses in households, markets and organizations remaina grey area waiting for behavioral researchers to shed lighton it.

A major paradigm shift is underway. Chances are that “thenew paradigm” will combine neoclassical and behavioralelements. It will replace unrealistic, heroic assumptions aboutthe optimality of individual behavior with descriptive insightstested in laboratory experiments. Asset pricing theory, wehope, will combine a new realism in assumptions withmethods and techniques first developed in neoclassicalfinance. (Behavioral mean-variance portfolios may explainrisk premiums. A unified SDF framework may also providethe basis for behavioral explanations of option pricing, theterm structure of interest rates, and other asset prices.) Finally,and more broadly, history requires that economic and financialsystems are continually updated, and that they are intelligentlyreconstructed to meet social changes and to take advantageof technological progress. It is clear that, if academicians areto succeed in understanding financial institutions and actors,and if the agents themselves, as well as policy-makers, wantto make wise decisions, they must take into account the truenature of people, that is to say their imperfections andbounded rationality.

References

Aït-Sahalia, Y. and A. Lo, 2000, “Nonparametric RiskManagement and Implied Risk Aversion,” Journal ofEconometrics 94, 9-51.

Baker, M. and J. Wurgler, 2007, “Investor Sentiment in theStock Market,” Journal of Economic Perspectives 21 (No.2), 129-151.

Barber, B. and T. Odean, 2007, “All that Glitters: The Effectof Attention and News on the Buying Behavior ofIndividual and Institutional Investors,” Review of FinancialStudies, forthcoming.

Barberis, N.A. Shleifer, and R. Vishny, 1998, “A Model ofInvestor Sentiment,” Journal of Financial Economics 49(No. 3), 307-344.

Barberis, N. and R. Thaler, 2003, “A Survey of BehavioralFinance,” in Ed. by G.M. Constandinides, M. Harris andR. Stulz, Handbook of the Economics of Finance Elsevier.

Barberis, N. and W. Xiong, 2009, “What Drives theDisposition Effect? An Analysis of a Long-StandingPreference-Based Explanation,” Journal of Finance 64(No. 2), 751-784.

Barone-Adesi, G. and P. Talwar, 1983, “Market Models andHeteroscedasticity of Residual Security Returns,” Journalof Business and Economic Statistics 1 (No. 2), 163-168.

Barone-Adesi, G., P. Gagliardini and G. Urga, 2004, “TestingAsset Pricing Models with Coskewness,” Journal ofBusiness and Economic Statistics 22, 474-485.

22Behavioral finance also offers a framework for understanding financialmarket regulation. See Shefrin and Statman (1993b).


Barone-Adesi, G., R. Engle and L. Mancini, 2008, “AGARCH Option Pricing Model with Filtered HistoricalSimulation,” Review of Financial Studies 21(No. 3), 1223-1258.

Basu, S., 1977, “Investment Performance of Common Stocksin Relation to their Price-Earnings Ratios: A Test of theEfficient Market Hypothesis,” Journal of Finance 32 (No.3), 663–682.

Benartzi, S. and R. Thaler, 1995, “Myopic Loss Aversionand the Equity Premium Puzzle,” Quarterly Journal ofEconomics 110 (No. 1), 75-92.

Bernard, V., 1992, “Stock Price Reactions to EarningsAnnouncements,” In Advances in Behavioral Finance,Edited by R.H. Thaler, Russell Sage Foundation.

Bernard, V. and J.K. Thomas, 1989, “Post-earningsAnnouncement Drift: Delayed Price Response or RiskPremium?” Journal of Accounting Research 27(Supplement), 1-48.

Bernard, V. and J.K. Thomas, 1990, “Evidence that StockPrices do not Fully Reflect the Implications of CurrentEarnings for Future Earnings,” Journal of Accounting andEconomics 13, 305–340.

Bartov, E., I. Krinsky and S. Radhakrishnan, 2000, “InvestorSophistication and Patterns in Stock Returns after EarningsAnnouncements,” Accounting Review 75, 43–63.

Black, F., 1986, “Noise,” Journal of Finance 41 (No. 3),529-543.

Chen, H.-L. and W. De Bondt, 2004, “Style MomentumWithin the S&P 500 Index,” Journal of Empirical Finance11(No. 4), 483-507.

Cochrane, J., 2005, Asset Pricing, Princeton University Press.

Cornell, B., and Q. Liu, 2001, “The Parent Company Puzzle:When is the Whole Worth Less Than One of its Parts?”Journal of Corporate Finance 7, 341-366.

Cutler, D., J. Poterba, and L. Summers, 1991, “SpeculativeDynamics,” Review of Economic Studies 58 (No. 3), 529–546.

Daniel, K.D., D. Hirshleifer and A. Subrahmanyam, 1998,“Investor Psychology and Security Market Under- andOver-Reactions,” Journal of Finance 53 (No. 6), 1839-1886.

Daniel, K.D., D. Hirshleifer and A. Subrahmanyam, 2001,“Overconfidence, Arbitrage, and Equilibrium AssetPricing,” Journal of Finance 56 (No. 3), 921-965.

Daniel, K.D., D. Hirshleifer and S.H. Teoh, 2002, “InvestorsPsychology in Capital Markets: Evidence and PolicyImplications”, Journal of Monetary Economics 49 (No.1), 139-209.

De Bondt, W.F.M., 1993, “Betting on Trends: IntuitiveForecasts of Financial Risk and Return,” InternationalJournal of Forecasting 9 (No. 3), 355–371.

De Bondt, W.F.M., 2002, “Bubble Psychology,” in W. Hunterand G. Kaufman (eds.), Asset Price Bubbles: Implicationsfor Monetary, Regulatory, and International Policies, MITPress.

De Bondt, W.F.M., 2005, The Psychology of World EquityMarkets, Edward Elgar.

De Bondt, W.F.M., 2008a, “Stock Prices: Insights fromBehavioral Finance,” in Alan Lewis (ed.), The CambridgeHandbook of Psychology and Economic Behavior,Cambridge University Press.

De Bondt, W.F.M., 2008b, “Investor Psychology andInternational Equity Markets,” Revue Economique etSociale 66 (No. 4), 107-121.

De Bondt, W.F.M. and R. Thaler, 1985, “Does the StockMarket Overreact?” Journal of Finance 40 (No. 3), 793-808.

De Bondt, W.F.M. and R. Thaler, 1987, “Further Evidenceon Investor Overreaction and Stock Market Seasonality,”Journal of Finance 42 (No. 3), 557-581.

De Bondt, W.F.M. and R. Thaler, 1990, “Do Security AnalystsOverreact?” American Economic Review 80 (No.2), 52–57.

Dittmar, R., 2002, “Nonlinear Pricing Kernels, KurtosisPreference, and Evidence from the Cross Section of EquityReturns,” Journal of Finance 57 (No. 1), 369-403.

Dreman, D.N., 1977, Psychology and the Stock Market,Amacom.

Dreman, D.N., 1980, The New Contrarian InvestmentStrategy, Random House.

Dreman, D.N., 1995, “Exploiting Behavioral Finance:Portfolio Strategy and Construction,” In BehavioralFinance and Decision Theory in Investment Management,Edited by Arnold S. Wood, AIMR.

Fama, E.F., 1965, “The Behavior of Stock Market Prices,”The Journal of Business 38 (No. 1), 34–106.


Fama, E.F., 1998, “Market Efficiency, Long-Term Returns,and Behavioral Finance,” Journal of Financial Economics49 (No. 3), 283–306.

Fama, E.F. and K. French, 1992, “The Cross-Section ofExpected Stock Returns,” Journal of Finance 47 (No. 2),427–465.

Fellner, W., 1961, “Distortion of Subjective Probabilities asa Reaction to Uncertainty,” Quarterly Journal ofEconomics 75, 670-690.

Frazzini, A., 2006, “The Disposition Effect and Underreactionto News,” Journal of Finance 41 (No. 6), 2017-2046

Friedman, M., 1953, “The Case for Flexible Exchange Rates,”Essays in Positive Economics, University of Chicago Press.

Froot, K. and E. Dabora, 1999, “How are Stock PricesAffected by the Location of Trade?” Journal of FinancialEconomics 53 (No. 2), 189-216.

Graham, B., 1949, The Intelligent Investor: A Book ofPractical Counsel, Harper.

Grinblatt, M., and T. Moskowitz, 1999, “The Cross-Sectionof Expected Returns and its Relation to Past Returns: NewEvidence,” UCLA Anderson School, Working Paper No.24-99.

Grinblatt, M., S. Titman and R. Wermers, 1995, “MomentumInvestment Strategies, Portfolio Performance and Herding:A Study of Mutual Funds Behavior,” The AmericanEconomic Review 85, 1088-1105.

Han, B. and M. Grinblatt,. 2005, “Prospect Theory, MentalAccounting and Momentum,” Journal of FinancialEconomics 78 (No. 2), 311-333.

Haigh, M.S. and J.A. List, 2005, “Do Professional TradersExhibit Myopic Loss Aversion? An ExperimentalAnalysis,” Journal of Finance 60 (No. 1), 523-534.

Harvey, C. and A. Siddique, 2000, “Conditional Skewness inAsset Pricing Tests,” Journal of Finance 55 (No. 3), 1263-1295.

Henderson, P. and P. Peterson, 1992, “Mental Accountingand Categorization,” Organizational Behavior and HumanDecision Processes 51 (No. 1), 92-117.

Hens, T. and M. Vlcek, 2005, “Does Prospect Theory Explainthe Disposition Effect?” University of Zurich workingpaper.

Keim, D., 1983, “Size-Related Anomalies and Stock ReturnSeasonality: Further Empirical Evidence,” Journal ofFinancial Economics 12, 13-32.

Keim, D., 1986, “Dividend Yields and the January Effect,”Journal of Portfolio Management 12 (Winter), 54-60.

Keynes, M., 1964, “The General Theory of Unemployment,Interest and Money,” Harcourt Brace Jovanovich, reprintof the 1936 edition.

Koening, J., 1999, “Behavioral Finance: Examining ThoughtProcesses for Better Investing,” Trust and Investments 69,17-23.

Lamont, O.A. and R.H. Thaler, 2003, “Can the Market Addand Subtract? Mispricing in Tech Stock Carve-outs,”Journal of Political Economy 111 (No. 2), 227-268.

Lease, R.C., W.G. Lewellen, and G.G. Schlarbaum, 1976,“Market Segmentation,” Financial Analysts Journal 32,53-60.

Lee, C.M. and B. Swaminathan, 2000, “Price Momentumand Trading Volume,” Journal of Finance 55 (No. 5),2017–2069.

LeRoy, S. and R.D. Porter, 1981, “Stock Price Volatility: ATest Based on Implied Variance Bounds,” Econometrica49, 97-113.

Lohr, S., 2008, “In Modeling Risk, the Human Factor WasLeft Out,” New York Times, November 5.

Ikenberry, D., J. Lakonishok, and T. Vermaelen, 1995,“Market Underreaction to Open Market ShareRepurchases,” Journal of Financial Economics 39 (No.2-3), 181–208.

Jegadeesh, N. and S. Titman, 1993, “Returns to BuyingWinners and Selling Losers: Implications for Stock MarketEfficiency,” Journal of Finance 48 (No. 1), 65-91.

Jegadeesh, N. and S. Titman, 2001, “Profitability ofMomentum Strategies: An Evaluation of AlternativeExplanations,” Journal of Finance 56 (No. 2), 699–720.

Johnson, E.J. and D. Goldstein, 2003, “Do Defaults SaveLives?” Science 302 (No. 5649), 1338-1339.

Kahneman, D. and A. Tversky, 1979, “Prospect Theory: AnAnalysis of Decision Making under Risk,” Econometrica47, 263-291.


Lopes, L., 1987, “Between Hope and Fear: The Psychologyof Risk,” Advances in Experimental Social Psychology 20,255-295.

March, J. and C. Heath, 1994, A Primer on Decision Making:How Decisions Happen, Free Press.

Michaely, R., R.H. Thaler, and K. Womack, 1995, “PriceReactions to Dividend Initiations and Omissions:Overreaction or Drift?” Journal of Finance 50 (No. 2),573–608.

Mitchell, M., T. Pulvino, and E. Stafford, 2002, “LimitedArbitrage in Equity Markets,” Journal. of Finance 57 (No.2), 551-584.

Muradoglu, G., 1989, “Factors Influencing Stock Demand inTurkey,” Unpublished thesis, Bogazici University, Istanbul,Turkey.

Muradoglu, G., 2000, “Turkish Stock Market: Anomalies andProfit Opportunities,” In Security Market Imperfectionsin Worldwide Equity Markets, Edited by D. Keim, and W.Ziemba, Cambridge University Press.

Muradoglu, G., 2002, “Portfolio Managers and Novices’Forecasts of Risk and Return: Are There PredictableForecast Errors?” Journal of Forecasting 21 (No. 6), 395-416.

Muradoglu, G,. A. Salih, and M.Mercan, 2005, “A BehavioralApproach To Efficient Portfolio Formation” The Journalof Behavioural Finance 6 (No. 4), 202-212.

Muradoglu, G. and D. Onkal, 1994, “An Exploratory Analysisof Portfolio Managers’ Probabilistic Forecasts of StockPrices,” Journal of Forecasting 13 (No. 7), 565–578.

Nicholson, F., 1968, “Price Ratios in Relation to InvestmentResults,” Financial Analysts Journal 24 (No. 1), 105-109.

Norman, D.A., 1993, Things That Make Us Smart, PerseusBooks.

Odean, T., 1998, “Are Investors Reluctant to Realize TheirLosses?”Journal of Finance 53 (No. 5), 1775-1798.

Odean, T., 1999, “Do Investors Trade Too Much?” AmericanEconomic Review 89 (No. 5), 1279-1298.

Poterba, J.M. and L.H. Summers, 1988, “Mean Reversion inStock Returns: Evidence and Implications,” Journal ofFinancial Economics 22, 27–59.

Reason, J., Human Error, Cambridge University Press,Cambridge, 1990.

Reinganum, M., 1983, “The Anomalous Stock MarketBehavior of Small Firms in January: Empirical Tests forTax-Loss Selling Effects,” Journal of Financial Economics12, 89-104.

Roll, R., 1983, “The Turn-of-the-Year Effect and the ReturnPremia of Small Firms,” Journal of Portfolio Management9 (Winter), 18-28.

Roll, R., 1984, “Orange Juice and Weather,” AmericanEconomic Review 74 (No. 5), 861–880.

Roll, R., 1988, “R2,” Journal of Finance 43 (No. 3), 541–566.

Rosenberg, J. and R. Engle, 2002, “Empirical PricingKernels,” Journal of Financial Economics 64 (No. 3), 341-372.

Rosenthal, L., and C. Young, 1990, “The SeeminglyAnomalous Price Behavior of Royal Dutch/Shell andUnilever N.V./PLC,” Journal of Financial Economics 26No. 1), 123–142.

Rouwenhorst, K.G., 1998, “International MomentumStrategies,” Journal of Finance 53 (No. 1), 267–284.

Rouwenhorst, K.G., 1999, “Local Return Factors andTurnover in Emerging Stock Markets,” Journal of Finance54 (No. 4), 1439–1464.

Scruggs, J.T., 2007, “Noise Trader Risk: Evidence from theSiamese twins,” Journal of Financial Markets 10 (No. 1),76-105.

Shefrin, H., 2001a, Behavioral Finance, Edward Elgar.

Shefrin, H., 2001b, “Behavioral Corporate Finance,” Journalof Applied Corporate Finance 14 (No. 3), 113-124.

Shefrin, H., 2002, Beyond Greed and Fear, Oxford UniversityPress.

Shefrin, H., 2005, Behavioral Corporate Finance, McGraw-Hill.

Shefrin, H., 2008a, A Behavioral Approach to Asset Pricing,Elsevier Academic Press (2nd Ed.).

Shefrin, H., 2008b, “Risk and Return in Behavioral SDF-Based Asset Pricing Models,” Journal of InvestmentManagement 6 (No. 3), 1-19.


Shefrin, H. and M. Statman, 1984, “Explaining InvestorPreference for Cash Dividends,” Journal of FinancialEconomics 13, 253-282.

Shefrin, H. and M. Statman, 1985, “The Disposition to SellWinners too Early and Ride Losers too Long: Theory andEvidence,” Journal of Finance 40 (No. 3), 777-790.

Shefrin, H. and M. Statman, 1993a, “Behavioral Aspects ofthe Design and Marketing of Financial Products,”Financial Management 22 (No. 2), 123-134.

Shefrin, H. and M. Statman, 1993b, “Ethics, Fairness andEfficiency in Financial Markets,” Financial AnalystsJournal 49 (No. 6), 21-29.

Shefrin, H. and M. Statman, 2000, “Behavioral PortfolioTheory,” Journal of Financial and Quantitative Analysis35 (No. 2), 127-151.

Shefrin, H. and R.H. Thaler, 1988, “The Behavioral Life-Cycle Hypothesis,” Economic Inquiry 26 (No. 4), 609-643.

Schill, M.J. and C. Zhou, 2001, “Pricing an EmergingIndustry: Evidence from Internet Subsidiary Carve-outs,”Financial Management 30 (No. 3), 5-34.

Shiller, R.J., 1981, “Do Stock Market Prices Move too Muchto be Justified by Subsequent Changes in Dividents?”American Economic Review 71, 421–436.

Shiller, R.J., 1993, Market Volatility, MIT Press.

Shleifer, A., 2000, Inefficient Markets, Oxford UniversityPress.

Siegel, J.J., 1998, Stocks for the Long Run, McGraw-Hill.

Slovic, P., 1972, “Psychological Study of Human Judgment:Implications for Investment Decision Making,” Journalof Finance 27 (No. 4), 779-799.

Smith, A., 1759, The Theory of Moral Sentiments, LibertyFund (published in 1982).

Statman, M., 2008, “The Expressive Nature of SociallyResponsible Investors,” Journal of Financial Planning,(February).

Taleb, N., 2007, The Black Swan: The Impact of the HighlyImprobable, New York, Random House.

Thaler, R., 1993, Advances in Behavioral Finance, RussellSage Foundation.

Thaler, R. and H. Shefrin, 1981, “An Economic Theory ofSelf Control,” Journal of Political Economy 89 (No. 2),392-410.

Thaler, R.H. and C.R. Sunstein, 2008, Nudge: ImprovingDecisions about Health, Wealth, and Happiness, YaleUniversity Press.

Tobin, J., 1958, “Liquidity Preference as Behavior TowardsRisk,” Review of Economic Studies 25 (No. 67), 65-86.

Tversky, A. and D. Kahneman, 1974, “Judgment underUncertainty: Heuristics and Biases,” Science 185 (No.4157), 1124-1131.

Tversky, A. and D. Kahneman, 1992, “Advances in ProspectTheory: Cumulative Representation of Uncertainty,”Journal of Risk and Uncertainty 5 (No. 4), 297-323.


The Effects of Institutional RiskControl on Trader Behavior

Ryan Garvey and Fei Wu

We examine how institutional risk control mechanismsinfluence proprietary stock trader behavior. When tradersare forced to liquidate their inventory at a pre-designated time,they often hold onto their losing trades until the very lastmoment. We find that the difference between losing andwinning round-trip holding times systematically widensleading up to an inventory liquidation deadline and tradingbecomes less driven by trading practices and more inducedby the firm’s control mechanism as the deadline draws near.When trade price is heavily controlled yet trade size isn’t, wefind that the difference between losing and winning round-trip holding times systematically widens with trade size. Thisresult suggests traders increase their risk-taking in areas whereinstitutional control mechanisms are weaker. Our findingshighlight the difficult balancing act firms face with gettingmarket professionals to realize their losses without impedingtheir trading strategies.

Ryan Garvey is an Associate Professor of Finance at Duquesne Universityin Pittsburgh, PA. Fei Wu is a Senior lecturer in Finance at MasseyUniversity in Palmerston North, New Zealand.

A considerable amount of research has uncoveredbehavioral biases among financial market participants.1 Inorder to circumvent these biases and help employees make

better decisions, financial institutions implement risk controlmechanisms with their employees who make trade decisionswith institutional capital. The contributions of this study are:1) to examine how effective control mechanisms are atmitigating psychological trading biases, and 2) to examinehow employees respond to control mechanisms.

While our results are useful for firms implementing orplanning to implement risk control mechanisms, our resultsalso provide a step forward for the academic literature. Muchof the academic literature has been devoted to uncoveringbehavioral biases in various settings and among different typesof market participants. Uncovering these biases is, of course,a necessary first step. Yet, some trading biases are well knownand firms have been grappling with ways to control them fordecades. Despite this, there is very little research thatexamines trader behavior in settings where preventiontechniques are actively implemented by firms to get theiremployees to recognize and refrain from biases in theirdecisions.2 In our paper, we examine such a setting.

We analyze how proprietary stock traders, who work onbehalf of a National Securities Dealer, react to institutionalcontrol mechanisms that are primarily intended to get themto realize their trading losses. It is well known that marketprofessionals often have difficulty coming to terms with theirlosses. Consequently, they have a tendency to hold their losingtrades too long because they want to recover from their losses.This desire to get even is quite persuasive in financial marketsettings, and it is inevitably ingrained in many of the everydaydecisions that traders make. Indeed, some of the greatest

1See, for example, Odean (1998), Grinblatt and Keloharju (2001), andCoval and Shumway (2005).

2Statman and Caldwell (1987) discuss risk control and behavioral biasesin the context of capital budgeting decisions.

We would like to thank the Editor, Ramesh Rao, and an anonymous refereefor helpful comments and suggestions on a prior draft. We are also gratefulto executives at the US Securities Firm for providing proprietary data.

22

23GARVEY & WU — THE EFFECTS OF INSTITUTIONAL RISK CONTROL ON TRADER BEHAVIOR

trading losses of all time have occurred because traders weresimply unwilling to take a loss, so they gambled in an attemptto recover from their loss.3

In order to help traders come to terms with their losses,our sample firm implemented several control mechanisms.The most binding of these control mechanisms was that theyrequired traders to liquidatetheir inventory by the end ofthe trading day in order toensure loss realization. Thefirm implemented othercontrol mechanisms includingan emphasis on price control.They implemented trainingsessions that often stressed thedangers of holding losses toolong. For example, the firmcites in their training manual that a trader’s inability to take aloss is the number one reason why traders fail. And theyemployed a trading manager who closely monitored tradingactivity throughout the day. The firm even hired an on-sitepsychologist who was readily available to meet with traders.4

Despite all of these control measures, traders still appearto have difficulty coming to terms with their losses. We findthat, on average, traders hold their losing trades significantlylonger than their winning trades, which is consistent with thebehavior underlying the disposition effect (see Shefrin andStatman, 1985). These longer holding times coincide withlower performance. While prior studies document thatprofessional traders have a tendency to hold their losing tradeslonger than their winning trades (see, for example, Lockeand Mann (2005) and Garvey and Murphy (2004)), theprofessional traders observed in prior studies were notrequired to close out of all of their positions by apredetermined time.5 In our setting, traders are forced to exit

their positions by a predetermined time (i.e. the end of thetrading day) and our main focus is on how traders holdingtimes and performance vary across the day leading up to theinventory liquidation deadline.

We find that the difference between losing and winninground-trip holding times systematically rises throughout the

day and that it rises to itshighest level just prior to theinventory liquidation deadline.Moreover, trading performancesignificantly declines as theliquidation deadline drawsnear. Our sample traders oftenhave difficulty realizing certainlosses and they have a tendencyto hold onto them until the verylast moment. While the firm’s

efforts do not statistically eliminate a trader’s tendency tohold losing trades longer than winning trades, they clearlydo have an influence on trader behavior.

Inventory liquidation requirements ensure losses getrealized, but firms (traders) also rely on price controlmechanisms to do the job. While trade price is often heavilycontrolled in institutional trading settings, trade size usuallyis not. Institutional market participants trade in large tradesizes and their ability to execute these large trade sizes intheir entirety is often driven by market conditions. Thus,professional traders need flexibility with respect to trade size.While we find that traders adhere to a highly disciplinedapproach with respect to their exit prices, trade sizeconsiderably varies and traders let their losses run longer onlarger size trades. Consequently, they are more unprofitablewhen they trade in larger trade sizes.

These traders were constantly being drilled on the dangersof holding losses too long, but they tended to hold onto theirlosses despite the warning. Their tendency to hold losses forlonger periods of time resulted in lower performance. If thefirm did not require traders to close out of their positions bythe end of the day, or if they did not implement any controlmeasures, presumably losses would be held for even longerperiods of time and performance would be a lot worse. Thisis why it is so important for financial firms to implementcontrol mechanisms.

Our results highlight the complexities involved withimplementing optimal risk control mechanisms to circumventtraders’ aversion to realizing losses. If firms implementcontrol mechanisms that are too stringent, they are likely to

We analyze how proprietary stocktraders, who work on behalf of aNational Securities Dealer, react toinstitutional control mechanismsthat are primarily intended to getthem to realize their trading losses.

3One of the more notable cases, or largest losses resulting from this behavior,occurred with Nicholas Leeson. Mr. Leeson incurred over $1.4 billion intrading losses in 1995, which led to the demise of his employer, 232 year-old Barings PLC.4We sat in on the firms training sessions for traders, reviewed tradermanuals, had several discussions with management and traders, andobserved traders trade so that we could better prepare this paper.

5For example, Garvey and Murphy (2004) examine proprietary stock traderswho mainly offset their positions intraday, but the traders can and do holdpositions overnight. Locke and Mann (2005) assume commodity tradersend each day flat when determining trader gains and losses. However, theauthors do not report, or are not aware of, any mandatory time period inwhich traders are required to liquidate their inventory.


conflict with traders’ overall strategies and objectives because,as our research shows, trader behavior is heavily influencedby control mechanisms. On the other hand, if firms do nothingthey open themselves up for considerable risks. Thus, ourfindings imply that firms do need to implement controlmechanisms, but they need to be very careful with how theyenforce this because traders align their strategies with controlmechanisms. Because psychological biases are so intertwinedin many people’s decision-making processes, firms areunlikely to eliminate these biases, but they can lessen thedamage caused by them.

I. Related Research

Kahneman and Tversky’s (1979) prospect theory providesa descriptive framework for decision-making under risk. Acentral theme in their research is the role of loss. Much of thebehavioral finance literature focuses around how marketparticipants make decisions when they are confronted withthe prospect of a loss. Shefrin and Statman (1985) were thefirst to apply prospect theory to a financial market setting.They also placed prospect theory in a wider theoreticalframework that includes mental accounting, regret aversion,and self-control. These factors together help explaintheoretically why traders have a tendency to hold their losingtrades much longer than their winning trades, a behavior thatis commonly known as the “disposition effect”.

Traders, who exhibit behavior that is consistent with thedisposition effect, think about stock purchases within separatemental accounts (see Thaler, 1985) then apply prospect theorydecision rules to each mental account. The disposition effecthas proven to be quite pervasive in US markets.6 For example,research shows that individual investors (e.g., Barber andOdean (1999) and Odean (1998)), mutual fund managers(e.g., Frazzini (2006) and Scherbina and Jin (2005)), andprofessional traders (e.g., Garvey and Murphy (2004) andHeisler (1994)) all exhibit signs of this behavior. Some ofthe more recent studies identify individual tradercharacteristics that are correlated with the disposition effect.For example, Dhar and Zhu (2007) find that individuals’income level, occupational status, etc. are important factorsin indicating who is more susceptible to the disposition effect.

While most researchers examine traders’ unwillingness totake a loss through holding times and focus on decision-making in a single period setting, some other studies havelooked elsewhere. They examine a trader’s reluctance torealize losses using other risk measures and focus on decision-making in a multi-period setting. For example, Coval andShumway (2005) and Garvey et al. (2007) find traders whohave experienced prior morning losses engage in subsequentafternoon risk-taking as measured through increased tradingactivity, larger size trading, etc. These findings are consistentwith the same behavioral tendency that leads traders to holdtheir losing trades too long.

Like much of the prior research, we examine traderresistance to loss realization through holding-time decisions,and we examine trader holding-time decisions in a single-period setting. Our motivation is not to examine if traderssuffer from the get even behavior that underlies the dispositioneffect, but rather our motivation is to examine how (if) tradersrespond to institutional control mechanisms to prevent thisbehavior.

II. Data

Our data originates from a National Securities Dealer. Thefirm had several trading operations and our focus is on thefirm’s proprietary trading operation for US equities. The datacovers June 3, 2002 through May 30, 2003. During this one-year period, the US stock markets were open for 251 days. Intotal, the 150 traders combined to execute 2.5 billion sharesthrough 1.3 million transactions on 693 securities. For everytransaction, the data reveals the identity of the trader, theexecution time, the type of trade (marketable versus limitorder), the action taken (buy, sell, short, and cover), thevolume, the price, the market where the order was sent, thecontra party on the trade (if given), the location of the trader(the traders were located in five branch offices), and variousother trade execution information.

Each trader’s sole objective was to generate intraday tradingprofits utilizing firm capital. Consequently, the traders tradeoften and they also trade in large trade sizes. The averagetrader executes 75 per day and the average executed tradesize is for 1,925 shares. This average trade size is more thanthree times the average trade size in US equity markets.7

Trading activity is concentrated in certain stocks (mostly6Researchers have also found strong evidence of the disposition effect inFinland (e.g. Grinblatt and Keloharju, 2001), Israel (e.g. Shapira andVenezia, 2001), China (e.g. Feng and Seasholes, 2005), Taiwan (e.g. Barberet al., 2007) and other countries.

7The average trade size for NYSE (Nasdaq) stocks was 488 (579) sharesin 2003 (Source: NYSE and Nasdaq data).


Nasdaq) on certain days, often accounting for a sizeableportion of a particular stock’s trading volume.8 While thetraders trade often, they also set market prices often.Approximately 65% of their trades provide liquidity, while35% of their trades take liquidity. The traders set marketprices in the various trading venues which US equities tradein.9

To get an idea of thetrading intensity differencebetween institutional andretail market participants,consider a sample of retailbrokerage accountsstudied by Barber andOdean (2000).10 Theyanalyze the tradingbehavior and performanceof retail marketparticipants, who tradethrough a US discountbrokerage firm. In total,66,465 households execute 1,969,701 stock trades over a six-year period ending in December, 1996. Our 150 tradersexecute 1,316,334 stock trades over just a one-year period.Institutional market participants dominate the tradinglandscape in US equity markets, yet much of the academicliterature examining trader decision-making in equity marketsis focused on retail market participants. While some studiesdo examine institutional trading (see, more recently, Conrad,Johnson, and Wahal, 2002), this literature largely focuses ontransaction costs, and prior studies do not examine howinstitutional traders respond to risk control mechanisms.

Our sample traders received continual training on varioustrading strategies from the firms’ management, yet they hadconsiderable freedom with selecting stocks to trade. Thecompensation of the traders is solely tied to their tradingperformance. Thus, the traders had a clear incentive tomaximize their trading performance. For our purposes, the

most intriguing aspect of this particular setting is the firm’sefforts to get traders to realize their trading losses. The firmis not alone with respect to their efforts in this regard. Gettingtraders to take losses is a common problem that securitiesfirms and their risk managers constantly grapple with.

III. Empirical Results

A. Methodology

The objective of our studyis to directly examine howtraders react to institutionalcontrol mechanisms that areprimarily intended to getthem to accept their losses.Specifically, we measure theeffects of institutional controlmechanisms on traders’holding-time decisions.Round-trip performance and

holding-time measures are not included in the raw transactiondata. In order to determine the gains and losses for eachround-trip, we use an intraday round-trip matching proceduresimilar to the one used in Garvey and Murphy (2005). Wepair off opening trades with the subsequent closing trade(s)in the same day. The traders did not always open and closepositions with two trades. A trader could combine a closingtransaction with an opening transaction, or they could lay offpart of an open position. Regardless of whether tradesopened, closed, or open and closed a position simultaneously,we searched forward in time each day until the openingposition was closed out, and we kept track of execution times,accumulated inventory, and corresponding prices paid orreceived. The matching procedure creates 730,417 round-trip trades from the 1.3 million trades. We then calculate outthe corresponding holding time for each round-trip. In orderto do this, we calculate the holding time between the intradayopening and closing transaction(s). If a trader accumulatesinventory before they eventually close out of their position,we use a weighted average between the various openingpositions.11

8The only stocks traded every day were Sun Microsystems (SUNW) andJDS Uniphase (JDSU). The traders accounted for 1.5% and 3.3% of theannual share volume of SUNW and JDSU respectively. SUNW and JDSUare two of the more actively traded US stocks.9The traders often traded on The Island ECN, which reported its tradesthrough the Cincinnati Stock Exchange. See Nguyen, Van Ness, and VanNess (2004) for a discussion on the reporting of Island trades on theCincinnati Stock Exchange.10The data has been used in several studies in finance literature.

11For example, suppose a trader opens up a 2,000 share position of Yahooat 10:30:00 a.m. and then purchases another 4,000 shares of Yahoo at10:30:10 a.m. If the next trade were a sell of 6,000 shares of Yahoo at10:30:20 a.m., the holding time on the round trip trade is 13.33 seconds(1/3 * 20 seconds + 2/3 * 10 seconds).

Traders hold losses longer than gains,but these holding time patterns do notremain constant throughout the day.The difference in holding timesbetween losing and winning round-trips systematically rises throughoutthe day and it dramatically increasesin the moments just prior to the firm’smandatory close-out period.


From the matching procedure, there are 290,248 winninground-trips (gross round-trip trading profits above $0),209,271 losing round-trips (gross round-trip trading profitsbelow $0), and 230,898 break-even round-trips (gross round-trip trading profits equal to zero). The frequency of break-even round-trips highlights how focused these traders are ontheir trade purchase price (the reference point). The tradersdo not hold their open positions for long. And when theyenter into a position on one side of the market, they generallyseek to quickly offset their position by trading on the oppositeside of the market. For example, the mean holding time perround-trip is 780 seconds and the median holding time perround-trip is 205 seconds.

The sample period we observe was a difficult time to tradeUS equities. Many securities firms were reporting steep losseson their equity trading desks.12 Our firm was not immune tothese difficult trading conditions, yet the traders didexperience many profitable (and unprofitable) round-tripswhich enable us to examine their holding time decisions inboth the domain of gain and loss.

B. Do Institutional Risk Control MechanismsEliminate Behavioral Biases?

Table I provides information on the overall holding timedifference between winning and losing round-trip trades.Despite the firm’s efforts to get traders to realize their lossesin a timely fashion, traders hold their losing round-trip tradesconsiderably longer than their winning round-trip trades (notethat the firm’s control mechanisms were in place over ourentire sample period). On average, traders hold their losinground-trip trades for 1,274 seconds and their winning round-trip trades for 568 seconds. The difference of 706 seconds isstatistically different from zero at the 1% level. Themagnitude of the overall holding time difference is rathersurprising given the firm’s continual efforts to get traders torealize their trading losses. It would be interesting to seehow this result would change if the firm did not engage inany efforts to get traders to realize their losses. Presumably,traders would hold their losses for even longer periods oftime.

In order to check the robustness of our initial result, weexamine the holding time differences for each individualtrader. This allows us to see if certain traders are driving our

overall result. On a mean holding time basis, 145 out of 150traders hold their losing round-trip trades longer than theirwinning round-trip trades (135 differences are statisticallysignificant). On a median holding time basis, 146 out of 150traders hold their losing round-trip trades longer than theirwinning round-trip trades (139 differences are statisticallysignificant). The individual trader results coincide with theaggregate trader results.13

In our setting, holding positions for longer periods of timeis generally undesirable. The objective of the traders is torapidly enter and exit positions in order to profit from smallprice changes. The trader’s information is short-lived, sowhen traders’ open positions are held for extended periodsof time, it is a good indicator that the position has movedagainst the trader and they are primarily holding it in order torecover from the loss. In Figure 1, we segregate the round-trip profits into five holding time categories. Round-triptrading profits systematically decline with longer holdingtimes. When traders hold their trades under (over) fiveminutes they are profitable (unprofitable). Trader profits arehighest when they hold their trades for under one minute,and trader profits are lowest when they hold their trades formore than 15 minutes.14 These results highlight the trader’sshort-term strategies. They also show how important holdingtimes are for a trader’s success. A professional trader’sdecision to hold trades for slightly longer periods of timecould make the difference between being profitable orunprofitable on an overall basis.

C. How Does Inventory Control Influence TraderBehavior?

While the firm’s efforts do not appear to eliminate biasesfrom traders’ decisions, their risk control measures do havean influence on trader behavior. This influence is quite strongin the moments just prior to the inventory liquidation deadline,which highlights how deadline effects coincide with a trader’sresistance to realize losses. Table II and Figure 2 provideinformation on round-trip holding times across the day.Traders hold losses longer than gains, but these holding timepatterns do not remain constant throughout the day. Thedifference in holding times between losing and winning round-

12For example, the revenues of broker dealers with a Nasdaq market makingoperation fell over 70% during 2000-2004 (GAO, 2005). Our firm had aNasdaq market making operation.

13 The average round-trip holding time among traders ranges from 122 to3,176 seconds.

14 The average round-trip gain is $13.70 and the average round-trip loss is$19.23. The absolute trading profit difference is statistically significantfrom zero at the 1% level.


Table I. Round-Trip Holding Times

This table reports the mean and median holding times for winning and losing round-trip trades. Results are reported on an aggregatebasis and on an individual basis. The results are based on the trading records of 150 proprietary stock traders, who traded the capital ofa National Securities Dealer from May 2002 through June 2003. The traders executed 1.3 million trades which resulted in 730,417round-trip trades over the 251 day sample period. The traders only traded US equities. Winning (losing) round-trips have a grosstrading profit above (below) zero and break-even round-trips are omitted. Holding times are calculated in seconds.

Figure 1. Performance and Holding Times

This figure displays the average round-trip trading profit for five holding time categories. The mean (median) holding timeis 780 (205) seconds. The results are based on the trading records of 150 proprietary stock traders who traded the capital ofa National Securities Dealer from May 2002 through June 2003. The traders executed 1.3 million trades which resulted in730,417 round-trip trades over the 251 day sample period.

Panel A. Aggregate Results

Mean Holding Time Median Holding Time

Losing round-trips 1,274 377

Winning round-trips 568 166

Difference 706* 141*

Panel B. Individual Results

Mean Median

Number of traders who hold their losses longer than their gains 145 146

Number of traders who hold their losses longer than their gains 135* 139*

*Significant at the 0.10 level.

($6.00)

($4.00)

($2.00)

$0.00

$2.00

$4.00

$6.00

Less than

1 minute

1-2

minutes

2-5

minutes

5-15

minutes

More than

15

minutes

Average Round-trip Profit and Holding Times


Table II. Trading by Time of Day

This table reports traders’ trading activity and the number (percentage) of round-trips exceeding the overall mean holding time (13minutes) for each half-hour period. The results are based on the trading records of 150 proprietary stock traders who traded the capitalof a National Securities Dealer from May 2002 through June 2003. The traders executed 1.3 million trades which resulted in 730,417round-trip trades over the 251 day sample period. The 1,428 round-trip trades executed before the open (9:30 a.m.) or after the close(4:00 p.m.) are omitted.

Figure 2. Winning vs. Losing Round-trip Holding Times by Time of Day

This figure displays the average holding time for winning and losing round-trips throughout the trading day. Winning round-trips havea gross trading profit above zero while losing round-trips have a gross trading profit below zero. Holding times are calculated inseconds. The results are based on the trading records of 150 proprietary stock traders who traded the capital of a National SecuritiesDealer from May 2002 through June 2003. The traders executed 1.3 million trades which resulted in 730,417 round-trip trades over the251 day sample period.

Intraday Time Number of Round-Trips

Percentage of Round-Trips

Number of Round-Trips above Mean

Holding Time

Percentage of Round-Trips above Mean Holding Time

9:30-10:00 a.m. 62,267 8.5% 2,093 3.4%

10:00-10:30 a.m. 75,921 10.4% 9,680 12.8%

10:30-11:00 a.m. 66,816 9.1% 12,128 18.2%

11:00-11:30 a.m. 61,874 8.5% 13,415 21.7%

11:30-12:00 a.m. 56,787 7.8% 13,485 23.7%

12:00-12:30 p.m. 50,537 6.9% 13,027 25.8%

12:30-1:00 p.m. 46,848 6.4% 13,094 28.0%

1:00-1:30 p.m. 47,207 6.5% 12,777 27.1%

1:30-2:00 p.m. 49,055 6.7% 12,390 25.3%

2:00-2:30 p.m. 51,597 7.1% 12,604 24.4%

2:30-3:00 p.m. 53,419 7.3% 13,124 24.6%

3:00-3:30 p.m. 53,107 7.3% 13,051 24.6%

3:30-4:00 p.m. 53,554 7.3% 16,625 31.0%

Average Winning vs. Losing Holding Times: Time

of Day

0

500

1000

1500

2000

2500

3000

3500

9:30-1

0:00 a

.m.

10:00-10:3

0 a.m.

10:30-11:0

0 a.m.

11:00-11:3

0 a.m.

11:30-12:0

0 a.m.

12:00-12:3

0 p.m.

12:30-1:0

0 p.m

.

1:00-1

:30 p

.m.

1:30-2

:00 p

.m.

2:00-2

:30 p

.m.

2:30-3

:00 p

.m.

3:00-3

:30 p

.m.

3:30-4

:00 p

.m.

Sec

onds

Winning round-trips Losing round-trips


trips systematically rises throughout the day and itdramatically increases in the moments just prior to the firm’smandatory close-out period. The results indicate that tradersoften hold losses up until the very last moment before theyare forced to realize them. The trader’s behavior is not adesirable reaction to what the firm is trying to accomplish,but the liquidation requirement appears necessary. Withoutthis control mechanism in place, these traders’ reluctance totake losses would have most likely resulted in larger tradinglosses.

In Table III, we report the gross round-trip trading profitsfor each half hour of the trading day. Trading profits arehighest in the initial opening period, or the period which isfurthest away from the liquidation deadline, and statisticallydifferent from zero at the 1% level. Trading profits steadilydecline until noon. Trading profits are positive after 12:00p.m., but they steadily decline again until the close of trading.The sharp decline in traders’ profits in the moments just beforethe close, along with the holding time patterns, indicates thattrading is significantly driven by the firm’s controlmechanism. The average round-trip profit in the last 30minutes is -$2.56, whereas the average round-trip profit atother intraday times is $0.18. The average round-trip tradingprofit in the closing 30 minutes is 1,522% less than the averageround-trip trading profit at other intraday times. The end-of-day trading losses can be broken down further. While thetraders lost $136,934 in the final 30-minute period,approximately 63% of this occurred in the final 5 minutes ofthe trading day (note that they lost money in each 5 minuteinterval in the final 30 minute period). The magnitude of thelosses that occurred in each five minute interval is highlightedin Figure 3.

Trading profits most likely continue to decline until noonbecause traders often break for lunch. When traders leavetheir trading terminals, they usually close out of theirpositions. For many traders, the lunch period serves asanother period for realizing losses. However, the middayclose-out period is not binding like the end-of-the-day periodis, and losses are generally held for shorter periods of timeleading up to the midday period. This is why losses are farmore pronounced at the close of the day than they are at themiddle of the day.

While the end of day closeout period induces trade andforces traders to realize their losses, it is interesting to theorizeabout what would occur if the firm did not have this controlmechanism in place. If the firm allowed traders to hold theirpositions overnight, would this give traders greater flexibility

with implementing their trading strategies and subsequentlyimprove performance? Or, would removing the liquidationdeadline result in traders holding their losses for significantlylonger periods of time resulting in catastrophic losses? Whileit is not possible to definitively answer these questions fromour available data, we compute a hypothetical performancemeasure assuming some trades were held overnight. Becausetrading in the very last moments of the day appears highlydriven by the firm’s control mechanism, we recalculate tradingprofits for positions closed out in the final 15 minutes of thetrading day. Contrary to using the round-trip trade price atthe end of the day to determine profits, we assume tradersclosed their positions at the opening price on the following(trading) day. In order to do this, we obtained opening pricedata on sample stocks traded from the Center for Research inSecurity Prices (CRSP) database and recalculated tradingprofits for positions closed out in the final minutes of theday. There are 24,332 round-trip observations, and theaverage closing position is for 1,674 shares. Under theadjusted closing price, the average round-trip trading profitdeclines from -$3.95 to -$4.17 (note that a few anomalousobservations are dropped from the analysis). While the firm’sliquidation requirement imposes a constraint on traders, mostof the losing positions traders were forced to realize at theend of the day continued to decline in value into the next dayof trading.

Our firm’s inventory liquidation requirement is not unique.Many securities firms require their market making,proprietary, arbitrage, and other types of traders to end theday flat, or they significantly restrict traders ability toaccumulate inventory from day to day.15 If traders at otherfirms exhibit behavioral tendencies similar to these traders,than our results provide insight into some factors that driveintraday order flow patterns. Researchers have long knownthat intraday trading activity in US equity markets exhibits aU-shaped pattern across the main trading hours (9:30 a.m. to4:00 p.m.). The Admati and Pfleiderer (1988) theory hasserved as a prominent explanation for these intraday volumepatterns. While our traders’ intraday trading activity closelyresembles a U-shape pattern, there are competing factors

15 There are many self-employed (retail) traders who adhere to this rule ona self-imposed basis. In the US, these retail day traders typically tradethrough direct access brokers. Bear Stearns finds that the more activetraders (25+ trades per day) who trade through direct access firms accountfor approximately 40% of Nasdaq/NYSE trading volume (Goldberg andLupercio, 2004). Thus, retail and institutional traders, who often end thetrading day flat, account for a very large percentage of overall daily tradingvolume in US equity markets.


Figure 3. Trading Losses in the Final 30 Minutes

This figure reports the percentage of the overall amount that was lost in the final 30 minutes of the trading day (see Table 3) across eachfive-minute category. The results are based on the trading records of 150 proprietary stock traders who traded the capital of a NationalSecurities Dealer from May 2002 through June 2003. The traders executed 1.3 million trades which resulted in 730,417 round-triptrades over the 251 day sample period.

Percentage of Losses in Closing Minutes

0%

10%

20%

30%

40%

50%

60%

70%

3:30-3:3

5 p.m

.

3:35-3:4

0 p.m

.

3:40-3:4

5 p.m

.

3:45-3:5

0 p.m

.

3:50-3:5

5 p.m

.

3:55-4:0

0 p.m

.

Table III. Performance by Time of Day

This table reports trading performance, average round-trip trade size and average round-trip holding time for each half-hour period. Theresults are based on the trading records of 150 proprietary stock traders who traded the capital of a National Securities Dealer from May2002 through June 2003. The traders executed 1.3 million trades which resulted in 730,417 round-trip trades over the 251 day sampleperiod. The 1,428 round-trip trades executed before the open (9:30 a.m.) or after the close (4:00 p.m.) are omitted.

Intraday Time Total Trading Profits

Average Round-Trip Trade Size

Average Round-Trip Trading

Profit

Average Holding Time per Round-Trip

9:30-10:00 a.m. $93,730 1,787 $1.51*** 187

10:00-10:30 a.m. $34,227 1,747 $0.45*** 341

10:30-11:00 a.m. $13,245 1,763 $0.20* 475

11:00-11:30 a.m. $12,488 1,771 $0.20* 591

11:30-12:00 a.m. -$11,718 1,782 -$0.21* 695

12:00-12:30 p.m. $22,187 1,757 $0.44*** 805

12:30-1:00 p.m. $12,656 1,740 $0.27*** 922

1:00-1:30 p.m. $2,846 1,749 $0.06 982

1:30-2:00 p.m. $5,004 1,713 $0.10 906

2:00-2:30 p.m. -$11,416 1,655 -$0.22* 924

2:30-3:00 p.m. -$19,853 1,674 -$0.37*** 973

3:00-3:30 p.m. -$32,833 1,698 -$0.62*** 1,084

3:30-4:00 p.m. -$136,934 1,717 -$2.56*** 1,645

***Significant at the 0.01 level.

*Significant at the 0.10 level.


17 Recently adopted Regulation National Market System (Reg. NMS)eliminated sub-penny pricing for securities priced above $0.99.

inducing this pattern across the day. Trading at the open, onaverage, appears to be motivated by short-term information(i.e. gross round-trip trading profits are statistically differentfrom zero). Trading activity rises in the second half of theday, but this time the rise in trading activity corresponds witha decrease in performance. The rise in trading activity in thesecond half of the day, on average, seems driven more by thefirm’s risk control mechanism rather than their traders’ normaltrading practices.

While our findings might beuseful for providing insight intointraday order flow patterns,they are also potentially usefulin understanding why tradingactivity levels may predictablyrise or decline on certain days,at certain times of the year, orin response to certain situations.Our results provide directevidence on how trader behavior varies within a set tradingtime horizon and when a deadline exists. We would expectsimilar loss-averse behavioral patterns to occur over varioustime horizon settings in which there is some type of deadlinebeing imposed or self-imposed on a decision-maker (e.g., afund manager, trader, retail investor, etc.), such as with aperformance evaluation period, compliance or audit period,a government tax period, a maintenance margin level, etc.Economics literature has found evidence of this “wait untilthe very last moment” approach in other settings (e.g., withbargaining negotiations), and our results provide someevidence on how psychological trading biases and deadlineeffects interact in a financial market setting. There has beenvery little direct research on this front because researchersstudying trader behavior in financial market settings oftenlack data on traders’ time horizons or the time horizon set bythe employee’s institution.16 Thus, it is difficult to measurehow trader behavior varies over a trading time horizon if thetrading time horizon is not truly known.

D. How Does Price Control Influence TraderBehavior?

While the end-of-day inventory liquidation requirement isthe most binding control mechanism the firm has in place,

the firm trained the traders to adhere to a disciplined exitstrategy (e.g., use of stop loss mechanisms) to ensure lossrealization. Stop loss mechanisms can be employed eitherexplicitly (attached with the opening order) or through a self-imposed rule. The firm attempted to monitor trader exit pricesthrough the trading manager and the traders uniformly exitedmost, but not all, of their positions within a very tight pricingrange. In Table IV, we report the distribution of round-trip

price changes for both winningand losing round-trips. Themedian price change for bothwinning and losing round-trips isone cent.17

The traders prefer trading inlarger trade sizes in order tomaximize their trading profits,but they are often forced to tradein smaller trade sizes due tofactors beyond their control. For

example, suppose a trader submits a 5,000 share limit orderat the underlying best bid price. If an incoming 1,000 shareorder executes against the trader’s order, but then the marketprice moves sharply away from the trader’s bid price, thetrader is left with 80% of the original order unfilled and willbe forced to reassess execution strategy.

While trade sizes vary with underlying market conditions,traders (firms) do not usually reset price control mechanismsin accordance with trade size (i.e. on a percentage basis). Ininstitutional trading settings such as ours, the emphasis is ondisciplined trading and adhering to a specific and well definedtrading strategy. Our traders are trained to enter and exittheir positions within a very tight price range and theytypically offset their positions within one or two cents. Ifthese traders were to constantly reset their exit prices on atrade size percentage basis, this would create a much lessdisciplined approach to trading, and it would give themincentives to deviate from their normal trading practices. Thetraders are not trained to capture large price changes. Instead,they are trained to capture small price changes while tradingfrequently on both sides of the market.

When price is heavily controlled and traders are givengreater leeway with respect to trade size, this leaves firmsvulnerable to heightened risk-taking with larger size trades.The existing stock price in relation to the opening stock price

16 For example, Benartzi and Thaler (1995) assume that the averageinvestors holding time period is one year. This assumption has been appliedin other research settings (e.g. Odean, 1998).

While losing round-trips areheld considerably longer thanwinning round-trips for eachtrade size category, thedifference systematically widenswith trade size.


determines whether a trade is for a capital gain or loss, buttrade size, along with trade price, determines the magnitudeof a trading gain or loss. When traders enter into larger tradesand the price moves against them, the magnitude of theirtrading losses will increase and according to decision-makingtheory, they will have an increasing desire to get even.

We segregate round-trip trade sizes into five trade sizecategories: 1) trade sizes less than 250 shares, 2) trade sizesgreater than 249 shares and less than 1,000 shares, 3) tradesizes greater than 999 shares and less than 2,000 shares, 4)trade sizes greater than 1,999 shares and less than 3,000shares, and 5) trade sizes greater than 2,999 shares. Theoverall holding time results for each trade size category, andfor gains and losses, are reported in Figure 4A. While losinground-trips are held considerably longer than winning round-trips for each trade size category, the difference systematicallywidens with trade size. We suspect this pattern is the resultof traders moving deeper into the red with their larger tradesizes. As losses surmount and traders move further awayfrom the break point, they will have an increasing desire togamble (hold trades longer) in order to get back to the breakeven point. Yet, control mechanisms are not in place or aremuch weaker to stop this undesirable behavior because the

emphasis is on disciplined trading with respect to uniformprice control.

In Panel A of Table V, we examine overall performanceand trade size. The absolute difference between the averagetrading gain and loss correspondingly widens with trade size.In general, we expect holding times to rise with larger tradesizes because it is more challenging to execute larger tradesthan smaller trades. However, this does not explain thewidening gap between losing and winning round-trips withrespect to trade size. Most of the trading losses can beattributed to trading in larger trade sizes (3,000 or moreshares), where the loss-gain holding time difference is mostpronounced. This suggests that traders’ decision to ride theirlosses longer with larger trade sizes is costly.

We check the robustness of our trade size results bycontrolling differences in liquidity across the stocks traded.How liquid a stock is can affect both holding times and tradingprofits. We expect, on average, holding times to be lower onmore liquid stocks. And, on average, we expect the priceimpact (if any) incurred executing a trade to be smaller onmore liquid stocks. Variations in price impacts will bereflected in trading profits. In order to control liquiditydifferences across stocks, we sort the stocks traded into two

Table IV. Round-Trip Price Changes and Trade Size Distributions

This table reports the distribution for winning round-trip price changes, losing round-trip price changes, and executed trade sizes. Theresults are based on the trading records of 150 proprietary stock traders who traded the capital of a National Securities Dealer from June2002 through May 2003. The traders executed 1.3 million trades which resulted in 730,417 round-trip trades over the 251 day sampleperiod.

Price Change on Gains Price Change on Losses Trade Size

Mean $0.0126 $0.0173 1,925

Percentiles

10th $0.0020 $0.0010 100

20th $0.0060 $0.0020 300

30th $0.0080 $0.0084 500

40th $0.0100 $0.0100 900

50th $0.0100 $0.0100 1,000

60th $0.0100 $0.0100 1,600

70th $0.0100 $0.0200 2,000

80th $0.0130 $0.0200 3,000

90th $0.0200 $0.0400 5,000

Obs. 290,248 209,271 1,316,334


Figure 4. Winning vs. Losing Round-trip Holding Times by Trade Size

This figure displays the average holding time for winning and losing round-trips based on trade size (shares). Winning round-trips havea gross trading profit above zero while losing round-trips have a gross trading profit below zero. Holding times are calculated inseconds. The results are based on the trading records of 150 proprietary stock traders who traded the capital of a National SecuritiesDealer from May 2002 through June 2003. The traders executed 1.3 million trades on 693 securities which resulted in 730,417 round-trip trades over the 251 day sample period. In Fig. 4B and 4C, holding times results are reported separately for trades occurring on liquidstocks and illiquid stocks. Liquid (illiquid) stocks are stocks with a turnover ratio in the top (bottom) 50% percentile of our sample.Stock turnover ratios are calculated using CRSP by averaging daily stock turnover (volume/shares outstanding) over the sample period.

Fig. 4A Fig. 4B

Fig. 4C


Table V. Performance and Trade Size

This table reports round-trip performance results segregated by five trade size categories. The results are based on the trading records of150 proprietary stock traders who traded the capital of a National Securities Dealer from May 2002 through June 2003. The tradersexecuted 1.3 million trades on 693 securities which resulted in 730,417 round-trip trades over the 251 day sample period. In Panel B &C, performance results are reported separately for trades occurring on liquid stocks and illiquid stocks. Liquid (illiquid) stocks arestocks with a turnover ratio in the top (bottom) 50% percentile of our sample. Stock turnover ratios are calculated using CRSP byaveraging daily stock turnover (volume/shares outstanding) over the sample period.

groups (liquid vs. illiquid stocks) based on their average dailyturnover ratios (volume / shares outstanding) over our oneyear sample period. Volume and share outstanding data isobtained from the CRSP database. We are able to usematching trade data from CRSP for more than 97% of thetrading activity in our data.

We compute holding time differences and performancedifferences for both liquid stocks and illiquid stocks accordingto our trade size classifications. The holding time results arereported in Figures 4B and 4C and the performance resultsare reported in Panels B and C of Table V. As expected,most trading activity occurs on liquid stocks and holding times

are much lower on liquid stocks. For both liquid stocks andilliquid stocks, losing round-trips are held considerably longerthan winning round-trips for each trade size category and thedifference systematically widens with trade size. In general,the absolute difference between the average trading gain andloss correspondingly widens with trade size too. Althoughthere is a sharp drop in the performance difference for illiquidstocks under the largest trade size category, there are veryfew large trade size observations on illiquid stocks.

The trade size results highlight the need to assessinstitutional market participants’ resistance to loss realization(and its associate costs) at the individual trade level. Our

Panel A. All Stocks

Trade Size (Shares)

Number of Round-Trips

Avg. Round-Trip Gain

Avg. Round-Trip Loss

Absolute Round-Trip Profit Difference

Total Trading Profits

<250 154,469 $1.66 -$2.69 $1.03*** -$18,652.07

250,<1000 182,982 $7.62 -$11.01 $3.39*** $102,517.69

1000,<2000 154,496 $14.47 -$19.46 $4.99*** $18,803.97

2000,<3000 89,044 $23.05 -$28.40 $5.35*** $41,808.65

3000 149,426 $38.78 -$44.51 $5.73*** -$189,269.11

Panel B. Liquid Stocks

Trade Size (Shares)






<250 139,271 $1.54 -$2.50 $0.96*** -$14,747.42

250,<1000 163,969 $7.00 -$9.99 $3.00*** $92,034.00

1000,<2000 138,732 $14.22 -$18.93 $4.71*** $27,821.93

2000,<3000 79,803 $22.85 -$28.09 $5.25*** $45,967.47

3000 130,802 $39.05 -$45.04 $5.99*** -$129,172.52

Panel C. Illiquid Stocks

Trade Size (Shares)






<250 12,221 $2.22 -$3.69 $1.47*** -$3,771.24

250,<1000 14,684 $9.66 -$15.12 $5.46*** $3,057.33

1000,<2000 12,018 $16.65 -$24.46 $7.82*** $8,629.62

2000,<3000 6,613 $26.00 -$33.24 $7.25*** $2,453.53

3000 12,104 $36.26 -$39.13 $2.88** -$29,340.84

***Significant at the 0.01 level.

**Significant at the 0.05 level.


results also highlight the need to devise control mechanisms,among other things, on a situational trade basis. For example,a firm may analyze the trading decisions of its fund managersand find that overall, their fund managers do not exhibit atendency to avoid realizing their losses. Consequently, thefirm might feel less of a need to implement controlmechanisms to prevent traders from taking excessive riskswhen they are confronted with the prospect of a loss. Whilethe fund managers may not exhibit a tendency to avoidrealizing their losses on an overall basis, they may exhibit atendency to do so with their larger holdings, which wouldpose a significant (preventable) risk that is not easilydetectable though casual analyses of the overall trading data.

IV. Conclusion

One of the more well known psychological tendencies thatpermeates Wall Street trading desks is the traders’ aversionto realizing losses. Traders have a tendency to hold theirlosing trades too long because they are predisposed to geteven with their losses. By all accounts, this behavior isundesirable and can be quite costly. Securities firms are wellaware of the costs that arise with this behavior and they oftenimplement risk control mechanisms to prevent (limit) it fromoccurring. In our paper, we examine whether such measuresactually work and how they influence proprietary stock traderbehavior.

Despite our sample firm’s efforts to get traders morecomfortable with realizing their trading losses throughtraining, managerial oversight, trader access to a licensedpsychologist, discipline price control, and inventoryliquidation, we find that professional traders still havedifficulties accepting their losses. For example, traders hold

losing trades more than twice as long as winning trades andthese longer holding times coincide with lower trading profits.

While our results highlight how difficult it is for institutionsto rid psychological biases from the traders’ decisions, ourresults also highlight the complexities involved withimplementing efficient control mechanisms to get traders torealize their losses. When firms force traders to liquidatetheir inventory and realize their losses, professional tradersrespond by holding their losses up until the very last moment.When firms heavily focus on disciplined trading and uniformprice control to ensure loss realization, professional tradersrespond by holding their losses longer on larger size trades.Clearly, these are not desirable behavioral responses to thefirms underlying objective. However, if financial institutionsimpose stricter control mechanisms to get their traders torealize their losses sooner, the additional trading constraintswill likely begin to start conflicting with the traders’ overallstrategies and trading practices. On the other hand, if controlmechanisms put in place are too lax, losing trades will beheld for longer periods of time and losses will surmount.

Securities firms implement control mechanisms to improveperformance and reduce risk, but their efforts to get employeesto accept their losses has much broader implications. Ourresults show that institutional risk control can have a stronginfluence on trader behavior. Future studies, which provideinstitution detail on the design of control mechanisms beingused at other financial institutions and how employeesrespond to them, would be insightful for both creating optimalrisk control mechanisms and also for determining their overalleffects in the marketplace.


References

Admati, A.R. and P. Pfleiderer, 1988, “A Theory of IntradayPatterns: Volume and Price Variability,” Review of FinancialStudies 1 (No. 1), 3-40.

Barber, B.M. and T. Odean, 1999, “The Courage of MisguidedConvictions: The Trading Behavior of Individual Investors,”Financial Analysts Journal 56 (No. 6), 41-55.

Barber, B.M. and T. Odean, 2000, “Trading is Hazardous to YourWealth: The Common Stock Investment Performance ofIndividual Investors,” Journal of Finance 55 (No. 2), 773-806.

Barber, B., Y. Lee, Y. Lu, and T. Odean, 2007, “Is the AggregateInvestor Reluctant to Realize Losses: Evidence from Taiwan,”European Financial Management, 13 (No. 3), 371-388.

Benartzi, S. and R. Thaler, 1995, “Myopic Loss Aversion and theEquity Premium Puzzle,” Quarterly Journal of Economics 110(No. 1), 73-92.

Conrad, J., K.M. Johnson, and S. Wahal, 2002, “The Trading ofInstitutional Investors: Theory and Evidence,” Journal of AppliedFinance 12 (No. 1), 7-14.

Coval, J. and T. Shumway, 2005, “Do Behavioral Biases AffectPrices?” Journal of Finance 60 (No. 1), 1-34.

Dhar, R. and N. Zhu, 2007, “Up Close and Personal: An IndividualLevel Analysis of the Disposition Effect,” Management Science,forthcoming.

Feng, L. and M. Seasholes, 2005, Do Investor Sophistication andTrading Experience Eliminate Behavioral Biases in FinancialMarkets?” Review of Finance 9 (No. 3), 305-351.

Frazzini, A., 2006, “The Disposition Effect and Underreaction toNews,” Journal of Finance 61 (No. 4), 2017-2046.

Garvey, R. and A. Murphy, 2004, “Are Professional Traders TooSlow to Realize Their Losses?” Financial Analysts Journal 60(No. 4), 35-43.

Garvey, R. and A. Murphy, 2005, “The Profitability of Active StockTraders,” Journal of Applied Finance 15 (No. 2), 93-100.

Garvey, R., A. Murphy, and F. Wu, 2007, “Do Losses Linger?Evidence from Proprietary Stock Traders,” Journal of PortfolioManagement 33 (No. 4), 75-83.

Goldberg, D.C. and A. Lupercio, 2004, “Cruising at 30,000: Semi-Pro Numbers Level Off, but Trading Volumes Rise,” Bear StearnsCompany Report, August.

Government Accountability Office (GAO), 2005, “Decimal Pricinghas Contributed to Lower Trading Costs and a More ChallengingTrading Environment,” GAO Report, May. Full report availableonline at: http://www.gao.gov/new.items/d05535.pdf.

Grinblatt, M. and M. Keloharju, 2001, “What Makes InvestorsTrade?” Journal of Finance 2 (No. 2), 589-616.

Heisler, J., 1994, “Loss Aversion in a Futures Market: An EmpiricalTest,” Review of Futures Markets 13 (No. 3), 793-822.

Locke, P.R. and S.C. Mann, 2005, “Professional Trader Disciplineand Trade Disposition,” Journal of Financial Economics, 76(No. 2), 401-44.

Nguyen, V.T., B.F. Van Ness, and R.A. Van Ness, 2004, “TheReporting of Island Trades on the Cincinnati Stock Exchange,”Journal of Applied Finance, 14 (No. 2), 30-39.

Odean, T., 1998, “Are Investors Reluctant to Realize Their Losses?”Journal of Finance 53 (No. 5), 1775-1798.

Scherbina, A. and L. Jin, 2005, “Change is Good or The DispositionEffect among Mutual Fund Managers,” Working Paper, HarvardBusiness School.

Shapira, Z. and I. Venezia, 2001, “Patterns of Behavior ofProfessionally Managed and Independent Investors,” Journalof Banking and Finance 25 (No. 8), 1573-1587.

Shefrin, H. and M. Statman, 1985, “The Disposition to Sell WinnersToo Early and Ride Losers Too Long: Theory and Evidence,”Journal of Finance 40 (No. 3), 777-790.

Statman, M. and D. Caldwell, 1987, “Applying Behavioral Financeto Capital Budgeting: Project Terminations,” FinancialManagement 16 (No. 4), 7-15.

Thaler, R., 1985, “Mental Accounting and Consumer Choice,”Marketing Science 4 (No. 3), 199-214.

Why Do People Trade?1

Anne Dorn, Daniel Dorn, and Paul Sengmueller

37

Besides trading to save, manage risk, and speculate,people trade simply because they find it entertaining. Ina survey of 1,300 German discount brokerage clients,respondents who indicate that they “enjoy investing” and“enjoy risky propositions” trade twice as much as theirpeers. In contrast, standard motives for trading such assaving and rebalancing explain little of the variation intrading activity across investors. Entertainment appearsto be a straightforward explanation for why some peopletrade much more than others and why active tradersunderperform their peers after transaction costs.

Anne Dorn is currently unaffiliated. Daniel Dorn is an Assistant Professorof Finance at Drexel University in Philadelphia, PA. Paul Sengmueller isan Assistant Professor of Finance at CentER-Tilburg University in Tilburg,the Netherlands.

1 This contribution is based on the article “Trading as Entertainment?” byDaniel Dorn and Paul Sengmueller which is forthcoming in ManagementScience. The copyright on “Trading as Entertainment?” is held by IN-FORMS.

Trading in financial markets is an important economicactivity. Trades are necessary to get into and out of the market,to put unneeded cash into the market, and to convert backinto cash when the money is wanted. They are also needed tomove money around within the market, to exchange one assetfor another, to manage risk, and to exploit information aboutfuture price movements. All this trading performs the

important social function of incorporating information intoasset prices. It also (viewed from certain quarters moreimportantly) provides a major source of revenue for securitiesfirms.

In this paper, we explore reasons investors trade that gobeyond assembling a portfolio with the best return-risk profile.We argue that in exchange for this major source of revenue,securities firms are also allowing traders to enjoy themselvesin the act of trading. Some investors view trading as a hobby,and hand over their large trading fees as happily as anaudiophile pays top dollar for the latest in speaker technology.Other investors, we find, are playing the market in a veryliteral sense, racking up trading costs like a casino patronsliding his chips across the table.

We make these characterizations in an attempt to explainthree stylized facts about trading that have caught the attentionof researchers and practitioners alike:

1. Trading volume in financial markets is high. Forexample, stocks in the US change hands roughly once peryear.

2. Trading volume is concentrated among a small numberof market participants. In a well-cited study of US discountbrokerage clients, Barber and Odean (2000) report that themost active investors turn over their portfolio several timesper year. In contrast, a substantial fraction of US individualinvestors with a brokerage account do not trade at all in agiven year, according to recent waves of the US Survey ofConsumer Finances.

3. Traders underperform buy-and-hold investors. Barberand Odean (2000) report that the most active investorsunderperform the least active investors by several percentagepoints per year and that the performance differential isessentially due to trading costs.

Standard economic theory appears to be at odds with thesefacts. The rational investor assumed by standard theory isonly interested in the return and risk attributes of his portfolio,


and only trades if the benefits from trading justify the costs.In general, return maximizing rational investors should bevery reluctant to trade with one other. The intuition is simple.Trading is a zero-sum game. If rational investor A offers totrade with rational investor B, B should be suspicious that Aknows something about the future price that B does not.Milgrom and Stokey (1982) and Tirole (1982) formalize thisintuition and show that indeed, rational investors should refuseto speculatively trade with each other.

The inadequacy of standard theory opens the door for thebehavioral approach. De Bondt and Thaler (1995) call theobserved trading volume in financial markets “perhaps thesingle most embarrassing fact to the standard financeparadigm.” The leading answer of the behavioral camp tothe question “Why do people trade?” is overconfidence,prominently advocated by Odean (1998). Essentially,overconfidence allows both parties to a trade to believe thatthey will win the zero-sum game of trading. For all its intuitiveappeal, the overconfidence hypothesis has received onlymixed empirical support. Barber and Odean (2001) reportthat male US discount brokerage clients trade more than theirfemale counterparts and interpret this finding as consistentwith the overconfidence hypothesis. Grinblatt and Keloharju(2008) use self-confidence assessments from a psychologicaltest administered by the Finnish military to inferoverconfidence of male Finnish investors. They report thatthe univariate correlation between the self-confidence scoreand trading activity is close to zero. Other things equal,however, the proxy for overconfidence derived from the self-confidence assessments is significantly positively related tothe number of trades, though not to portfolio turnover. Glaserand Weber (2003) use a questionnaire to elicit nine proxiesfor overconfidence in a sample of German discount brokeragecustomers and relate the proxies to actual portfolio turnover.None of the proxies help explain cross-sectional variation inportfolio turnover. In trading experiments with students fromdifferent countries, Deaves, Lüders, and Luo (2004), andBiais, Hilton, Mazurier, and Pouget, (2005) report little orno relation between proxies for overconfidence and observedtrading activity.

This paper explores a different explanation of why peopletrade, anticipated by Black (1986) who notes that “[w]e mayneed to introduce direct utility of trading to explain theexistence of speculative markets.” For people who tradebecause they like to do so, the monetary cost of trading isoffset by non-pecuniary benefits from researching, executing,talking about, anticipating the outcome of, or experiencingthe outcome of a trade.

Motives for entertainment trading can be classified in threedistinct groups: recreation, sensation seeking, and anaspiration for riches.

Recreational trading can be motivated by a feeling of

accomplishment (similar to a homeowner who decides to doit himself rather than hiring a contractor), camaraderie (amongmembers of an investment club, for example), or it can emergeas a by-product of following the financial markets as a hobby(like a technophile who likes to read reviews of the latestgadgets, and is then tempted to go out and buy them).

Perceiving investing as a diversion rather than a chore,hobby investors have less of a psychological hurdle toovercome when executing changes to their portfolio, directlylowering their marginal cost of trading. By actively followingthe financial markets, they also expose themselves to moretrading signals and should hence be expected to trade morethan their peers.2

Entertainment trading can also be motivated by sensationseeking in the financial domain. According to Zuckerman(1994), “Sensation seeking is a trait defined by the seekingof varied, novel, complex, and intense sensations andexperiences, and the willingness to take [...] financial risksfor the sake of such experience.” In a (perhaps subconscious)quest for arousal, sensation seekers look for both intensityand novelty in experience. An undiversified portfolio ofvolatile stocks exposes its holder to intense stimuli in theform of extreme returns. Exposure to such stimuli by itselfmay trigger trading as argued by Dorn and Huberman (2007).In addition, as pointed out by Grinblatt and Keloharju (2008),sensation seekers in the financial domain may value the actof trading in and of itself because a trade — a new bet —affords the desired novelty of experience. Grinblatt andKeloharju (2008) use traffic violations to proxy for thrillseeking behavior; they report that variation in the number ofspeeding tickets explains variation in trading activity in alarge sample of Finnish investors.

Alternatively, trading can be motivated by an aspirationfor riches as suggested by Statman (2002). A trade can beseen as a bet that carries a “dream value,” that is, the joy ofimagining what a handsome payoff will buy. Such aspirationshave been used to explain lottery participation (Conlisk, 1993)and exploited in advertising by retail brokers (Barber andOdean, 2002). Aspiration-driven investors should holdportfolios with volatile and positively skewed returns toincrease the chance of reaching an aspiration level far abovetheir current wealth (see Kumar, 2008). The exposure totrading stimuli in the form of extreme returns, coupled withan inherent impatience to reach their desired wealth level,may lead aspiration-driven investors to pick up and abandontrading ideas more quickly than their peers.

2This argument is reminiscent of Merton (1987) who motivates his 1986Presidential Address to the American Finance Association by the simpleobservation that an investor needs to know about a stock before he cantrade it.

39DORN, DORN, & SENGMUELLER — WHY DO PEOPLE TRADE?

The paper examines the hypothesis that entertainmentmotives drive trading by combining survey responses andtransaction records for a sample of more than 1,000 clientsat one of the top three discount brokers in Germany. Thesurvey offers responses to statements that elicit whether arespondent enjoys investing and statements that have beenused to identify compulsive gamblers. The responses to thesestatements serve as proxiesfor the entertainment benefitsderived from trading.

The main findings are asfollows:

1. Standard motives fortrading fail to explain much ofthe observed trading activity.Turnover due to savings,dissavings, liquidity, andrebalancing considerationsaccounts only for about one third of total turnover.

2. Standard motives for trading fail to explain why somepeople trade much more than others. Turnover due to savings,dissavings, liquidity, and rebalancing considerations variesmuch less across investors than turnover that cannot bejustified by standard trading motives.

3. Entertainment appears to be a major driver of portfolioholdings. Entertainment-driven investors hold moreconcentrated portfolios, riskier portfolio components, andportfolios with more positively skewed returns.

4. Entertainment appears to be a major driver of portfolioturnover, especially turnover that cannot be justified bystandard trading motives. Entertainment-driven investors turnover their portfolio of stocks, bonds, funds, and options atroughly twice the rate of their peers.

5. Proxies for overconfidence are at best weakly correlatedwith trading activity.

The remainder of the paper proceeds as follows: Section Idescribes the data and the construction of the variables.Section II discusses the main findings in more detail. SectionIII concludes.

I. The Data

A. Brokerage Records

The analysis is based on a complete history of dailytransaction records in individual stocks, term deposits, bonds,mutual funds, and options obtained for a sample of 1,345current and former clients at one of Germany’s three largestdiscount brokers between January 1, 1995 and May 31, 2000.3

In addition to the standard trade attributes, the records includea channel variable that indicates whether the order was placedover the phone, over the internet, or within an automaticinvestment or withdrawal plan that exist for dozens ofindividual stocks and mutual funds. Such plans allowinvestors to gradually build or reduce positions at four dateseach month (similar to ShareBuilder in the US).

In July and August 2000,after the sample period, eachinvestor participated in asurvey that elicited a widerange of objective andsubjective investor attributesdetailed below. Table Isummarizes the clientportfolios and trading activityduring the sample periodJanuary 1995 to May 2000.

Average monthly turnover, defined as one half the sum ofthe absolute values of purchases and sales during a givenmonth divided by the average portfolio value during thatmonth averaged first across time for each investor and thenacross investors, is 15%. In our turnover calculation, weconsider purchases and sales of individual stocks, individualbonds, mutual funds, options, and term deposits. Individualstock trades account for 62%, fund trades account for 18%,and option trades account for 15% of the total trading volumeduring the sample period. The average portfolio size overthe entire account life is roughly 90,000 Deutsche Mark[DEM] or 50,000 US dollars [USD] at the average USD/DEM exchange rate of 1.7 during the sample period.

We analyze the portfolios using a measure of concentrationknown as the Herfindahl-Hirschmann Index (HHI).4 Themedian HHI of the stock and fund portfolios during the sampleperiod is 31%; that is, the typical client holds the equivalentof an equally-weighted portfolio of three individual positions.

From the information provided by the client to the brokerat account opening, we can infer the gender of all mainaccount holders and the age of those who choose to reporttheir birth date. The typical respondent is male, young, andhas held the account for three years. Judging from a surveyof Germans who hold stocks, either directly or through mutualfunds (see Deutsches Aktieninstitut, 2000), our sampleinvestors are more predominantly male and younger than the

3The broker is labeled as a discount broker because no investment adviceis given.

4The HHI is defined as the sum of squared portfolio weights. A portfolioconsisting of n equally-weighted stocks would have an HHI of 1/n. Stockmutual funds are assumed to consist of one hundred equally-weightedpositions that do not overlap with other holdings of the investor. That is,the HHI of portfolio of an investor holding one stock mutual fund is 1%and that of an investor splitting his money equally between two stockmutual funds is 0.5%.

Standard motives for trading suchas saving and rebalancing fail toexplain both the level of observedtrading activity and the variationof such activity across individuals.


typical German stock market participant. Relative to thepopulation of German stock market participants, the sampleinvestors also appear to be more highly educated and earnhigher incomes (see Dorn and Huberman, 2005).

B. Survey Variables

To gauge which investors likely derive non-pecuniarybenefits from their trading activities, we use their self-reportedattitudes towards investing and gambling gleaned from asurvey administered in July and August 2000. The surveyelicited information on the investors’ investment objectives,risk attitudes and perceptions, investment experience andknowledge, portfolio structure, and demographic and socio-economic status. Dorn and Huberman (2005) describe thesurvey in detail.

To pin down the importance of entertainment motives fordifferent investors, we focus on the survey items that makean explicit reference to whether or not respondents enjoydealing with their investments or enjoy gambling. This focusyields responses to a total of four statements (reproducedbelow in translation from the original German). The investorsare asked to indicate their agreement with the four statements

on a five-point scale ranging between (1) strongly disagree,(2) tend to disagree, (3) tend to agree, (4) strongly agree, and(5) don’t know:

1. I enjoy investing.2. I enjoy risky propositions.3. Games are only fun when money is involved.4. In gambling, the fascination increases with the size of

the bet.

Agreement with statement one defines a hobby investor.Agreement with statements two to four identifies respondentswho enjoy risky propositions, in general, and gambling, inparticular; in fact statements three and four are taken from astudy on identifying compulsive gamblers (Nadler, 1985).Hobbyists and gamblers appear to form distinct groups; theresponse to the first statement is only weakly correlated withthe responses to the other statements. Statements two throughfour flag investors as gamblers more or less consistently; thepairwise correlation between the responses to statements twoto four is quite high and reaches 0.46 between statementsthree and four.

Table II summarizes objective demographic and socio-economic attributes of investors grouped by their responses

Table I: Summary Statistics

Portfolio characteristics are calculated from the complete daily transaction history available for each of the 1,345 sample investors fromthe day when the account was opened until May 31, 2000 or the day when the account was closed, whichever comes first. Turnover ina given month is the sum of the absolute value of purchases and sales of stocks, bonds, mutual funds, and options divided by twice thehigher of the portfolio value at the beginning or at the end of the month (to avoid extreme values). Average portfolio value is calculatedat the end of every month across all individual stocks, funds, options, bonds, and term deposits in the client’s portfolio. During thesample period, one US Dollar [USD] corresponds to roughly Deutsche Mark [DEM] 1.7. The Herfindahl-Hirschmann Index (HHI) iscalculated using only stocks and stock mutual funds for which Datastream offers a complete history of non-stale prices and returns.Higher values of the HHI indicate less diversification. “Gender” is a dummy variable that is one if the respondent reports to be male andzero otherwise (if missing, we replace the missing value with the gender recorded for the main account holder in the brokerage database).“Age” is the age of the respondent (if missing, we replace the missing value with the age recorded for the main account holder in thebrokerage database). “College” is a dummy that is one if a respondent has a college degree and zero otherwise. “Self-employed” is adummy that is one if the respondent reports to be self-employed and zero otherwise. “Income” is the self-reported gross annual income.“Wealth” is the self-reported total net worth (including all financial assets and real estate).

Portfolio Characteristics Mean Median

Average monthly portfolio turnover 15% 7%

Average portfolio value [DEM] 86,000 38,000

Average Herfindahl-Hirschmann Index 31% 25%

Investor Characteristics

Gender [fraction male] 88%

Age 39 36

Education [fraction with college education] 70%

Self-employed 17%

Gross annual income [DEM} 93,000 88,000

Net worth [DEM] 373,000 325,000


Table II: Characteristics of Entertainment-Driven Investors

Panels A through D characterize investors grouped by their responses to four survey statements designed to elicit whether the respondentsenjoy investing or gambling with money. The investors are asked to indicate their agreement with the four statements on an ordinal scaleof (1) strongly disagree, (2) tend to disagree, (3) tend to agree, (4) strongly agree. In Panel A, we have combined the categories (1) and(2) to “disagree” since only four respondents choose to “strongly disagree.” In Panel D, we have combined the categories (3) and (4) to“agree” since only thirty-eight respondents choose to “strongly agree.” “Nobs” is the number of respondents in each category. Thedemographic and socio-economic variables are defined as in Table I.

***Significant at the 0.01 level **Significant at the 0.05 level. *Significant at the 0.10 level.

to the above statements. We exclude the few investors withmissing responses and investors who respond with “don’tknow” — out of a total of 1,345 respondents, the number ofmissing responses ranges from 10 (for statement three) to 15(for statement one); the number of respondents who respondwith “don’t know” ranges from 11 (for statement one) to 56(for statement four). To be able to make meaningful statisticalcomparisons across groups, we group investors who “stronglydisagree” with statement one together with those who “tendto disagree” as there are only four investors who “stronglydisagree.” For the same reason, we combine the “strongly

agree” and “tend to agree” categories for statement four asonly 38 investors “strongly agree.”

Male investors and wealthier investors appear to enjoydealing with investments more than their female and lesswealthy counterparts. Those who enjoy games only whenmoney is involved, in particular, tend to be younger, less welleducated, and less wealthy. Although we have no directinformation about whether our sample investors engage ingambling outside the stock market, it is interesting to notethat younger age, a lower level of education, and less wealthhave been linked to a higher propensity to participate in legal

Nobs Gender Age College Self- Income Wealth employed

Panel A - Statement 1: “I enjoy investing.”

Disagree 84 76% 40 73% 15% 90 262

Tend to agree 403 87% 41 72% 17% 94 358

Strongly agree 822 91%*** 41 69% 16% 94 396***

Panel B - Statement 2: “I enjoy risky propositions.”

Strongly disagree 148 82% 48 69% 13% 85 421

Tend to disagree 571 88% 41 70% 16% 93 385

Tend to agree 492 90% 39 70% 17% 97 357

Strongly agree 87 95%*** 37*** 75% 21% 102** 330*

Panel C - Statement 3: “Games are only fun when money is involved.”


Tend to disagree 487 89% 41 66% 17% 94 381

Tend to agree 277 92% 40 67% 19% 97 351

Strongly agree 71 87% 38* 62%** 22% 84 282**

Panel D - Statement 4: `Ìn gambling, the fascination increases with the size of the bet."


Tend to disagree 396 88% 41 68% 18% 90 364

Agree 199 90% 39** 65%* 18% 94 329**


forms of gambling in Germany (see Albers and Hübl, 1997),the UK (see Farrell and Walker, 1999), and the US (seeClotfelter and Cook, 1989).

II. Main Results

It is our task in this section to show that the high tradinglevels laid out in the stylized facts with which we began thepaper are due to investorstrading for entertainmentmotives. We will demonstratethat that small group ofinvestors who exhibit highturnover and lagging returnsis composed of thoseindividuals who find tradingentertaining, those individualswhose survey responses labelthem as hobby or gamblinginvestors. Dorn and Sengmueller (2008) go on to documentthat these results are not driven by survey response bias, pastreturns, or small accounts.

A. Standard Motives Inadequately Explain theObserved Trading Activity

We divide the turnover we observe in our sample into twofiner measures: normal turnover and excess turnover. Normalturnover consists of trading that can be explained by standardmotives for trading such as savings, dissavings, liquidity, orrebalancing considerations; excess turnover is the portion oftotal turnover that cannot be explained by these motives.

Similar to Barber and Odean (2002), we define an excesssale as a sale of a complete position of an individual stock,mutual fund, or option that is followed by one or more stock,fund, or option purchases within three weeks of the sale. Wedefine excess purchases as all stock, fund, and optionpurchases made within three weeks of an excess sale. Allother trades are classified as normal trades.5 In particular,all trades in term deposits and automatic investment andwithdrawal plans — plans that allow investors to graduallybuild or reduce positions in dozens of stocks and funds atfour predetermined dates per month — are classified as

normal as they are likely motivated by liquidity and savingsconsiderations.6

Our trade classification likely overstates normal turnover,in part because we only observe part of the portfolio for someinvestors. For example, an investor might sell off a completeposition of stock A in an unobserved account and invest theproceeds in stock B in the observed account because heexpects stock B to outperform stock A; such a purchase wouldbe classified as a normal trade even though it is not driven by

savings, liquidity, orrebalancing motives.

Table III reports summarystatistics for normal andexcess turnover. Across thesample respondents, theaverage monthly total turnoverof 15% consists of 5% normalturnover and 10% excessturnover — in other words,

only one third of the observed trading volume can beexplained by savings, liquidity, and rebalancing motives.

The standard deviation of normal turnover across thesample respondents is 4% as opposed to 29% for excessturnover. Therefore, the challenge in explaining theheterogeneity in trading activity across investors appears tolie in understanding excess turnover; investors appear to befairly homogenous in their desire to trade due to savings,liquidity, or rebalancing motives.

B. Cross-sectional Differences in PortfolioCharacteristics Are Consistent With theEntertainment Hypothesis

Table IV sets out the portfolio characteristics of investors,separated by their responses to our four statements. We findthat those investors who are most excited by risk indeed holdmore risky portfolios.

Self-professed gamblers in our sample hold moreconcentrated equity portfolios. For example, those whostrongly agree with the statement “I enjoy risky propositions”hold equity portfolios with an average HHI of 0.39 whichcorresponds to an equally weighted position in two to threeindividual stocks; by contrast, their peers who stronglydisagree with this statement hold the equivalent of an equally

6Barber and Odean (2002) use the terms “non-speculative” and“speculative” trades instead of normal and excess trades. Substantively,our classification differs from theirs in three ways. First, they restrict theiranalysis to trades in common stocks. Second, they require that sales be fora profit to rule out tax-loss motivated trading (capital gains from sales offinancial securities are essentially not taxed in Germany). Third, they donot distinguish between savings plan and non-plan trades.

Entertainment-driven investorsturn over their portfolio of stocks,bonds, funds, and options atroughly twice the rate of theirpeers.

5We have explored variations of this definition in unreported robustnesschecks. For example, one could argue that the complete sale of anindividual stock position followed by the purchase of a stock mutual fundconstitutes a diversifying and hence normal trade. One could also arguethat put purchases are used for portfolio insurance purposes. Since mosttrading occurs in individual stocks and call options, these variations inthe definition of excess turnover have little effect on our results.


Table III: Normal Turnover Versus Excess Turnover

Normal turnover in a given month is defined as one half the sum of the absolute values of normal purchases and normal sales during agiven month divided by the average portfolio value during that month. Excess turnover is defined similarly, but using excess purchasesand excess sales. An excess sale is defined as a sale of a complete position of an individual stock, mutual fund, or option that is followedby one or more stock, fund, or option purchases within three weeks of the sale. An excess purchase is defined as a stock, fund, or optionpurchase made within three weeks of an excess sale. All other trades are classified as normal trades. In particular, all trades in termdeposits and automatic investment and withdrawal plans are classified as normal.

Table IV: Portfolio Choices of Entertainment-Driven Investors

HHI is the average Herfindahl-Hirschmann Index across the portfolios in the group, higher values indicate less diversification. Averagecomponent volatility (ACV) is the value-weighted average volatility of the portfolio components in an investor’s portfolio. Realizedskewness is calculated from daily portfolio returns as in Chen et al. (2001). “Options” is the fraction of respondents in a group that havetraded options at some point during the sample period. HHI, ACV, and skewness are calculated using only the individual stocks andstock mutual funds for which Datastream provides daily total return data.

Mean Std Median

Average monthly portfolio turnover 15% 32% 7.4%

thereof:

normal turnover 5% 4% 3.9%

excess turnover 10% 29% 2.9%

***Significant at the 0.01 level **Significant at the 0.05 level.

Nobs HHI ACV Realized Options

Skewness

Panel A - `Ì enjoy investing."

Disagree 84 30% 42% 0.92 18%

Tend to agree 403 30% 42% 0.50 29%

Strongly agree 822 31% 45% 0.74 41%***

Panel B - `Ì enjoy risky propositions."

Strongly disagree 148 26% 39% 0.45 22%

Tend to disagree 571 28% 41% 0.45 29%

Tend to agree 492 33% 48% 0.91 46%

Strongly agree 87 39%*** 52%*** 1.32*** 51%***

Panel C - ``Games are only fun when money is involved."


Tend to disagree 487 30% 43% 0.52 34%

Tend to agree 277 34% 48% 0.87 46%

Strongly agree 71 38%*** 53%*** 1.07 52%***

Panel D - `Ìn gambling, the fascination increases with the size of the bet."


Tend to disagree 396 32% 45% 0.77 42%

Agree 199 37%*** 50%*** 0.96** 47%***


weighted portfolio of four stocks.Not only are the portfolios of gamblers more concentrated,

but they also consist of individually riskier securities. Forexample, the average component volatility — the value-weighted average of the annualized volatility of the stockportfolio components — of investors who strongly agree withthe statement “I enjoy risky propositions” averages 52%relative to 39% for investorswho strongly disagree withthis statement.

Consistent with gamblerspreferring skewness (seeGolec and Tamarkin, 1998),people classified asgamblers in our data set holdportfolios of stocks andmutual funds that exhibitmore positively skewed returns. We exclude holdings ofindividual bonds and options when calculating portfoliostatistics, in part because of a lack of high-frequency pricedata. However, options holdings and trades also point toentertainment-motivated investors preferring securities withpositively skewed payoffs. For example, half of the investorswho strongly agree with the statement “I enjoy riskypropositions” trade options during our sample period; incontrast, only one out of five investors who strongly disagreewith this statement also trade options.

C. Cross-sectional Differences in TurnoverAre Consistent With the EntertainmentHypothesis

Trading is costly. The typical respondent in the paper’ssample spends 0.5% of his self-reported gross annual incomeon trading commissions. The main hypothesis entertainedhere is that some investors derive non-pecuniary benefits fromresearching, executing, talking about, anticipating, orexperiencing the outcome of a trade. These benefits helpoffset the cost of trading.

We group the survey respondents by their responses to eachof the four entertainment statements. Figures 1-4 illustratethe equally-weighted average monthly turnover rates for themembers of each group. Investors who report enjoyinginvesting also trade more aggressively than their peers. Figure1 shows that investors who strongly agree with “I enjoyinvesting” exhibit an average monthly turnover of 17% —significantly higher than the average turnover rate of 10%for the investors who disagree with the statement. Similarturnover patterns are obtained for investors grouped by theirresponses to statements that elicit the investor’s affinity togambling (see Figures 2-4). For example, investors whostrongly agree with “Games are only fun when money is

involved” turn over their portfolios at an average monthlyrate of 24% — twice the rate of those who strongly disagreewith the statement (see Figure 3).

If differences in trading activity were indeed driven byentertainment, one would expect such differences to manifestthemselves in terms of excess turnover; that is, turnoverunlikely due to savings, liquidity, or rebalancing

considerations. Indeed,Figures 1-4 also show thatvirtually the entire differencein total turnover betweenthose who enjoy investing orgambling and their peers isdue to the higher excessturnover of thee n t e r t a i n m e n t - d r i v e ninvestors. For example,

investors who strongly agree with “Games are only fun whenmoney is involved” exhibit normal turnover rates averaging5% — similar to the average normal turnover of 4.5% oftheir peers who strongly disagree with the statement.However, the average excess turnover rate of the self-professed gamblers, 19%, is almost thrice the correspondingrate of their peers (see Figure 3).

In Dorn and Sengmueller (2008) we investigate whetherthese correlations between turnover and entertainmentmotives hold up under multivariate statistical analysis. Andindeed we find that even after controlling for gender, age,education, income, employment status, and wealth, investorswho enjoy investing or gambling trade more than those whoenjoy neither, and investors who enjoy both trade the most ofall.

These results suggest that investors appear to derivepleasure from trading both as a pastime and as a form ofgambling. Figure 5 illustrates that respondents who enjoyinvesting (that is, they strongly agree with the statement “Ienjoy investing”) but not gambling (that is, they disagree orstrongly disagree with the statement “Games are only funwhen money is involved”), and those who enjoy gamblingbut not investing trade more than their peers who enjoy neitherinvesting nor gambling; those who enjoy both investing andgambling trade the most.

D. Differences in Overconfidence Fail toExplain Turnover Differences

Overconfidence might explain the paper’s results ifoverconfident investors report enjoying trading because theyenjoy doing what they wrongly perceive themselves to begood at. Alternatively, entertainment might amplify the effectsof overconfidence or vice versa.

The wealth of survey responses allows us to construct three

Entertainment as a motive for tradingis distinct from overconfidence.Proxies for overconfidence are at bestweakly correlated with tradingactivity.


Figure 1: Turnover as a Function of Enjoyment of Investing

Figure 2: Turnover as a Function of Enjoyment of Risky Propositions

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

Tend to disagree or strongly

disagree

Tend to agree Strongly agree

Agreement with "I enjoy investing."

Total turnover

Normal turnover

Excess turnover

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

Strongly disagree Tend to disagree Tend to agree Strongly agree

Agreement with "I enjoy risky propositions."

Total turnover

Normal turnover

Excess turnover

.


Figure 3: Turnover as a Function of Affinity for Gambling (I)

Figure 4: Turnover as a Function of Affinity for Gambling (II)

0%

5%

10%

15%

20%

25%

30%

Strongly disagree Tend to disagree Tend to agree Strongly agree

Agreement with "Games are only fun when money is involved."

Total turnover

Normal turnover

Excess turnover

.

0%

5%

10%

15%

20%

25%

30%

Strongly disagree Tend to disagree Tend to agree or strongly

agree

Agreement with "In gambling, the fascination increases with the size of the bet."

Total turnover

Normal turnover

Excess turnover


Figure 5: Excess Turnover of Investors Sorted by Enjoyment of Investing and Gambling

proxies that capture different aspects of overconfidence: thetendency to overestimate one’s knowledge, the tendency tooverly attribute successes to skill in conjunction with pastreturns (known as the self-enhancing attribution bias), andthe erroneous expectation of being able to affect chanceoutcomes (known as the illusion of control; see also Barberand Odean, 2002; Daniel et al., 1998; and Gervais and Odean,2001).

We use the investor’s agreement with the statement “I’mmuch better informed than the average investor” as a proxyfor the tendency to overestimate one’s knowledge, or relativeknowledge. To estimate the self-enhancing attribution bias,we consider the extent to which survey participants agreewith the statement “My past investment successes were, aboveall, due to my specific skills.” To construct a proxy for theillusion of control, we compute an aggregate score using theinvestors’ responses to four statements:

1. When I make plans, I am certain that they will workout.2. I always know the status of my personal finances.3. I am in control of my personal finances.4. I control and am fully responsible for the results of myinvestment decisions.

In regressions reported in Dorn and Sengmueller (2008),we find that none of the overconfidence proxies issignificantly related to excess turnover. Moreover, thesignificance of our hobby and gambler proxies is maintainedeven while controlling for overconfidence.

We also investigated how our hobbyist and gamblerdesignations interact with overconfidence. The results ofinteracting the investor’s agreement with “I enjoy investing”with the three overconfidence proxies are shown in Figures6-8. To simplify the presentation, and to ensure that theresulting groups consist of enough members, we createdbinary entertainment and overconfidence proxies. In general,no additional insights were found in this interaction,overconfident hobby or gambler investors turned over theirportfolios at similar rates as underconfident investors withthe same hobby or gambling affinities.7

III. Conclusion

Some investors derive enjoyment from trading which offset

7In Figure 6, it appears that the underconfident who do not enjoy investingtrade less than their overconfident peers. However, the difference in tradingactivity is not statistically significant.

0%

5%

10%

15%

20%

25%

30%

35%

Do not enjoy investing Enjoy investing

Do not enjoy gambling

Enjoy gambling


Figure 6: Excess Turnover of Investors Sorted by Enjoyment of Investing and Relative Knowledge

Figure 7: Excess Turnover of Investors Sorted by Enjoyment of Investing and Self-Attribution ofSuccess

0%

5%

10%

15%

20%

25%

30%

35%


Less knowledge than average investor

More knowledge than average investor

0%

5%

10%

15%

20%

25%

30%


Low self-attribution of success

High self-attribution of success


Figure 8: Excess Turnover of Investors Sorted by Enjoyment of Investing and Self-Control

the costs of churning. Like lottery players who buy ticketswith negative expected values, entertainment-driven investorstrade even though trading diminishes the expected monetarypayoff of their portfolio. Consistent with this conjecture,variation in the self-reported enjoyment of investing andgambling explains variation in trading intensity even aftercontrolling for competing explanations such asoverconfidence.

The most entertainment-driven investors trade about twiceas much as those who fail to take pleasure in gambling orinvesting. Relying solely on transaction records (that is,

independently of the survey responses), we estimate that morethan half of the observed portfolio turnover is excess turnover— turnover in excess of what can be justified by standardtrading motives such as savings/dissavings, liquidity, andrebalancing. Most of the variation in trading activity acrossindividuals is variation in excess turnover. Variation in excessturnover is highly correlated with our proxies for non-pecuniary benefits derived from trading. In sum,entertainment trading appears to be quantitatively important— at least for this sample of discount brokerage customersduring the late 1990s.

0%

5%

10%

15%

20%

25%

30%


Low self-control

High self-control


References

Albers, N. and L. Hübl, 1997, “Gambling Market and IndividualPatterns of Gambling in Germany,” Journal of GamblingStudies 13 (No.2), 125-144.

Barber, B. and T. Odean, 2000, “Trading is Hazardous to YourWealth: The Common Stock Investment Performance ofIndividual Investors,” Journal of Finance 55 (No.2), 773-806.

Barber, B. and T. Odean, 2001, “Boys Will Be Boys: Gender,Overconfidence, and Common Stock Investment,” QuarterlyJournal of Economics 116 (No. 1), 261-292.

Barber, B. and T. Odean, 2002, “Online Investors: Do the SlowDie First?” Review of Financial Studies 15 (No. 2), 455-487.

Biais, B., D. Hilton, K. Mazurier, and S. Pouget, 2005, “JudgmentalOverconfidence, Self-monitoring and Trading Performance inan Experimental Financial Market,” Review of Economic Studies72 (No. 2), 287-312.

Black, F., 1986, “Noise,” Journal of Finance 41 (No. 3), 529-543.

Chen, J., H. Hong, and J.C. Stein, 2001, “Forecasting Crashes:Trading Volume, Past Returns and Conditional Skewness inStock Prices,” Journal of Financial Economics 61, 345-381.

Clotfelter, C. T. and P. J. Cook, 1989, Selling Hope: State Lotteriesin America, Harvard University Press, Cambridge, MA.

Conlisk, J., 1993, “The Utility of Gambling,” Journal of Risk andUncertainty 6, 255-275.

Daniel, K.D., D. Hirshleifer, and A. Subrahmanyam, 1998, “InvestorPsychology and Security Market Under- and Overreactions,”Journal of Finance 53 (No. 6), 1839-1886.

Deaves, R., E. Lüders, and G. Y. Luo, 2004, “An ExperimentalTest of the Impact of Overconfidence and Gender on TradingActivity,” Working Paper.

DeBondt, W.F.M. and R. H. Thaler, 1995, “Financial Decision-making in Markets and Firms: A Behavioral Perspective,” inR.A. Jarrow, V. Maksimovic and W.T. Ziemba (eds), Finance,Handbooks in Operations Research and Management Science,Vol. 9, North Holland, Amsterdam, chapter 13, 385-410.

Deutsches Aktieninstitut, 2000, Factbook 1999, Frankfurt am Main.

Dorn, D. and G. Huberman, 2005, “Talk and Action: WhatIndividual Investors Say and What They Do, Review of Finance9 (No. 4), 437-481.

Dorn, D. and G. Huberman, 2007, “Turnover and Volatility,” DrexelUniversity Working Paper.

Dorn, D. and P. Sengmueller, 2008, “Trading as Entertainment?”Management Science, Forthcoming.

Farrell, L. and I. Walker, 1999, “The Welfare Effects of Lotto:Evidence from the UK,” Journal of Public Economics 72, 99-120.

Gervais, S. and T. Odean, 2001, Learning to Be Overconfident,”Review of Financial Studies 14 (No. 1), 1-27.

Glaser, M. and M. Weber, 2003, “Overconfidence and TradingVolume,” University of Mannheim Working paper.

Golec, J. and M. Tamarkin, 1998, “Bettors Love Skewness, NotRisk, at the Horse Track,” Journal of Political Economy 106(No. 1), 205-225.

Grinblatt, M. and M. Keloharju, 2008, “Sensation Seeking,Overconfidence, and Trading Activity,” Journal of Finance,Forthcoming.

Kumar, A., 2008, “Who Gambles in the Stock Market?” Journal ofFinance, Forthcoming.

Merton, R.C., 1987, “A Simple Model of Capital MarketEquilibrium With Incomplete Information,” Journal of Finance42 (No. 3), 483-510.

Milgrom, P. and N. Stokey, 1982, “Information, Trade and CommonKnowledge,” Journal of Economic Theory 26 (No. 1), 17-27.

Nadler, L., 1985, “The Epidemiology of Pathological Gambling:Critique of Existing Research and Alternative Strategies,”Journal of Gambling Behavior 1 (No. 1), 35-50.

Odean, T., 1998, “Volume, Volatility, Price and Profit When AllTraders are Above Average,” Journal of Finance 53 (No. 6),1887-1934.

Statman, M., 2002, “Lottery Players/Stock Traders,” FinancialAnalysts Journal 14-21.

Tirole, J., 1982, “On the Possibility of Speculation under RationalExpectations,” Econometrica 50 (No. 5), 1163-1182.

Zuckerman, M., 1994, Behavioral Expressions and Biosocial Basesof Sensation Seeking, Cambridge University Press.

The Long-Term Variation of TradeInformativeness

Michel Rakotomavo

This paper analyzes the time variation of the informativenessof trades for NYSE-listed stocks between 1998 and 2004. Tradeinformativeness is defined as the percentage of efficient pricevariance that is attributable to trades. The evidence suggeststhat trade informativeness was related to institutional buyingand both uninformed and informed trading. The resultsindicate a positive relation between institutional buying andtrade informativeness before Regulation Fair Disclosure anddecimal pricing. After these events, the evidence is consistentwith both a rise in uninformed trading and a fall in informedtrading. Similar results are found for the period followingthe enactment of the Sarbanes-Oxley Act (SOA). While thedecrease in informed trading may be a continuation of thedecimalization effect, there is evidence pointing to a relationbetween uninformed trading and preceding-quarterinstitutional buying, a phenomenon that does not seem to bepresent before SOA.

Asymmetric information can have an important impacton microstructure, as illustrated by Whitcomb (2003). Bydecomposing the variance of changes in the efficient priceinto its trade-correlated and uncorrelated components,Hasbrouck (1991) proposed the ratio, trade informativeness,of the trade-correlated component to the total variance as a

measure of the degree of asymmetric information (relative tototal information) in the market for the security under study.Trade informativeness is defined as the percentage of efficientprice variance that is attributable to trades (Hasbrouck(1991)). For example, this percentage for Boeing was, onaverage, 26.22 between the first quarter of 1998 and the lastquarter of 2000. Therefore, about 26.22% of the publicinformation on Boeing stock came from transactions on itsshares during that period.

While it is well known that some traders have superiorinformation,1 studies of the dynamics of asymmetricinformation among investors have focused on short timeperiods. For example, Zhao and Chung (2006) analyze theeffect of NYSE decimal pricing on the probability of informedtrading. They define November 1, 2000-January 28, 2001 asthe pre-decimal period, and June 1, 2001-August 31, 2001as the post-decimal period. They find that the post-decimalprobability of informed trading is greater than its pre-decimalequivalent. Chakravarty, Van Ness, and Van Ness (2005)examine the effect of the same event on adverse selectioncosts. They cover the months of January and February, 2001.They conclude that the percentage adverse selection cost hasincreased and the dollar adverse selection cost has decreased.

This paper complements the above studies by using adifferent information asymmetry metric, implementing tradeinformativeness, and focusing on a longer term (1998-2004)that includes the enactment of the Sarbanes-Oxley Act (SOA).Collver (2007) uses August 1, 1999-January 31, 2002 NYSEpanel data and finds a significant decrease in daily tradeinformativeness after both the implementation of Regulation

Michel T.J. Rakotomavo is an Assistant Professor of InternationalBusiness Administration at the American University of Paris in Paris,France.

I thank an anonymous referee and Betty Simkins (Editor) for theirconstructive comments. Logistical support was provided by the AndrewBatinovich Trading Room and Research Center.

1 see Golbe and Shranz (1994) and Karpoff and Lee (1991) for illustra-tions of informed traders.

51


Fair Disclosure (RFD) and the switch to decimal pricing.2

Collver assumes that the reduction is caused by less informedtrading. This paper extends his finding by 1.) analyzing thetemporal variation of trade informativeness in terms ofmicrostructural variables, thus enabling a distinction betweenthe various causes of such variation, and 2.) considering alonger term that includes the post-SOA period. For Boeing,the average quarterly trade informativeness of 26.22%, aspreviously mentioned, decreased to 16.77% after RegulationFair Disclosure and decimal pricing, and went further downto 3.66% after SOA. The main results of this paper areillustrated with Boeing data in Section II.

Section I develops the paper’s hypotheses on the temporalcovariation between the informativeness of trades and somemicrostructure variables. Section II discusses the data andresults. Section III concludes the paper.

I. Hypotheses

Amihud and Li (2006) find that the abnormal returns atdividend change announcements is a decreasing function ofinstitutional holdings. They also show that institutionalinvestors use their information advantage to buy beforedividend increases. Interestingly, both Nissim and Ziv (2001)and Garrett and Priestley (2000) find that only dividendincreases are directly correlated with earnings. Nissim andZiv (2001) note that managers may elect to take a “big bath”when faced with bad news by reflecting that news directlyonto current earnings. Therefore, this paper hypothesizes thatinstitutional buying may increase trade informativenessbecause of institutions’ information advantage. Bothinstitutional holdings and changes in institutional holdingsare used as proxies for the level of institutional buying in astock for the reasons that follow. Sias, Starks, and Titman(2006) find that changes in institutional holdings, whichrepresent net institutional demand, are correlated withcontemporaneous stock returns because of informationeffects. This conclusion is consistent with holdings changesbeing a measure of informed buying intensity. This paperassumes that any residual institutional buying of a stock,which is not captured by net institutional demand, iscorrelated with the level of institutional holdings. The basisof this assumption is the herding behavior of institutionsevidenced, for example, in Nofsinger and Sias (1999).Therefore, holdings levels would capture the intensity of anylagged herd net buying. Hence, if institutional buying affectsthe time variation of trade informativeness and institutionshave an information advantage then:

Hypothesis 1: An increase in the level of institutionalholdings, over time, implies an increase in theinformativeness of trades, ceteris paribus.

Hypothesis 2: An increase in the change of institutionalholdings, over time, implies an increase in theinformativeness of trades, ceteris paribus.

Chiyachantana, Jain, Jiang, and Wood provide international“evidence of an increased use of order-breaking strategy”(2004, p.878) by institutions over time.3 This implies thatinstitutions have decreased the size and increased thefrequency of their trades. Therefore, if institutional buyingaffects the time variation of trade informativeness andinstitutions have an information advantage, then:

Hypothesis 3: An increase in trade size, over time, impliesa decrease in the informativeness of trades, ceteris paribus.

Hypothesis 4: An increase in trade frequency, over time,implies an increase in the informativeness of trades, ceterisparibus.

Both Nofsinger and Sias (1999) and Sias, Starks, andTitman (2006) report a positive correlation between changesin institutional holdings and contemporaneous stock returns.If institutional buying affects the time variation of tradeinformativeness, then:

Hypothesis 5: An increase in price, over time, implies anincrease in the informativeness of trades, ceteris paribus.

The Kyle model (1985) suggests that a greater volatilitymay increase the profit of informed traders (therefore, it mayattract more informed trading). Similarly, a greater marketdepth may increase profit. Therefore, if informed tradingaffects the time variation of trade informativeness, then:

Hypothesis 6: An increase in the percent price range, overtime, implies an increase in the informativeness of trades,ceteris paribus.

Hypothesis 7: An increase in depth, over time, implies anincrease in the informativeness of trades, ceteris paribus.

Coughenour and Deli argue that “to the extent trading offof the NYSE represents purchased order flow and to theextent purchased order flow dries up during periods ofincreased informed trading, the percent of dollar volumeexecuted at the NYSE could reflect the degree of informedtrading.”4 Hence:

2 see Jorgensen and Wingender (2004) for the evidence on the reaction oflarge corporations to RFD.

3 see Conrad, Johnson, and Wahal (2002) for a review of the theory andevidence on institutional trading.

4 See Coughenour and Deli (2002) p. 857.

53RAKOTOMAVO — THE LONG-TERM VARIATION OF TRADE INFORMATIVENESS

Hypothesis 8: An increase in the percentage of NYSEexecutions, over time, implies an increase in theinformativeness of trades, ceteris paribus.

II. Data and Results

Our sample includes 2,296 quarterly observations of 82randomly chosen firms for which 1998-2004 NYSE TAQintraday trade (time, size and price) and quote (time, bid andask) data, as well as1998-2004 ThomsonFinancial institutionalshareholding data areavailable. For acomparison, Brooks(1996) used a randomsample of 90 dividend-paying stocks to studythe variation of tradei n f o r m a t i v e n e s saround earnings andd i v i d e n dannouncements. Morerecently, Al-Suhaibaniand Kryzanowski(2000) used a sampleof 56 stocks to study the informativeness of orders on theSaudi stock market.

A. Trade Informativeness

In Hasbrouck (1991), trades and price changes aremodeled in a vector autoregression (VAR):

rt = a1 rt-1 + … + a5 rt-5 + b0 xt + … + b5 xt-5 + v1t

xt = c1 rt-1 + … + c5 rt-5 + d1 xt-1 + … + d5 xt-5 + v2t , (1)

where rt is the mid-quote return (logarithm differentials),xt is the column vector of trade attributes, and t is the time ofa transaction or a change of quote. Three trade attributes areconsidered: the sign of the trade, the signed trade size, andthe signed squared trade size. A transaction that has a priceabove the prevailing quote midpoint is assigned a positivesign; the opposite holds for a negative sign. The return is setto zero if no quote revision follows a trade within 5 seconds.Transactions occurring within 5 seconds of each other withoutany intervening quote are aggregated. A quote posted withinless than 5 seconds prior to a trade is resequenced. A movingaverage representation of the VAR in equation (1) iscomputed as follows:

rt = v1,t + e2 v1,t-1 + … + e11 v1,t-10 + f1 v2,t

+ f2 v2,t-1 + … + f11 v2,t-10 . (2)

If w represents the innovation in the efficient price whichis assumed to evolve as a random walk, trade informativeness(TINFO) is the ratio vw,x/vw, where:

vw,x = (i=1,11 fi).var (v2,t). (i=1,11 fi T) , (3)

and

vw = vw,x + (1+i=2,11 ei)2.var (v1,t) . (4)

Therefore, TINFO is a ratiowhere the efficient pricevariance attributable to tradesis divided by the full efficientprice variance. Tradeinformativeness is estimatedfrom 1998-2004 NYSE TAQintraday data for each of the82 firms and each of the 28quarters between 1998 and2004, resulting in 2,296values.

B. Other Variables

Institutional holdings aremeasured as the percentages of shares outstanding held byfinancial institutions, and are computed for each of the 82firms and each of the 28 quarters between 1998 and 2004.The data are from the Thomson Financial base. The changein institutional holdings in quarter q for each stock is thedifference in institutional holdings between quarter q andquarter q-1.

Price, trade size, price range (as a percentage of theminimum price of the day), trade frequency, volume, thepercent of trades executed at NYSE, and depth are estimateddaily for each stock before the quartely averages (of the dailyvalues) are computed. This allows a comparison with thedaily averages reported in the literature. The data are fromthe NYSE TAQ intraday trade and quote database.

C. Sample Description and Preliminary Results

Figure 1 shows the evolution of the median values of thepreviously mentioned variables from 1998 to 2004.5 Twonotable patterns are the declining time trend for tradeinformativeness and the clear break in the depth data after2001 Q1. These, and other patterns, will be investigatedshortly. To aggregate values and test hypotheses, the sample

5These values are available at http://ac.aup.edu/~mrakotomavo/JAFdata.htm.

Trade informativeness is defined as thepercentage of efficient price variance thatis attributable to trades. For example, thispercentage for Boeing was, on average,26.22 between the first quarter of 1998and the last quarter of 2000. Therefore,about 26.22% of the public informationon Boeing stock came from transactionson its shares during that period.


Figu

re 1

. Sam

ple

Stat

istic

s B

etw

een

1998

And

200

4

TIN

FO is

trad

e in

form

ativ

enes

s. IH

LDG

is in

stitu

tiona

l hol

ding

s. D

IHLD

G is

cha

nge

in in

stitu

tiona

l hol

ding

s. TR

AD

EFR

EQ is

dai

ly tr

ade

freq

uenc

y.PC

NTN

YSE

EXEC

is th

e pe

rcen

t of t

rade

s ex

ecut

ed a

t NY

SE.

.0.1.2.3.4.5

199

8199

9200

0200

1200

2200

3200

4

Med

ian

of T

INF

O

.52

.56

.60

.64

.68

.72

199

8199

9200

0200

1200

2200

3200

4

Med

ian

of IH

LDG

-.06

-.04

-.02

.00

.02

.04

.06

.08

199

8199

9200

0200

1200

2200

3200

4

Med

ian

of D

IHLD

G

400

600

800

100

0

120

0

140

0

160

0

199

8199

9200

0200

1200

2200

3200

4

Med

ian

of T

RA

DE

SIZ

E

0

200

400

600

800

100

0

120

0

140

0

160

0

199

8199

9200

0200

1200

2200

3200

4

Med

ian

of T

RA

DE

FR

EQ

812

16

20

24

28

32

36

199

8199

9200

0200

1200

2200

3200

4

Med

ian

of D

EP

TH

1.2

0E

+07

1.4

0E

+07

1.6

0E

+07

1.8

0E

+07

2.0

0E

+07

2.2

0E

+07

2.4

0E

+07

2.6

0E

+07

199

8199

9200

0200

1200

2200

3200

4

Med

ian

of V

OLU

ME

.82

.83

.84

.85

.86

.87

.88

.89

199

8199

9200

0200

1200

2200

3200

4

Med

ian

of P

CN

TN

YS

EE

XE

C

16

20

24

28

32

36

40

44

199

8199

9200

0200

1200

2200

3200

4

Med

ian

of P

RIC

E

2.0

2.4

2.8

3.2

3.6

4.0

4.4

4.8

199

8199

9200

0200

1200

2200

3200

4

Med

ian

of R

AN

GE


is divided into subperiods. Regulation Fair Disclosure (RFD)became effective on October 23, 2000. The NYSE’sdecimalization was fully implemented on January 29, 2001.The Sarbanes-Oxley Act (SOA) was enacted on July 30,2002. Therefore, three periods are considered:

1.) 1998 Q1-2000 Q4 (P1), before decimal pricing, RFD,and SOA.

2.) 2001 Q1-2002 Q2(P2), after decimalpricing and RFD, butbefore SOA.

3.) 2002 Q3-2004 Q4(P3) after decimalpricing, RFD and SOA.The Jarque-Bera testindicates a significantdeparture from normalityof the panel data. Thisleads to the use ofnonparametric teststhroughout the paper.

Panel A of Table Icontains the periodicmedian values and PanelB contains the periodictime trend (rankcorrelation with a quarterindex running from 1 to28) for each variable. Thehypothesis of equality of median values across each pair ofperiods is tested, using the Wilcoxon/Mann-Whitney, MedianChi-square, Kruskal-Wallis, and Van der Waerden statistics.The hypothesis reports rejection when at least 3 out of the 4statistics are significant at the 10%, or lower significancelevel.

The median trade informativeness is 33.15% for P1. Thisvalue is comparable with Hasbrouck (1991)’s averageestimate of 34.3% for a sample of 177 firms on the NYSEfor 1989 Q1. Panel B suggests that trade informativenesshas not moved over time within P1. However,informativeness has decreased from 33.15% to 15.71% inP2, after both RFD and decimalization, and the difference issignificant. Furthermore, its time trend has changed frominsignificant in P1 to negative in P2. These results areconsistent with Collver (2007)’s finding of a significantdecrease in trade informativeness after both theimplementation of RFD and decimal pricing. They also agreewith Chakravarty, Van Ness, and Van Ness (2005) who finda reduction in dollar adverse selection after decimalization.The same pattern is observed in P3, after SOA is enacted:the median trade informativeness drops from 15.71% in P2

to 1.87% in P3, and the time trend is still negative in P3.Therefore, SOA may have lowered trade informativeness.No other study seems to be available on this result, althoughJain and Rezaee (2008) find an improvement in marketliquidity after SOA.

Institutional holdings have increased from a median of58.73% in P1 to 68.07% in P2. For comparison, Grinstein

and Michaely (2005)report increasing medianholdings for their sample,culminating at 57.78%for 1991-1996, beforeP1. Similarly, Amihudand Li (2006) have themedian value climbing at54.47% in 1998 (thebeginning of P1) for theirsample. The medianchange in these holdingsseems to have stayedconstant over time; inparticular, the P1 and P2values of .37% and .38%are comparable with the.35% average that Sias,Starks, and Titman(2006) report for their1979 Q4-2000 Q4sample.

The evidence in PanelsA and B suggests that trade size has decreased and tradefrequency has increased after RFD/decimalization, and afterSOA. The same time variations are observed within eachperiod; most notably, trade size was decreasing and tradefrequency was increasing over time before these events . TheP1 median trade size of 1296.51 shares per day is consistentwith the 1,500 and 1,596 group means reported inCoughenour and Deli (2002) for their September-November1997 sample. The P1 median trade frequency of 294.19transactions per day is consistent with the 282 average(116+153+13) reported in Table 4 of Chakravarty, Van Ness,and Van Ness (2005) for their January 1, 2001-January 26,2001 pre-decimalization sample. The same table providesevidence of an increase in the frequency of small trades thatis not offset by the observed decrease in the frequency ofmedium and large trades, after decimalization; the changesin trade size and frequency between P1 and P2, shown inFigure 1, are consistent with this evidence. However, anotherstudy documenting a decrease in trade size and an increasein trade frequency after SOA could not be found.

The P1 median depth of 23.84 round lots is smaller thanthe subsample averages of 36.32 and 38.30 in Table II of

This paper presents evidence suggestingthat the quarterly variation of theinformativeness of trades for NYSE-listedstocks between 1998 and 2004 was relatedto institutional buying, uninformedtrading, and informed trading. The resultsindicate a positive relation betweeninstitutional buying and tradeinformativeness before Regulation FairDisclosure and decimal pricing. Afterthese events, the evidence is consistent withboth a rise in uninformed trading and afall in informed trading.


Tabl

e I.

Des

crip

tive

Stat

istic

sTh

is ta

ble

show

s the

des

crip

tive

stat

istic

s on

the

data

use

d in

the

pape

r. Th

e sa

mpl

e co

ntai

ns 2

,296

qua

rterly

obs

erva

tions

, bet

wee

n 19

98 a

nd 2

004,

of 8

2 st

ocks

list

ed o

n th

eN

YSE

. Tra

de in

form

ativ

enes

s, is

the

perc

enta

ge o

f the

eff

icie

nt-p

rice

varia

nce

attri

buta

ble

to tr

ade

inno

vatio

ns. I

t is c

ompu

ted

by fo

llow

ing

Has

brou

ck (1

991)

and

est

imat

edfo

r eve

ry q

uarte

r bet

wee

n 19

98 a

nd 2

004

from

NY

SE T

AQ

intra

day

trade

and

quo

te d

ata.

The

inst

itutio

nal h

oldi

ngs l

evel

is th

e nu

mbe

r of s

hare

s hel

d by

fina

ncia

l ins

titut

ions

divi

ded

by th

e nu

mbe

r of

sha

res

outs

tand

ing

from

the

Thom

son

Fina

ncia

l dat

abas

e. T

he c

hang

e in

inst

itutio

nal h

oldi

ngs

in q

uarte

r q

for

each

sto

ck is

the

diffe

renc

e in

inst

itutio

nal h

oldi

ngs b

etw

een

quar

ter q

and

qua

rter q

-1. P

rice,

trad

e si

ze, p

rice

rang

e (a

s a p

erce

ntag

e of

the

min

imum

pric

e of

the

day)

, tra

de fr

eque

ncy,

vol

ume,

the

perc

ent

of tr

ades

exe

cute

d at

NY

SE, a

nd d

epth

are

est

imat

ed d

aily

, for

eac

h st

ock,

bef

ore

the

quar

tely

ave

rage

s ar

e co

mpu

ted.

The

dat

a ar

e fr

om th

e N

YSE

TA

Q in

trada

y tra

de a

ndqu

ote

data

base

. Pan

el A

use

s th

e W

ilcox

on/M

ann-

Whi

tney

, Med

ian

Chi

-squ

are,

Kru

skal

-Wal

lis, a

nd V

an d

er W

aerd

en te

sts

of e

qual

ity o

f m

edia

n va

lues

acr

oss

diff

eren

tpe

riods

. The

hyp

othe

sis o

f equ

ality

is re

ject

ed w

hen

at le

ast 3

out

of t

he 4

stat

istic

s are

sign

ifica

nt a

t the

10%

or l

ower

leve

l. Th

e tim

e in

dex

equa

ls 1

at t

he st

art o

f eac

h pe

riod.

The

rank

ing

of v

alue

s is r

epea

ted

over

all

firm

s. 19

98 Q

1-20

00 Q

4 (P

1) is

a p

re-d

ecim

al p

ricin

g/R

egul

atio

n Fa

ir D

iscl

osur

e, p

re-S

arba

nnes

/Oxl

ey p

erio

d. 2

001

Q1-

2002

Q2

(P2)

is a

pos

t-dec

imal

pric

ing/

Reg

ulat

ion

Fair

Dis

clos

ure,

pre

-Sar

bann

es/O

xley

per

iod.

200

2 Q

3-20

04 Q

4 (P

3) is

a p

ost-d

ecim

al p

ricin

g/R

egul

atio

n Fa

ir D

iscl

osur

e, p

ost-

Sarb

anne

s/O

xley

per

iod.

Pan

el A

. M

edia

n V

alue

s by

Per

iod

Pe

riod

T

rade

In

form

ativ

enes

s In

stit

utio

nal

Hol

ding

s

Cha

nge

in

Inst

itutio

nal

Hol

ding

s

Trad

e Si

ze

Tra

de

Fre

quen

cy

Dep

th

(Rou

nd

Lots

)

Vol

ume

($M

illio

ns)

NY

SE

Exec

utio

ns

Pric

e R

ange

(%

)

19

98 Q

1-20

00

Q4

0.33

15

0.58

73

0.00

53

1296

.51

294

.19

23.8

4 17

.06

0.86

53

33.8

4 3.

25

20

01 Q

1-20

02

Q2

0.15

71

0.61

49

0.00

37

935

.52

673

.06

11.5

6 18

.30

0.87

20

27.2

5 3.

01

20

02 Q

3-20

04

Q4

0.01

87

0.68

07

0.00

38

643

.12

1263

.79

8.9

5 19

.87

0.84

46

26.1

4 2.

74

Hyp

othe

sis

P1

=P2

No

No

Yes

N

o N

o N

o Y

es

Yes

N

o N

o P2

=P3

No

No

Yes

N

o N

o N

o N

o N

o Y

es

No

P1=P

3

N

o

N

o Y

es

No

No

No

No

No

No

No

Pa

nel B

. Ran

k C

orre

latio

n w

ith th

e Ti

me

Inde

x by

Per

iods

Peri

od

Tra

de

Info

rmat

iven

ess

Inst

itutio

nal

Hol

ding

s

Cha

nge

in

Inst

itutio

nal

Hol

ding

s

Trad

e Si

ze

Trad

e Fr

eque

ncy

Dep

th

Vol

ume

N

YSE

E

xecu

tion

s

Pric

e R

ange

19

98 Q

1-20

00

Q4

0.0

126

0.11

60**

* -0

.048

1 -0

.086

9***

0.

3377

***

0.3

345*

**

0.03

73

-0.0

908*

**

-0.3

512*

**

0.4

906*

**

2001

Q1-

2002

Q

2 -0

.258

1***

0.

0715

0

.128

2***

-0

.224

9***

0.

3623

***

-0.3

338*

**

0.06

08

-0.2

007*

**

0.0

651

-0.3

765*

**

2002

Q3-

2004

Q

4 -0

.209

7***

0.

3766

***

0.0

948*

**

-0.1

864*

**

0.38

91**

* 0

.078

3**

0.27

14**

* -0

.251

4***

0

.384

3***

-0

.571

3***

All

perio

ds

-0.7

210*

**

0.52

19**

* 0

.019

1 -0

.499

5***

0.

7862

***

-0.5

944*

**

0.27

36**

* -0

.232

7***

-0

.172

0***

-0

.221

9***

***S

igni

fican

t at t

he 0

.01

leve

l. *

*Sig

nific

ant a

t the

0.0

5 le

vel.


Coughenour and Deli (2002). However, the significant dropto a median of 11.56 in P2 is consistent with Bessembinder(2003)’s finding of a depth reduction after decimalization.The evidence in Panel A points to an increase of depth withinP1, followed by a decrease in P2. P3 shows a reversal ofthat trend, with an increase at a small rate.

The price range has increased over time within P1, forthis sample. Therefore, the P1 median range of 3.25% isconsistent with the 2.66% reported in Table II of Coughenourand Deli (2002) for 1997. The decrease in range shown inPanels A and B, between P1 and P2, is in step with thedecrease in return volatility documented by Bessembinder(2003) after decimalization. There seems to be a decreasealso after SOA.

The P1 volume of $17.06 million per day is higher thanthe group means of $11.99 million and $8.58 million foundin Table I of Coughenour and Deli (2002). The evidencesuggests that volume has not changed after RFD/decimalization, but has increased after SOA.

The median percent of trades executed at the NYSE in P1is 86.53%. This figure is comparable with 84.01% and85.05% subgroup averages for the 806 stocks sampled byCoughenour and Deli (2002). This percentage seems to havenot changed after RFD/decimalization, but it appears to havedecreased after SOA.

The P1 median price of $33.84 is similar to the subgroupaverages of $32.09 and $31.37 for the 806 stocks inCoughenour and Deli (2002). The median price level hasgone down from P1 to P2, but does not seem to have changedfrom P2 to P3.

D. Multivariate Results

Because of departure from normality, the rank, not thevalue, of trade informativeness is used in the multivariateanalyses that follow. Trade informativeness rank, which goesfrom 1 to 28, is computed for each stock. Ordered probitmodels are used throughout this section. All Jarque-Bera testson residuals cannot reject the hypothesis of normality.

The first model uses all observations and the followingexplanatory variables:

• a period 2 (post-RFD/decimalization, pre-SOA) dummyvariable.

• a period 3 (post-RFD/decimalization/SOA) dummyvariable.

• institutional holdings.• change in institutional holdings.• log(trade size).• log(trade frequency).• log(price).

• range.• Percent of NYSE trade executions.• log(depth).The use of log mirrors Coughenour and Deli (2002). The

results are shown in the first column of Table II. The P2variable has a negative and significant coefficient. Thisconfirms the univariate result showing a decrease in tradeinformativeness after RFD/decimalization. It also agrees withCollver (2007)’s finding of a significant decrease in tradeinformativeness and Chakravarty, Van Ness, and Van Ness(2005)’s finding of a reduction in dollar adverse selectionafter decimalization. The P3 variable has also a negative andsignificant coefficient, again confirming the univariate result.This seems to indicate a drop in trade informativeness afterthe enactment of the Sarbanes-Oxley Act. Change ininstitutional holdings has a positive and significantcoefficient. This supports the hypothesis that institutionalbuying leads to greater informativeness of trades. The Kyle(1985) finding that volatility is positively correlated withinformed trading profit is strengthened by the positivecoefficient for price range. The coefficient for trade frequencyis negative and significant. This result does not support theinstitutional order-breaking hypothesis, which predicted apositive relation. Instead, when combined with the decreasein trade informativeness and the rise in trade frequencydocumented for the whole period in Panel B of Table I, it isconsistent with a rise in uninformed trading between 1998and 2004. No other coefficient is statistically significant.

The previous results are further detailed by restricting theobservations to each of the three periods (and removing thedummy variables). The results are shown in the last threecolumns of Table II. For P1, the coefficients for both leveland change in institutional holdings are positive andsignificant. By hypotheses 1 and 2, this suggests that thetime variation of trade informativeness before RFD/decimalization has been related to institutional buying. ForP2, the negative coefficient for trade frequency, the rise intrade frequency (see Panel B, Table I), and the decrease intrade informativeness (see Panel B, Table I) suggest that arise in uninformed trading may have taken place after RFD/decimalization. The coefficients for price level, range, anddepth are positive and significant, indicating that the timevariation of trade informativeness may also be related to thetime variation of institutional trading (see hypothesis 5 onprice level) and informed trading (see hypotheses 6 and 7 onprice range and depth) after RFD/decimalization.Specifically, combined with the decline of both range anddepth within that period as shown in Panel B of Table I andthe decline in trade informativeness in P2, these results pointto a decline in informed trading. The evidence for P2 isconsistent with Chakravarty, Van Ness, and Van Ness (2005),


Table II. 1998-2004 Ordered Probit Analysis of Trade InformativenessThis table shows the results of ordered-probit models of trade informativeness against the variables listed below. The sample contains2,296 quarterly observations, between 1998 and 2004, of 82 stocks listed on the NYSE. Trade informativeness (TINFO), is the percentageof the efficient-price variance attributable to trade innovations. It is computed by following Hasbrouck (1991) and estimated for everyquarter between 1998 and 2004 from NYSE TAQ intraday trade and quote data. Price, trade size, price range (as a percentage of theminimum price of the day), trade frequency, the percent of trades executed at NYSE, and depth are estimated daily, for each stock, beforethe quartely averages are computed. The data are from the NYSE TAQ intraday trade and quote database. The institutional holdings levelis the number of shares held by financial institutions divided by the number of shares outstanding from the Thomson Financial database.The change in institutional holdings in quarter q for each stock is the difference in institutional holdings between quarter q and quarterq-1. Price, trade size, price range, trade frequency, volume, the percent of trades executed at NYSE, and depth are estimated daily, foreach stock, before the quartely averages are computed. The data are from the NYSE TAQ intraday trade and quote database. The rankingof values is repeated over all firms. 1998 Q1-2000 Q4 (Period 1) is a pre-decimal pricing/Regulation Fair Disclosure, pre-Sarbannes/Oxley period. 2001 Q1-2002 Q2 (Period 2) is a post-decimal pricing/Regulation Fair Disclosure, pre-Sarbannes/Oxley period. 2002 Q3-2004 Q4 (Period 3) is a post-decimal pricing/Regulation Fair Disclosure, post-Sarbannes/Oxley period. z-statistics are in parentheses.

***Significant at the 0.01 level. **Significant at the 0.05 level. *Significant at the 0.10 level.

Trade Informativeness Rank (All Periods)

Trade Informativeness Rank (Period 1)



Period 2 Dummy

Variable

-1.17***

(-14.22)

Period 3 Dummy

Variable

-2.21***

(-19.94)

Institutional Holdings -0.0335

(-0.2684)

0.5248***

(2.76)

-0.4406

(-1.38)

-0.7113***

(-3.02)

Change in Institutional

Holdings

0.8179**

(2.04)

0.9850*

(1.73)

-0.0511

(-0.0553)

0.8107

(1.12)

Ln(trade size) 0.0219

(0.6160)

-0.0166

(-0.3256)

0.1171

(1.47)

0.0145

(0.1761)

Ln(trade frequency) -0.0711**

(-2.21)

0.0285

(0.5340)

-0.2710***

(-3.73)

-0.0414

(-0.7020)

Ln(price) 0.0574

(1.10)

0.0855

(0.9125)

0.3328***

(2.93)

-0.1686**

(-2.07)

Range (%) 0.0242**

(2.39)

-0.0289

(-1.59)

0.0410*

(1.91)

0.0588***

(3.79)

NYSE Executions (%) 0.0081

(0.3259)

-0.0103

(-0.4097)

0.4141

(0.9124)

0.4749

(1.38)

Ln(depth) 0.0580

(1.22)

0.0373

(0.4941)

0.3214**

(2.52)

-0.1207

(-1.58)


who find a reduction in dollar adverse selection cost afterdecimalization on the NYSE, an increase in the frequencyof small trades (which they interpret to mean a greaterparticipation by retail customers), smaller decreases in thefrequency of medium and large trades (which they interpretto mean less institutional trading), and the strongest evidenceof a decrease in adverse selection cost for trades of mediumsize. They suggest that institutions trade less because of lowerliquidity supply (as evidenced by the smaller depths andsmaller limit-order sizes), which may explain the reductionin adverse selection cost.

For P3, the coefficient for price range is positive andsignificant, suggesting that the hypothesized link betweeninformed trading and trade informativeness exists. Since bothrange and trade informativeness decreased during this period(see Panel B, Table I), this result is consistent with a decreasein informed trading after SOA. Both coefficients for priceand institutional holdings are negative and significant. Thisresult is inconsistent with hypotheses 1 (positive institutionalholdings coefficient) and 5 (positive price coefficient), which

assume that institutional buying (including lagged herdbuying captured by the level of institutional holdings) drivesthe time variation of trade informativeness. Instead, sinceboth institutional holdings and price increased, while tradeinformativeness decreased, during that period, the negativecoefficients are consistent with 1.) a rise in uninformedtrading, and 2.) a relation between uninformed trading andlags of institutional buying after SOA. To investigate thesehypotheses, a probit model, where various lags of change ininstitutional holdings are included as explanatory variables,is estimated for each of the three periods. Only the significantlags are reported in Table III. The results suggest that tradeinformativeness was not affected by lags of institutionalbuying in the first two periods. However, after SOA, therewas a negative association between trade informativeness,and 1-quarter (and, at a weaker significance level, 2-quarter)lagged institutional buying. The negative coefficients supportthe hypotheses that there was a post-SOA increase inuniformed trading and that this uninformed trading wasrelated to the 1-quarter lag of institutional buying.

The Case of BoeingThe following discussion is only meant to illustrate, not prove, the main results of the paper.As previously mentioned, the quarterly informativeness of trades for Boeing’s stock averaged 26.22% in 1998 Q1-2000

Q4 (P1), before decimal pricing, Regulation Fair Disclosure (RFD), and the Sarbanes-Oxley Act (SOA), 16.77% in 2001Q1-2002 Q2 (P2), after decimal pricing and RFD, but before SOA, and 3.66% in 2002 Q3-2004 Q4 (P3) after decimalpricing, RFD and SOA. Therefore, trade informativeness for Boeing, like that of the average stock in this study, hasdecreased over time.

In P1, the level of net institutional buying of Boeing shares per quarter went from an average of 0.79% of sharesoutstanding in the first half of the period to 1.59% in the second half. Meanwhile, Boeing’s trade informativeness wentfrom 23.31% to 29.14%, thus illustrating the effect of institutional buying during this period.

After decimal pricing and RFD, in P2, informed trading of Boeing shares became relatively less attractive, as quoteddepth for Boeing stock went from an average of 17.33 round lots in the first half of P2, to 15.23 round lots in the secondhalf (and the average was 57.32 round lots in P1). At the same time, the average number of daily trades increased from2,471.15 to 3,366.26 (while the evidence indicates that trade size did not affect trade informativeness). As Boeing’s tradeinformativeness went from 21.68% to 11.86%, these figures imply that there was both an increase in the level of uninformedtrading and a decrease in the level of informed trading of Boeing shares during this period.

After SOA, in P3, the evidence indicates that the level of institutional holdings affected trade informativeness negatively.Since institutional holdings reflect past institutional buying, their level is a proxy for the amount of trades that are relatedto such past institutional actions, with a lag (the evidence in the paper points to a lag of one quarter). For Boeing, this levelwent from 61.65% in the first half of P3 to 63.17% in the second half. Since Boeing’s informativeness of trades droppedfrom 4.55% to 2.77%, the trading related to lagged institutional buying previously described was mostly uninformed.Concurrently, informed trading became relatively less attractive as the daily price range narrowed from 3.43% (of theminimum price of the day) to 2.25%, thus reducing the informed investor’s profit potential.


Table III. 1998-2004 Ordered Probit Relation Between Trade Informativeness and Lagged InstitutionalHoldings Changes

This table shows the results of ordered-probit models of trade informativeness against lags of change in institutional holdings. The datadefinitions in Table II apply here. z-statistics are in parentheses.




Lag 1 of the Change in

Institutional Holdings

-0.6821

(-1.04)

-0.7942

(-0.8589)

-1.91***

(-2.59)

Lag 2 of the Change in

Institutional Holdings

0.4435

(0.6939)

-0.9743

(-1.03)

-1.35*

(-1.81)

Ln(trade size) 0.0128

(0.2434)

0.0846

(1.13)

-0.0173

(-0.2122)

Ln(trade frequency) 0.0083

(0.1431)

-0.2713***

(-3.73)

-0.0627

(-1.07)

Ln(price) 0.1681*

(1.70)

0.3182***

(2.81)

-0.1793**

(-2.20)

Range (%) -0.0236

(-1.25)

0.0386*

(1.79)

0.0583***

(3.75)

NYSE Executions (%) -0.0118

(-0.4622)

0.1525

(0.3802)

0.1129

(0.3551)

Ln(depth) 0.0555

(0.6806)

0.3449***

(2.72)

-0.0820

(-1.09)

***Significant at the 0.01 level. **Significant at the 0.05 level. *Significant at the 0.10 level.

III. Conclusion

This paper presents evidence suggesting that the quarterlyvariation of the informativeness of trades for NYSE-listedstocks between 1998 and 2004 was related to institutionalbuying, uninformed trading, and informed trading. The resultsindicate a positive relation between institutional buying andtrade informativeness before Regulation Fair Disclosure anddecimal pricing. After these events, the evidence is consistentwith both a rise in uninformed trading and a fall in informed

trading. Similar evidence is found for the period followingthe enactment of the Sarbanes-Oxley Act (SOA). The decreasein informed trading may be a continuation of thedecimalization effect. However, there is evidence pointingto a relation between uninformed trading and institutionalbuying in the previous quarter, a phenomenon that does notseem to be present before SOA.


References

Al-Suhaibani, M. and L. Kryzanowski, 2000, “The InformationContent of Orders on the Saudi Stock Market,” Journal ofFinancial Research 23 (No.2), 145-156.

Amihud, Y. and K. Li, 2006, “The Declining Information Contentof Dividend Announcements and the Effects of InstitutionalHoldings,” Journal of Financial and Quantitative Analysis41(No.3), 637-660.

Bessembinder, H., 2003, “Trade Execution Costs and MarketQuality after Decimalization,” Journal of Financial andQuantitative Analysis 38(No .4), 747-777.

Brooks, R. M., 1996, “Changes in Asymmetric Information atEarnings and Dividend Announcements,” Journal of BusinessFinance and Accounting 23 (No. 3), 359-378.

Chakravarty, S., B.F. Van Ness, and R.A. Van Ness, 2005, “TheEffect of Decimalization on Trade Size and Adverse SelectionCosts,” Journal of Business Finance and Accounting 32 (No. 5,6), 1063-1081.

Chiyachantana, C.N., P.K. Jain, C. X.Jiang, and R. A. Wood, 2004,“International Evidence on Institutional Trading Behavior andPrice Impact,” Journal of Finance 59 (No. 2), 869-898.

Collver, C. D., 2007, “Is There Less Informed Trading AfterRegulation Fair Disclosure?” Journal of Corporate Finance 13(No. 2), 270-281.

Conrad, J.. K. M. Johnson and S. Wahal, 2002, “The Trading ofInstitutional Investors: Theory and Evidence,” Journal of AppliedFinance 12 (No. 1), 7-14.

Coughenour, J. F. and D. N. Deli, 2002, “Liquidity Provision andthe Organizational Form of NYSE Specialist Firms,” Journal ofFinance 57 (No. 2), 841-869.

Garrett, I. and R. Priestley, 2000, “Dividend Behavior and DividendSignaling,” Journal of Financial and Quantitative Analysis 35(No. 2), 173-189.

Golbe, D. L. and M. S. Shranz, 1994, “Bidder Incentives forInformed Trading Before Hostile Tender Offer Announcements,”Financial Management 23 (No. 4), 57-68.

Hasbrouck, J., 1991, “The Summary Informativeness of StockTrades: an Econometric Analysis,” Review of Financial Studies4 (No. 3), 571 605.

Jain, P.K. and Z. Rezaee, 2006, “The Sarbanes-Oxley Act of 2002and Capital-Market Behavior: Early Evidence,” ContemporaryAccounting Research 23 (No. 3), 629-654.

Jain, P.K., J.C. Kim, and Z. Rezaee, 2008, “The Sarbanes-OxleyAct of 2002 and Market Liquidity,” Financial Review 43 (No.3), 361-382.

Jorgensen, R D. and J.R. Wingender, Jr., 2004, “A Survey on theDissemination of Earnings Information by Large Firms,” Journalof Applied Finance 14 (No. 1), 77-84.

Karpoff, J.M. and D. Lee, 1991, “Insider Trading Before New IssueAnnouncements,” Financial Management 20 (No. 1), 18-26.

Kyle, A.S., 1985, “Continuous Auctions and Insider Trading,”Econometrica 53 (No. 6), 1315-1335.

Lee, C. M.C. and M.J. Ready, 1991, “Inferring Trade Directionfrom Intradaily Data,” Journal of Finance 46 (No. 2), 733-746.

Nissim, D. and A. Ziv, 2001, “Dividend Changes and FutureProfitability,” Journal of Finance 56 (No. 6), 2111-2133.

Nofsinger, J.R. and R.W. Sias, 1999, “Herding and FeedbackTrading by Institutional and Individual Investors,” Journal ofFinance 54 (No. 6), 2263-2295.

Sias, R.W., L.T. Starks, and S. Titman, 2006, “Changes inInstitutional Ownership and Stock Returns: Assessment andMethodology,” Journal of Business 79 (No. 6), 2869-2910.

Whitcomb, D.K., 2003, “Applied Market Microstructure,” Journalof Applied Finance 13 (No. 2), 77-80.

Zhao, X. and K.H. Chung, 2006, “Decimal Pricing and Information-Based Trading: Tick Size and Informational Efficiency of AssetPrice,” Journal of Business Finance and Accounting 33 (No.5,6), 753-766.


Shareholder Theory – HowOpponents and Proponents Both

Get It Wrong

Morris G. Danielson, Jean L. Heck, and David R. Shaffer

Shareholder wealth maximization is accepted by mostfinancial economists as the appropriate objective for financialdecision-making. Recently, wealth maximization has beencriticized by a growing array of opponents for condoningthe exploitation of employees, customers, and otherstakeholders, and encouraging short-term managerialthinking. Although these critics are misguided, proponentsof shareholder theory have helped to create this confusion byexhorting managers to maximize the firm’s current stock price.Because a firm’s stock price can be manipulated in the short-term, incentives to increase a firm’s current stock price candistort operating and investment decisions. When wealthmaximization is properly defined as a long-term goal, it isnot as narrowly focused as critics believe. The mainprescription of shareholder theory—invest in all positive netpresent value projects—benefits not only shareholders, butalso key stakeholders including employees and customers.

is a fundamental building block of corporate financial theory.However, the shareholder model has been criticized forencouraging short-term managerial thinking and condoningunethical behavior. Smith (2003) notes that critics believeshareholder theory is “. . . geared toward short-term profitmaximization at the expense of the long run.”1 Freeman,Wicks, and Parmar (2004) assert that shareholder theory “. .. involves using the prima facie rights claims of one group—shareholders—to excuse violating the rights of others.”

This paper explains why such critiques of shareholdertheory are misguided yet understandable. They are misguidedbecause wealth maximization is inherently a long term goal—the firm must maximize the value of all future cash flows—and does not condone the exploitation of other stakeholders(Jensen, 2002; Sundaram and Inkpen, 2004a). The criticismsare understandable because many proponents of shareholdertheory, in a stylized version of the model, exhort managers tomaximize the firm’s current stock price (Keown, Martin, andPetty, 2008; Lasher 2008; Ross, Westerfield, and Jordan,2008; Brealey, Myers, and Marcus, 2007; Melicher andNorton, 2007). This notion underlies the formal (e.g., stockoptions) and informal (e.g., pressure from the investmentcommunity and corporate boards) incentives that rewardmanagers if a firm’s stock price continually increases.2 By

1For example, Freeman, Wicks, and Parmar (2004) criticize managers forpursuing policies designed to continually increase a firm’s stock price.Fuller and Jensen (2002) criticize mangers for focusing undue attentionon whether a firm meets analyst earnings forecasts each quarter, to avoidstock price declines.Morris G. Danielson is an Associate Professor of Finance at Saint Joseph’s

University in Philadelphia, PA. Jean L. Heck is an Associate Professor ofFinance at Saint Joseph’s University in Philadelphia, PA. David R. Shafferis an Associate Professor of Finance at Villanova University in Villanova,PA.

Danielson and Heck gratefully acknowledge financial support from thePedro Arupe Center for Business Ethics at Saint Joseph’s University.

62

Shareholder theory defines the primary duty of a firm’smanagers as the maximization of shareholder wealth (Berleand Means, 1932; Friedman, 1962). The theory enjoyswidespread support in the academic finance community and

2Although incentive stock options typically vest over several years andcan have long maturities, the presence of stock options also encouragesmanagers to pursue policies designed to increase the stock price in theshort-term (especially as the expiration date approaches). Danielson andPress (2006) argue that these incentives can create agency costs wheneverthe stock price falls below the option exercise price.

63DANIELSON, HECK, AND SHAFFER — SHAREHOLDER THEORY

focusing on the current stock price, which can be manipulatedin the short-term by unscrupulous managers, proponents ofshareholder theory open up the model to criticism.

Opponents of shareholder theory often recommend thatfirms balance the interests of shareholders against those ofemployees, customers, and other stakeholders when makingbusiness decisions (Freeman, 1984). However, unless theinterests of future stakeholders are explicitly considered, thestakeholder model can lead to the same type of short-termthinking that shareholder theory has been accused ofencouraging. Indeed, the shareholder model, when viewedfrom a long-term perspective, provides a better frameworkthan stakeholder theory in which to protect the interests ofboth current and future stakeholders. Thus, stakeholder theoryis not superior to shareholder theory from an ethicalperspective.

I. Should Firms Maximize the CurrentStock Price?

In the shareholder model, the goal of the firm is to maximizethe present value of future cash flows. If the cash flow a firmis expected to pay shareholders (in the form of dividends orstock repurchases) in year n is CFn, and the required returnon equity is r, the intrinsic (per share) value of the firm’sequity today (V0) is defined by Equation (1).

VCF

r

nn

n0

1 1=

+( )=

∞

∑ . (1)

To maximize the value of Equation (1), managers shouldinvest in all positive net present value (NPV) projects (Brealeyand Myers, 2003). The right-hand side of Equation (1)highlights the long-term nature of this goal: shareholderwealth depends on the firm’s cash flows in all future years.3

The shareholder model is difficult to implement becausethe estimated cash flow stream on the right-hand side ofEquation (1) cannot be observed. Thus, proponents ofshareholder theory often assert that a firm’s current stock price(P0) equals its intrinsic value (V0) and instruct managers tomaximize the firm’s current stock price. This is the stylizedform of the shareholder model.

Although shareholder theory directs managers to maximizeshareholder wealth, managers face formal and informalincentives to increase the firm’s current stock price. Forexample, incentive stock options will provide a positivepayoff to managers only if the firm’s stock price increasesfrom the grant date level. In addition, some managers facepressure from corporate boards and the investmentcommunity to continually increase firm value (Jensen, 2005).However, maximizing and increasing shareholder wealth aretwo very different objectives. If the business conditions facinga firm change unfavorably (through perhaps no fault ofmanagement), a firm’s maximum possible value can decrease.This is not an unusual or unlikely occurrence; Jensen (2005)notes that future events could reveal that the stock prices ofperhaps 50% of all firms are too high (because a stock priceis a function of a distribution of possible outcomes).

As the business conditions facing a firm change, a firm’sstock price can diverge from its intrinsic value becauseinformation is not instantaneously and continuouslycommunicated to the market. If business conditions changeunfavorably, P0 will exceed V0 and the stock will be(temporarily) overvalued.4 To implement the shareholdermodel correctly, the firm should continue to invest in allpositive NPV projects (which are now less valuable than themarket originally expected), and the stock price willeventually decrease to the new intrinsic value. However, ifmanagers (who will typically know that business conditionshave changed before the rest of the market) are incentivizedto increase the stock price, Jensen (2005) and Danielson andPress (2006) argue that efforts to further inflate (or tomaintain) the stock price may destroy long-term value. Theseactions could include delaying new investments (even if theNPV is positive), reducing discretionary spending (e.g.,advertising, R&D, maintenance, quality control, etc.),accounting manipulation, or adopting fraudulent businesspractices.

Jensen (2005) uses Enron to illustrate the agency costs ofovervalued equity. At Enron’s peak market value of $70billion, Jensen estimates the company was only worth $30billion. He notes that Enron’s managers tried to justify theexcess valuation of $40 billion by “. . . trying to fool themarkets through accounting manipulations, hiding debtthrough off-balance sheet partnerships, and over hyped newventures such as their broadband futures effort.” Clearly,these efforts were not designed with the long-term interestsof the firm in mind, and they did not pay off for Enron’sshareholders. Thus, the case of Enron does not provideevidence against shareholder theory. But this experience does

3A large portion of shareholder wealth is often tied to cash flows to bereceived in the distant future. For example, if the firm is expected to paya $1 dividend next year, and the dividend is expected to grow at a 4% rateper year (forever), the stock price today is $25 if the required return is 8%(= $1/(0.08 – 0.04)). In this example, dividends during the next 10 yearsonly account for 31.4% of the stock price (= $7.86/25), leaving 68.6% ofthe value to be realized in years 11 through infinity. Clearly, shareholderwealth maximization is not a short-term goal.

4Although deviations between P0 and V0 can arise in the short-term evenin efficient markets, evidence in Summers (1986) and Cornell (2001)suggest that such deviations can persist for prolonged periods.


show that efforts to increase a firm’s current stock price canbe harmful if these policies are detached from strategiesdesigned to maximize the firm’s long term cash flows.

II. Does Stakeholder Theory Promote aLong-Term Focus?

Because of the perceiveddeficiencies of shareholdertheory, stakeholder theoryhas gained popularity inrecent years and is now usedto guide the businessdecisions of a wide range offirms (Donaldson andPreston, 1995; Jorg,Loderer, and Roth, 2004;and Kaler, 2006). One of thegoals of stakeholder theoryis to promote “anenhancement of distributivejustice within the confines ofa basically capitaliststructure . . . .” (Kaler, 2006).Along these lines, the 1988Sloan Colloquy in its “Consensus Statement on StakeholderModel of the Corporation” recommends that firms “attemptto distribute the benefits of their activities as equitably aspossible among stakeholders, in light of their respectivecontributions, costs, and risks.”5 To do this, Blair and Stout(1999) argue that the board of directors should split a firm’seconomic surplus (i.e., investment returns in excess of therisk-adjusted cost of capital) between shareholders,employees, customers, and other stakeholders.

If a firm is forced to allocate a portion of its economicsurplus to employees (by paying wages in excess of theemployees’ marginal productivity) or to customers (byreducing prices), these stakeholders will benefit in the short-term. However, these policies could stifle future innovation,hurting shareholders, stakeholders, and society in the long-run. For example, US employees in the steel industry, theauto industry, and the airline industry benefited in the short-term from lucrative union contracts negotiated in the latterhalf of the twentieth century. But these contracts ultimatelycontributed to financial difficulties at the firms, reducing jobsecurity and compensation for today’s employees. Similarly,the current customers of pharmaceutical companies wouldbenefit greatly if patent laws were revoked, and all drugswere then sold at a price equal to production costs plus, for

example, 10%. However, this policy would reduce both thefunds available to invest in research and development andthe incentive for firms to do so. Thus, future customers wouldnot benefit from potential life-saving products that mightotherwise have been developed.

The following example illustrates the potential problem.Assume that a firm operates in a simple one-period world.The entrepreneur invested $100 in the firm at t = 0, and the

firm produces a cash flow of$160 at t = 1. If the requiredreturn is 10%, the economicsurplus of the firm is $50 (=$160 – $100(1.10)). Becausethe firm has a realizedinvestment return of 60%,stakeholder advocates mightargue that the shareholders’profits are excessive. Fromtheir perspective, an equitabledistribution of the economicsurplus might be to increasewages or decrease prices,reducing the investmentreturn toward the requiredreturn of 10%. But, thisoutcome would not be fair to

the entrepreneur unless the policy were known before theinvestment decision was made.

Most investments offer risky outcomes; it is likely that theentrepreneur did not know with certainty that the project’s t= 1 payoff would be $160 when the initial $100 investmentwas made at t =0. Assume that at time t = 0, the investmenthad an equal 50% probability of paying either $160 or $60 att = 1. If so, the expected payoff at t = 1 was $110, and theproject had a net present value of 0 (= 110/1.10 – 100). Onan ex-ante basis the project was acceptable, but it did notcreate an economic surplus.

Once the future outcome is revealed, it would not be ethicalto change the rules of the game and split the excess return($160 – $110) between shareholders and other stakeholders.If the entrepreneur had known at t = 0 that the project wouldonly yield, for example, $150 in the good outcome, theentrepreneur would not have made the $100 investment.Thus, proposals to split the realized economic surplus amongvarious stakeholder constituencies have the potential forreducing future investment, harming society (and potentialfuture stakeholders) in the long run.

Stakeholder theory, of course, does not advocate that firmsbe managed in the interests of current stakeholders at theexpense of future ones. Instead, Freeman (1994) recommendsthat a corporation “. . . be managed as if it can continue toserve the interests of stakeholders through time.” Similarly,

If a firm is forced to allocate a portionof its economic surplus to employees(by paying wages in excess of theemployees’ marginal productivity) orto customers (by reducing prices),these stakeholders will benefit in theshort-term. However, these policiescould stifle future innovation, hurtingshareholders, stakeholders, andsociety in the long-run.

5 This statement is reprinted in the appendix to Marcoux (2000).

65DANIELSON, HECK, AND SHAFFER — SHAREHOLDER THEORY

DesJardins and McCall (2005) argue that a corporation shouldbe managed as a social institution, providing benefits tostakeholders both now and in the future.

However, the question of how a manager might balancethe interests of current and future stakeholders has receivedvery little attention in the stakeholder literature. One notableexception is Mitchell, Agle, and Wood (1997), who arguethat managers should consider the urgency of variousstakeholder claims when making decisions. But this approachwould encourage managers to adopt a short-term focus whenimplementing stakeholder theory: the needs and requirementsof current stakeholders will always be more “urgent” thanthose of future stakeholders.

III. The Shareholder Model and Long-TermStakeholder Interests

One drawback of stakeholder theory is that the identity ofthe individual stakeholders is constantly changing. Thus, thecustomer or employee who extracts excess benefits from afirm during the current period is not the same person wholoses future benefits. The identity of shareholders will alsochange over time, but there is a key difference. A large portionof any investor’s return (even a short-term trader) will dependon the firm’s stock price on the date of sale. Because aninvestor must find a person who believes the firm will producesufficient cash flows to justify the prevailing market price,shareholder wealth maximization (when defined properly asa function of all future cash flows) is inherently a long-termgoal. And, because a firm must continue creating value foremployees and customers to generate future cash flows, themaximization of a firm’s long-term cash flow stream shouldnot harm the firm’s stakeholders. Indeed, the interests offuture stakeholders can only be satisfied if the firm remainsfinancially strong.

IV. Conclusion

In the aftermath of financial scandals at Enron, Worldcom,and Global Crossing, shareholder theory faces increasedscrutiny and criticism. As stated by Freeman, Wicks, and

Parmar (2004), “It is hard to imagine how anyone can look atthe recent wave of business scandals, all of which are orientedtoward ever increasing shareholder value at the expense ofother stakeholders, and argue that this philosophy is a goodidea.” However, proponents of shareholder theory point outthat policies adopted by Enron, Worldcom, and GlobalCrossing clearly did not benefit the firms’ shareholders inthe long-run, and thus are not evidence against shareholdertheory (Sundaram and Inkpen, 2004b).

Before dismissing critics of shareholder theory outright, itis important to recognize that supporters of shareholder theoryoften emphasize the model’s short-term implications whendefining the theory. Indeed, many leading finance texts equateshareholder theory with the maximization of a firm’s currentstock price, and executive compensation (e.g., incentive stockoptions) frequently rewards managers for increasing the stockprice. Thus, it should not be surprising that some critics ofthe shareholder theory might (incorrectly) view it as being ashort-term goal.

We disagree, however, with those who would use thedeficiencies of the stylized model as a reason to abandonshareholder theory in favor of stakeholder theory. Despiteits current popularity, stakeholder theory provides littleguidance about how to balance the often competing interestsof various stakeholder groups (Marcoux, 2000; Jensen, 2002).In addition, stakeholder theory can encourage managers toadopt a short-term focus (much like the stylized version ofthe shareholder model) to the detriment of a firm’s long-termhealth.

The shareholder model—when viewed from a long termperspective—still provides the best framework in which tobalance the competing interests of various stakeholders(including both current and future stakeholders) when makingbusiness decisions. However, proponents of shareholdertheory must recognize that it matters how the theory is definedand implemented. In particular, the goal of financial managersshould be to invest in all positive net present value projects,regardless of whether these decisions will cause an immediateincrease in the firm’s stock price. To focus managerialattention on this goal, corporate incentive structures shouldreward managers for maximizing a firm’s value in the longrun rather than increasing its stock price in the short term.


References

Berle, A.A. and G.C. Means, 1932, The Modern Corporationand Private Property, Macmillan, New York.

Blair, M.M. and L.A. Stout, 1999, “A Team ProductionTheory of Corporate Law,” Virginia Law Review 85 (No.4),247-328.

Brealey, R.A. and S.C. Myers, 2003, Principles of CorporateFinance, McGraw-Hill/Irwin, New York..

Brealey, R.A., S.C. Myers, and A.J. Marcus, 2007,Fundamentals of Corporate Finance, McGraw-Hill/Irwin,New York.

Cornell, B., 2001, “Is the Response of Analysts to InformationConsistent with Fundamental Valuation? The Case ofIntel,” Financial Management 30 (No.1), 113-136.

DesJardins, J.R. and J.J. McCall, 2005, “The Corporationas a Social Institution,” In Contemporary Issues in BusinessEthics, 5th edition, Edited by J.R. DesJardins and J.J.McCall, Wadworth, Belmont, CA.

Danielson, M. and E. Press, 2006, “Do Stock Options AlwaysAlign Manager and Shareholders’ Interests? An AlternativePerspective,” Advances in Financial Education 4 (No.2),1-16.

Donaldson, T. and L.E. Preston, 1995, “The StakeholderTheory of the Corporation: Concepts, Evidence, andImplications,” Academy of Management Review 20 (No.1),65–91.

Freeman, R.E., 1984, Strategic Management: A StakeholderApproach, Pitman, Boston.

Freeman, R.E., 1994, “The Politics of Stakeholder Theory:Some Future Directions,” Business Ethics Quarterly 4(No.4), 409–421.

Freeman, R. E., A. C. Wicks, and B. Parmar, 2004,“Stakeholder Theory and ‘The Corporate ObjectiveRevisited’,” Organization Science 15 (No.3), 364–369.

Friedman, M., 1962, Capitalism and Freedom, Universityof Chicago Press, Chicago.

Fuller, J. and M. Jensen, 2002, “Just Say No to Wall Street:Putting A Stop to the Earnings Game,” Journal of AppliedCorporate Finance 14 (No.4), 41-46.

Jensen, M.C., 2002, “Value Maximization, StakeholderTheory, and the Corporate Objective Function,” BusinessEthics Quarterly 12 (No.2), 235-256.

Jensen, M.C., 2005, “Agency Costs of Overvalued Equity,”Financial Management 34 (No.1), 5-19.

Jorg, P., C. Loderer, and L. Roth, 2004, “Shareholder ValueMaximization: What Managers Say and What they Do,”DBW Die Betriebswirtschaft 64 (No. 3), 357-378.

Kaler, J., 2006, “Evaluating Stakeholder Theory,” Journalof Business Ethics 69 (No.2), 249–268.

Keown, A.J., J.D. Martin, and J.W. Petty, 2008, Foundationsof Finance, Pearson Prentice Hall, Upper Saddle River.

Lasher, W.R., 2008, Practical Financial Management,Thomson South-Western, Mason.

Marcoux, A.M., 2000, “Balancing Act,” In ContemporaryIssues in Business Ethics, 4th edition, edited by J.R.DesJardins and J.J. McCall, Wadworth, Belmont, CA.

Melicher, R.W. and E.A. Norton, 2007, Introduction toFinance, John Wiley and Sons, Hoboken.

Mitchell, R.K., B.R. Agle, and D.J. Wood, 1997, “Toward aTheory of Stakeholder Identification and Influence:Defining the Principle of Who and What Really Counts,”Academy of Management Review 22 (No.4), 853-886.

Ross, S.A., R.W. Westerfield, and B.D. Jordan, 2008,Fundamentals of Corporate Finance, McGraw-Hill/Irwin,New York.

Smith, H.J., 2003, “The Shareholders vs. StakeholdersDebate,” MIT Sloan Management Review 44 (No. 4), 85–90.

Summers, L.H., 1986, “Does the Stock Market RationallyReflect Fundamental Values?” Journal of Finance 41 (No.3), 591-601.

Sundaram, A. and A. Inkpen, 2004a., “The CorporateObjective Revisited,” Organization Science 15 (No.3),350–363.

Sundaram, A. and A. Inkpen, 2004b., “Stakeholder Theoryand ‘The Corporate Objective Revisited’: A Reply,”Organization Science 15 (No.3), 370–371.

Student Managed InvestmentFunds: An International Perspective

Edward C. Lawrence

The most comprehensive survey ever conducted on studentmanaged investment funds shows there are now 314universities worldwide that offer students the chance to learnabout portfolio management by investing real money. Inaggregate, students are directly managing more than $407million in assets in 2007. Most of these programs supplementthe more traditional investment courses, which are offered byevery institution with a business college. Over the last twodecades, student-managed investment funds have grown inboth size and complexity as universities have tried to mirrorreal world experiences. The career success of students comingout of these programs demonstrates the benefits of providingstudents with as much hands-on experience as possible. Thispaper should be of interest to faculty, students, employers,and practitioners in the financial community who desire basicknowledge about state-of the-art teaching investments andportfolio management.

Edward C. Lawrence is a Professor of Finance and Department Chair atthe University of Missouri - St. Louis in St. Louis, MO.

The author would like to acknowledge the assistance of Kerry Sallee,Ken Locke, Karen Wagster, Anthony Lerro, Brian Bruce, Larry Belcher,and all of the university faculty participants who generously gave theirtime to complete this survey.

In the early 1970s, there was a strong movement inWestern countries for universities to start providing studentswith both academic knowledge and the ability to apply newskills on the job. Employers were often critical of newgraduates who had difficulty stepping into employmentwithout first receiving extensive on-the-job training. To

overcome this obstacle, many business colleges beganpartnerships with major companies to offer students co-opand internship programs while the students were still pursuingtheir degrees. Deans also started encouraging their faculty toinvite more guest speakers from government and industry toaddress classes on issues of the day. It also became commonfor professors to take classes on field trips to local employersto gain greater insights as to what it was like to work in aparticular field. Finally, with the development of computers,interactive software became more prevalent, allowing studentsto simulate starting a new company, managing a bank, orinvesting in the stock market with play money.1 Although allof these approaches were a significant improvement over whateducational institutions had done historically, there was stilla need to offer students even greater realism and morepractical experience.2

In the field of finance, student managed investment funds(SMIFs) were created to take investment education to thenext level. These funds allow students to invest real moneyin the stock and bond markets. The vast majority of SMIFshave close faculty involvement to provide oversight andstructure to student activities. Nevertheless, students aregenerally responsible for making all investment decisions and

1Most basic investment courses today use simulations or play money toallow students some practice with security analysis, stock selection,portfolio composition and market timing. However, it is well recognizedthat play money often leads to excessive risk taking as students try tooutperform the market without realistic penalties.

2 Even now, Pfeffer (2007) still argues that business schools are still notdoing enough to ensure students can “translate business knowledge intoapplicable business skills” in real world situations. However, he does notspecifically address the effectiveness of many of the initiatives mentionedabove.

67


managing the portfolio. Some funds have outsideprofessionals that serve in an advisory capacity to enrich theexperience for students and showcase their programs. WhileSMIFs were first started in the US and have been aroundsince 1950, only 12 colleges had them by 1969. Unfortunately,those programs were not widely known outside of thosecampuses. Today, there are 314 funds worldwide ranging insize from $2,000 to $62 million. In aggregate, students aredirectly managing $407 million in assets. The purpose of thispaper is to discuss the evolution of SMIFs and the impactthey are having on teaching investments around the globe.Surprisingly, there has not been a single paper presenting dataon funds from outside the United States.

There is much to be gained within academia and theinvestment community through the sharing of ideas andinformation on financial education in a variety of culturesand environments. Not only do new and existing SMIFs learnfrom the innovations of other successful programs, facultyand administrators benefit from considering a broader arrayof approaches to solving specific constraints faced by aparticular school. Students desiring to embark on investmentcareers will be able to more fully evaluate the differentprograms being offered and select the one that most closelyfits their learning style. Finally, employers and practitionersalso need to become more knowledgeable of the various typesof SMIFs. Besides the opportunity to hire highly trainedstudents coming out of these programs, professionals shouldbecome educational partners by providing guest speakers,serving on boards, providing funding, sharing technicalresources, etc.

I. Previous Studies

One of the primary reasons student investment funds wereslow to be established in the 1970s was the lack oforganizational information and data on the benefits and costsof these programs. Until 1990, there was not even a list ofuniversities in the US that had such funds. In fact, many ofthe faculty who were closely involved in SMIFs prior to thistime had limited knowledge of other programs and almost nocommunication with their colleagues. As a result, it was verydifficult for finance faculty to start new funds given the lackof operational data and instructional inexperience with suchprograms. Recognizing this problem, Lawrence (1990)conducted the first survey to profile and discuss thecharacteristics of almost two dozen established programs.Until then, it was common for paper authors to only describea single fund.3 While these efforts were insightful, it was

impossible to fully understand the scope of this movement ininvestment education without a broader database.Furthermore, many senior university officials were stillreluctant to commit their scarce resources in such fundswithout convincing hard data showing the clear benefits andcosts of such programs. By the early 1990s, with so manyleading business schools embracing the basic concept, itbecame an “easy sell” for finance faculty and alumni in NorthAmerica. This led to an explosion of programs, which havespread to other continents including Asia and Europe.

In 1994, Lawrence expanded his study to include 34programs in order to better describe their operations andfunding sources. Johnson, Alexander and Allen (1996)investigated alternative decision making environments instudent managed funds. By 2003, Neely and Cooley (2004)reported 134 funds had been established in the US alone.Ammermann and Runyon (2003) investigated risk aversionand group dynamics among students making portfoliodecisions at California State University in Long Beach. Allof these papers served as a major catalyst for the rapid growthin the number and size of SMIFs worldwide, especially inNorth America.

II. The Survey

From June 2007 to April 2008, a written survey waselectronically sent to all universities in the US and abroadwith both known and possible SMIFs.4 For US participants,survey participants were also asked to share their knowledgeof other programs in their states. In the case of foreigncountries, participants were solicited for information onexisting and potential funds in their own country orneighboring countries. It was assumed that faculty involvedwith current programs would most likely know of other SMIFsfrom their professional contacts at conferences. Since locatingforeign programs would be more challenging, the meetingroster of attendees at the 2007 Financial ManagementAssociation meeting in Florida was used to contact a largenumber of faculty from South America, Europe, Australiaand Asia. This database was supplemented by the author whoattended major academic conferences in both Europe and Asiaduring this time period in order to make personal contactwith other finance faculty who may have had knowledge offoreign funds. Finally, a significant number of contacts weremade with finance department chairs and deans at major

3Some of the earliest case studies included Belt (1975), Hirt (1977), Bearand Boyd (1984), Markese (1984), Kester (1986), Tatar (1987), Blockand French (1991), Bhattacharya and McClung (1994) and Kahl (1997).

4Anthony Lerro of the Association of Student Managed Investment Funds(ASMIF) graciously provided an initial list of American universities thatwere members of the association. While the list provided a good startingpoint for the US, it also contained a fair number of schools that did not yethave a program and was missing a large number of other universities thatdid.

69LAWRENCE — STUDENT MANAGED INVESTMENT FUNDS

business schools (not included in any of the previouslymentioned screens) in Europe and Asia to make sure therewas as much international exposure as possible, given theobvious language barriers.

This protocol resulted in locating 314 SMIFs from aroundthe world. Appendices A and B provide a complete list of allUS and non-US programs, respectively. Each program wascontacted up to 7 times byemail and/or phone toencourage their participationin the survey. With facultyadvisors frequently rotatingin and out of the programs, itwas sometimes very difficultto find the appropriate personwith sufficient knowledge tocomplete the 48 questionsurvey. In addition, a smallnumber of the funds operatemore like investment clubswith little or no faculty involvement. Yet, these funds givethe students much of the same experience of investing realmoney, but as an extracurricular activity. Of the 314 funds,224 programs returned the completed survey for a 71%participation rate. For the remaining 90 SMIFs who declinedto participate, summary information about their programs wasobtained from external sources including the university’s website and media sources.

III. The Growth and Size of Programs

After 40 years of very slow growth, the number of SMIFsin the US exploded in the 1990s and 2000s as real moneyfunds began to supplement the more traditional methods ofteaching investments. As reported in Table I, the 1990s wasthe turning point. To remain competitive in the marketplace,most business schools had to offer students the opportunityto invest in the stock market with real money.5 The currentdecade has experienced the highest number of new programscreated despite the data only including seven years. With1,680 business schools in the US according to the AACSB,the country may be a long way from reaching a saturationpoint. However, with the exception of Canada, SMIFs arejust in the early stages of development in the rest of the world.The first non-US fund was established at the University ofBritish Columbia in 1987. Canada actually has a higherconcentration of funds within institutions of higher learning

than in the US. In other parts of the world, it could easilytake another 20 years to catch up to North America.

The size of SMIFs today has expanded beyond what manypeople would have thought possible only a decade ago. Thereare 78 universities worldwide with more than $1 million undermanagement by students. As reported in Table II, the largestfund is at the University of Wisconsin - Madison which has

$62 million being invested insome form by students.6 Thereare 8 SMIFs with more than$10 million including 2 funds,Ohio State and the Universityof Minnesota, with $25 millioneach in assets.7 The largestnon-US fund is Canada’sSimon Fraser University with$10 million. However, fewfaculty members would arguethat a multi-million dollarportfolio is necessary to have

a successful program. The incredible expansion in fund sizeis even more impressive when one considers that most of thefunds started with only $100,000 or $200,000 in initial seedcapital. Of course, almost all of the SMIFs continued toreceive additional investment capital from various sourcesas they demonstrated an ability to manage the moneysuccessfully.

One of the more interesting growth patterns for SMIFs ishow widely they are being used in a broad range of educationalenvironments. While 99% of all current SMIFs are housedwithin business schools, there are a few exceptions. Forexample, Tufts University, without a business college, has a$1 million student fund that focuses on investing inbiomedical companies. Besides the traditional universityundergraduate and graduate business students, there areseveral high schools that broke new ground by adapting thesame learning principles with students less prepared in financeand business.8 This is part of a broader trend where subjects

In the field of finance, studentmanaged investment funds (SMIFs)were created to take investmenteducation to the next level. Thesefunds allow students to invest realmoney in the stock and bondmarkets.

5 A university having a student investment fund has become the goldstandard for investment programs at all levels. As one faculty memberstated, “One cannot have a top 10 MBA program today without it.”

6The University of Wisconsin - Madison actually has five distinct funds.Only one portfolio is invested in equity securities. The other four portfoliosare fixed income portfolios, which provide a different set of learningexperiences depending on the objectives of the fund. The largest portionof the fund is private money managed for clients based on set investmentcriteria.

7The average size for US funds was $1.4 million and $1.2 million for non-US programs.

8 It has been reported that the following American schools at one time hadactive funds: Dominican High (WI), Groton High (MA), WisconsinLutheran (WI), Jenks High (OK), Ariel Community Academy (IL), WestAllis Central (WI), Gaithersburg High (MD), and Burnsville High (MN).These scaled down programs closely mirror those established at theuniversity level in the way they are structured and operate.


Table I. Growth in New Funds

0

20

40

60

80

100

120

140

160

Number of New

Funds

1950s 1960s 1970s 1980s 1990s 2000s

Decade

Table II. The Largest Funds (in US Dollars) Panel A. US Universities (2007)

Rank Institution Country Total Assets

1 University of Wisconsin US $62.0 Million 2 Ohio State University US $25.8 Million 3 University of Minnesota US $25.0 Million 4 University of Utah US $18.2 Million 5 University of Texas US $17.0 Million 6 Cornell University US $13.5 Million 7 University of Arkansas US $12.0 Million 8 University of Houston US $9.2 Million 9 & 10 Tie Baylor University US $6.5 Million Southern Methodist US $6.5 Million

Panel B. Non- US Universities

Rank Institution Country Total Assets

1 Simon Fraser University Canada $10.0 Million 2 HEC Montreal Canada $3.8 Million 3 Univ. of British Columbia Canada $3.5 Million 4 Queens University Canada $3.0 Million 5 Univ. Of New Brunswick Canada $2.2 Million 6 Concordia University Canada $1.4 Million 7 University of Alberta Canada $1.3 Million


that used to be taught only at the college level are now beingintroduced in high schools. According to McInerny (2003),some of the stimulus is being provided by Paul O’Neill, aformer US Treasury Secretary. O’Neill has been aggressivelypromoting greater financial skills at the primary and secondaryeducation level. It seems clear to faculty participating in theseprograms that real money portfolios have dramaticallyincreased student interest in majoring in finance or businessin college.

A surprising finding of this survey is that at least sixinstitutions in the US have allowed their SMIFs to becomeinactive over the past few years.9 Given how difficult it is formany schools to establish SMIFs and their popularity withstudents and employers, this was unexpected. In personaldiscussions with faculty and administrators at these schools,the most common reason the program became inactive wasthe loss of the key faculty member (often due to retirement)who advised the students. Other finance faculty wereunwilling to take over the fund, partially due to the greateramount of time it takes to stay abreast of the financial marketson a daily basis. With many schools emphasizing researchproductivity for promotion and raises, it is most onerous fortenure track faculty to lead the fund activities for more than afew years. Of course, many schools have dealt with this issueby hiring nontenure track faculty or adjuncts to run the SMIFs.

IV. SMIFs versus Professionally ManagedFunds

The central goal of SMIFs is to create a realistic learningenvironment for training the next generation of portfoliomanagers. Unlike professionally managed funds, which aresolely focused on generating the highest risk-adjusted ratesof return possible, SMIF returns are secondary in nature tothe educational mission.10 Faculty advisors generallyrecognize that some of the best learning experiences comefrom failures, not successes per se. As any experiencedinvestor knows, there is always an element of luck andincomplete data behind any decision. Thus, a very carefullyanalyzed opportunity with great potential can fail for an almost

unlimited number of reasons that could not have beenaccurately forecasted a year or more in advance. However,unlike practitioners in real life, students lack a strong incentivesystem of monetary rewards for beating benchmarks orpenalties for poor performance (being fired).

Student portfolios often have constraints that mostprofessionals do not have. For example, most SMIFs arestructured to rely on group or committee decisions rather thanthose of a single portfolio manager. Sometimes there are morethan 30 students involved with various levels of skill andknowledge. All have an equal vote in the ultimate decision.Depending on the quality of the student research and grouppresentation skills, decisions are not always based entirelyon objective analysis. In addition, the majority of funds offera one or two semester class that may encourage students touse short-term planning horizons since they may not be aroundto witness the final outcome of any particular investment.Obviously, it takes several economic cycles to really evaluatethe success of any investment strategy. The fund at theUniversity of Missouri - St. Louis is one of the exceptionswith a credit program structured to allow students toparticipate for several years (see Table III for a summary).

Professional managers can react almost instantaneously torapidly changing market conditions without the need toassemble the group for a vote. This factor alone should favorpractitioner performance, provided they are not trading onunfounded rumors and there really are fundamental marketchanges taking place. Professionals can also trade on marginor use derivatives to enhance returns which are not widelyavailable techniques for the majority of SMIFs. Offsettingsome of these advantages, professionally managed funds mustabsorb all of their own operating expenses, whereas mostSMIFs get subsidized resources (e.g. facilities, computers,faculty salary) from the universities and rarely pay allexpenses related to the fund.

V. Funding Sources and OrganizationalStructure

The majority of older SMIFs received ear-marked moneyfrom alumni and other private donors to establish the funds.Twenty-eight percent of the funds got all of the money fromthe university’s own endowment. Another 23% of schoolshad only a single large donor. The balance of programs was acombination of capital sources, including many small donorsand corporate donations. For universities wishing to establisha new fund, the average program in the US during the 2000swas started with approximately $414,000 in initial capital.

The most common form of organizational structure ishaving the SMIF be part of the university endowment. About62% of all funds are structured this way. Another 14% are set

9These universities include the University of Central Florida, SouthernIllinois University - Edwardsville, the University of Florida, the Universityof Missouri - Kansas City, Winthrop University, and the University ofLouisiana.

10Although there has been no systematic data collected on SMIFperformance, the limited anecdotal evidence suggests students generallydo as well and sometimes better than investment professionals or the marketas a whole. For example, the Tennessee Valley Authority reported thatover a three year period, the 19 universities participating in its program in2002 outperformed the S&P 500 benchmark by 5.3% (Mansfield, 2002).


Table III. Profile of the University of Missouri - St.Louis Fund

up as a separate entity, like a nonprofit foundation or trust toprovide more autonomy from the university. It is alsobecoming more popular for programs to establish profitmaking companies (e.g. LLCs or partnerships) where studentsare managing the portfolio for private companies or otherinvestors. At least ten of the largest funds, including theUniversity of Wisconsin, University of Minnesota,Pennsylvania State University, University of Houston, andUniversity of Texas are all managing some private investormoney.11 It should be noted that this is a more complexstructure in the US, which requires government reporting (e.g.partnership tax returns with K-1 forms) due to the taxablenature of the investments.

Several innovative companies in the US have long providedmoney to support financial education at institutions in themarkets they serve. The largest is the Tennessee ValleyAuthority (TVA), a large electric utility company, whichsponsors 25 universities in its service area. Two brokeragefirms followed the TVA’s lead with D. A. Davidson &

Panel A. Fund Characteristics

Date Established 1988 Size in June 2007 $125,000 Annual Student Participation 45 Fund structure Part of endowment Funding source Small private donations Faculty Member Full-time regular Credit hours per semester 1 credit hour per semester Max credit hours 3 hours (may continue without credit) Student level Undergraduate Application None Decision process Majority vote of students Investment style Growth and value Investment types allowed Equities, fixed income and options Equity strategy Bottom-up approach Diversification required? Yes Income Distributions Scholarships

Panel B. Actual 5 Year Historical Annual Returns (Including Dividends)

Year Actual fund performance S&P 500

2003 36.78% 28.68% 2004 17.67% 10.88% 2005 5.76% 4.91% 2006 6.93% 15.80% 2007 12.25% 5.49%

Company sponsoring 20 schools and Stern Agee Group, Inc.supporting 5 universities. The basic model at these programsis for the company to provide all funding ($400,000 each forthe TVA) with the company and universities sharing theprofits. In case of a falling stock market, the company absorbsall losses and fully replenishes the money the following year.About 58% of universities have an advisory board associatedwith their programs. All of these boards have outsideinvestment professionals and alumni serving as a valuableresource in a counseling capacity. This allows students tointeract with professionals and showcases the program to thelocal community. In many cases, students make formalpresentations to the boards to sharpen their presentation andanalytical skills.

VI. Student Participation

Just over 5,000 students participate in SMIFs in the USeach year, with another 500 students being trained at foreignuniversities. Approximately 71% of the programs in the US(45% of foreign programs) are structured as part of a formalclass.12 The number of credit hours a student can earn rangesfrom 1 to 12. Of those providing credit, 44% allow studentsto earn 6 or more semester credit hours over 2 or moresemesters. Another 39% of schools limit students to amaximum of 3 credit hours. Only 22% of programs limit thestudent learning experience to a single semester. The SMIFsthat are not part of a formal class allow students to participateas an extra curricular activity. This less structured formatpermits greater inclusion since almost any student enrolledin the university, regardless of major field or prior coursework, can join the group. In contrast, formal classes oftenrestrict the quality of students usually through an applicationprocess (59% of schools have a formal application processto screen students).

Unlike many other university programs, most SMIFscarefully control the level of student participation. Althougha few schools allowed more than 100 students to manage theportfolios each year, the average fund in the US had only 29student managers per year (23 students for foreign funds).For approximately 90% of SMIFs, students were responsiblefor making all investment decisions. In the other 10% ofprograms, advisory boards or a faculty member also sharedin the decision making. At the London Business School,students performed all of the usual research on securities butthey had to make formal presentations to a professional board,which actually made the final investment selections. While

11At the time of this survey, Penn State had 68 private investors and wasplanning to expand the number to 99. The faculty advisor also reports theadditional burden from accounting and tax costs for the LLC run between$25,000 and $50,000 per year.

12In France, universities have legal barriers for incorporating SMIFs intothe curriculum. Thus, only a few informal clubs not sponsored by the schoolexist, which are advised by outside professionals.


the major goal of these programs is to strengthen studentdecision making through actual investment experience, 64%of the funds have guidelines that allow a faculty member oran advisory board to veto student recommendations if aninvestment is deemed inappropriate for the portfolio.However, this is a power that is rarely used. In the previousfive years for those programs with a veto power, the higherauthority vetoed less than 4% of all student decisions. Thisresult indicates that studentstake their fiduciary roles asportfolio managers seriouslyand act prudently.

To facilitate investmentdecisions, 65% of the programsassign students to groups tomanage the portfolios.Depending on the specificportfolio and its goals, thesegroups are often based on typesof securities being traded,industries, etc. For 70% of theSMIFs, investment decisionsare determined by a simple majority vote by the students.Individual portfolio managers make the decisions 6% of thetime with the remaining funds using a combination of students,faculty advisors and/or boards to reach a consensus. Of theclasses, 42% of programs allow only undergraduate students,10% permit only graduate students and 48% allow both levelsof students.

VII. Faculty and Professional Involvement

With a small number of exceptions, faculty are closelyinvolved with SMIFs at all levels.13 Because so many of theprograms are relatively new, many of the faculty involvedtoday with SMIFs worked hard to obtain the original funding.Of the universities requiring students to participate througha formal class, 58% of these professors believe these classestake substantially more faculty time than a regular course.The average assessment was that the instructional load was50% higher than a traditional class, or the equivalent toteaching a 4.5 credit hour course rather than a 3 hour course.Another 33% of faculty felt they spent about the same amountof time as any other course. Only 9% of faculty thought itactually took less time than their normal instructional duties.

Despite a majority of participating faculty believing SMIFcourses take significantly more time, 63% of the schools paidthe same level of compensation as for a regular course. Thebalance of the universities provided additional compensationin the form of research funding, supplemental pay, or areduced service load. This may partially explain why someprograms have become inactive because key faculty membersfeel the compensation levels are not commensurate with the

time involved. This wouldsuggest college deans need totake a closer look at the cost/benefit ratio of SMIFs andmake a conscious effort toadequately reward facultyinvolvement in these highprofile programs. It would bea shame if much of the progressin financial education of thepast three decades would beallowed to erode based solelyon short-term economicsavings. One possible solution

would be to allow the fund itself to provide additionalcompensation to the participating faculty members. Thiscould be in the form of added salary, research grants, a courserelease, a graduate assistant, etc.

Given the nature of SMIFs, many of the programs havelocal investment professionals closely involved. The mostdirect role for outside professionals is to serve as an adjunctfaculty member and run the actual program. A small butgrowing number of schools take this approach, drawing onthe general finance community to provide the courseinstructor. In most cases, the adjunct faculty member is retiredand thus has time available to staff day sections of the class.It works less effectively for active professionals who mayhave difficulty finding the free time during normal work hourswhen the financial markets are open. Even when the programis being taught by a full-time faculty member, it iscommonplace to have professionals serve on advisory boardsand be guest speakers. There is no question that professionalinvolvement enriches the experience of students, faculty andprofessionals. Frequent contact between the various partiesensures that current practice is quickly incorporated into theclassroom and students leave better prepared to apply theirknowledge and skills.

VIII. Investment Activity

It is interesting that 28% of schools with a SMIF have morethan one fund. Many of these funds have different investmentobjectives and are specifically designed to give students a

The size of SMIFs today hasexpanded beyond what manypeople would have thoughtpossible only a decade ago. Thereare 78 universities worldwide withmore than $1 million undermanagement by students.

13A number of universities have real money funds that operate more as anextracurricular activity with no direct faculty involvement. In some cases,the students invest their own money. Schools that operate this way includethe University of Edinburgh, Harvard University, Auburn University, theCalifornia Institute of Technology, Princeton University, DartmouthUniversity, Vanderbilt University and the Georgia Institute of Technology.


broader investment experience than a single fund couldprovide. For schools with more than one fund, the mostcommon number was to have three distinct portfolios. It alsomakes sense to have more than one SMIF where there aremultiple bodies of students including undergraduate/graduate,day/evening classes, etc.

With the growing size of the average portfolio, it is notsurprising to find 92% of universities have formal writteninvestment guidelines. Diversification is a principle widelyemphasized by most SMIFs. Counting a mutual fund as asingle holding, the average portfolio contained about 30individual securities with three SMIFs exceeding 75 differentissues. Approximately 80% of the programs have clearguidelines that specifically require the funds be diversified.It is interesting that 19% of US SMIFs are not prohibitedfrom becoming hedge funds to increase returns by taking onmore risk. Only 10% of non-US SMIFs had this samefreedom. In the first reported case of an actual hedge fund ona university campus, Cornell University changed its SMIF’sinvestment strategy from a indexed styled fund to a “market-neutral” hedge fund in 2002.14 The stated goal was to producepositive returns regardless of which direction the market wasmoving. The $3 million fund uses investors’ money (alumniand friends) which allow students to manage the money forthe experience without any fees.

As to types of security investments, some schools, such asthe University of Toledo, are very restrictive and require thatall investments must be in domestic markets. On the otherextreme, Roger Williams University limits domesticinvestments to a maximum of 20% of the portfolio with theother 80% being comprised of international securities. MostSMIFs focused on traditional securities with common stockdominating portfolios, regardless of where the companieswere domiciled. Corporate and Treasury bonds were alsowidely used to balance the portfolios. For alternativeinvestments, real estate investment trusts (REITs) were themost popular followed at a distance by limited partnerships.

For actual trading activities, full service brokerage firmswere most often used by SMIFs, followed closely by discountbrokerage firms. Bank trust companies were used only abouthalf as much as either type of brokerage firm. The funds arelarge enough to negotiate some very favorable rates with fullservice brokerage firms. In addition, since the majority ofthe programs are charitable and tax-free by design, some ofthe brokerage firms are donating their services to theuniversities.

For SMIFs operating a single fund, only 10% of theprograms characterized their investment style as focusing ongrowth stocks. Another 23% considered themselves to be

more value investors. But the vast majority of the fundscharacterized their investment style to be more of a blend.The most employed equity strategy was the bottom-upapproach with 37% of SMIFs primarily using this method. Aclose second was the top-down approach used by 27% offunds, followed at a distance by the buy-and-hold strategyreported by 11% of respondents. The balance of the SMIFsused a combination of these strategies and others (e.g. pricemomentum and contrarian) in making stock selections. Twiceas many programs thought asset allocation was a veryimportant consideration compared to those who deemed itnot very important (38% v. 18%). Individual security selectionwas rated very important by 58% of SMIFs. In contrast, 65%of funds believed market timing was not important, most likelyreflecting their long-term investment horizons.

IX. Comparisons of US and Foreign Funds

With the widely divergent political and economic climatesabroad, it was not unexpected to find that SMIFs outside ofNorth America evolved slower and in a somewhat differentdirection than those in the US. For the most part, Canadianand American universities share a similar educationalenvironment. Thus the Canadians adopted SMIForganizational structures and operating procedures modeledafter those already successfully employed in the US for severaldecades. Outside of North America, however, European andAsian schools were far more likely (by a ratio of 2 to 1) tohave extra-curricular programs rather than formal classes.One result of a less formal structure is that fewer foreignuniversities (55%) have anyone with veto power over studentinvestment decisions compared to US schools (64%). Ofcourse, the funds outside of North America are much youngerand smaller, averaging only $142,000. With maturity andmore money at risk, stricter university controls may eventuallydevelop. Unlike the US, none of the European or Asianprograms had a taxable structure similar to a LLC orpartnership.

When it comes to investing, most American SMIFs focuson investing in common stock with an average portfolio in2007 containing 82% of their money invested in thesesecurities. Foreign programs generally leaned more towardbalanced portfolios with greater allocations of fixed incomesecurities. Canadian SMIFs averaged 70% of their moneyinvested in common stock, while European and Asian fundswere much lower at 59%. On the lower end of the scale,Hebrew University had the most balance with 25% in commonstock, 20% in preferred stock, 15% in corporate bonds, 25%in Treasury bonds, and the remainder in other securities. Interms of investment policies, almost the same percentage offoreign SMIFs required diversification in their written policiesas in the US (79% v. 81%). As to investment styles, only 14Myers (2004).


10% of American funds with a single fund consideredthemselves “growth” investors in 2007 compared to 20% ofthe foreign funds. Regardless of where the fund wasdomiciled, the predominant investment style was a “blend”rather than a single focus. The equity strategies most used inthe US and abroad was a bottom-up approach, closelyfollowed by the top-down approach, or a combination of thetwo.

X. Benefits to the University Community

It has long been recognized that students learn more byhands-on experience than simply reading about a topic in atextbook. Besides learning the intricacies of portfoliomanagement and trading, students also benefit in manyprograms by going on field trips to Wall Street and otherfinancial markets. Schools like Virginia Tech and GannonUniversity have a long tradition of taking students annuallyto Wall Street to view the financial markets first hand. RogerWilliams University has taken it to the next level by takingstudents abroad to the London and Frankfurt stock exchanges,which would not have been possible without a SMIF togenerate the student interest and fund the activity.

Eighty-one percent of all program directors cited bettertrained students as a major benefit of having a SMIF. Almosta third of faculty believed having a real money fund providedsynergy and significantly improved the quality of the overallfinance program. Conversations with faculty also indicatedgreater job opportunities for students participating in SMIFs.Many employers, including private equity and hedge funds,bank trust companies, and mutual funds have beenaggressively recruiting students who have these experiencesto draw on. These are highly competitive jobs that can bedifficult for a new college graduate to obtain without sufficientexperience.

As anyone involved in a SMIF can attest, having a realmoney portfolio generates a substantial amount of mediaattention. This activity not only showcases the students andthe finance discipline but also the business school and theuniversity. Alumni in particular are highly supportive ofSMIFs, which creates new opportunities for guest speakers,field trips, internships, student recruitment, etc. Finally, theprograms provide badly needed financial support for studentscholarships, visits to financial markets, operating tradingrooms, and other university programs. For 2006, sixty-six ofthe American funds made cash distributions totaling morethan $1.9 million to support academic programs, or an averageof $29,381 per school. However, an even larger number ofSMIFs reported making no cash distributions in the previousyear. Many of these were still relatively new and thereforewere still in the capital building years. Nevertheless, it wasnot all that unusual for the larger SMIFs to spin-off several

hundred thousand dollars in cash flow while providingstudents with a valuable learning experience. Very fewuniversity programs have such a high benefit/cost ratio,especially since almost all universities have endowment fundsthat must be managed by someone. Historically, the limitedevidence shows SMIFs have performed as well and sometimesbetter than funds managed by professional investmentadvisors (Mansfield, 2002). Of course, since SMIFs do notnormally charge management fees, this saving alone favorsstudent-managed funds even without the educational benefits.

XI. Recent Developments

A. Trading Rooms

A growing contingent of programs are operating tradingrooms to add even more realism to student learning. Many ofthe universities with SMIFs have invested up to $1 million tofully furnish and equip trading rooms. The expanded programsinclude Pennsylvania State University, Iowa State University,Rice University, Michigan State University, StetsonUniversity, Texas Christian University, University ofMichigan, and the University of Missouri - Columbia. Theseschools believe this development has raised the bar inattracting top students and community financial support. Ofcourse, there are other universities with trading rooms thatdo not have SMIFs and simply simulate trading activities.

B. Social Responsibility Funds

Social responsibility funds are becoming more popular inacademia and with investors in general. Bluffton Universityin Ohio has had an investment policy since 1956 of avoiding“sin stocks”, which include tobacco, alcohol, and defensecompanies. Villanova University follows a similar investmentguideline with two of its funds. The University of Californiaat Berkeley started a new social responsibility fund inFebruary of 2008 with $1.2 million as part of its MBAprogram. Students will hold long positions in firms that aresocially responsible and take short positions in firms withpoor social records. The director of its program maintainsone does not have to sacrifice financial returns for a goodrecord of social responsibility (Alsop, 2007). Establishingsocially responsible funds can be highly controversial inacademia. Some opponents argue that it is pushing a politicalagenda. In 1997, Stanford University rejected a studentproposal for such a fund, noting their endowment alreadyhad substantial stock investments in socially responsiblecompanies and industries and thus it was not needed.


C. Using Investors’ Money

Several of the largest funds (including the universities ofTexas, Minnesota, Houston, Wisconsin, and PennsylvaniaState) manage investor money in one or more of their funds.The University of Texas was the first large for-profit fundwhen it raised $1.6 million of private investor money inDecember 1994 for their MBA students to manage. By 2007,they had three distinct funds with different investmentobjectives totaling $17million from 60 investors.These funds are usuallystructured as a LimitedLiability Company (LLC)where the income is taxableto investors similar to apartnership distribution. Thisstructure limits the numberof investors, so they mustmake large contributions. Inexchange for managing themoney, the students anduniversity sometimes get a management fee of between .5%and 1.5% of the assets.

D. Venture Capital Funds

One of the most exciting developments over the last decadehas been the emergence of venture capital funds managed bystudents. Given the success of SMIFs, it is only natural thatthe programs would evolve in new directions. The Universityof Michigan created the first student-managed venture capitalfund in the US in 1997 with about $3 million in capital. YaleUniversity, the University of North Dakota, the Universityof Utah, Cornell University, the University of Wyoming, andMiami University of Ohio followed in Michigan’s footstepswith venture capital funds of their own dedicated to investingin emerging companies.15 The University of Utah sponsoredthe largest venture capital fund with $18 million, which alsopermits students from more than 15 other universities toparticipate in the activities. Although not a student-managedfund, the University of Maryland in 2003 worked withinvestors to establish the New Markets Growth Fund with$20 million in capital run by professional managers butassisted by students and faculty. Others, including theUniversity of Queensland and the University of Melbourne(both in Australia), have similar funds run by professionals.

Many of these programs are designed to provide seedcapital for businesses started by students, recent graduates,

15For further information, see Rombel (2007), Yale Bulletin (2001), DailyHerald (2006) and the Business Wire (2006).

faculty, or the general community at large. Doing so mayspeed technology transfer from universities and fulfill onemission of higher education. All of these innovative programsexpand the practical training offered to finance students byconventional SMIFs in new dimensions. They encouragestudents to take a more entrepreneurial approach to raisingcapital in the private equity market (often with partner orinvestor money). Students benefit by evaluating businessplans and performing due diligence before actually making

the investment decision oncompanies with little or nofinancial performance record.This focus provides a nicecomplement to a regular SMIFwhere the focus is onestablished, publicly-tradedinvestment opportunities.16

Finally, the venture capitalfunds offer an excellent vehiclefor the College of Business toprovide value added support toother units within the

university community. For example, the Colleges of Medicine,Science and Engineering produce a continuing stream ofinnovative research and technology but have great difficultyproving the commercial viability of their inventions andpatents. Along with business schools, law schools can assistnew startups with legal issues to further reach another segmentof the community. A student managed venture capital fundoffers the best opportunity in years to capitalize and profitfrom the research strengths of universities while enhancingthe teaching mission.

E. Micro Finance Funds

There is an amazing amount of creativity surroundingSMIFs in the way the programs are being re-engineered toaccomplish more than simply teaching students the basics ofinvesting. Several universities, including ColumbiaUniversity, are starting micro finance funds to make smallentrepreneurial investments in third world countries. Anorganization called PlaNet Finance (a microfinanceorganization based in Paris) is working with Columbia andseveral European universities to sponsor these programs. IfColumbia’s program is a success, it will provide another venue

16The 2008 state budget for the Commonwealth of Massachusetts containedan amendment for establishing a student investment fund to encouragestudent entrepreneurship. There were to be three students on the governingboard and it was specifically designed to fund new student businessescreated within the Commonwealth. (See Section XX, Chapter 23A of thebudget)

One of the most excitingdevelopments over the last decadehas been the emergence of venturecapital funds managed by students.Given the success of SMIFs, it is onlynatural that the programs wouldevolve in new directions.


for students to learn about business while benefitting socialwelfare initiatives around the world.

XII. Conclusion

Over the past 50 years, student-managed investment fundshave revolutionized the way in which investment educationis taught in universities. These programs have expanded to314 worldwide today from only a few funds in the 1950s. Inthe process, SMIFs are evolving in exciting new directions.These include managing money for private clients,establishing hedge funds, or venture capital funds and microlending initiatives. While it is much too early to evaluate thesuccess of these new programs, it does seem clear that

business schools are becoming even more relevant byaddressing important issues in both the financial markets andsociety.

The benefits of providing students with greater practicalexperience and technical skills in finance are widelyrecognized in the job market. Students graduating today fromuniversities with SMIFs already have at least 1 or 2 semestersof actual trading and research experience. Although theuniversity experience is not as intense as in a professionaljob, it still provides a solid foundation for the knowledgeneeded in portfolio management. The skills and techniqueslearned here can be further refined in the workplace over amuch shorter period of time than what would have beenpossible in the 1960s or 1970s.

Appendix A. Listing of All US Funds

University Name City State Year Started Funds 2007 $000 Abilene Christian University * Abilene TX n/a 319 Adelphi University Garden City NY 2007 100 Alabama A&M University Normal AL 1998 330 Alaska Pacific University * Anchorage AK 2000 185 Alfred University * Alfred NY 1995 200 American University Washington DC 2002 100 Anderson University Anderson IN 2007 10 Appalachian State University Boone NC 2000 116 Arizona State University * Tempe AZ 1996 515 Ashland University Ashland OH 2000 375 Auburn University Auburn AL 1999 50 Austin College * Sherman TX 2007 1,000 Austin Peay State University * Clarksville TN 1998 400 Babson College Babson Park MA 1997 1,300 Baldwin-Wallace College * Berea OH 2006 175 Ball State University Muncie IN 2005 577 Bates College * Lewiston ME 2004 100 Baylor University Waco TX 2001 6,500 Belmont University * Nashville TN 2003 400 Bentley College * Waltham MA 1997 555 Binghamton University - SUNY * Binghamton NY 2003 130 Bluffton University Bluffton OH 1956 174 Boise State University Boise ID 1995 149 Boston College * Boston MA 1983 360 Boston University Boston MA 2001 25 Bowling Green State University Bowling Green OH 2006 265 Brandeis University * Waltham MA 1998 13 Brigham Young University Provo UT 1984 1,866 Bryant University Smithfield RI 2005 425 Bryn Mawr College * Bryn Mawr PA 1975 100 Bucknell University Lewisburg PA 2000 750 Butler University Indianapolis IN 2007 1,000 California Institute of Technology Pasadena CA 1978 490 California Polytechnic State Univ. San Luis Obispo CA 1992 453 California State University Fresno Fresno CA 1999 90 California State University - Long Beach Long Beach CA 1995 100


University Name City State Year Started Funds 2007 $000 California State University - Northridge Northridge CA 1994 2,000 California State University - Los Angeles* Los Angeles CA 2001 100 Cameron University Lawton OK 1988 800 Canisius College * Buffalo NY 2003 300 Carnegie Mellon University * Pittsburgh PA 2006 64 Carroll College * Helena MT 2004 50 Cedar Crest College Allentown PA 1997 52 Cedarville University* Cedarville OH 2008 75 Centenary College of Louisiana * Shreveport LA 2003 120 Central Michigan University Mt. Pleasant MI 1997 469 Christian Brothers College Memphis TN 2003 400 Christian Brothers University Memphis TN 2003 400 Claremont Graduate School * Claremont CA 2001 381 Clemson University * Clemson SC 2004 300 Cleveland State University * Cleveland OH 2007 100 College of New Jersey Ewing NJ 2000 170 College of William & Mary Williamsburg VA 1999 590 College of Wooster Wooster OH 1955 1,300 Colorado College Colorado Springs CO 2004 24 Colorado State University Fort Collins CO 1998 190 Connecticut College * New London CT 2002 77 Cornell University * Ithaca NY 1998 13,500 Creighton University Omaha NE 1993 2,500 Culver Stockton College * Canton MO 1996 55 Dartmouth Hanover NH 1996 400 DePaul University * Chicago IL 1982 341 Drake University * Des Moines IA 1999 239 Drexel University Philadelphia PA 2007 250 Duke University * Durham NC 1952 162 East Tennessee State University Johnson City TN 2000 370 Eastern Illinois University Charleston IL 1994 136 Eastern Washington University * Cheney WA 2004 50 Elizabethtown College Elizabethtown PA 2007 130 Emory University Atlanta GA 2006 1,200 Fairfield University Fairfield CT 2006 300 Florida Gulf Coast University Fort Myers FL 2005 220 Franklin and Marshall College Lancaster PA 1999 204 Gannon University Erie PA 1952 126 Gardner Webb University * Boiling Springs NC 2000 25 George Washington University Washington DC 2005 1,500 Georgetown University * Washington DC 1999 200 Georgia Institute of Technology Atlanta GA 1986 810 Georgia State University Atlanta GA 2005 368 Gonzaga University Spokane WA 2000 200 Grinnell College * Grinnell IA 2000 122 Gustavus Adolphus College * ST. Peter MN 1998 123 Harvard University * Cambridge MA na na Henderson State University Arkadelphia AR 2001 343 Humboldt State University Arcata CA 2006 7 Idaho State University Pocatello ID 2005 59 Illinois College Jacksonville IL 1995 458 Illinois State University Normal IL 1982 383 Illinois Wesleyan University Bloomington IL 1993 740 Indiana State University * Terre Haute IN 2000 437 Indiana University * Bloomington IN 1986 500 Indiana University of Pennsylvania Indiana PA 2005 223 Iowa State University Ames IA 1999 195

Appendix A. Listing of All US Funds (Continued)


University Name City State Year Started Funds 2007 $000 Ithaca College * Ithaca NY 2005 24 Jacksonville University Jacksonville FL 1987 454 James Madison University * Harrisonburg VA 1999 146 John Carroll University Cleveland OH 1996 170 Kansas State University Manhattan KS 2002 1,100 Kennesaw State University Kennesaw GA 2006 100 Kutztown University of Pennsylvania Kutztown PA 2005 190 Lafayette College Easton PA 1950 455 Lehigh University Bethlehem PA 1962 360 Lipscomb University Nashville TN 2003 450 Longwood University Farmville VA 2002 430 Loras College Dubuque IA 1998 172 Louisiana State University * Baton Rouge LA 2005 1,000 Loyola College * Baltimore MD 2006 500 Marquette University Milwaukee WI 2005 1,200 Marywood University Scranton PA 2006 75 Massachusetts Institute of Technology * Cambridge MA 1964 27 McNeese State University Lake Charles LA 2007 21 Miami University Oxford OH 1996 375 Michigan State University East Lansing MI 2003 4,200 Michigan Technological University Houghton MI 1998 1,300 Middle Tennessee State University Murfreesboro TN 1998 325 Middlebury College Middlebury VT 1987 275 Millsaps College Jackson MS 1989 200 Mississippi State University * Mississippi State MS 1998 400 Mississippi University for Women Columbus MS 1999 385 Montana State University - Bozeman* Bozeman MT 1985 50 Montana State University - Billings * Billings MT 1985 50 Moravian College Bethlehem PA 1962 1,442 Murray State University * Murray KY 1998 440 Nebraska Wesleyan University Lincoln NE 2005 250 New Mexico State University Las Cruces NM 2007 5,013 New York University New York City NY 2000 2,001 North Carolina State University Raleigh NC 2004 135 North Dakota State University Fargo ND 2007 110 Northeastern University Boston MA 2007 50 Northern Arizona University Flagstaff AZ 2000 997 Northern Illinois University DeKalb IL 2000 230 Northern Michigan University Marquette MI 2006 210 Northwest Nazarene University Nampa ID 2003 70 Northwestern University Evanston IL 1964 2,375 Oberlin College* Oberlin OH 2004 281 Ohio Northern University Ada OH 1989 128 Ohio State University Columbus OH 1990 25,810 Ohio University Athens OH 2001 2,000 Oregon State University Corvallis OR 2005 60 Ouachita Baptist University Arkadelphia AR 2000 20 Pace University Pleasantville NY 2002 280 Pacific Lutheran University * Tacoma WA 1982 92 Pennsylvania State University University Park PA 2005 5,000 Portland State University Portland OR 1997 251 Princeton University * Princeton NJ 2006 10 Purdue University West Lafayette IN 2000 400 Radford University Radford VA 2003 495 Rice University Houston TX 1996 900 Roanoke College * Salem VA 2004 500 Roger Williams University Bristol RI 2004 122



University Name City State Year Started Funds 2007 $000 Rollins College Winter Park FL 1999 750 Rutgers University * New Brunswick NJ 2000 1,605 Saint Bonaventure University St. Bonaventure NY 2003 45 Saint Cloud State University St. Cloud MN 1999 115 Saint John's University New York NY 2001 2,700 Saint Joseph's University Philadelphia PA 2004 117 Saint Louis University St. Louis MO 2002 916 Saint Mary's University San Antonio TX 2007 1,000 Salisbury University Salisbury MD 2000 388 Samford University Birmingham AL 2008 500 San Diego State University * San Diego CA 1992 100 Santa Clara University Santa Clara CA 2000 350 Scripps College * Claremont CA na 200 Seattle University * Seattle WA 2004 50 Shippensburg University Shippensburg PA 1994 81 Southeast Missouri State University Cape Girardeau MO 1990 835 Southern Illinois University Carbondale IL 2000 360 Southern Methodist University Dallas TX 1980 6,500 Southern New Hampshire University Manchester NH 2004 59 Southwestern University Georgetown TX 1999 349 Spring Arbor University Spring Arbor MI 2005 12 Stanford University * Stanford CA 1978 180 State University of New York - Geneseo Geneseo NY 2007 18 Stetson University DeLand FL 1980 3,100 Syracuse University * Syracuse NY 2006 1,100 Tennessee State University * Nashville TN 1998 400 Tennessee Tech University Cookeville TN 2000 500 Texas A & M University * College Station TX 2000 250 Texas Christian University Ft. Worth TX 1973 1,500 Texas Tech University Lubbock TX 1997 2,200 Texas Wesleyan University Ft. Worth TX 1998 776 Trevecca Nazarene University Nashville TN 2003 405 Trinity University San Antonio TX 1998 1,340 Tufts University * Medford MA na 1,059 Tulane University * New Orleans LA 1999 2,419 Union University * Jackson TN 2003 400 University of Akron Akron OH 1996 100 University of Alabama - Huntsville Huntsville AL 1998 428 University of Alabama - Birmingham * Birmingham AL 2007 385 University of Alabama - Tuscaloosa* Tuscaloosa AL 1998 50 University of Alaska Fairbanks AK 1995 550 University of Arizona Tucson AZ 2000 930 University of Arkansas-Fayetteville Fayetteville AR 1971 12,000 University of California - Los Angeles Los Angeles CA 1987 2,000 University of California - Berkeley* Berkeley CA 1999 120 University of Chicago * Chicago IL 2005 1,000 University of Cincinnati * Cincinnati OH 2000 350 University of Colorado - Boulder Boulder CO 2002 300 University of Colorado - Colorado Springs Colorado Springs CO 2004 58 University of Connecticut * Storrs CT 2000 2,300 University of Dayton * Dayton OH 1994 6,300 University of Delaware * Newark DE 1996 800 University of Denver * Denver CO 1999 550 University of Georgia Athens GA 2007 101 University of Houston Houston TX 2002 9,177 University of Idaho * Moscow ID 1989 400 University of Illinois Champaign IL 1999 390 University of Iowa Iowa City IA 1994 536



University Name City State Year Started Funds 2007 $000 University of Kansas Lawrence KS 1994 1,523 University of Kentucky Lexington KY 1999 400 University of Louisville Louisville KY 2004 50 University of Maine Orono ME 1993 1,253 University of Maryland College Park MD 1993 1,350 University of Memphis Memphis TN 1999 475 University of Michigan Ann Arbor MI 2000 3,700 University of Minnesota - Minneapolis Minneapolis MN 1998 25,000 University of Minnesota - Duluth Duluth MN 2003 440 University of Mississippi University MS 2001 335 University of Missouri-Columbia Columbia MO 1967 1,354 University of Missouri-St. Louis St. Louis MO 1988 125 University of Montana * Missoula MT 1985 50 University of Nebraska - Lincoln Lincoln NE 1981 1,300 University of Nebraska - Omaha* Omaha NE 2000 1,400 University of Nevada * Reno NV 2004 107 University of New Hampshire * Durham NH 1995 50 University of New Mexico Albuquerque NM 2006 2,400 University of North Alabama * Florence AL 2003 400 University of North Carolina - Chapel Hill Chapel Hill NC 1952 1,424 University of North Carolina - Wilmington * Wilmington NC 2007 1,000 University of North Carolina - Charlotte * Charlotte NC 1997 235 University of North Dakota Grand Forks ND 2005 676 University of North Florida * Jacksonville FL 1999 772 University of North Texas * Denton TX 2003 277 University of Northern Colorado * Greeley CO 1992 1,100 University of Northern Illinois * DeKalb IL 1999 200 University of Northern Iowa * Cedar Falls IA 1999 115 University of Notre Dame Notre Dame IN 1998 5,000 University of Oklahoma Norman OK 1996 505 University of Oregon Eugene OR 1999 900 University of Pennsylvania * Philadelphia PA 1996 700 University of Pittsburgh Pittsburgh PA 1999 351 University of Portland Portland OR 2003 65 University of Rhode Island Kingston RI 2001 151 University of Richmond Richmond VA 1993 325 University of Rochester Rochester NY 1995 200 University of South Dakota Vermillion SD 2001 520 University of Southern California * Los Angeles CA 1986 2,600 University of Southern Mississippi Hattiesburg MS 2002 308 University of St. Thomas St. Paul MN 1999 3,000 University of Tampa Tampa FL 2003 155 University of Tennessee - Martin Martin TN 1998 460 University of Tennessee- Knoxville Knoxville TN 1998 1,000 University of Tennessee - Chattanooga Chattanooga TN 1998 510 University of Texas Austin TX 1994 17,000 University of the Pacific Stockton CA 2007 1,100 University of Toledo * Toledo OH 2005 1,000 University of Tulsa Tulsa OK 1998 1,577 University of Utah Salt Lake City UT 1998 18,173 University of Virginia - McIntire School Charlottesville VA 1994 500 University of Virginia - Darden Graduate * Charlottesville VA 1990 6,200 University of Washington * Seattle WA na 50 University of Wisconsin-Eau Claire Eau Claire WI 2003 250 University of Wisconsin-Madison Madison WI 1970 62,000 University of Wisconsin-Whitewater Whitewater WI 1999 85 University of Wisconsin- Oshkosh Oshkosh WI 2000 135 University of Wisconsin - Platteville* Platteville WI 2001 190 University of Wyoming Laramie WY 2005 1,700 Utah State University * Logan UT 1985 50


* Did not respond to survey. Information collected from media and institution’s web site.


Appendix B. Listing of All Non-US Funds

* Did not respond to survey. Information collected from media and institution’s web site.

University Name City Province Country Year Started Funds 2007 $000 Birla Institute of Tech. & Sciences Pilani Rajasthan India 2007 3 Bishop's University Sherbrooke Quebec Canada 1996 485 Bond University Gold Coast Queensland Australia na 28 Brock University St. Catharines Ontario Canada 1995 17 Concordia University Montreal Quebec Canada 1999 1,378 Hebrew University of Jerusalem Jerusalem Israel 1999 580 HEC Montreal Montreal Quebec Canada 1999 3,810 London Business School London United Kingdom 2003 300 Maastricht University Maastricht Limburg The Netherlands 1994 70 Massey University Auckland New Zealand 1995 15 McGill University * Montreal Quebec Canada na 10 Punjab College of Technical Ed Ludhiana India na 3 Queens University * Kingston Ontario Canada 2001 3,000 Simon Fraser University Vancouver British Columbia Canada 2003 9,983 St. Francis Xavier University Antigonish Nova Scotia Canada 2000 2 St. Mary's University Halifax Nova Scotia Canada 2005 184 University of Alberta Edmonton Alberta Canada 1998 1,292 University of British Columbia * Vancouver British Columbia Canada 1987 3,514 University of Calgary Calgary Alberta Canada 1996 361 University of Edinburgh Edinburgh United Kingdom 1997 na University of Guam Mangilao Guam US Territory 2006 53 University of Manitoba Winnipeg Manitoba Canada 1997 11 University of New Brunswick Fredericton New Brunswick Canada 1998 2,200 University of Toronto * Toronto Ontario Canada 2007 17 Wilfrid Laurier University Waterloo Ontario Canada 2001 340

References

1997, “Students Press for Socially Responsible EndowmentFund,” Stanford University News Service (February 5).

2001, “Yale SOM Launches Student-Managed VentureCapital Fund,” Yale Bulletin 29, 1 (February 16).

2006, “Local News,” Daily Herald (April 12).

2006, “Record $18 million Closing for Largest Student-RunVenture Capital Fund,” Business Wire (June 19), 1.

Alsop, R., 2007, “MBA Students to Run Socially ResponsibleFund,” The Wall Street Journal Online (September 14).

Alsop, R., 2007, “Talking b-School: Haas Takes New Tackon Investing,” The Wall Street Journal (September 18,2007), B8.

Ammermann, P. A. and L. R. Runyon, 2003, “Risk Aversionand Group Dynamics in the Management of StudentManaged Investment Fund”, Journal of the Academy ofBusiness and Economics 1 (No. 1).

Ary, E. J. and R. L. Webster, 2000, “A Survey of UniversityStudent Investment Funds,” Midwest Review of Financeand Insurance 14 (No. 1), 9-18.

Bear, T. and G. M. Boyd, 1984, “An Applied Course inInvestment Analysis and Portfolio Management,” Journalof Financial Education 13 (Fall), 68-71.

Belt, B., 1975, “A Securities Portfolio Managed by GraduateStudents,” Journal of Financial Education 4 (Fall), 77-81.


Bhattacharya, T. K. and J. J. McClung, 1994, “CameronUniversity’s Unique Student-Managed InvestmentPortfolios,” Financial Practice and Education 4 (No. 1),55-59.

Block, S. B. and D. W. French, 1991, “The Student-ManagedInvestment Fund: A Special Opportunity in Learning,”Financial Practice and Education 1 (No. 1), 55-60.

Dolan, R. C. and J. L. Stevens, 2006, “Business Conditionsand Economic Analysis: An Experiential Learning Programfor Economic Students,” Journal of Economic Education37 (No. 4), 395-405.

Hirt, G. A., 1977, “Real Dollar Portfolios Managed byStudents - An Evaluation,” Journal of Financial Education6 (Fall), 57-61.

Johnson, D. W., J. F. Alexander and G. H. Allen, 1996,“Student-Managed Investment Funds: A Comparison ofAlternative Decision-Making Environments,” FinancialPractice and Education 6 (No. 1), 97-101.

Kahl, D. R., 1997, “The Challenges and Opportunities ofStudent-Managed Investment Funds at MetropolitanUniversities,” Financial Services Review 6 (No. 3), 197-200.

Kester, G. W., 1986, “Extending Investments Beyond theClassroom Through Investment Clubs,” FinancialManagement Collection 1 (No. 3), 9-10.

Lawrence, E. C., 1990, “Learning Portfolio Management byExperience: University Student Investment Funds,”Financial Review 25 (No. 1), 165-173.

Lawrence, E. C., 1994, “Financial Innovation: The Case ofStudent Investment Funds at United States Universities,”Financial Practice and Education 4 (No. 1), 47-53.

Mansfield, D., 2002, “TVA’s Student Investors OutperformMarket, TVA Managers,” The Florida Times Union (April5).

Markese, J. D., 1984, “Applied Security Analysis andPortfolio Management,” Journal of Financial Education13 (Fall), 65-67.

McInerny, P. M., 2003, “The Student-Managed InvestmentFund at the High School Level,” The Clearing House 76(No. 5), 252-254.

Myers, L., 2004, “CU Student-Managed Investment FundThrives After Change in Strategy,” Cornell Chronicle 35(No. 24), 5

Neely, W. P., 2002, “A Survey of Student-Managed Funds:The Way They Work,” presented at the FinancialManagement Association meeting, (October 17).

Neely, W. P. and P. L. Cooley, 2004, “A Survey of StudentManaged Funds,” Advances in Financial Education 2(Spring), 1-9.

Pfeffer, J., 2007, “What’s Right and Still Wrong with BusinessSchools,” Biz Ed 6 (No. 1), 42-48.

Rombel, A., 2007, “Student-Run Venture Capital Firm atCornell Invests in Local Tech Firm,” Business Journal -Central New York 21 (No. 12), 8.

Tatar, D. D., 1987, “Teaching Securities Analysis with RealFunds,” Journal of Financial Education 16 (Fall), 40-45.


Behavioral Basis of the Financial Crisis

Joseph V. Rizzi

84

Financial institutions suffered large losses following thecollapse of the credit markets despite making huge riskmanagement investments. Major risks are frequentlyignored due to behavioral biases resulting in incorrectdecisions. These biases are reinforced by organizationalobstacles, such as misaligned compensation systems. Thisarticle outlines a supplemental behavioral risk framework,and applies it to the structure finance market. Behavioralfinance can improve how risk decisions are made. Youignore behavioral risk at your peril.

Major strides were made in quantitative risk managementduring the 1990s. Yet despite these advances, financialinstitutions suffered large losses following the collapse ofthe subprime and structured products markets. How this couldhave occurred given sophisticated tools and massive risksystem investments is a concern. A further concern is thelikelihood of repeating this experience during the next cycle.Although we know how risk decisions should be made, lessis known on how these decisions are actually made.

Risk management should encourage profitable risk takingwhile discouraging unprofitable and catastrophic risk. In mostinstitutions, however, political power and capital flows tosuccessful individuals. Unfortunately, it is difficult todetermine whether they are truly successful or just lucky. Ourexisting risk measures account for perhaps 95% of what

occurs. The major catastrophic risks lurk in the fat tails ofthe remaining 5%. We tend to underestimate these improbablerisks due to behavioral biases.

Institutions and regulators are changing their risk systemsand personnel to address this issue. The problem, however,is not only with the systems or the quality of the personnelbut within the individuals themselves. Most individuals havea model of how the world works. When challenged by events,we try to explain away the events. Behavioral economicsprovides insight into risk-assessment errors and possibleremedies.

This article outlines a behavioral risk framework to addressjudgment bias and develop appropriate responses. Behavioralfinance recognizes that decision processes influenceperception and shape our behavior. The frameworksupplements current quantitative risk management byimproving responses to risk changes over time. Theframework will then be applied to the structured finance crisis.

I. Behavioral Finance Framework

Risk can be classified along two dimensions. The firstconcerns high-frequency events with relatively clear cause-effect relationships. Other risks occur infrequently.Consequently, the cause-effect relationship is unclear. Thesecond dimension is impact severity. No matter how remote,high-impact events cannot be ignored because they canthreaten an institution’s existence as was demonstrated in thecurrent market crisis. The dimensions are reflected in therisk map in Figure 1.

Quadrant A events include retail credit products includingcredit cards. Many small defaults are expected. Screeninghelps identify groups with higher default probabilities. Thesegroups are charged higher rates to offset the risk. QuadrantB represents many internal operational risks such as checkprocessing errors. The costs are absorbed and the focus is

J.V. Rizzi is a Senior Investment Strategist at CapGen Financial inNew York, NY.

The views expressed represent those of the author, and not CapGenFinancial.

85RIZZI — BEHAVIORAL BASIS OF THE FINANCIAL CRISIS

on mitigation and prevention through improved processingand training.

Type C events include concentrated exposures to high riskborrowers. These well known risks are managed by constantmanagement monitoring and control. Type D events arefrequently ignored due to a low frequency. Examples includemany of the structured finance products which representedshort positions in an option. They offered long period ofsteady income punctuated with occasional large losses.

Cyclical risks are low-frequency-high-impact eventscharacterized by their negative skew and “fat-tailed” lossdistributions. Investors incurring such risk can expect mainlysmall positive events but are subject to a few cases of extremeloss. These risks are difficult to understand. The difficultystems from two factors. First, there is insufficient data todetermine meaningful probability distributions. In this case,the statistics are descriptive, not predictive. Consequently,no amount of mathematics can tease out certainty fromuncertainty.1 Second, and perhaps more important,infrequency clouds hazard perception. Risk estimates becomeanchored on recent events. Overemphasis on recent eventscan also produce disaster myopia during a bull market, asinstruments are priced without regard to the possibility of acrash. These facts lead to risk mispricing and the procyclicalnature of risk appetite.

Quantitative risk-management models are based onportfolio and option pricing theory and provide a framework

on how risk managers should act. These models build onexpected utility theory (EUT), which views individuals asexpected utility maximizers.2 Empirical support of EUT ismixed with numerous reported anomalies.3 Examples ofanomalies include holding losers, selling winners, excesstrading, and herding.

An alternative, prospect theory,4 can explain these facts.Instead of being expected utility (E(U)) maximizers, investorsare viewed as expected regret (E(r)) minimizers focusing moreon losses than gains. This is reflected in Figure 2.

EUT focuses on wealth changes. The value function inprospect theory is based on gains or losses relative to areference point, usually par or the original purchase price.

Behavioral finance examines how risk managers gather,interpret, and process information. Specifically, itconcentrates on perception and cognitive bias. It recognizesmodels can influence behavior and shape decisions. Thesebiases can corrupt the decision process, leading to suboptimalresults as emotions override self-control.

Market signals are complex. They include both informationand noise. Information concerns facts affecting fundamentalvalues. Noise is a random blip erroneously interpreted as asignal.5 Risk managers have developed shortcuts, rules ofthumb, or heuristics to process market signals. These belief-based heuristics incorporate biases or cognitive constraints,which will now be investigated.

A. Regret

Risk is forward looking. Regret, however, is backwardlooking. It focuses on responsibility for what we could havedone but did not do. Regret underlies several biases. We tryto minimize regret by seeking confirming data, suppressingdisconfirming information, and taking comfort that othersmade the same decision. Consequently, regret can inhibitlearning from past experiences.

Sunk costs are the first regret bias considered. Sunk-costbias involves avoiding recognizing a loss despite evidencethe loss has already occurred and a further loss is likely.

I

Impact

A C

B D

Frequency

A: High frequency/low impact events: reflected in risk pricing.

B: Low frequency/low impact events: treated as a cost of business.

C: High frequency/high impact events: managed through control.

D: Low frequency/high impact events: frequently ignored.

Figure 1. Risk Map

1This is the Knightian distinction between risk, randomness with knowableprobabilities and uncertainty, randomness with unknowable probabilities.See F. Knight 1921, Risk, Uncertainty and Profit, Houghton Mifflin, 1921.

2 M. Friedman and L. Savage, 1948, “The Utility Analysis of ChoicesInvolving Risk,” Journal of Political Economy.

3 D. Ellsberg, 1961,“Risk, Ambiguity and the Savage Axioms,” QuarterlyJournal of Economics 643.

4 A. Tversky and D. Kahmeman, 1992, “Advances in Prospect Theory:Cumulative Representation of Uncertainty,” Journal of Risk andUncertainty, which builds on their earlier work. Prospect theory is a keycomponent of Behavioral Economics. Behavioral finance is a subset ofBehavioral Economics, applying its concepts to asset pricing. This articleuses the terms interchangeably.

5 See E. Black 1986, “Noise,” Journal of Finance July.


Examples include the reluctance to sell impaired assets atreduced prices. Usually this is defended as the market pricesbeing too low. Most institutions, however, reject the logicalalternative of acquiring additional exposure at the marketprice to exploit the alleged under pricing; thus, illustrating inthis instance, price is of secondary importance relative toregret.

Panic conditions are also based on a combination of regretand herding. In a crisis, the reference is pessimism, and weactively seek bad news to confirm our belief. Thus, tominimize regret, we follow the herd not to be left behind andengage in panic selling. This further depresses prices leadingto continued forced selling and the creation of a negativefeedback loop as occurred in the fourth quarter 2008.

Another regret-related bias is the house money effect. Riskmanagers will assume greater risks when they are up in a bullmarket and lower risk in a bear market. Regret is perceivedto be less when risk of winnings is involved, than risk ofinitial capital. This procyclical phenomenon leads to “buyhigh and sell low” behavior, reflected in Figure 3.

This illustrates the George Soros reflexivity or feedbackprinciple, whereby markets affect psychology and psychologyaffects markets. Positive feedback is self amplifying, whilenegative feedback is self corrective. For example, collateralvalues rise during a bull market. This increases their accessto lower priced funding and liquidity, which fuels furthergains.

Finally, regret leads to confusing risk with wealth. Larger,better-capitalized financial institutions can absorb more riskthan smaller institutions. Their greater risk tolerance lessenstheir downside sensitivity, especially during a bull market

when income levels are high. Thus, risk appetite increaseswith wealth. Risk and return are, however, scale invariant.Larger institutions confuse the ability to absorb risk providedby capital with the desirability of the risk position. Therefore,they acquire underpriced, higher-yielding, higher-risk assetsin bull markets.6

B. Overconfidence

Overconfidence occurs when we exaggerate our predictiveskills and ignore the impact of chance or outsidecircumstances. It results in an underestimation of outcomevariability.7 Overconfidence is reinforced by self-attributionand hindsight. Self-attribution involves internalizing successwhile externalizing failure. Structured finance bankers andquantitative risk managers took credit for results during theboom, failing to consider the impact of randomness and meanreversion creating an illusion of control.8 Hindsight involvesselective recall of confirming information to overestimatetheir ability to predict the correct outcome, which inhibits

. Convex slope indicates pain of

loss (regret exceeds value of gain

. The conflict between E(u) maximizing

and E(r) minimizing underlies many

anomalies

. Investment decisions involve 3 Rs:

return, risk and regret

+

-

Reference point

Utility

Value function value

Losses Gain

Figure 2. Investors Minimize Expected Regret

6This is consistent with the H. Minsky financial instability hypothesis.Investors increase their risk exposures driving bull markets until they havetaken on too much. See H. Minsky, 2008, Stabilizing an Unstable EconomyMcGraw-Hill.

7This is magnified by the naïve use of market-based risk-managementtools.

8Studies indicated the underestimate at 15 %-25 %. The direction of theoverconfidence is usually positive reflecting a related optimism bias.


learning. Disappointment and surprise are characteristics ofprocesses subject to overconfidence.

Industry and product experts are especially prone tooverconfidence based on knowledge and control illusions.Knowledge is frequently confused with familiarity. This isreflected in the number of industry experts including mostfamously the former Federal Chairman who missed thecollapse of the housing and structured credit bottom.9 Thisis due, in part, to misguided overreliance on quantitative creditscoring models without understanding their limitations. Keymodel limitations include the following:

• Homogenous populations: Statistical models require largehomogenous populations with a long history of observations.The new structured finance credit portfolios were small,heterogeneous, and concentrated with limited histories.

• Statistical Loss Distribution: Loss distributions for creditare skewed, with unexpected event losses hidden in thedistribution’s fat tails. Models tend to be blinded by the meanand underestimate extreme events.

• Historical basis: History is a guide, not the answer. Thepast represents but one possible outcome from an event

sequence and is not an independent observation. Historybecomes less relevant as markets and underwriting practiceschange. This was especially true for mortgage default models.They ignored the impact of securitization of mortgageoriginator underwriting practices.10

• Uncertainty: Decisions involve both risk, knownunknowns, and uncertainty, unknown unknowns, elements.Financial models adequately contemplate the former butinadequately deal with the later. Managing uncertaintyrequires judgment, not calculation.

Control reflects the unfounded belief of our ability toinfluence or structure around risk. Risk is accepted becausewe believe we can escape its consequences due to our abilityto control it. Examples include the perceived ability todistribute or hedge risk, independent of the likelihood of beingbetter or faster at identifying risk than the market.

This reflects an optimistic underestimate of costs whileoverestimating gains. Optimism is heightened by anchoringwhen disportionate weight is given to the first informationreceived. This is usually based on the original plan, whichtends to support the transaction.

H

Risk Appetite

Low High

BullHigh

Risk Level

Low

HighLow

Bear

Market State

Financial Institution Profitability

Housing bubble

2004-1H07

Contrarian

2H07 - ?

Figure 3. Risk/Market Appetite, Structured Finance

9Inappropriately designed incentive compensation reinforcesoverconfidence.

10U. Rajan, A. Seru, V. Vig, “The Failure of Models that Predict Models:Distance, Incentives and Defaults,” Working Paper, September, 2008.


Time-delayed consequences magnify overconfidence asindividuals weigh short-term performance at a higher levelthan longer-term consequences. These occur whenever short-term benefits clash with long-term effects. Although we knowof the potential negative long-term effects, we believe thatthey will not happen to us, at least during the currentaccounting period. An example is dropping creditunderwriting standards to increase short-term income, marketshare, or league table statusas occurred during the heightof the boom.

C. Statistical

Statistical bias involvesconfusing beliefs forprobability and skill forchance by selecting evidencein accordance with ourexpectations.11 Economics isa social science based on human behavior. Prices are notdetermined by random number machines.12 Rather, they comefrom trades by real people. Feedback loops, prices, tradesand people complicate statistical modeling, and invalidatethe use of normal distributions as used in the physical sciences.

Institutions find it difficult to accept chance and arefrequently fooled by randomness. A manifestation is therepresentative bias, whereby we see patterns in randomevents. We interpret short-term success as “hot hands” by askilled banker. Risk-adjusted return on capital and othermeasures are unable to distinguish results based on luck versusskill.

Statistically based risk management practices are inherentlylimited. They are unable to reflect the hidden risk that thestate of the world may change rendering current state dataobsolete. For example, switching from a boom to a bust cycleimpacts correlations. Formerly diversified positions beginmoving together, triggering unexpected losses. They areunexpected because such movements are unfamiliar. We tendto view the unfamiliar as improbable, and the improbable isfrequently ignored.

Actions and outcomes can be unrelated. Consequently, itbecomes important to examine the decision process and notjust the outcome.13 As Scholes notes, to value risk or pricereserves you must reflect the values of the options not

purchased to hedge the position. Since this is not priced, itcreates incorrect capital allocation incentives.14 Thus, the“lucky fool” is rewarded and encouraged with bonuses andincreased capital until luck turns and losses are incurred.Examples include the numerous apparently lucky real estateexperts at institutions like Bear Stearns and Lehman.Eventually, all lucky streaks come to an end as this one didduring the summer of 2007.

Another statistical errorprevalent during a boom isextrapolation bias. Thisoccurs when current events ortrends are assumed tocontinue into the foreseeablefuture, independent ofhistorical experience, samplesize or mean reversion.Undoubtedly, this resulted inmany of the projectionsunderlying structured credit

proposals. The major error focused on the belief that housingprices would not decline nationwide in the US

Perhaps the most dangerous statistical bias is disastermyopia. This occurs whenever low-frequency but high-impact events are underestimated. Since the subjectiveprobability of an event depends on recent experience,expectations of low-frequency events, like a market or firmcollapse, are very small. These types of events are ignoredor deemed impossible, particularly when recent occurrencesare lacking. This causes a false sense of security as risk isunderestimated, or assumed away, and capital is misallocated.Unlikely events are neither impossible or remote. In fact,unlikely events are likely to occur because there are so manyunlikely events that can occur.15 Thus, the longer the timeperiod, the higher the likelihood of a “Black Swan” eventoccurring.16

D. Herding

The previous discussion concerned individualpsychological aspects of risk decision making. There arealso social aspects to decision making when individuals areinfluenced by the decisions of others as reflected in herdingand ‘group think’.

11See P. Bernstein, 1996, “The New Religion of Risk Management,”Harvard Business Review (March-April).

12W. Sharpe, 2007, Investors and Markets, Princeton University Press 11.

13P. Rosenweig, The Halo Effect, Free Press, 2007.

15P. Bak, 1996, How Nature Works Springer-Verlay.

16Black swans are high impact unexpected rare events. The term waspopularized by N. Taleb in The Black Swan: The Impact of the HighlyImprobable (Random House, 2007).

Behavioral finance examines howrisk managers gather, interpret,and process information. Itrecognizes that models caninfluence behavior and shapedecisions.

14M. Scholes, “Crisis and Risk Management,” AEA Papers and Proceedings,May, 2000


Herding occurs when a group of individuals mimic thedecisions of others. Through herding, individuals avoidfalling behind and looking bad if they pursue an alternativeaction. It is based on the social pressure to conform, andreflects safety by hiding in the crowd.17 In so doing, you canblame any failing on the collective action and maintain yourreputation and job. Even though you recognize market risk,it pays to follow the crowd. Managers learn to manage careerrisk by clinging to an index. Essentially, principal loss isconverted into benchmark risk.

Herding reduces regret by rationalizing that you did noworse than your peers. It constrains both envy during anupswing and panic in a down market. This is critical inbanking when performance contracts are based on relativeperformance measures tied to peer groups.18 Herdingunderlies why banking experts’ forecasting abilities are poor.The experts tend to play it safe by staying close to the crowdand extrapolating past performance.19

A related effect is an informational cascade. A cascade isa series of self-reinforcing signals obtained from the directobservation of others. Individuals perceive these signals asinformation even though they may be reacting to noise. Thisis referred to as a positive feedback loop or momentuminvesting, which can produce short-term self-fulfillingprophecies.

Herding amplifies credit cycle effects, as decisions becomemore uniform. The cycle begins with a credit expansionleading to an asset price increase. Investors rush in to avoidbeing left behind using rising asset values to support evenmore credit. This explains why bankers continued riskpractices even though they feared this was unsustainable andleading to a crisis. Eventually, an event occurs, such as amove by the central bank, which triggers an asset pricedecline. This causes losses, a decline in credit, and an exit ofinvestors, which strains market liquidity.

E. Group Think

Group think, or organizational pressure, enhances cognitivebiases. It occurs when individuals identify with theorganization and uncritically accept its actions. Once thecommitment is made, inconsistent information is suppressed.

Consequently, mutually reinforcing individual biases andunrealistic views are validated.20

Experts are prone to group think. They tend to limitinformation from all but other expert sources. Thus, theyrepeat statements until they become accepted dogmaregardless of their validity, due to a lack of critical thinking.

The recent subprime collapse illustrates this fact. Theindustry participants used the same consultants and modelsfor their projections. The consultants based their reports andrecommendations on the surveys of industry participants.Once the perception of a bull market took hold, it wasreinforced and accepted uncritically. When the crashoccurred, the experts were taken by surprise by a supposedperfect storm.

This is illustrated in the 2006 Business Week cover story inwhich risk officers at numerous institutions, including BearStearns and Lehman, are surveyed.21 They believed thatdespite the risks taken they were safer than ever. This beliefwas based on complex risk models and market diversification.The faith in risk management encouraged institutions toincrease their risk exposures, believing they were undercontrol.

F. Sentiment Risk

The aggregate investor error based on biases is sentimentrisk. It can be either optimistic or pessimistic and is timevarying as reflected in Figure 4.

Sentiment risk is zero in an efficient market. PaulSamuelson has noted markets in the short-term can be microefficient concerning individual instruments, but macroinefficient regarding the market as a whole. Additionally,during the short-term the direction of the inefficiency is likelyto widen due to momentum and herding.22 Most risk modelsignored sentiment risk. This causes losses when sentimentchanges leading to closed markets and mark to market losseswhich has threatened the basis of originate to distribute model.

Investor responses based on the interplay of sentiment andmarket valuation is reflected in Figure 5.

During a late stage boom with high sentiment levels, A,behavioral risk factors will dominate and quantitative riskmeasures will be unreliable. This is reflected in the famouscomment “As long as the music is playing, you have to getup and dance”. This is characterized as irrational exuberancewhere prices are driven principally by momentum and herding

18The industry expert impact is significant, as most large financialinstitutions adopted best practices based on similar experts.

19Relative performance measures are a form of sophisticated “me-too”metrics. Rather than focus on absolute value creation, they focus onarbitrary market silos that may be in a downturn.

21E. Thornton, D. Henry and A. Carter, 2006, “Inside Wall Street’s Cultureof Risk,” Business Week (June 12).

22H. Shefrin, 2008, Ending the Management Illusion, McGraw-Hill.

17This is reflected in Keynes” famous statement that it is better for abanker ’s reputation to fail conventionally, than to succeedunconventionally.

20See J. Chevalier and G. Ellison, 1997, “Risk Taking by Mutual Funds asA Response to Incentives,” Journal of Political Economy.


reflected in high liquidity levels. When sentiment is low,fundamentals will rule as in B and C. Price may divergefrom fundamentals23, but are quickly eliminated byarbitrageurs in B and C. D represents irrational despondencyfound in market bottoms reflected in tight liquidity.

II. Remedies

Behavioral finance demonstrates how biases influence riskperception leading to underestimation of improbable events.We base our actions on experience of what has happened.This ignores beyond the data exposures leading to futureblindness. Consequently, we misjudge actual risk leading tosurprise losses.

Recognizing and dealing with biases is complicated by threefactors. First, bias can be amplified within organizations dueto incentive misalignment and group think. Next, we diminishinformation inconsistent with our existing views, whilesearching for conforming information. Finally, this leads to

Time

EC

D

BA

EA: fundamental value

EB: financial Price

C: credit bubble: optimism

D: liquidity trap: pessimism

Figure 4. Senitment Risk

Market

ValuationBoom

Bust

A B

D C

High Low

Sentiment

Figure 5.

23See T. Debels, Behavioral Finance (Garant Uitgevers, 2006) 183 for adiscussion of various forms of behavioral finances that can occur inmarkets.


a false sense of security and reduced vigilance. Potentialresponses to reduce biases will now be explored.

A. Principles

Historically, risk management has been primarily anexposure accounting and control system. While thiscontrollership reportingfunction is important,protection is needed againstit becoming a regulatory andquantitative ritual. Theemphasis should be onforward-looking dynamicmanagement, which involvesthree components:

1. Risk appetite involvesdetermining how much aninstitution is prepared tolose, and is frequently defined by earnings at risk or a potentialratings downgrade.

2. The actual institutional risk profile is monitored throughscenario analysis and stress testing applied against actualportfolio movements.

3. Corrective action is taken when a mismatch between theorganization’s risk profile and appetite occurs. The correctiveaction can be at the transaction, when involving majorexposures, or portfolio level.

Investors have difficulty in processing market signals. Thiscomes from a failure to distinguish noise, price movementswithout meaningful changes in economic prospects, from trueinformation. This difficulty places decisions at risk (DAR)as reflected in Figure 6.

Uncertainty, as opposed to risk, is difficult to manage dueto several biases. Chief among these are optimism andoverconfidence based on an illusion of control based onflawed models. These biases are amplified in certainorganizations by compensation and governance problems.Bureaucracy and opaqueness inhibits responses until it is toolate, leading to massive losses.

B. Alternatives

Practical alternatives to biased based bounded rationalityexist. Some options include:

• New markets and product limits: Behavioral bias isstrongest in areas where inexperience reigns, such as in thestructured finance area with its new technology and limitedhistory. Thus, strict limits to control exposure in these areasare needed.

• Obtain a second opinion: Anticipation of review by anunrelated third party encourages greater care.

• Beware of experts: Seek diversity to avoid myopic focuson issues within an expert’s area of interest. Expertsfrequently ignore the benefits of alternatives.

• List a wide variety of possible scenarios: Focus onunpopular and unlikelypossibilities. View the futureas a collection of eventualitiesrather than as a singleprediction. The preferreddecision is one that worksacross several possibleeventualities, and not just thecurrent market state.

• Avoid herding: Developindependent analysis. Thisrequires encouraging

contrarian views supported by compensation programs.• Postmortem: Review both successful and unsuccessful

decisions. The focus should be on whether the results wereluck or skill-based. The key is to avoid rationalizing andhindsight bias and to learn from the experience.

• Directly engage the environment: Independentinvestigation is needed to verify and to avoid filteredinformation.

• Heterogeneous risk team: Construct a diverseindependent risk team. Rotations can be used to maintaindiversity.

• Lengthen risk horizon: Given the long-tail nature of creditrisk, increase the evaluation horizon beyond the traditionalaccounting-based yearly horizon.

No process is foolproof. A backup procedure is needed toremove the temptation to accept unintended catastrophic risk.This is provided by portfolio control and enforced by theboard of directors. Excess concentrations must be reducedor covered by additional capital. Also, strategies are neededfor each position, allowing adaptation to random changes inmarket states. Institutions should invest in portfolio strategies,not in illiquid excess concentrations, which have twocomponents.

The first concerns risk acceptance or transactional approvalbased on the institution’s risk underwriting criteria. Secondis risk reduction through diversification. Diversification,however, does not prevent losses. Rather, it prevents losingeverything at one time. The focus is on position size to avoidover betting. Institutions must guard against unproductivenaïve diversification, which emphasizes the number ofportfolio assets instead of their asset-class diversity. This iscritical because correlations are scenario specific and

No matter how remote, high-impact events cannot be ignoredbecause they can threaten aninstitution’s existence, as wasdemonstrated in the currentmarket crisis.


approach one during a crisis. Thus, the key is the overallasset-class allocation and not necessarily the number of assetsin an asset class.

III. Current Crisis

A. Setting

A herding among financial institutions occurred during thelast several years. Consequently, they invested too much atthe same time in the same areas. This was done under guiseof the “originate to distribute” model. This model allowedinstitutions to rationalize poorly structured, underpricedproducts by selling them to others. They failed, however, toconsider the impact of markets closing down, leaving themwith large high risk exposures.

Some institutions like Wells Fargo and Pittsburgh Nationalescaped the herding. It is difficult to determine if they werelucky or smart. The pressure to herd is illustrated by MorganStanley. Under its previous management, Morgan Stanleyrefused to participate in structured products. Its performancesuffered relative to its peers. Consequently, in 2005 it wasreplaced by a new team. They vowed to regain market shareby matching its peers, which it achieved in 2008 by recordingrecord losses.

A declining economy and falling markets triggeredaggressive Federal Reserve interest rate cutting and liquidityinjections in 2001 to 2002. Liquidity-driven technicalsimproved, resulting in falling risk premiums increasing creditasset prices. Institutions responded by adopting an asset-intensive carry trade strategy, which involves borrowing short-term to invest in longer-term risk assets.

A credit bubble formed as liquidity-driven technicalssurpassed fundamentals. This was reflected in historically lowcredit-risk spreads in the real estate, leveraged buyout andstructured credit markets. Spread narrowing and a flattening

yield curve reduced the attractiveness of the carry trade,putting pressure on institutional accrual and trading budgets.

In the search for yield, institutions adopted a procyclicalasset heavy 5Ls strategy:

• Longer duration (e.g., mortgage backed securities)• Long tailed option type risk found in the AAA tranches

of structured securities• Large positions (e.g., multibillion dollar mortgage

warehouse facilities)• Leverage levels approaching 30:1• Less liquid assets (e.g., collateralized debt obligations)• The 5Ls strategy involves going long on higher-risk assets

for the institution’s own account instead of distribution. Thestrategy is reflected in principal finance, merchant banking,bridge loans and warehousing activities. These activitiesrepresented up to 75% of revenues at some institutions.

The risk inherent in the 5Ls strategy were obscured byjudgment biases in the following areas:

• Unproven business models were justified based onoptimistic plans while down-playing the negative possibilitiesdue to group pressures.

• Institutional overconfidence in risk management modelsbased on the illusion of control. This causes an observationof safety, which creates as illustrated by the Peraud Paradox,risk.24

• Peer pressure that was not based on independent economicreasoning. Nonetheless, as the Morgan Stanley exampleillustrates, ignoring peer pressure can be hazardous to yourcareer.

24Persaud Paradox is the observation of safety created by using the samemodels as your peers, which creates model risk.

Figure 6. Decisions at Risk (DAR)

Uncertainty Bias Amplifiers

Beyond the data events Optimistic Incentives Experiences Overconfident Bureaucracy Exposures Illusion of control Opaqueness Black Swans Rare Large Impact


This caused a major credit boom. During the boom, pricescan exceed underlying fundamental economic values asillustrated in Figure 4. Such cycles, while predictable, aredifficult to manage for several reasons. First, financialinstitution compensation is tied to peer group comparisons.Thus, firms and individuals not following their peers suffer.Next, organizations frequently discourage pessimism.Therefore, conservative risk managers and bankers arepressured to become optimistic or leave. Finally, institutionsrisk losing bankers if their risk activities are curtailed.

Frequently, positive short-term results mask long-term risks.Seemingly high returns can reflect the subjective probabilityof an event that has not occurred in the time period studied.Investing in such instruments is profitable most of the time.Eventually, a beyond the data event occurs. The housingevent occurred in mid-2007 and has continued for more thaneighteen months, costing billions in provisions. Individualsand institutions succumbed to a bias of “assuming the absenceof evidence implied evidence of risk absence”.

B. Concerns

The appropriateness of the 5Ls portfolio strategy dependson several factors. First, it works best early in the cycle beforethe opportunities are exploited by the competition and spreadsnarrow.

Next, the strategy involves incurring increased systematicor beta risk exposure versus value-adding alpha returns.Structured products are less liquid than market investments.Consequently, the return on structured products reflectscompensation for liquidity risk. This risk was poorly reflectedin risk management models. The liquidity premium wasmischaracterized as alpha. Thus, liquidity risk was underreported. This was subsequently discovered during the crisis.

Finally, pricing and trading discipline is needed to ensurean adequate risk premium is earned. Maintaining disciplinebecomes increasingly difficult as the cycle continues.

Warning signs began to form during the first six months of2007:

• Continued Federal Reserve tightening• Rating agency downgrades• Flattening yield curve• Increased mortgage defaults

Unfortunately, apparent success breeds an inability toimagine the possibility of failure, and the warnings wereignored. Firms continued to underestimate the likelihoodand impact of unlikely events. Widespread credit risk underpricing existed due to an emphasis on nominal returns. Thissuggests a correction when investor emphasis shifts fromreturn on capital to return of capital.

It is difficult to price rationally when risk seems remoteand hard to measure and conditions seem favorable.25 Thelast market correction had occurred more than three yearsago and was largely forgotten by the first half of 2007. Thus,risk sensitivity had diminished. This recognition problem isrooted in the complex nature of cyclical risk.

C. Regime Changes

Procyclical risk appetite and feedback loops underlie creditcycles.26 As risk appetite increases, credit extension expands.Investors use the increased debt capacity to bid asset priceshigher. The higher asset prices increase collateral values,which supports additional credit expansion creating a virtuouscredit cycle with increasing liquidity. A tipping point or eventcan, however, prompt investors to adjust simultaneously theirpositions triggering a decline in asset prices.27 Figure 7 showsthat the decline can trigger a vicious cycle leading to reducedcollateral values, curtailed credit, declining investor demand,falling asset prices and reduced liquidity.

The tipping point represents a change in investor sentimentbased on an awareness of the risk that investors have assumed.Once the tipping point is reached, feedback overwhelmsfundamentals and the trend dominates. Tipping pointsrepresents triggers, not causes of the change in investoractions. Overvalued assets, which are vulnerable to bad news,are prone to volatile investor sentiments. Thus, tipping pointsare unexpected and occur during the height of an over-valuedbull market.

D. Reinforcement

The complexity of low-frequency/high-impact cyclical riskis compounded by institutional factors such as budgets andcompensation systems that reinforce the behavioral biaseffect. These systems favor “consistent” earnings and misreadlow –frequency/high-impact risk “profitability.” Such risk issimilar to underwriting out-of-the-money put options. Thepremiums appear profitable until the put event occurs.

Risk models also contribute to the problem by presentingthe illusion of safety and control, leading to over optimism.

26Procyclical risk appetite is aggravated by ratings-based regulatory capitalrequirements. The regulatory rules reduce capital requirements during abull market as ratings increase, thereby encouraging credit expansion.Conversely, they increase capital requirements during a bear market asratings are pressured, leading to a credit contraction.

27See D. Sornette, 2003, Why Markets Crash, Princeton, NJ, PrincetonUniversity Press. This is similar to physical events such as forest firesand earthquakes arising from “criticality.”

25President of New York Federal Reserve T. Geither, as quoted in theFinancial Times, May 12, 2005.


Backward-looking risk models confuse history and science.Unfortunately, financial markets are not actuarial tables. Someimportant model issues include the following:

• Inadequate consideration of the cyclical effect on, andcorrelation among, probability of default, loss given defaultand exposure at default

• Poor understanding of the interaction between liquidityand credit risk as bull markets create their own liquidity, whichcan evaporate in a downturn

• Feedback impact of models on markets is ignored. Theobservation of safety created by using the same models asyour peers creates model risk.28

• Difficulty reflecting out-of-sample, beyond-the-datapossible effects.

Consequently, models underestimated low frequency/high-impact cyclical risk. The underlying exposure builds duringa bull market as apparent risk declines, while the lossesmaterialize in the bear market cycle. This anomaly is due tosocial and psychological biases resulting in boundedrationality. Ignoring these facts substitutes an inaccuratenormative model for the real world.

The objective is to supplement existing quantitative riskmanagement with developments taken from the evolving fieldof behavioral finance. In so doing, it can reduce future losses

Virtuous Cycle

(Disaster Myopia)

2004-1H07

Vicious Cycle

(Disaster Magnification)

2H07 - ?

Asset Prices

Collateral

Values

Credit

Investors

Tipping Point

Investors

Asset Prices

Collateral

Values

Credit

Figure 7. Credit Cycle

during the credit cycle as risk management evolves to a morebalanced system, incorporating human behavior. Thisrequires taking low-probability-worst-case scenariosseriously, and developing appropriate responses. The processis similar to earthquake engineering, which does not attemptto predict a shock. Rather, the focus is on constructing astructure to withstand a certain shock level.

Currently, counter cyclical capital charges decrease duringbull markets as ratings improve as demonstrated in Figure 3.This fact underlies the procyclical bias in portfolio strategiesas lower bull market capital requirements increase returns,encouraging an inappropriate, asset-heavy, 5L portfoliostrategy. Supplementing currently determined capital chargeswith a requirement tied to asset prices would encourage ashift to a counter cyclical portfolio strategy. Capital levelswould relate to changes in asset prices. The higher capitalallocation serves as a risk-taking budget constraint duringbull markets by dampening compensation-related returns. 29

E. Lessons Learned

Risk decisions are at a risk from behavioral bias. This isespecially true when dealing with high impact low probabilityrisks. Governance mechanisms represent possible control overthe bias by introducing outside viewpoints. The specificaction taken depends on the source of the bias. Figure 8

29See C. Goodhart, 2005, “Financial Regulation, Credit Risk, and FinancialStability,” National Institute of Economic Review ( Apr.il), 118.

28This is the “Persaud Paradox,” where the observation of safety createsrisk. See A. Persaud, 2003, Market-Liquidity and Risk management,Liquidity Black Holes, London: Risk Books, Ch. 11.


reflects the possible combination of market and managerialbias.

Biased managers operating in an efficient market, A, needto be protected by their boards of directors and regulatorsfrom overreacting to market noise through tight controls.Classical financial theory is represented in B with efficientmarkets and rational managers who require limited oversight.Rational managers operating in a biased market, C, willexploit market inefficiencies by selling over priced claims.When both managers and markets are biased, D, whichcharacterized the late stage of the boom, the situation becomesproblematic. Boards and regulators are likely to fall prey tothe same behavioral biases as affecting mangers and controlsare likely to fail.

IV. Conclusion

Presently, it is difficult to consider the end to the bearmarket. Defaults are increasing and liquidity remains fragile.This difficulty is compounded by behavioral bias reinforced

30R.C. Merton, interviewed by N. Nickerson, 2008, “On Markets andComplexity,” Technology Review (April 2).

Figure 8. Bias Response Choices

by institutional factors. While no two cycles are identical,we must resist the temptation to say, “This time is different.”The deeper we are into illiquid credits, products andstructures, the more difficult it becomes to manage risk. Thekey is to identify potential adverse scenarios, stress-test todetermine their impact, compare the test results to our riskappetite and take appropriate portfolio decisions. This entailsadopting counter cyclical portfolio strategies despite negativeshort-term revenue implications. This also requires adoptingdifficult infrastructure changes.

Organizational obstacles inhibit appropriate responses tohigh-impact low-probability risks. Chief among the obstaclesare short-term compensation systems which reinforcebehavioral biases. This leads to a potentially fatal neglect ofthe longer-term build of risk. As Robert Merton noted “Theamount of risk we take personally, individually, or collectivelyis not a physical given constant. We chose it.”30 Behavioralfinance offers a means to choose wisely, as it affects bothindividual decision making and market efficiency. You ignorebehavioral risk at your own peril.

A

A B

CD

Biased Rational

Biased

Rational

Investor

Markets

Managemen


References

Bak, P., 1999, How Nature Works: The Science of Self-organized Criticality, Springer Verlag, New York.

Bernstein, P., 1996, “The New Religion of RiskManagement,” Harvard Business Review 72 (No. 2), 47-51.

Black, F., 1986, “Noise,” Journal of Finance 41 (No. 3),529-543.

Chevalier, J. and G. Ellison, 1997, “Risk Taking by MutualFunds as A Response to Incentives,” Journal of PoliticalEconomy 105 (No. 6), 1167-1200.

Debels, T., 2006, Behavioral Finance, Garant Uitgevers,Antwerpen.

Ellsberg, D., 1961, “Risk, Ambiguity and the SavageAxioms,” Quarterly Journal of Economics 75 (No. 4), 643-669.

Friedman, M. and L. Savage, 1948, “The Utility Analysis ofChoices Involving Risk,” Journal of Political Economy56 (No. 4), 279-304.

Goodhart, C., 2007, “Financial Regulation, Credit Risk andFinancial Stability,” National Insitiute of Economic Review192 (No. 1), 118-127.

Merton, R.C., 2008, “On Markets and Complexity,” Interviewby Nate Nickerson, Technology Review (April 2).

Minsky, H.P., 2008, Stabilizing an Unstable Economy,McGraw Hill Professional, Columbus, OH.

Persaud, A., 2003, Liquidity Black Holes: Understanding,Quantifying and Managing Financial Liquidity Risk, RiskBooks, London.

Rajan, U., A. Seru, and V. Vig, 2008, “The Failure of Modelsthat Predict Failure: Distance, Incentives and Defaults,”Chicago GSB Research Paper No. 08-19, Ross School ofBusiness Paper No. 1122.

Rosenzweig, P.M., 2007, The Halo Effect—and the EightOther Business Delusions that Deceive Managers, Simonand Schuster, New York.

Scholes, M., 2000, “Crisis and Risk Management,” AmericanEconomic Review 90 (No. 2), 17-21.

Sharpe, W. F., 2006, Investors and Markets: PortfolioChoices, Assets Prices, and Investment Advice, PrincetonUniversity Press, Princeton, NJ.

Shefrin, H., 2008, Ending the Management Illusion,McGraw-Hill, Columbus, OH.

Sornette, D., 2003, Why Markets Crash, Princeton UniversityPress, Princeton, NJ.

Taleb, N., 2007, The Black Swan: The Impact of the HighlyImprobable, Random House, New York.

Thornton, E., D. Henry, and A. Carter, 2006, “Inside WallStreet’s Culture of Risk,” Business Week (June 12).

Tversky, A. and D. Kahmeman, 1992, “Advances in ProspectTheory: Cumulative Representation of Uncertainty,”Journal of Risk and Uncertainty 5 (No. 4), 297-323.

University of Rochester Roundtable onBankruptcy and Bailouts: The Case of the

US Auto IndustryThe GeVa Theatre Rochester, NY February 2, 2009

97

Panelists: Thomas Jackson, Charles Hughes, James Brickley, Joel Tabas, and CliffordSmith

Moderator: Mark Zupan

Mark Zupan: Good evening, and welcome to thisdiscussion of a very topical and pressing issue: today’sproblems with the US auto industry, and the potential role ofbankruptcy in dealing with them. I’m Mark Zupan, Dean ofthe University of Rochester’s Simon School of Business, andI will be serving as moderator.

I’m not going say much about the topic itself—I’ll leavethat to our panelists, who are the experts. What I will tellyou is that bankruptcy, like business school applications, is a“negative beta” activity. In other words, when the market’sup, both business school applications and bankruptcy casestend to go down. But when the market’s down, ourapplications go in the reverse direction, and so does theamount of attention and effort devoted to bankruptcy.

We have five panelists tonight. Three of them—TomJackson, Cliff Smith, and Jim Brickley—are distinguishedacademics from the Simon School faculty. The other two—Charlie Hughes and Joel Tabas—are both Simon Schoolalums who have gone on to become accomplished“practitioners” in their own fields, Charlie as an auto companyexecutive and Joel as a bankruptcy lawyer. I’ve asked eachof our five panelists to provide a brief statement of theirthoughts on the problems of the US auto industry, and possiblesolutions, including Chapter 11, to those problems. After wehear from each of them, we’ll open up the discussion toquestions from the audience.

Our first speaker will be Tom Jackson. Tom—along with

his former student, Douglas Baird, former dean of theUniversity of Chicago Law School—is widely regarded asone of the world’s top two authorities on US bankruptcy law.We feel very privileged to have him at the University ofRochester. From 1995-2005, Tom served as President of theUniversity. Since stepping down from that position, he hasheld joint appointments at both the Simon School and in theUniversity’s political science and economics departments.Before coming to Rochester in ’95, he was the provost anddean of the University of Virginia Law School.

So, Tom, would you please start things off for us?

I. The Social Function of Bankruptcy:Uses and Limitations

Tom Jackson: Thanks, Mark, for the kind words. Let mestart by saying how much I appreciate this Depression-erastage set that GeVa has provided as the backdrop for ourdiscussion tonight—it seems very appropriate for the topic.

I want to begin this discussion by providing a broadeconomic framework for this issue of bankruptcy vs. bailoutsbecause I suspect we haven’t seen the last of businesses—orindustries—facing such choices. My field, as Mark told you,is bankruptcy—and bankruptcy is a process for reorganizingtroubled companies that is rooted in the economic goal ofincreasing efficiency. Bailouts, by contrast, are a means of


rescuing troubled companies where, for good or ill, politicstend to mix with and override fundamental economicconsiderations. So I’d like to talk about what bankruptcycan do, and perhaps what it can’t—and I’ll do so in the contextof the recent controversy over about what the Detroitautomotive manufacturers should have done.

Chapter 11 is designed to do one thing well—and, for themost part, I think it does so. And that is to rearrange thecapital structure of companies with more debt than assets toallow those that should survive to survive—and allow thosethat should fail to fail. The criteria for survival in such casesare economic ones: can the troubled company, if properlyreorganized and recapitalized, be made profitable enough forits new investors to earn a fair rate of return on their money?If the answer is yes—in which case, presumably, the newcapital will be provided—the company gets reorganized underChapter 11. But if the answer is no, the best outcome for theoriginal investors is to shut down the business and sell theassets piecemeal to the highest bidders, either in Chapter 11or after converting to Chapter 7.

Whether bankruptcy or bailouts, however, it’s importantto recognize that there is a difference between financialfailures and business failures. Financial failures are caseswhere the assets, although valuable when kept together aspart of a going concern, are worth less than the liabilities—and these companies, as a general rule, get reorganized inand come out of Chapter 11. Business failures, by contrast,are cases where the assets themselves are worth less whencontinued as part of a firm—even if the firm were to berecapitalized or given new money—than sold off piecemealto new owners. In practice, of course, we often see elementsof financial and business failure mixed together. But Chapter11 is premised on the idea of separating these two ideas insuch a way that companies facing a financial but not a businessfailure will be reorganized and continued—and businessfailures will be sold off in parts.

To see this distinction, consider the case of Johns Manvillein the 1970s, a company that appears to have been a veryefficient manufacturer of building supplies. The companybecame hopelessly insolvent not because of any problemswith its then-current business line, but because of the tortliability associated with its manufacturing of asbestos 20, 30,and 40 years earlier. Keeping the company going—whichrequired writing down the claims against it and convertingmany of them to equity interests—was the right outcome sinceManville’s was a financial and not a business failure. And,again, Chapter 11 is designed to do just that.

Conversely, one can have a business failure without afinancial failure. My family had a business in Kalamazoo,Michigan that made gas lights at the turn of the century—abusiness that was not a growth industry in a world of electriclight bulbs. Now, because it made very little use of debt, the

business was able to survive and be converted, over the courseof 50 years, into one that makes pneumatic air cylinders—which it continues to do to this day. But if that business hadinstead been financed with debt, it would almost certainlyhave filed for bankruptcy. Unless new owners and investorscould be convinced that the existing management couldeffectively make the transition to a new business, the assetswould have been sold off in a Chapter 7-type proceeding—and, sooner or later, someone else would have entered thebusiness of pneumatic air cylinders.

But as I suggested earlier, most corporate failures—eventhose in very large companies—tend to result from a mix offinancial and business failure. Part of the blame in such casescan be laid to having the wrong business model, and thecurrent management team may not be quite up to the task.But much of the current problem can also be attributed topast business mistakes in combination with accumulated debtsand liabilities that the current management may or may notbe responsible for.

And before one can discuss Detroit—and bankruptcy—one needs to figure out which model it fits: Is it mainly afinancial failure, a problem that can be addressed largely byrewriting claims and contracts and providing new capital? Isit really at bottom a business failure? Or does it have elementsof each that need to be addressed? And, I hasten to add, thesame questions need to be asked when designing government“bailouts” as well. It makes no sense to bail out a failedrestaurant that was operated by mom and pop. Mom andpop will leave the scene, and someone else will take theirplace. Any intervention by government will only make thingsworse.

Detroit has a 40-year—perhaps longer—history ofdecisions and actions that, in retrospect, have turned out tobe wrong. Some, though by no means all, can be blamed onpast management. As a result, one or more of themanufacturers in Detroit are almost certainly insolvent in theclassic sense: that is, their liabilities exceed their assets. Anysolution to Detroit’s problems has to figure out how to getthese things back in line. I suppose giving them money fromthe government is one way to do it. But is it the best way?

And when it comes to addressing the question of businessfailure, one or more of the manufacturers in Detroit areprobably also not “efficient” producers any more. But, again,that’s not necessarily because its current management isincompetent, but because the accretion of mistakes over thepast 40 years has produced manufacturing operations thatare not as efficient as its competitors’.

But other than noting the consequences for operationstoday, the real need here isn’t to explain the past. The mostimportant, and often overlooked, question is how to deal withthe future. How do we identify and save those parts of theUS auto industry that are worth saving? And how do we

99JACKSON, HUGHES, BRICKLEY, TABAS, AND SMITH — BANKRUPTCY AND BAILOUTS

ensure that whatever companies emerge from the current messare profitable enough to stand on their own, and so avoidcreating permanent corporate dependents?

And I think it’s important here to begin by identifying thefundamental issue, one that often seems to be ignored in thecurrent debate: Is there too much manufacturing capacitygoing forward in the USauto industry? I’d say“yes, without question.”Rather than a baseline of16 million cars, we needto contemplate a baselineof 12-13 million cars.Auto manufacturing, tobe sure, has always beena cyclical business—again, I know first-hand,having grown up inMichigan. Cyclicalbusinesses will fluctuate.But there are a lot ofreasons—cars that lastlonger, perhaps a shift incars as a “status symbol,”and the reality that, evenin the early years of thisdecade, demand seemedto be kept artificially highthrough a number ofdevices such as “rebates”and fleet sales—to thinkthat the fluctuations arelikely to be around amedian level that is two or three million vehicles smaller thanit had become over the past decade.

Now, if these estimates are correct, then that is the gorillain the corner. It means that we need to pull huge capacity outof the system. We can take it out of one or more of the Detroitmanufacturers, or we can take it across the board—but eitherway, the capacity needs to come out. We need to deal withthe consequences of doing that. It won’t be pretty. It’s goingto mean shutting down plants, car dealers, and suppliers—and putting people out of work. Once you start with thispremise, you then have to ask which method, bailout orbankruptcy, is likely to accomplish this downsizing in themost cost-effective way.

Now, it’s probably true that if you decide to take capacityout of the automobile industry as a whole rather than justDetroit, you will “save” jobs. But that is true preciselybecause Detroit is less efficient than the rest of the industry;any time you take jobs out of companies that are moreefficient, you probably save jobs. But this is as perverse as it

sounds: by trying to prop up less efficient enterprises, youimpose large costs on the rest of the economy—on USconsumers, who end up paying higher prices for cars; on UStaxpayers, who foot the bill for today’s (as well as tomorrow’s)subsidies. You also impose costs on other, more efficientcompetitors who, although they may be foreign companies,

employ lots of USworkers—and thesecompanies will getdragged down by theexcess capacitypreserved by any bailout.

What basis do I havefor my claim that Detroitis less efficient? Thereare many ways to countit, but let me name just afew. Let’s start with thenumber of differentkinds of vehicles. GM,which has well over ahalf-dozen major“brands” of cars in theUS alone—not countingdistinct brands such asHolden in Australia orOpel in Europe—is theonly manufacturer in theworld I can think of withmore than three lines inone country. Along withtoo many models, GMalso has far too many

dealers. Both of these are the consequence of early- to mid-20th-century mergers and an earlier strategy that is reflectedin the company’s name—General Motors. With a businessmodel that seemed to work in the 1950s, GM encouragednew buyers to start by buying Chevys and, as they workedtheir way up the economic ladder, to move to Pontiacs, thenOldsmobiles, then Buicks, and finally Cadillacs. But thatmodel made less and less sense as we entered the latter partof the 20th century and the first part of the 21st century.Changing strategies, under the best of circumstances, wouldhave been difficult—although that doesn’t explain why GMcontinued to add brands, such as Saturn and Hummer. Andchange was made much more difficult by a franchise systemfor dealers that, with the help of state politicians and law,was effectively frozen in place—and ensured the continuedexistence of too many brands.

Besides too many models and dealers that cannot bedropped without major expense, another cause of Detroit’scurrent problems was their successful efforts to persuade

We need to pull huge capacity out of thesystem. We can take it out of one or moreof the Detroit manufacturers, or we cantake it across the board—but either way,the capacity needs to come out. We needto deal with the consequences of doingthat. It won’t be pretty. It’s going tomean shutting down plants, car dealers,and suppliers—and putting people out ofwork. Once you start with this premise,you then have to ask which method,bailout or bankruptcy, is likely toaccomplish this downsizing in the mostcost-effective way.-Tom Jackson


lawmakers to limit foreign competition in the 1960s and1970s. Part of the justification for allowing Detroit to beprotected by such barriers to entry were many of the samearguments that we hear today for a bailout, including thedesirability of protecting Detroit’s ways of doing businessand the high wages that came with them. But such wages ofcourse translated into the high labor costs that plague theindustry today, as well as its continuing reputation—fair ornot—for producing a lower-quality product. In other words,by succeeding in its efforts to limit foreign imports, Detroitnot only preserved its high-cost wage structure but effectivelyguaranteed its own failure to respond effectively to productinnovations by its foreign competitors. After all, why changeunless you’re forced to? Although the difference betweenDetroit’s and other carmakers’ US labor costs has beenexaggerated—the oft-cited $70 an hour versus $45 an hourmistakenly includes retiree pensions as a wage rather than afixed cost—the reality is something like $55 an hour versus$45 an hour, or a 20% difference, which isn’t small potatoes.

So, with industry excess capacity and Detroit’sinefficiencies as the problem that should be addressed by anyintervention—bankruptcy or bailout—the question I’d liketo focus on is: What can bankruptcy do to fix the problem?

In the case of the automotive industry, bankruptcy—Chapter 11 in particular—does several things extraordinarilywell. But it also faces a couple of serious hurdles.

Let’s start with how bankruptcy can “help” Detroit. First,bankruptcy law allows the rejection of what lawyers call“executory contracts”—things such as leases, franchiseagreements, supply contracts, and labor contracts. That abilitywould allow Detroit to convert many obligations tofranchisees that are imposed by state law into unsecuredclaims against the company. To give you some idea of thecost of eliminating those franchise agreements outside ofbankruptcy, when GM shut down Oldsmobile it reportedlypaid as much as $2 billion to Olds dealers pursuant to thesestate laws. So that’s Plus 1 for bankruptcy.

Bankruptcy would also probably allow the industry to turnits unfunded pension obligations to retirees into unsecuredclaims. Unlike current wages, which represent marginal costs,pension obligations to retired workers are fixed costs thathave contributed to one or more of Detroit’s manufacturersbeing insolvent. Bankruptcy’s ability to deal with accruedpension obligations is Plus 2 for bankruptcy. Now, it’s truethat the net effect would be to shift those liabilities to thePension Benefit Guaranty Corporation, and thus to us thetaxpayers—and so the end result would be a governmentsubsidy no matter what Congress does. But as I will suggestlater, shifting these kinds of one-time “social costs” from theprivate sector to the government is a better use of subsidiesthan propping up businesses that need to shrink to survive.By removing the burden of their pension costs, we can get a

much clearer picture of what it will take to turn them intoviable standalone enterprises.

Bankruptcy will also allow a manufacturer to reject itscurrent labor contracts, although the union might—andprobably would—strike. Still, over all, a Plus 3 forbankruptcy.

A fourth and final benefit of bankruptcy is that someoneother than current shareholders and their representatives willbe deciding on the appropriate size of these companies goingforward. I think this is an important benefit that hasn’treceived much attention. Once a company is insolvent, itsmanagement—put in place by the equity interests that arenow under water—are effectively playing with other people’smoney. Since the equity interests are already under water,they cannot be made any worse off, and so they have a naturaltendency both to take greater risks and to drag out any “dayof reckoning” in which they will be firmly shut out withnothing. Chapter 11 will transfer that equity ownership tonew people, whose money—or financial recovery—will beat risk, and who are thus much more likely to make the bestdecisions about what to do with the assets going forward.Under Chapter 11, the current management could remain inplace; but the decision to keep them there will be in the handsof the new owners—that is to say, the existing creditors whoseinterests are converted into equity in any reorganizedcompany, as well as the investors that agree to provide fundingfor the new, slimmed-down companies.

But having discussed the potential benefits of bankruptcyin this setting, what are its limitations—what do we need toworry about?

The biggest question mark for bankruptcy has to do withwhether Chapter 11 is a self-fulfilling prophecy in the sensethat no one will buy cars from a GM or Ford or Chrysler inbankruptcy. Most of the time when we buy something, wepay little or no attention to the fact that the selling companyis in Chapter 11. We don’t stop flying on United because it isreorganizing. We don’t stop shopping at Bloomingdale’sbecause it is reorganizing. (In fact, Bloomingdale’s reportedlyachieved new levels of profitability and efficiency whenoperating in Chapter 11 under Allen Questrom in the ’90s.)But that’s because we care only about the immediate “thing”we are purchasing. For the most part, if the company ceasesto exist after we buy or fly, we don’t care.

But that’s not true for cars. We care about the warranty. Itisn’t whether we’ll get parts or service—I have little doubtthat businesses will spring up to provide that stuff. Thequestion is whether we will get those parts and services “forfree”—as our original deal provided—for a period of, say,five years. This right—the warranty—has a certain economicvalue to the buyer, one that, just to put a number on it, mightbe estimated at around $1,000. The problem here is that ifyou buy a car from GM after it files for Chapter 11, your


warranty claim, while having “administrative expense”priority in GM’s Chapter 11, would be only an unsecuredclaim in any subsequent liquidation of GM. So unless youare confident that GM will “make it” for the five years forwhich your warranty is good, you won’t value the warrantyat its full $1,000.

Someone needs to figure out how to deal with this problem.Government guarantees have been held up as a “solution,”but that has a major moral hazard problem—that is to say, ifthe government guarantees warranties, GM has an incentiveto build lousy cars. Another possibility, which to me is morepalatable, might be to raise the priority of the warranty claimsabove those of unsecured creditors in any subsequentliquidation. This solution is likely to be better because itwould entrust the question of GM’s reorganizing—andoptimal size—to those people whose money would be on theline in the Chapter 11 proceeding.

But even if this issue is solved, the problem of warrantiesfor people who bought GM cars before bankruptcy needs tobe addressed as well. Those warranty claims would beunsecured claims in Chapter 11. The outcry over that wouldalmost certainly require GM to “assume” those claims as anexpense of Chapter 11 as well. Concerns have also beenvoiced about auto parts being made available—though I tendto think this problem is relatively minor since suppliers willcontinue, or will spring up, to provide the parts.

At any rate, these are serious issues that require carefulthought and responses—indeed, the kind of response that GM(and others) should have been working on in terms of a“prepackaged” bankruptcy instead of putting all their eggsin the bailout basket. (And, by the way, the statements madeby GM’s management and board that they “never considered”bankruptcy as an option make sense only in one scenario—aworld where Chapter 11 would spell the end of the currentequity owners’ interests and where the political branchappeared to hold out the only hope of postponing, if notavoiding, any such day of reckoning.) And if I’m right aboutthe overcapacity problem, Chapter 11 has a lot going for it,and perhaps a lot more than a government bailout.

This isn’t an exercise of imagining a perfect world; it is anexercise of comparing bankruptcy to alternatives and,specifically, to a bailout. If nothing else, bankruptcy—bythe “self-selective” nature of the companies that will be usingit—is much more likely to focus the solution to the excesscapacity problem on that part of the industry the excesscapacity should come out of—namely, the less efficientproducers that are more likely to become insolvent (in part,because such companies tend to find it more expensive toraise new equity). A bailout, on the other hand, which is farmore likely to tolerate (or ignore) the excess capacityproblem—because taking it seriously requires one to talkabout and focus on shutting plants and putting people out of

work—is likely not only to extend the problem, but to makeit far worse and even intractable. Even with conditions putaround them, bailouts will continue the existence of thosecompanies within an industry that are least deserving ofcontinuation on almost any scale. If you think I’mexaggerating, consider that many of today’s bailoutproponents view the proper role of government as returningthe industry to its “normal” production of something like 16million cars a year. This is a clear prescription for an industrythat will face “permanent” overcapacity and a predictableseries of future crises—and perhaps permanent governmentsupport.

Of course, bankruptcy can’t do it all. There is no denyingthe seriousness of the dislocations and hardship that will beproduced—not so much because of bankruptcy but becauseof the underlying need to pull capacity out of the system, oneway or another. Dealing with such dislocations seems to mea useful role for governments—and one that isn’t talked aboutenough. The government, in my view, would be far betteroff figuring out a good way of providing relief to those harmedby the transition than propping up companies in industrieswith excess capacity. Doing so will only make the temporarysupport permanent.

So, my suggestion is to let bankruptcy work, and deal withthe issues of overcapacity through a thoughtful governmentresponse. This way, we avoid sliding into a “solution” thateither ignores the underlying issue of overcapacity or respondsto it by spreading the solution around and dragging down allmanufacturers.

And even if you are unpersuaded by my proposal, let meleave you with one last point: one can’t understand bailoutswithout understanding bankruptcy. Bankruptcy is anincredibly important and useful tool, one that plays anessential function in a healthy free-market economy—and Ithink we all understand that such an economy is the underlyingsource of our collective wealth. Even though it operatescompany-by-company, bankruptcy can be used to pull excesscapacity out of entire industries. It has accomplished as muchwith the airlines and steel industries. We hardly give it asecond thought any more when it is used to take out a Linens‘n Things—because less efficient than Bed, Bath andBeyond—or a Circuit City—because less efficient than BestBuy. Of course, there has always been a lot of mystiquesurrounding automobiles—and “what’s good for GeneralMotors is good for the US” But I wonder if the trend towardbailouts—and I do see it as a trend, not just a once-in-alifetime response—is the reflection not only of politicians’perceived demand for immediate government “action,” butalso of the public’s and policymakers’ failure to understandthe positive role of bankruptcy. Bankruptcy may not alwaysproduce the right result, but it most certainly cannot if it isnot understood—and therefore not given the chance.


Thank you.

II. The Case for Bailouts

Zupan: Thanks, Tom. One of the great pleasures of myjob is getting to see where a degree from the Simon Schoolends up taking people.Our next speaker,Charlie Hughes, who’san alum from our classof 1970, is arguably oneof the foremostbranding experts in theautomobile industry. Ina career that hasincluded stints as theCEO of both MazdaNorth America andLand Rover NorthAmerica, Charlie hasmanaged or represented11 different brands,domestic as well asinternational. Whilerunning Land Rover, forexample, he introducedtheir sport utility line—built it from scratch inthe late ’80s during aperiod of a year and ahalf, developing asupplier and distributornetwork, and eventuallygrowing sales of thatline to 22,000 per year.

Charlie has alsorecently co-authored abook called BrandingIron, and appeared onnational news networks,including Bloomberg, to discuss the auto industry bailout—and, as you might have guessed, he has views on the subjectthat are going to differ from Tom Jackson’s. So, having heardfrom one of the world’s foremost authorities on bankruptcy,let’s now hear from someone who has spent his most of hiscareer in the auto industry—someone who can share hisfirsthand knowledge of not only the industry’s weaknessesand vulnerabilities, but also its strengths andaccomplishments.

Charlie Hughes: Thanks, Mark, and good evening. Myrole tonight is to play “Joe the Plumber”—or maybe “Joe the

Mechanic” would be more apt. When I was going to graduateschool here in Rochester in the late ’60s, I used some of myspare time to modify a 1963 Chevrolet Impala for drag racingon the street. One of my crowning achievements was teachingmy wife to beat all the high school kids in that car.

I’ve been thinking of this evening in terms of three words:bailout, bankruptcy, or bust. When I say that, I’m thinking

not in terms of the Detroitcar companies, but ratherin terms of our nation.We are a mess. It’s notjust the banking industry,the housing industry, thecar industry; it’s the entirecountry. We are at acrossroads. What kind ofa future do we want tohave? And, yes, Iunderstand that servicesare playing a growing rolein our economy relativeto manufacturing. But arewe going to continue tobe a nation of makers andbuilders, or will we endup a nation of moneychangers?

Like you, I have highhopes for PresidentObama. Yet one can’thelp but wonder if he willbe pragmatic and toughenough. We are inuncharted waters, and thestimulus package—atleast what I’ve seen ofit—is a troubling start.But let’s look at how wegot here.

We have a failedPresidency behind us, a Congress with approval ratings thatwould shame a child molester, a financial crisis born ofslipshod government oversight, and a widespread ethicalmeltdown in our financial industry. We have both states anda federal government that are dominated by special interests.I don’t know how many of you watched the Congressionalhearings where the car companies were taken in hand andtaught a few lessons—some of which they deserved. ButNancy Pelosi, our Speaker of the House, couldn’t restrainherself from using that occasion to push her green agenda,even if it means sinking our domestic car industry.

As a nation, we are behaving like fourth-generation heirs;

If we are determined to push the socialagendas of energy independence andclimate control, let’s make sure we do itwith street smarts and guts. Let’s raisethe federal gas tax. Let’s have one set ofregulations for emissions and fuel economynationwide. (How shortsighted andarrogant is it for people in each state todemand their own emissions standards?)And let’s rush—and I do mean rush—toharmonize those standards with Europe.Think of the powerful platform we couldachieve if we could get agreement onstandards for what’s basically 70% of theglobal car market worldwide—and wecould then take that agreement to Indiaand China where the real pollution isoccurring and get them on board.-Charlie Hughes


we don’t understand how the business that made us wealthyreally works. We apparently don’t understand how today’sworld operates. On too many issues, we are appallinglyignorant.

We are mad at Detroit. During the recent hearings inWashington, six out of ten constituents told their Congressmento let them die. But these people are clearly unaware of someimportant realities. The US car companies are far better thanyou think—though admittedly not as good as they need tobe. Let me cite a few facts to make my point:

What company runs the most efficient plant in NorthAmerica?

The answer is the Chrysler Jeep plant in Toledo. I mightalso add that Chrysler, which is the company that’s in themost trouble of the Big Three, is viewed by industry expertsas the equal of Toyota in running the most efficient plantsthroughout North America.

What line of cars had the best JD Power rating for initialquality in mid-size cars, which is the largest and mostcompetitive segment?

What company had the most cars with IIHS highest ratingfor crash safety?

The answer is Ford, with 16 cars. Number two was Honda,with 13 cars.

Who builds a large SUV hybrid that gets bettermileage than the Toyota Camry?

The answer is General Motors, and the car is the CadillacEscalade Hybrid. You can’t get much bigger than that—andthe car gets 20 miles a gallon around town.

And, finally, what is the real difference in pay for factoryworkers between Toyota and Ford? It’s $9 an hour if you dothe calculation the conventional way. But if you factor in thetypical bonus the Toyota workers have gotten during the goodyears—though not this year—the difference is less than $4an hour.

But here’s the irony I see in what’s going on today. Whatgot GM in trouble were its hubris and quick-fix mentality.As our government tosses around trillion dollar fixes, do thewords hubris and quick fix come to mind? Starting with thecredit crisis, to the Wagner Labor Act, CAFÉ, and transplantfactory tax subsidies, our government has played no smallrole in creating the problems of our auto industry in Detroit.

So what’s to be done? Here is my short list of suggestions:First, do no harm. We will debate tonight whether bailout

or bankruptcy is the better course. But this is not a labexperiment; and if we get it wrong we are in real trouble.Chapter 11 has never been tested on an industry that is sointertwined with our entire economy.

Second, treat each of the Detroit car companies accordingto their degree of distress and specific circumstances. Ford,for example, is in reasonably good shape: they have a goodplan, a solid cash base, and they haven’t taken any money

yet. If they end up needing it, we should support them.GM is a different story. It’s got too much debt, too many

brands, and they’ve already borrowed money—and it needsto demonstrate its long-term viability to receive more. Butwith that said, I can’t imagine this country without them.

And, finally, there’s the case of Chrysler, which I think weneed to help find an international partner. Fiat hasvolunteered—and we should see whether that marriage canwork. I think our government should continue to supportChrysler until we find out.

Now to my third prescription: if we are determined topush the social agendas of energy independence and climatecontrol, let’s make sure we do it with street smarts and guts.Let’s raise the federal gas tax. Let’s have one set of regulationsfor emissions and fuel economy nationwide. (Howshortsighted and arrogant is it for people in each state todemand their own emissions standards?) And let’s rush—and I do mean rush—to harmonize those standards withEurope. Think of the powerful platform we could achieve ifwe could get agreement on standards for what’s basically 70%of the global car market worldwide—and we could then takethat agreement to India and China where the real pollution isoccurring and get them on board.

Fourth, let’s make sure that when we think of our autoindustry, we believe in fair trade, not one-way free trade.Since World War II, every economy that we would considerto be an economic powerhouse has cultivated a strong, home-based car industry. Germany, France, Japan, Korea, and nowChina all view their auto industries as springboards toeconomic growth. Not just for the jobs, or the exports, butbecause the foundation of technological development in thesecountries—and ours as well—is the auto industry. You maybe surprised to know this, but during the Congressionalhearings in December, Silicon Valley came out in support ofDetroit saying that if one or two of the Detroit auto companieswere to go out of business, at least two big names intechnology would follow into Chapter 11.

I’ve worked for eight different car companies, and six wereimporters—from Germany, Italy, Britain, and Japan—and Ihave consulted for the Koreans. All those countries fightfiercely for the success of their homegrown car companies,and in ways we don’t fully appreciate. Their car companiesare vitally important to them, and they play the game as ateam sport.

Sad to say, we are a world champion athlete going to seed.We have gambled our money away and are left staring at ourgambling debts. We are at a crossroads; do we want to be anation of builders—or money changers?

Thank you.


III. The Difference between Automakersand Banks

Zupan: Thanks, Charlie. Now let’s hear from Jim Brickley,who is the Gleason Professor of Business Administration atthe Simon School. He’s an accomplished scholar inorganizational economics, competitive policy, corporategovernance, and compensation policy. He’s also, along withCliff Smith, one of the co-authors of the leading textbook onorganizational architecture. Jim is also a highly regardedteacher on our campus, having been a past recipient of ourhighest teaching award. He has published extensively on thetopics of franchising and vertical organization, is widelyregarded as an expert on distribution systems, and has doneextensive consulting to law firms and a variety of corporationson topics like organizational design and governance issuesas well as franchising and distribution systems.

Jim Brickley: Thanks, Mark. Let me start by saying thatthe auto industry is clearly very important to the US economy.It employs roughly two million people in manufacturing andin sales and service jobs, and it helps to support many otherjobs throughout the economy. It is thus an importantcontributor to our national GDP, and to our R&D effort aswell.

But the American auto companies also, of course, haveproblems, and they are problems that unfortunately run deeperthan the current economic recession. Given the importanceof this industry, we all hope that productive solutions to theseproblems can be found. The question we are discussing heretonight is whether these problems are best addressed thoughgovernment bailouts or reorganizations using the Chapter 11bankruptcy process.

But before we get into the case of the auto companies, let’stalk briefly about the problems with US banks and financialinstitutions. People often ask why the government has beenso quick to bail out banks and other troubled financialinstitutions, while at the same time being resistant to the ideaof bailing out the auto industry. Aren’t both industriesimportant to the economy, and weren’t there just as manymanagement blunders in banking as in the auto companies?

The answer is that the banking and auto industries havefundamentally different effects on our overall economy.While policy makers might view bankruptcy as a workableoption for auto companies, the use of a similar process in thecase of large banks—one that would put a freeze on allcreditors’ claims—could have far more serious effects on theoverall economy. The banking and financial system in aneconomy is like the circulatory system in a human being; justas people can’t survive if their hearts fail and blood doesn’tget to vital organs, economies can’t function with majordisruptions in the flow of credit. Virtually every business of

any size in this country depends on financial institutions tofinance its operations and investments. Consumers dependon banks to provide a relatively risk-free place to hold theirsavings—not to mention their mortgage and auto loans,insurance, and other financial services.

Because of their importance to both businesses andindividual savers—and their role in linking the two groups—the failure of major banks and financial institutions wouldsend shockwaves throughout the economy, leading towidespread lack of confidence in the banking system and evenfinancial panic. The bankruptcy of Lehman Brothers gavepolicymakers a frightening glimpse of the potential for a largedomino effect when a big, well-known financial institutiondefaults on its agreements. That event, along with the nearbankruptcy of AIG, resulted in a literal “run on the banks”that threatened Goldman Sachs, Morgan Stanley, and justabout every major financial institution in the US As I alreadysuggested, the financial panic triggered by the failure ofleading financial institutions would have restricted the flowof funds to the rest of the economy—even more than it alreadyhas—as investors pulled their funds out of the banks, and thebanks became increasingly reluctant to lend to consumers, tothe business community, and even to one another.

Now, to come back to where I started, the auto industry isvery important. Failures in the industry will have harmfuleffects on many people—including people who work for otherauto-related companies—and the overall economy. Buthaving said that, allowing a large manufacturing company tofile for bankruptcy, even one as large as GM, would not havethe devastating system-wide effects that would occur if thegovernment allowed large financial institutions like Chaseor Bank of America to default on their obligations. As TomJackson was just suggesting, Chapter 11 could well help theauto industry address some of its most pressing problems.

But let’s take a closer look at the challenges now facingthe auto industry. Wall Street analysts, when discussing theproblems of the Big Three auto companies, tend to focus onunions, and on their labor costs and debt. But another criticalproblem is the inefficiency stemming from their number ofbrands and models and from their distribution systems, ordealer networks. I think that these issues of corporate strategyand structure are likely to be addressed more effectivelythrough bankruptcy than bailouts.

As Tom told us earlier, the Big Three auto companiesdeveloped much of their product lines and dealer networksstarting back in the 1950s and ’60s, when they dominatedthe US auto market. It is widely acknowledged that thesecompanies now have far too many brands, models, anddealerships, given their current market shares, which arecollectively less than 50% of the US market. The Big Threenow market 112 different car and truck models in the USthrough 15 distinct brands. In contrast, their major


competitors—the top three Japanese companies—offer onlyabout half the choices, with 58 models and seven brands.GM by itself has eight brands and 70 models, and thus morebrands and models than the Japanese companies combined.

And as Tom also told you, the Big Three also have far toomany dealers. GM currently has some 6,700 dealers thatoperate 14,000franchises for its eightbrands. Its closestcompetitor, Toyota, hasonly 1,200 dealers withjust 1,600 franchises,and thus nearly 90%fewer. Now, the autocompanies have allrecognized the need toreduce their brands,models, anddealerships. But, asTom said earlier, this isgoing to be difficult,and very expensive, toaccomplish outside ofbankruptcy. Autodealers are a well-organized and powerfulpolitical force in theirlocal communities.Over time, they havesecured protectivelegislation in almost allstates that makes it very costly for the auto companies todiscontinue brands or close or combine dealerships. Forexample, it reportedly cost GM over $1 billion to settledisputes with dealers when they stopped making Oldsmobilesa few years ago.

Now, as Tom also said, in the case of bankruptcy, all of thecompany’s dealer contracts become subject to cancellationand reworking. As a result, the auto companies would havemuch more flexibility to reconfigure their brands anddealership systems in a quick and efficient way. Of course,some restructuring is going on as we speak. The number ofAmerican car dealerships has been falling almost daily asthese businesses fail. But relying on local business failuresto reduce the number of dealers—thanks to all their legalrecourse to and demands on the Big Three for life support—is a very protracted and costly way of addressing the basicproblem. What is needed instead are systematic andcoordinated changes in these companies’ product lines anddealership systems.

State laws not only make it expensive to alter dealershipcontracts, they also prevent manufacturers from owning their

own dealerships in many states and prohibit direct marketingto consumers through other media such as the Internet. Infact, a number of attempts by the Big Three to introduce newmarketing channels have been blocked by dealer-initiatedlawsuits or regulatory actions.

I have studied the effects of franchise and dealer protectionlaws across a broad rangeof industries. My researchindicates that such lawslead to less efficientdistribution systems andthe destruction ofcorporate values.Consistent with thesefindings, a study by theFTC has concluded thatstate laws preventing automanufacturers fromowning their owndealerships has cost USconsumers billions ofdollars a year in the formof higher auto prices.

How do we address thisproblem? It is unrealisticto expect 50 statelegislatures to reformthese laws in the face ofopposition from the localcar dealers. Mysuggestion is that the US

federal government consider national legislation that wouldsupersede state laws and grant the auto companies moreflexibility to design efficient distribution systems.

And let me leave you with one final thought: Inefficientfranchise laws are but one example of how politicalconsiderations often trump economics in legislative orregulatory solutions. Restructuring and consolidating theautomobile industry will require many tough choices—andthere will be losers as well as winners. Bankruptcyproceedings are much more likely to focus on economicconsiderations in making these tough choices than a bailoutprocess that involves politicians and politically-motivated“car czars.” In the long run, the industry will be much strongerif we allow economics rather than politics to drive theoutcome.

IV. A Bankruptcy Practitioner’sPerspective on Chapter 11

Zupan: Thanks, Jim. Now let’s hear from Joel Tabas, a

While policy makers might viewbankruptcy as a workable option for autocompanies, the use of a similar process inthe case of large banks—one that wouldput a freeze on all creditors’ claims—couldhave far more serious effects on the overalleconomy. The banking and financialsystem in an economy is like thecirculatory system in a human being; justas people can’t survive if their hearts failand blood doesn’t get to vital organs,economies can’t function with majordisruptions in the flow of credit.-Jim Brickley


Simon alum from the class of 1980 and the managing partnerof Tabas, Freedman, Soloff and Miller, a Miami-based lawfirm that specializes in reorganization and bankruptcy. Aspart of his practice, Joel has dealt with Ponzi schemes, realestate reorganizations, and healthcare workouts andbankruptcies. He has found himself operating airlines,retailers, and restaurants—and participated on creditors’committees in complex reorganization cases involving suchnames as Planet Hollywood, Brothers Gourmet Coffee, andThe Discovery Zone. Joel has graciously agreed to join ustonight in the midst of what are pretty busy times for hisbusiness.

Joel Tabas: Thank you, Mark. And let me start by sayingthat it’s a great honor to be taking part in this discussion.Cliff Smith was my finance professor when I was in the MBAprogram in the late ’70s. Tom Jackson’s classic article onreform of the US bankruptcy system was required readingwhen I went to law school. And, like President Jackson, bythe way, I too was struck by the stage backdrop behind us. InMiami, we’re dealing with an incredibly distressed real estatemarket—and this Depression-era stage set looks very familiar,makes me feel right at home.

As Tom started out by saying, when evaluating any kind ofdistressed corporate situation and the range of possiblesolutions, it’s very important to understand what can beaccomplished in Chapter 11. Most people have an instinctiveaversion to the word “bankruptcy”; they think of it as a deathsentence for companies. There’s good reason for this: Historytell us that about 90% of all companies that enter into aChapter 11 proceeding for reorganization do not emerge asgoing concerns; instead they are sold to outside investors orend up liquidating in a Chapter 7 or similar proceeding.

That’s the bad news about bankruptcy—but there is somegood news here as well. After all, 10% of the companies thatfile Chapter 11 do emerge as independent viable enterprises.One of the main distinguishing features of such successfulreorganizations is planning and preparation. The companiesthat come out of Chapter 11 tend to be those that carefullyexplore the potential benefits of a bankruptcy before goinginto it—they don’t just passively react. I would argue thatthe 90% failure rate is in large part the result of inadequatepre-bankruptcy planning, of the tendency of many companiesto wait until it is too late to rehabilitate the business. In thissense, the high rate of failure is not really attributable to theChapter 11 process itself, but rather to the fact that so manypatients arrive in bankruptcy almost “DOA”—in which casethey tend to get put on artificial life support for a short periodbefore going into liquidation.

I have represented both debtors and creditors in thereorganization process. If you’re helping a debtor negotiatewith creditors in a distressed situation, you have tounderstand—and to make sure that the creditors understand—

the likely outcome of a bankruptcy proceeding. Just theprospect of Chapter 11, with its “automatic stay” provisionand the potential rejection of “executor” contracts, is veryhelpful in getting concessions from lenders and other majorclaimholders. As Tom mentioned earlier, such claims tendto be reduced significantly in Chapter 11, and are oftenconverted to equity interests. In the case of the auto industry,as Tom also said, Chapter 11 could be very effective in gettingconcessions from not just creditors, but from the franchiseesor dealers and the unions as well.

Another important advantage of bankruptcy—one thatcould be especially helpful in the case of the US automakers—is its role in centralizing and coordinating thereorganization process. When dealing with large numbers ofcreditors that are dispersed around the country and have theoption of seeking different venues and courts, a private, out-of-court workout process would be a nightmare—the legalfees and expenses would be astronomical. The beauty of thebankruptcy proceeding is that the debtor files a bankruptcyin one particular forum—and all of the disputes are focusedfor the most part in that forum. So, instead of General Motorsfacing litigation throughout the country on franchise disputes,in Chapter 11 it would be handling the litigation involvingall of those franchisees in the one forum where the bankruptcyis filed.

So, that is an extraordinary benefit that bankruptcy bringsto a situation like this. It focuses the efforts and avoids thepotential for inconsistent consequences. Avoiding thispossibility is likely to mean some cost savings for thefranchisees. One of the things that happens early on in manybig bankruptcy cases is the formation of “committees” ofcreditors or other claimants with similar situations. That washow I got involved in the Planet Hollywood case that DeanZupan mentioned. My client was a creditor, and we wereinvited to become part of the committee of unsecuredcreditors. The role of such committees in such cases is to actpretty much as the boards of directors of public companiesare supposed to act. They have fiduciary obligations to theirconstituents—namely, all the similarly situated, unsecuredcreditors—that resemble the obligations of corporate directorsto the company’s shareholders. In other words, they are notsupposed to be using the platform for personal gain, or tobenefit their clients at the expense of other claimants. They’resupposed to be trying to maximize the recovery of all thecreditors. They have the right, and are given the resources,to hire professionals—accountants and other financial typesas well as accountants—to help them make the managerialdecisions that have to be made.

Now, it’s true that maximizing the recovery of creditors isnot necessarily the same thing as maximizing the health andfuture viability of the entire enterprise; there is some potentialfor conflict here, and for a premature liquidation of the


business. But even so, I would argue that the formation andfunctioning of such creditor committees is a critical featureof the bankruptcy process—one that does not exist at alloutside of Chapter 11. I’m a believer in having people withthe economic interests involved in the key decisions aboutthe future of the business, especially if a big portion of theirclaims is going to be converted into equity. I’m convincedthat such people are far better able to help fashion how thecompany will go forward than the typical regulator, who isbeholden to all the various constituents of the enterprise.

Moreover, in determining the company’s future, debtorsare greatly aided by the automatic stay provision I mentionedearlier. By putting a halt to all the disputes and lawsuits, theautomatic stay provides a breathing spell that enables all ofthe constituents—all of the parties to the process—to makeimportant decisions: Can the company be reorganized andrestructured in a way that will allow it to succeed? Or is itworth more dead than alive and a candidate for liquidation?

Still another advantage of Chapter 11—and this one is verytimely—is its ability to restrain excessive or unearnedexecutive pay. Early on in bankruptcy proceedings, all ofthe top executives basically have to submit their compensationpackages for approval by the court and vetting by thecreditors. So, this brings all compensation arrangements outinto the daylight. Earlier in this decade, we used to see peoplefiling for compensation packages with golden parachutes. Butthat practice has now been largely ended by the courts.

Now, let’s come back to this issue of franchises thateverybody has identified as a big problem for the auto makers.As has already been noted, most states have passed laws thatmake it very difficult and expensive for the manufacturers toshut down their franchisees. We’ve been involved with afew Ford franchisees in the Miami area that have recentlyfiled bankruptcy and shut down. I can tell you that they’reall struggling—and it’s going to be a widespread situation ifthe economy stays the way it is now, and there are likely tobe significant damages to the manufacturers associatedshutting down franchises.

But, as Tom pointed out earlier, if a manufacturer filesbankruptcy, it could deal with its franchisees’ claims in oneforum—and everyone could be treated the same. There couldeven be a committee for the franchisees so that they too couldhave an economic voice about the firm’s future. In fact it’smore than likely that, at the end of any successfulreorganization process, the franchisees will becomesignificant equity holders in the auto makers—and if thishappens, they’ll actually have a stake in the health of theunderlying business. The same comment also holds, by theway, for the unions: Only after becoming major equity holdersare they likely to act in ways designed to preserve the going-concern value of the enterprise.

Another valuable aspect of bankruptcy is its ability to

increase disclosure and transparency. As already mentioned,executive compensation is typically submitted to courts forapproval. But professional fees also have to be submitted ona periodic basis for approval with the courts as well. WhileI’ve seen studies suggesting that the costs of a bankruptcyproceeding in terms of professional fees would be muchhigher than in a private workout, I think that there are certainaspects of private workouts that have not been incorporatedinto the analysis. My guess is that, especially in a case likeGM or Chrysler, there would be significant cost savings notonly on the debtor’s side, but for the creditors as well—because of their coordinated representation by the committeesI mentioned.

Another aspect of a bankruptcy proceeding that willfacilitate information flow is the provision—specifically rule2004—that gives any party “in interest”—be it a creditor, anequity holder, or the government in its role as The UnitedStates Trustee—the right to obtain financial information fromthe debtors, including information about their plans torestructure and rehabilitate the debtors. Bankruptcyeffectively gives such parties the right to take depositionsfrom the debtor—a right that would not be available outsideof a bankruptcy in an out-of-court workout or a bailoutsituation.

I would also argue that, thanks to years of litigation in highprofile cases involving many of the complex issues now facingour auto makers, there is a very well established set of caselaw and dynamics and parameters that are used by the courtsin arriving at the judgments they make about whether toreorganize companies or let them fail. In bankruptcy courts,you will be dealing with jurists who handle reorganizationsand feasibility determinations on a regular basis. So you havea very well-developed area of the law that will not be availablein an out-of-court situation—where you’re likely to see a raceby all creditors to a state courthouse instead.

And let me come back to the point about the creditorscommittees that I made earlier. The committees and otherconstituents with financial interests are going to determinethrough a process of negotiation the important features ofthe company that emerges from a bankruptcy—what productsit will continue to make and sell, and how the company willbe financed. My own experience suggests that Chapter 11can provide a cost-effective process for restructuring thecompanies that are deemed by the court to be worth saving.For one thing, it provides a very effective way of eliminatingobstacles to private workouts. One obstacle is holdoutsamong creditors to a negotiated solution—and the Chapter11 can be used to “cram down” such a solution. Anotherobstacle is entrenched managers or owners. It’s always toughfor someone to admit they’ve taken the wrong tack—that theirmanagement strategies haven’t worked and they should notbe given another chance. While the process can sometimes


get a bit heated and hostile, I’ve found that the adversarialprocess that leads up to confirmation of a plan generally tendsto yield a good outcome—one that typically reflects theconcerns and interests of all the major constituencies.

Before I close, let me mention one other importantadvantage of Chapter 11—a feature designed to help debtorsraise new capital.This feature is likelyto be most valuable,of course, in caseswhere the capitalmarkets are otherwiseunwilling to providenew capital. Thatappears to be the casefor the US automakers, which is whythe government iscontemplating anexpansion of thebailout money alreadyprovided. The capitalmarkets are not goingto be giving money tothe Big Three—they’re unable to raiseequity or debt—and so they’re going to the government. Butif one of the auto makers were instead to file for Chapter 11,it could go to the court and say, “To raise new capital, I needto be able to issue super-priority debt financing—debt that isgoing to come ahead of the other secured creditors in mycapital structure.” And to the extent they were successful inraising private capital on those terms—which is hard to predictunder the current circumstances—the further bailout of theindustry could effectively be financed by private investors.If that fails, the other option would be to have the governmentprovide the super-priority financing.

So, there are a number of features of the US bankruptcycode that, in my view, could be used to help US auto makersto work their way out from under their current burdens. And,as President Jackson suggested, they should be weighing alltheir options very carefully. One reason they should beweighing those options—and its one that I’ve haven’t heardmentioned tonight—is that if the officers and directors ofthese companies do not consider bankruptcy, and thecompanies end up in liquidation, the directors could be facingdirector and officer suits, which is a fertile area of law rightnow. What those suits are alleging is that is when a companyenters what is known as “the zone of insolvency,” directorshave fiduciary duties that are supposed to shift from theshareholders to the creditors. If directors have failed toconsider bankruptcy as a means of preserving the enterprise

value of their companies, they could be facing a D&O suit.In sum, the auto manufacturers need to carefully consider

the possibility that Chapter 11 is the low-cost way of workingthrough their problems and preserving their companies asviable—though likely much smaller—going concerns.Bankruptcy, for all its flaws and bad press, may have a lot to

offer under thesecircumstances. Thankyou.

V. The (Long-Run) Costs ofBailouts

Zupan: Thanks,Joel. Batting cleanuptonight on our panel isCliff Smith, who is theLouise and HenryEpstein Professor ofFinance at the SimonSchool. Cliff is, first ofall, an accomplishedscholar. He has longbeen one of the main

editors of the Journal of Financial Economics, which isheadquartered at the Simon School and, along with theJournal of Finance, is one of the top two journals in the field.He’s published 16 books and some 90 articles. He won amajor prize a year ago for his impact on the field of insurance.He is also a very dedicated and talented teacher. In a careerat the Simon School that is now in his 35th year, Cliff hasreceived our full-time MBA Teaching Award ten times andour Executive MBA Teaching Award an amazing 19 times!

Cliff Smith: Thanks, Mark. It’s good to be here. As along-time subscriber, I appreciate what GeVa has done forthe local arts community. I want to thank them for letting ususe this wonderful facility.

It’s become an old saying that people who do not studyhistory are doomed to repeat mistakes that have already beenmade. I thought it might be useful to look at precedents toour current circumstances, and to try and glean lessons fromthe past.

When you talk about bailouts in the auto industry, peoplein the US tend to point to Chrysler as an example of a successstory. They will say, “Chrysler got their act together andthings worked out wonderfully. Let’s just do it again?” Now,if you say that fast enough, and don’t think about it very hard,it sounds good. But it’s important to remember that Chryslerwas not the only bailout that we lived through during the’70s and ’80s.

I’m a believer in having people with theeconomic interests involved in the keydecisions about the future of the business,especially if a big portion of their claims isgoing to be converted into equity. I’mconvinced that such people are far betterable to help fashion how the company willgo forward than the typical regulator, whois beholden to all the various constituents ofthe enterprise.-Joel Tabas


Remember the US savings and loan industry and whathappened to it? In the early ’80s, when interest rates onTreasury notes and bonds got into double digits, executivesfrom many S&Ls went to Congress looking for help. Sincemost of these S&Ls were holding mainly long-term fixed-rate mortgages with rates around 5-8%, they were effectivelyinsolvent. My dad was a banker in Greensboro, Georgia inthose days, and he liked to tell people, “You can’t write 8%,30-year mortgages, fund them with CDs paying 12%, andexpect to make it up on volume.”

What happened next? Well, Congress effectively changedthe bank accounting standards in such a way that the S&Lscould maintain at least the appearance of solvency andcontinue to stay in business. So, for the next few years, wehad lots of “zombie” S&Ls—they were dead, economicallyspeaking, but were still walking around underwriting riskymortgages and investing in risky commercial real estate. Itwas those transactions that ended up doing most of thedamage. The net result of this regulatory “forbearance” wasthat, despite the best efforts of the Resolution TrustCorporation ten years later, US taxpayers ended up footing abill that has been estimated at about $130 billion.

My point here, then, is that although the S&L bailout istoday widely viewed as having been a good thing, what seemsto have largely vanished from the collective memory is anysense of the eventual cost of that initial act of forbearance.By failing to deal with the troubled S&Ls effectively in theearly ’80s, our government turned what would have likelybeen relatively modest losses into much larger ones.

So, the first lesson from history is that bailouts are a riskybusiness—and not only is the outcome uncertain, but bailoutscan have the effect of increasing risk within the system. Ifyou go back and look at accounts in AutoWeek of Chrysler’spost-bailout success story, you will see articles in the late’70s and early ’80s about Chrysler’s bold, new, innovativemodels. As a finance professor, when companies use wordslike “bold,” “new,” and “innovative,” what I hear is “risky,”“risky,” “risky.” And that leads to an interesting problem forregulators—and of course the rest of us as taxpayers. As thepolitical process is unfolding and people are saying, “Well,the cost that we’re forecasting for this bailout is X dollars,and the US auto industry is clearly worth more than that,” Iwould recommend a fair amount of skepticism because thosecosts are regularly understated by what can turn into scaryamounts.

One of the big reasons these cost estimates turn out to beunderstated is that the behavior of the companies that arebailed out tends to change. They are being given theopportunity, in a sense, to play poker with someone else’smoney. If you’re ever invited to a poker game and allowedto play with someone else’s money, I’ve got a piece of advice:increase your bets.

Bringing out risky new products is one way automakerscan do it—but there are others. Before Chrysler got its bailoutpackage in the ’70s, product warranties in the industrycovered 12 months or 12,000 miles. After Chrysler’s debtwas guaranteed by us, the taxpayers, Chrysler managementdecided to expand Chrysler warranties to five years or 50,000miles.

Now, as things turned out, those bold new productsgenerally were well-received and well-produced. So theresulting warranty claims didn’t eat us out of house and home.But think about this from Chrysler’s perspective. “We’regoing to try something that is bold, new, and innovative. If itworks, we’re heroes. If it doesn’t work, we’re giving thecompany to the Treasury.” It is like flipping a coin whereheads I win tails you lose.

Thus, my second history lesson is that bailouts allowcompanies to play poker with the taxpayers’ money. That iswhat both Chrysler and the S&Ls did when the governmentgave them a second chance—and that is what I would expectUS automakers to do this time around. We are going to seelots of outsized bets being funded not by private investors,but by taxpayer dollars—bets that are going to be initiatedby corporate managers with little to lose and overseen bygovernment officials with limited expertise, and perhaps evenless to lose.

My third point is that the forecasted duration of this bailoutis something that can easily expand. Think about the historyof US agriculture since World War II. During the War, mostEuropean wheat fields were turned into battlefields. Inresponse, Roosevelt granted draft deferments to US farmersalong with instructions to “crank up production and feed theAllies.” And they did a marvelous job.

But what happened after V.E. Day? The swords were turnedinto plowshares, the European battlefields back into wheatfields, and there was a massive increase in the global supplyof agriculture products. The resulting oversupply and plungein crop prices meant that the US agricultural industry facedhard times.

This huge increase in supply and crash in prices put theUS at a political crossroads with respect to its agricultureindustry. What was to be done? One option was to do nothing.If the government did nothing, agricultural prices would likelyhave remained low for two years, or maybe three—and USfarmers would have had a tough row to hoe. You would haveseen many leaving that industry. Who would have been mostlikely to leave? Well, the people with the most opportunitiesother places, those with the most flexibility. So you wouldhave seen younger farmers leaving while older farmers stayed.People with college degrees and more opportunities in otherindustries would be more likely to go. But after a few years,the wrenching adjustments would have been behind us, andwe would have been back in normal operation, though with


far fewer people working in the industry. (And the same, bythe way, would likely have been true if the government hadnot bailed out Chrysler in the early ’80s. Had we made thetough choice back then, we would not now be facing themagnitude of problems Detroit is forced to deal with—because the industry, and the overcapacity problem, wouldlikely never have reached their current levels.)

The other choice facing US policymakers back then wasof course to bail outthe US agricultureindustry. And we allknow how that oneturned out. Wedecided to pay ourfarmers not toproduce. Thathappened in the ’40sand then in the ’50s—and then again in ’60s,’70s, ’80s, and ’90s.We’re still doing ittoday. If you believethat this bailout of theauto industry issomething that we’regoing to do once andbe done with, perhapsyou need to thinkagain. The costs of the bailout is likely to turn out to bemassively understated—and it could well turn into a kind ofperpetual annuity. Thus, the third history lesson is thatbailouts can persist—sometimes for decades.

Bailout advocates in Congress regularly announce, “We’renot planning on just handing suitcases full of money toGeneral Motors, Ford and Chrysler. We’re going to putconstraints on what they can do. We’re going to put constraintson how they can pay people. Nancy Pelosi is talking aboutforcing them to start making “green” cars—and she’s nottalking about her favorite paint color. To me, this begins tosound like allowing the government to run the industry.Unfortunately, the government’s track record in runningbusinesses is not the best. Think about Fannie Mae andFreddie Mac—not to mention the US Postal Service. Thusmy fourth history lesson is that the government is unlikely tobe especially good at running businesses.

We were told earlier that we’ve never had a bankruptcyapplied to an industry that is as large and important to oureconomy as the US auto industry—and that this crisis is justtoo big to be managed as an experiment. Yet this same logicshould also rule out a bailout: we have never bailed out anindustry that is this large and important either (unless we countthe agricultural industry). But if we look overseas, history

has provided us with an example. In the 1970s, the U.K.government engineered a bailout of British Leyland, themaker of Austin, Morris, Mini, MG, Rover, and Jaguar.Leyland had a weak balance sheet, contentious labor relations,and inefficient manufacturing: moreover, it had suffered asubstantial loss in market share. The U.K. government pouredabout $16.5 billion (in current dollars) into the companyduring the ’70s and ’80s. The bailout ended up lasting longer

and costing more thanhad been forecast—andit ultimately failed tosave the company:British Leylandeventually went out ofbusiness, with selectpieces being sold toforeign auto makers.

I think we all agreethat we are discussingan incredibly importantset of problems for theUS auto industry. Inmaking our policychoices, we need tothink carefully about thelong-run consequencesof whatever policychoices get made—

about whether and how these companies can be made to standon their own, and how many of our taxpayer dollars we arewilling to use to see if we can make it happen.

VI. Bank Bailouts and the Credit Crunch

Zupan: Thanks, Cliff. I’ll now invite the other paneliststo join us on the stage, and we will take some questions fromthe audience. Here’s the first one: “Should Lehman Brothershave been forced to go bankrupt?” Tom, can you start us offon that one?

Jackson: Any time you’re looking at a large financialinstitution, there are many more linkages with the rest of theeconomy, and things are much more complicated.Commercial banks can’t use bankruptcy; they need to gothrough some other regulatory process. In the case of LehmanBrothers, what I’ve been told by people suggests that it’s tiedin such an important way to the financial infrastructure that Ithink they probably should have rescued it instead of lettingit go. I think our regulators learned a lesson from that failure.My guess is that they were too quick to believe that this wouldbe the last failure and that we could survive it—and whenthey quickly saw there would be huge problems unwinding

Here in the US, it’s always been a very largenumber of people putting their own intuitioninto their business models and strategies, andputting their own capital on the line to backtheir bets. What you wind up with when youallow that kind of experimentation is a verylarge portfolio of options. As any financeprofessor will you, a portfolio of options isdramatically more valuable than an optionon a single portfolio.-Cliff Smith


all the claims, we went back to a model of stepping in. AndI suspect it was probably the right thing to do under thecircumstances.

Zupan: Next question: “Banks have taken federal moniesyet don’t appear to have increased their lending in a significantway? When do you expect the banks to start lending to othercompanies?” Cliff, can you take a shot at that one?

Smith: Banks are making loans right now to companieswith lots of tangible assets and established credit histories.Wall Street will take your debt to public markets if you’vegot a triple A credit rating. But what has happened, and whatalmost always happens during these kinds of financialdislocations, is that credit spreads have risen dramatically.So the curtailing of access to credit has been most pronouncedfor businesses with weaker credits.

Now the real problem here is all the uncertainty about howlong it is going to take before the economy recoveries and,as a result, about how much collateral lies behind the business,and how much debt it is really capable of supporting. Thus,if you’re a start-up company with little in the way of tangibleassets and not much of a track record, you’re going to havetrouble persuading a commercial bank to make you a loan,or an investment bank to help you raise debt capital.

Hughes: I’d like to jump in here, since I think we’reavoiding the biggest issue with the banks—namely, theirunwillingness to lend to each other because they don’t trusteach others’ balance sheets. I think there are two main waysout of this problem: the Japan model and the Swedish model.The Japanese approach was to accommodate the banks, toallow them to continue to operate and make more loans whilecleaning up their balance sheets very gradually. The Swedessaid, “We’ve got to clean up the balance sheets right awayand we’ll nationalize the banks—take temporary ownershipand control of them—to accomplish that.” Sweden cameback pretty quickly while Japan was in a recession for overten years.

So, while I think it was good that the government pumpedin some cash and kept other institutions from collapsing, Ithink we’re avoiding the big issue. You can’t have a bankingsystem where institutions can’t trust each others’ balancesheets. It’s like a game of liar’s poker.

Tabas: I represent some local banks in Miami, and theamount of new loans—particularly real estate loans—aredown as much 90% in some cases. One of my best friends, awell-known appraiser in Miami, is refusing to appraiseresidential real estate values because the prices on single-family homes have plummeted about 40% on average—andcondominiums are down 50% or more. Because of thissituation, banks are being forced to write down their assets.One local bank recently wrote down its real estate-based assetsfrom about $6 billion to $4 billion—and because of theircapital requirements, it’s very hard for them to make new

loans.And this is a kind of a self-perpetuating problem in the

sense that the markdowns and capital requirements seem tobe compounding the difficulties, creating a downward spiral.Our real estate market clearly overshot on the way up ’04and ’05. Now I think it has overshot on the way down. Butmarket participants tend to overreact—and in some casesperhaps bank regulators, too. The result is that right nowpeople in Miami are not able to borrow money for real estatefrom banks.

Jackson: I think that cleaning up the banks’ balance sheetsis a necessary but not a sufficient step in dealing with ourpresent problems. Even if you clean up their balance sheets,the banks have to make sure that the people who are trying toborrow the money are capable of repaying the loans—becauseif they’re not, then we’ve only added to the existing troubles.Things look awfully murky out there. As Joel said, they’rehaving a tough time getting people to step up and makeappraisals on the properties. So it hasn’t been a big surpriseto me that the bailouts have failed to produce an immediateincrease in bank lending. That’s going to take time.

So, this is a very complicated and multi-faceted problem—and cleaning up the balance sheets is, as I said, a necessarypart of the process of getting credit flowing again. But otherthings have to happen too.

Smith: Well, in thinking about this question, I think it’simportant to start with an understanding of what banks havea comparative advantage in doing. If you are a fairly largebusiness with a good track record of producing earnings andcash flow, your first choice will typically be to go to WallStreet and have them package your debt as a public issue.Banks, on the other hand, tend to finance smaller companiesthat, even if publicly traded, have substantially lessinformation produced about them. For regional andcommunity banks in particular, it’s these kinds of smaller,more opaque enterprises that have always been their breadand butter. Another way of saying this is that banks acquirea lot of what’s known as “specific knowledge” about theircorporate clients—the kind that is not easily transferred fromone lender to another.

And that suggests that this idea of cleaning up bank balancesheets so they can start trusting each other has some importantlimits. Financial institutions—and particularly smallerbanks—are by their nature somewhat opaque institutions thathold many assets that are difficult for outsiders to value.That’s why I’m frankly skeptical about the government’s planto buy troubled assets. In cases where insiders have anadvantage over outsiders in valuing bank-originated assets—and as I say, that’s especially been true of the smaller regionalbanks—I think it makes more sense to recapitalize those banks


with infusions of equity than to buy individual assets.

VII. Global Competition and Jobs

Zupan: Next question: “Does reduction of capacity inUS industries imply that American workers are supposed torelocate to foreign countries to work?”

Hughes: I don’t think many people are aware of this, butbefore the credit crisis began to set in, both General Motorsand Ford went through massive restructurings that took outalmost half of their production capacity. They were forcedto buy out thousands of workers at $140,000 a shot becauseof contracts with the UAW. I think they were pretty smartand decisive in doing that. Had they not done that, thecompanies would be in much more trouble than they are now.

But, if I can be a little patriotic, I find it bizarre to say,“We’ve got three million units worth of excess capacity; let’stake it out of the US producers.” If we were to do this, wewould be the only country in the world to take that approach.

Jackson: I don’t think this question of domestic versusforeign production is nearly as simple as you make it out.Some of G.M.’s most efficient operations are manufacturersin other countries, such as Holden in Australia. Obviously alot of the foreign companies have now built US plants thatemploy US workers. So distinguishing between US versusforeign production is not straightforward. The real questionhere is whether we are going to continue to have the capacityto produce 16 million cars when we don’t need it. I thinkthat using taxpayer dollars to subsidize that overproductionis a terrible idea, and that we have to figure out some way totake capacity out of the system. I don’t believe that the jobslost by Detroit are necessarily going overseas—they’re justgoing to be shifted to more efficient producers here in theUS, most of them, I would guess, in the service sector ratherthan manufacturing.

Brickley: To expand on Tom’s point, something like 60-70% of the Toyotas that are sold in this country are alsoassembled in this country. Since there are lots of Americaninvestors who own shares in Toyota, it’s no longer even clearwhat it means to be a foreign company. As Tom said, Honda,Toyota, and the other Japanese companies employ lots of USworkers here in the US And since GM now imports partsthat are made all over the world, I’m not sure it even makessense to talk about a US-produced car.

Hughes: That’s all true. But we still import a huge numberof cars. Again, I find it very odd that we would be havingany conversation where people say, “We should be supportingcars that are built somewhere else over cars that are builthere.” I’m not talking about putting tariffs on imports. Mypoint is that, in the past few years, the Big Three have alreadymade huge efforts to take out excess capacity; and althoughwe may well have three million units of excess capacity in

the US, not all of that capacity is sitting in the United States.So if we are talking about supporting our US producers—and there now seems to be a national and political will to dothat—then it seems to me that we should be willing to providethe capital needed to rehabilitate them. This way, and givensome time, they can become the efficient producers that wewant.

VIII. The Role of Greed

Zupan: Another question: “It seems that all the problemswe’re currently dealing with can ultimately be traced to greed.When will we learn how to deal with this?

Smith: I’ll tell you when. When the physicists figure outhow to repeal the law of gravity, the economists will be rightbehind them repealing the law of demand and abolishinggreed. All you can do is to recognize greed—or what weeconomists call “self interest”—and then try to set up ourinstitutions so that self interest becomes mainly a force forgood. That’s a matter of getting the incentives right insideorganizations—something that I believe is incrediblyimportant.

Brickley: Greed is a pretty loaded term, I agree. Whenyou hear it, it’s important to keep in mind what another guynamed Smith—not an economics professor, but a Professorof Moral Philosophy—told us over 200 years ago. AdamSmith’s message was that self interest plays a very importantrole in creating lots of the good things that we all take forgranted. It drives innovation, all the new products andservices that are the real source of prosperity.

Now, one question we are asking is whether people areany more self-interested now than they were, say, in thecaveman era. But, as the environment becomes morecomplex, there are new and sometimes destructive ways topursue self-interest—things like the off balance sheetpartnerships that brought down Enron and some of the morespeculative uses of derivatives by companies that we’ve seenin recent years. You couldn’t have done these things 20 yearsago because the financial instruments just weren’t available.

Smith: And to add to what Jim’s just said, I think it makesense to view our entire financial system as engaged in akind of Darwinian process of trial and error. We keep tryingdifferent things, we make mistakes—and then we learn fromour mistakes and make adjustments. One of the strengths ofcapitalism is that it tends to prevent people from persistingin error, making the same mistakes over and over again. Wewill no doubt make mistakes in the future. We will continueto have boom and bust cycles of the kind we’re now goingthrough.

Now, one important lesson underscored by recentexperience is that problems are going to arise wheneverindividuals and companies are granted a lot of “free


options”—that is, whenever they can acquire assets or dodeals without putting any of their own capital at risk. Wesaw that kind of behavior by people getting mortgages—andalso by banks that originated the mortgages with the idea ofsecuritizing and selling off as much as they could. That’s aclear prescription for too many mortgages and too manysecuritized deals.

Jackson: True, but it wasn’t just opportunistic or greedylenders and homeowners at work here; government policyclearly had a hand in producing the housing and mortgagebubble. It was government policy, pushed by Republicansand Democrats alike, that effectively encouraged lenders todrop standard downpayment requirements and come up withcreative financing—all with the idea of realizing a bipartisangovernment notion that everybody should own a home. Whenpeople and institutions respond in predictable ways to thosepolicy initiatives, I’m not sure we learn much from identifyingthe source of such behavior as “greed.”

Hughes: That’s all true. On the other hand, I tend to thinkthat behavior crosses the line from financial incentive to greedwhen you have a financial community that’s willing to sink aglobal economy. When you look at how the banks bundledthese mortgages into securities—bringing in the best and thebrightest from places like MIT to do the statistical analysisto put these packages together, and ending up with leverageratios of 40 to one—you have to ask how that all came about.I don’t know many bankers that are comfortable with the ideaof operating with that kind of leverage. I think that at thatpoint you can say that the driving force was greed.

IX. Solving the Dealer Problem

Zupan: Ok, we have time for one more question, and hereit is: “Instead of relying on bankruptcy, wouldn’t it be betterto deal directly with the adverse effects of franchising anddealer protection laws just by changing state and Federallaw?” Jim, you’re the expert on franchising, why don’t youtake this one?

Brickley: Well, I see two different issues here. One has todo with the states, almost all of which have these laws thatmake it difficult for the auto companies to operate efficiently.Now, the dealers have to worry about protecting theirinvestments—and I think much if not all of this protectioncould be provided by private contracts with the manufacturer.I think it’s important for the government to back thesecontracts. But the way things are now, the automakers areprohibited by state laws from owning dealerships—and theyare also prevented from selling cars directly to consumersover the Internet. I think both of these prohibitions are sourcesof inefficiency that increase the cost of automobiles—and,in my view, they should be overridden by federal legislation.

The second issue raised by the dealers—by, say, General

Motors’ need to deal with 14,000 franchise contracts—is onethat I don’t think can be addressed effectively by legislativeaction. To have a chance of becoming a competitive producer,GM must renegotiate these contracts. But, as Tom said before,this renegotiation is going to be very difficult outside ofbankruptcy. If they try to accomplish this outside Chapter11, people are going to be fighting over pieces of the pieinstead of trying to preserve the overall operating value ofthe firm.

So I think that the federal government can address some ofthe restrictions on the auto makers’ dealings with their dealers.There are more and very urgent problems that cannot behandled through legislation.

Jackson: Like Jim, I think it would be great if we couldremove some of these inefficiencies through legislation—andwithout resorting to Chapter 11. The history of the last 20years of General Motors would probably look very differentif the company hadn’t been forced to contend with the statefranchise laws. But getting political action on this is likelyto be difficult. As Jim mentioned, there’s no doubt that suchchanges would be blocked at the state level. But whetherthey could be accomplished at the federal level is also highlyquestionable. It’s this uncertainty about the political processthat makes me think that bankruptcy is the right way to go.As I said earlier, the rejection of executor contracts inbankruptcy suggests that Chapter 11 is the ready-madesolution to these franchise problems.

So, I agree with the premise of the question that a legislated,across-the-board solution would be preferred if possible. Butgiven the realities of the political process, I don’t think wecan get it done.

Smith: Let me add to Tom’s point. It’s a fairly well-established principle in political science that these kinds of“collective action” problems are generally likely to beintractable. You’re extraordinarily unlikely to get a politicalsolution in this case simply because the people who benefitfrom these franchise laws represent a small number of well-organized people with large concentrated benefits—namely,the profits from the dealerships. At the same time, the peoplehurt by these laws—namely, anybody who ever bought a car—are a widely dispersed group of individuals, each bearing arelatively small cost and having little interest in the issue.

So this is the collective action problem at work. It’s hardto get millions of people excited about being mugged for afew hundred bucks each when that winds up transferringsuitcases full of money to people who get big benefits andmake big political contributions. That problem keeps a lotof politicians from forgetting about their commitment to thepublic good.

Hughes: I want to jump in here. In talking about thedealers, I agree that we probably don’t have the will to makea lot of changes that we should. I agree that we should have


a federal franchise law. If a dealer goes out of business, orthe company wants to stop doing business with a certaindealer, there should be a contract that says, “This is whatwe’re going to pay you.”

We talked about example of GM’s shutting downOldsmobile earlier. No one knows, or is willing to reveal,the actual costs of ending relationships with the dealers—but in that case it was reportedly over $1 billion, and maybeas high as $2 billion. That subject’s got Rick Wagoner soafraid he won’t even touch it anymore. There are now some440 Saturn dealers that, although excellent dealers, are notmaking any money. They should be put to rest. Since theyalso own a lot of other franchises, you would not be puttingthem out of business.

Now, if it was merely a matter of General Motors goingout and saying, “We will buy back the parts and tools, andpay you all the money that we owe in accordance with ourcontracts,” then we wouldn’t be talking about anything like$1.2 billion to $2 billion. But the dealers are asking for a lotmore than that—they want “Blue Sky.” The problem,however, is that there is no longer any Blue Sky in the Saturnfranchise; it hasn’t made any money in the last dozen years.But the dealers are still asking for it—and that’s where theproblem becomes intractable.

So if we did pass a federal law—though I realize it’sunlikely to happen, like a lot of other things we talked abouttonight—we could solve that problem. I don’t thinkbankruptcy, by the way, would be the solution to thisproblem—though when you’ve called on and negotiated withas many dealers as I have, it sure sounds sweet to be able todo that. But there are other issues that also need to berecognized and addressed.

Let me mention one other interesting piece of auto industryhistory. There’s no question there are some hidden costs andinefficiencies in the system, but at one point in the past, themanufacturers once had the right to their own car dealerships.When the dealers were getting their way with stategovernments, they succeeded in passing legislation thatprevented the automakers from owning dealerships. Theinteresting thing here is that, behind the scenes, it was peoplefrom the manufacturers who were working to get thisprovision passed—because their own dealers were losing somuch money that they wanted a way out.

Smith: You mean the manufacturers needed a law toprotect them from themselves?

Hughes: That’s basically right. There are few things morecommon than believing you can do something as well assomebody else.

Brickley: Well, let me weigh in on this one. If you look atunregulated or less regulated distributor relationships in otherindustries, you almost never see a so-called “corner solution”where you have either all independent dealerships or 100%

company owned-stores.Hughes: Right.Brickley: Most companies use a mix of both

arrangements—say, 80% dealerships and 20% dealer-ownedstores—depending on variables such as location, and theprobability of repeat business. But the state governmentshave taken that option away from the auto companies. Infact, the governments have even prevented the auto companiesfrom writing their own contracts with the dealers in the sensethat the provisions in state law effectively override thecontractual agreements where they come into conflict.

Hughes: That isn’t the real obstacle. People do buy carsover the Internet every day.

Smith: From the manufacturer?Hughes: Not from the manufacturer. But I think you’re

making the assumption that it would be more efficient for theconsumer to buy directly from the manufacturer than fromthe dealer. I think that’s a mistake.

Smith: I wasn’t making that assumption. I’m assumingthat allowing people to experiment with a different model issomething that has a lot of value. That by putting a regulatorystop sign at the intersection that says, “You can’t turn downthat street,” you take away that opportunity to learn somethingyou didn’t know.

Hughes: Well, let me tell you a bit more about what thedealers actually do. One thing we know is that, when youbuy a car, it’s probably not the last time you have to go into adealership. Even Toyotas sometimes have to go back. Sothere’s a whole array of services in a car transaction that gobeyond just buying a car. And, at the moment, the industryhas a network of dealers that in most instances has been willingto give the cars away for almost nothing, but is there to serviceand trade them and help buyers sort out their finances in away that manufacturer cannot do. It does seem to work.

Brickley: Well, let me give you an example of somethingFord tried and then got blocked by regulation. In Texasaround the year 2000, Ford had a bunch of used cars thatthey wanted to be able to market directly to buyers over theInternet. The idea was that if they sold the cars, they wouldthen have to contract with some of their dealers to deliverthem to the buyers.

But this experiment never got off the ground. The dealerswho were not part of these arrangements went to the Texascourts and argued that such arrangements were a violation ofTexas law. I agree with Cliff that, by tying your hands behindyour back and saying you can’t try something, you will neverknow what might have worked best.

Smith: For those of you here who are old enough toremember, this all reminds me of those discussions back inthe ’80s about Japanese industrial policy. In those days,publications like Business Week and Fortune and the HarvardBusiness Review were all talking about how Japan, Inc. was


competing the US right off the map. It was an Al Gore kindof national industrial policy in which the future developmentof the entire economy was orchestrated by the JapaneseMinistry of Finance. What happened in this case is that avery small number of admittedly really smart people madehuge coordinated bets with the Japanese manufacturingindustry. When those bets turned out well, Japan’sproductivity soared—and the country ended up moving fromground zero after World War II to being the world’s secondlargest economic power. But that approach seems to havelost its magic in the last two decades.

And that’s not the way we do things in the US Here it’salways been a very large number of people putting their ownintuition into their business models and strategies, and puttingtheir own capital on the line to back their bets. What wewind up with is a tremendously robust and resilient economyin which literally millions and millions of these small betsare being made all the time. Some of these bets turn outwonderfully—take Google for example. But a lot of themcrash and burn—and you rarely hear about them.

Now, the problem with these dealer laws we’re talking aboutis that they absolutely prevent certain kinds ofexperimentation. You are legally prohibited from tryingcertain business models and practices. I just want to say thatstopping that kind of experimentation is not without costs.

I’m not arguing that if Ford had been allowed to sell carsdirectly on the Internet, it would have been a multibilliondollar product line for them. In fact, it may well have blownup in their face. My point is more narrow: The problem hereis that we will never know. I’d much rather have the Americanbusiness community continue to make thousands of calculatedbets, putting their money where their mouth is, than havingsomebody in Washington or Albany say, “As a regulatorymatter, we’re not going to let you see if that would work ornot.”

What you wind up with when you allow that kind ofexperimentation is a very large portfolio of options. As anyfinance professor will you, a portfolio of options isdramatically more valuable than an option on a singleportfolio. The value of the successes is almost sure tooutweigh the losses from the failures for a pretty simplereason: options give you right to keep the upside, but cutyour losses and move on when you’re failing. That’ssomething the US economy has been pretty good at—cuttingits losses when necessary and moving on to something morepromising.

Zupan: Well, let’s leave it at that—and let me thank all ofthe panelists for taking part in an instructive and entertainingdiscussion.


Pioneers of Finance

116

An Interview with Vernon L. Smith: 2002Nobel Laureate in Economic Sciences and

Father of Experimental EconomicsTerrance Odean and Betty J. Simkins

“For having established laboratory experiments as a tool in empirical economic analysis, especially in the study ofalternative market mechanisms” – 2002 Nobel

On January 9th 2009, Terry Odean and Betty Simkins interviewed Vernon L. Smith for this issue of the Journal ofApplied Finance.1 Vernon Smith is widely regarded as the “father of experimental economics” for his pathbreaking work inthis area. After decades of research, the once novel field of ‘experimental economics’ has become a recognized strand ofthe literature that contributes to our understanding of market mechanisms and more broadly, to the field of behavioralfinancial economics.

In 2002, Vernon Smith was a co-recipient for the Nobel Prize in Economics “for having established laboratory experimentsas a tool in empirical economic analysis, especially in the study of alternative market mechanisms”.2 He has written or co-written more than 200 articles and books on capital theory, finance, experimental economics, and natural resource economics.Dr. Smith remains very active in the economics profession, currently serving as Professor of Economics and Law at ChapmanUniversity School of Law.3

1A video of the interview is available at the Journal of Applied Finance website.

2 Daniel Kahneman was the other co-recipient of the 2002 Nobel Prize “for having integrated insights from psychological research into economicscience, especially concerning human judgment and decision-making under uncertainty”.

3He has previously held faculty positions at George Mason University, the University of Arizona, Purdue University, Brown University, and theUniversity of Massachusetts. He serves or has served on the board of editors of the American Economic Review; The Cato Journal; Journal ofEconomic Behavior and Organization; the Journal of Risk and Uncertainty, Science, Economic Theory, Economic Design, Games, and EconomicBehavior; and the Journal of Economic Methodology. In addition, he is past president of the Public Choice Society, the Economic Science Association,the Western Economic Association and the Association for Private Enterprise Education. Professor Smith is a distinguished fellow of the AmericanEconomic Association, the 1995 recipient of the Adam Smith award, and an elected member of the National Academy of Sciences. Furthermore, he hasbeen a Ford Foundation Fellow, Fellow of the Center for Advanced Study in the Behavioral Sciences, Sherman Fairchild Distinguished Scholar at theCalifornia Institute of Technology, Fellow of the Econometric Society, Fellow of the American Association for the Advancement of Science, and Fellowof the American Academy of Arts and Sciences.

117ODEAN & SIMKINS — PIONEERS OF FINANCE: VERNON L. SMITH INTERVIEW

In this interview, Vernon Smith shares his insights on markets. Among the issues he addresses are:* the relation between experimental economics and behavioral economics,* the insights his research on speculative bubbles in experimental markets provides for understanding the recent bubble is US residential real estate, and* his view as to reasons for the dramatic rise and fall in oil prices last year.

Terry Odean: You received the 2002 Nobel Prize in Economics for your work in experimental economics and you sharedthe prize with Daniel Kahneman for his contribution to behavioral economics. How are experimental economics and behavioraleconomics related?

Vernon Smith: Well, experimental and behavioral economics are actually complimentary. I see behavioral economics ashaving evolved out of the early cognitive psychology work with the emphasis being on fundamental decision making,decision making under uncertainty, and related issues. It is also characteristic of some of the work in experimental economics.Most experimental economists that came in the 1960s and 1970s were not primarily focusing on individual decision makingunder uncertainty. They were first focusing on decision making in markets and market exchange situations and the work wasprimarily interested in the performance of markets. If you go back before Danny Kahneman, you find people like SidneySiegel Ward Edwards, who were early psychologists but were not part of the cognitive psychology development. They wereseen as the “Skinner behaviorists” – that view the human mind as a kind of black box. The idea was “Let’s just do experimentswith it (the human mind) and just see what comes out of it.”

Even though I have interacted with Danny Kahneman over the years, we’ve always been interested in different questions andissues. Danny is more utilitarian than I am and that sounds odd – because we economists ought to be utilitarians. Right? Ithought the work I did with auctions and auction theory organized the data so well that I was pretty fond of utility theory.What lay behind it ought to be an important part of the picture.

Odean: Your early experiments confirmed that markets can work surprisingly well for setting prices that maximize socialwelfare when participants have private valuations or costs. Can you tell us about these early findings?

Smith: In my early work, we began to see and realize in experiments that what we got from utilitarian and formal analysis,the predictions of behaviors in different auction formats, wasn’t holding up. In particular, if you had multiple units, peoplewould get into what they called “jump bidding” and in that jump bidding process, they would miss out on all of the payments.4

It was that sequential nature of the bidding that led to the problem. You can’t believe how many experiments we did tounderstand that. We ended up just letting a clock raise the price and on each round asked: “Are you in or are you out?” Thatbasically solved the problem.

Now in English: This gives you outcomes that are completely efficient. You’re not suppose to jump bid. A person behavingrationally in the English auction should always raise the bid, if the starting bid is less than the value; never bid again yourself– that is, don’t raise your own bid. You should only bid by the minimum increment. If you bid more than that, you have thedanger of paying more than you need to. People don’t follow these rules. They have various rationalizations about it.They’ll tell you: “My desire to bid was, by jumping the bid, I would advertise — They (other traders) had better get outbecause I was going to win.” But as far as we can see, that is completely ineffective. It doesn’t work. It doesn’t bluff anyoneout. There is no way this is going to work.

With the Federal Communication Commission (FCC) auctions, the increments were $5 and $6 million. Small stakes.5 Somepeople tried to come up with elaborate theories of what was going on. They thought there is some sort of signaling going on.

Experimental and behavioral economics are actually complimentary.

4A jump bid is a bid higher than necessary to reach the next bidding level.

5This refers to FCC auctions of the electromagnetic spectrum. See: wireless.fcc.gov/auctions/


It was irrational behavior. As far as I could see, it led nowhere. It was jump bidders. The companies participating in theauctions had behaviorists advising them.

In my early work on markets, the first ones I ran performed so well I didn’t believe the results.

Odean: Did you do that in your classes at Purdue?

Smith: Yes. It was what eventually got me hooked into experimental economics. It turns out that people without any

training in economics, without any understanding of supply and demand, without any sophistication at all walk into a roomand they find the equilibrium of these markets. This is robust all types of subjects. They don’t have any idea of that’s whatthey did, because they don’t have a concept of what there means to be an equilibrium. They don’t have the vocabulary. Theywill deny there is any kind of model that can predict the convergence properties. It is hard for them to imagine anyonemodeling what it is they have done. They also will report, if you ask them: “Is there anything they could have done toincrease their earnings in this market?” They are certain there is some way they could, even though by definition of anequilibrium, they can’t. From what everyone else is doing, you can’t improve your position.

Many people will say this is a great victory of economic theory. Not exactly, because economic theory never predicted theweak conditions under which it prevailed. The early work on isolated single markets extended to multiple markets – marketsin which what you are willing to pay for commodity A depends on the price of B and vice versa. You can only describe theequilibrium with simultaneous nonlinear equations. People find those equilibria too. People have this ability, using institutionsthat somehow survived in our society using those rule systems. They have a capacity to do well for the group while doingwell for themselves. Given my traditional economic training, this came as an astonishing surprise.

Odean: In contrast to the early experiments in which markets performed surprisingly well, you’ve studied experimentalasset pricing markets that lead to speculative bubbles. Can you tell us more about your work in this area and what conditionslead to bubbles?

Smith: We didn’t really start to look at multiple commodity trading markets until we became computerized to handle themechanics. This led to the idea of looking at asset trading markets in the laboratory by the early 1980’s. Traditional theoryfrom financial economics was that all the information in the markets gets quickly incorporated into the price. One of thethings we were interested in doing was studying the possibility of price bubbles. There are a lot of stories about bubbles inhistory going back to the South Sea Bubble and the Tulip Bubble. We see what happens to the bubbles in the stock marketall the time. We thought: “Let’s see what we can learn in the laboratory?” The original idea was to begin with an environmentthat was transparent; that people would trade at the fundamental value because of the information we gave them. And wewould disturb them to see if we could produce a bubble. Well, that research program never got off the ground. Because rightfrom the beginning, we were getting bubbles.

I think the important thing we learned was that the common information wasn’t sufficient to give you common expectations.That is actually an experiential process. We brought people back a 2nd and a 3rd time, where the same people would comeback before we got convergence — observations where people were trading at fundamental value (the intrinsic value basedon the dividend and information that you had from the drawings of that distribution). We had this incredible contrast, in onecase – markets worked far better than you expected based on the theory.

It really goes back to William Stanley Jevons’ writing in the early 1870’s that contributed the basic supply and demand

People have this ability, using institutions that somehow survived in our societyusing those rule systems. They have a capacity to do well for the group whiledoing well for themselves. Given my traditional economic training, this cameas an astonishing surprise.


theory. He posed the question — How would the real market actually approach such an equilibrium? People would have toknow basic principles of supply and demand. There was no theorem, no result that said if people have complete information,the market converges; and if they don’t, the market doesn’t converge. It was, I think, simply the theorist saying: How mightthis actually happen in the world? Someone would have to know what I do and do pencil and paper calculations. That neverinspired confidence in it being a very believable model. Essentially, I began doing experiments in the mid to late 1950’s tothe mid 1980’s in which we found that this would extend to the asset markets, that performed very badly compared to ourexpectations. By then, we expected markets to work better than the theory.

Odean: So your early work confirmed how markets worked better than we expected but then your later work in assetmarkets worked worse than expected, and you got bubbles. Can you share more on this work?

Smith: We thought somehow, they (the traders) did not understand the instructions. What we did was to compute for themand then remind them in each period, what the remaining dividends left were and the holding period. They ignored that. Itis a beautiful example that if subjects want to dissatisfy the experimenter, they completely falsify his result. People didn’tbelieve those results (results that were too bad to be true) like my earlier results (results that were too good to be true). Westill don’t know what it is about asset trading that leads to the contagion of bubbles that we observe, even though we haveways of modeling them. You can model it by postulating there are two kinds of investment: fundamentalists and momentumtraders. The fundamentalists buy in proportion to the discount from true value (fundamental value) and proportion to thepremium. The momentum traders are the type of trader who simply buys in proportion to the rate of change in the price.This gives you a differential equation model that gives you bubbles.

We are able to come up with models that lead to testable propositions. The momentum traders are very much influenced bythe amount of money that is around them. If you model that we have two groups, and group A has a larger endowment ofcash, you get bigger bubbles. The model predicts this. This helps us understand why the market is sensitive to monetarypolicy and what the Federal Reserve is going to do. We’ve tried all kinds. Since people have different endowments of cashand shares, some may be willing to sell for less than fundamental value because they get a more balanced portfolio and areless exposed. We quickly shot that down with people who had exactly the same endowments. It is really common in stockmarket bubbles to put on price change limit rules.

Odean: Many of the world’s exchanges have these such as Taiwan and the Chinese exchanges.

Smith: Most of the time, they are wide enough that they are not binding.

Odean: They do get hit occasionally. I’ve seen a couple of papers looking at what happens when you hit the limit.

Smith: In experimental markets, what we think is going on is people feel there is a downside limit. It is still is an experimentalprocess and the traders have to come back for a 3rd time to finally create near fundamental value. The bubbles can go longerand can carry further.

We still don’t know what it is about asset trading that leads to the contagionof bubbles that we observe, even though we have ways of modeling them. Youcan model it by postulating there are two kinds of investment: fundamentalistsand momentum traders. The fundamentalists buy in proportion to the discountfrom true value (fundamental value) and proportion to the premium. Themomentum traders are the type of trader who simply buys in proportion tothe rate of change in the price. This gives you a differential equation modelthat gives you bubbles.


Odean: In experimental markets, people sometimes believe that there is some downside protection. That is similar to thehousing market in 2005 and 2006. People talked about how you can’t lose money in the real estate market. Real estateprices just went up. Or people talked about real estate and said “There is only so much land, so real estate has to go up inprice.” You get this belief in the markets.

Smith: I think it is a little different in the sense that there is not actually a ceiling or floor on the amount that the price willchange. I think what you are getting there is self-reinforcing beliefs about price change. In past housing bubbles, the

bubbles have gone on for maybe 3 or 4 years and then the prices decline. This is by far the largest bubble on record. Theindex started up in 1997.

Odean: This was the largest bubble here in the U.S. There was Japan.Smith: You have the phenomena of boom towns where the real estate goes crazy, such as local ones as in Alaska where theoil pipeline was built. That was a speculative bubble and these types of bubbles have been local.

The current bubble has hit national proportions and is astonishing — the movement from 1997 to the peak in the Case-Shiller index during 2006. My colleague, Steve Gjerstad, had been looking at the structure of those prices — at the lowpriced, mid-tier, and high priced homes. The low priced homes consistently in market after market, went up faster andfurther and more sharply. So the housing bubble had components. The lower tier had the biggest bubble component. Theseare exactly the people that are much more vulnerable to the decline in prices because whatever the wealth they had, it was inthose homes.

Odean: I think of this in terms of liquidity. At the low end, the ability to buy or not buy was most effected by liquidity. Sobasically the subprime mortgages created liquidity for people who otherwise wouldn’t have been able to buy and who didn’thave any margin of safety when things went badly.

Smith: Yes, I don’t see the financial magic that was being done in the subprime as so much a cause, but as a thing that wasa collateral effect. People had been expecting prices to rise and then the subprime lending industry created so much difficultyfor the whole financial system. I really believe that an important distinction in this housing bubble than previous ones, is thatwe had a tax law change in 1997. This tax law change allowed home buyers to receive capital gains from their home thatwere tax free up to $500,000. This capital gain wasn’t restricted to once; you could do it more than once. There was a twoyear holding period. This is one of the things I want to look at incidentally, of course. My conjecture is, that if you givefavorable tax treatment, say to 1/3rd of U.S. wealth (1/3rd of all US wealth is in a form of homes, about another 1/3 is in alllisted securities) and you are taking one of them and giving special capital gains tax-free treatment. And you can’t take theloss on homes against income as you can with stocks

Simkins: So you think that the tax law change in 1997 is partially responsible for the crisis we are in now?

Smith: Yes. I think if there was a spark that explains why this bubble was bigger than our previous ones, that would be oneof the things that I would want to look at. In national statistics, it is without a question the largest housing bubble. There ismuch more than the tax change of course. Now – you never had as much trouble with FreddieMae and FannieMac before,than this last bubble this time around. There is evidence that there is pressure on FM&FM to take substandard mortgages,both from above (politically) and below (industry), because a lot of people in the industry realize that some of these assetsare risky enough…. If we could just get the government to take them over. I think these two things together. We’ve longtried to create conditions for people to own their own wealth through home ownership. It has never succeeded so dramaticallyas in this last bubble.

I really believe that an important distinction in this housing bubble than previousones, is that we had a tax law change in 1997.


So what you have is simply people who have a home on the market longer – that happens well before the price – break inprices. Actually, you see that in our experimental markets. Because you see in experimental markets The boom phase,there’s lots more bids than there are asks and you can just look at excess bids (the difference between bids and asks), the bidsbecome an uncertain forecaster of the term and what will happen is that the asks will start to thicken up and prices start to riseand it takes awhile for the bid-ask activity to get back into balance. In experimental markets, it turns again if it over reacts.The bids start to become thicker and the asks start to thin out. The markets actually generate information that people don’tpick up on and immediately incorporate into some sort of rational calculation. This sort of thing can make short selling veryhazardous because it is not enough to know that something is overvalued, you must be good at timing.

Odean: I learned that the hard way.

Smith: You short a stock and it is too soon. It continues to go up and you get a bad stomach ache and you get people buyingthe cover, and now the market is about to turn down. When we allow short selling in our experimental markets and peoplecan cover their shorts, it can exeaberate bubbles. We see this in our experimental markets.

Odean: So you find that you’ve rekindled some bubbles?

Smith: I already know that when people had larger endowments in cash, they tended to get bigger bubbles. We couldreignite the bubbles which we did, but there weren’t quite as many believers. It took pretty extreme treatments to do that(reignite the bubbles). Which tells you that once people had this experience (the same group) they are reluctant to get into– to repeat that behavior. Of course in the real world, you continue to have new assets coming in.

Odean: In the real world, there are lots of investment options — internet stocks in the late 1990’s and real estate stocks afew years later.

Smith: People can always tell stories that “This case is different. This situation is different.” But all our evidence indicatesthat “It’s not!”. We haven’t seen the last housing bubble but we’re just not going to see another one for awhile.

I remember being concerned about the change in the tax laws. Not because I was opposed to reduce capital gains. I willmake no distinction between capital gains and income. If I would make anything deductible, it would be savings andinvestments. But there is no way you’re going to see that today, when people are concerned that there is adequate consumptionspending. On the other hand, a negative consumption tax, so people with lower income would get rebates; this is anincentive for the poor to accumulate.

Odean: You mention the change in the tax law may have contributed to the bubble. Do you see policy implications to yourresearch?

Smith: No, but I think it is important that the investment decision not be biased by differential tax treatment. It may turn outthat this tax law change is less important (or will be thought to be less important) than I think it is. But it is certainlysomething I would point out right away. Somehow, we have to account for why it is we are now crashing from the mother ofall housing bubbles. The policy right now is trying to keep prices from falling – but prices have to fall. Unless prices getback to some reasonable levels, we are going to have the problem that buyers are not attracted back – and buyers are the onlysource of sustainability in this market.

Somehow, we have to account for why it is we are now crashing from the motherof all housing bubbles. The policy right now is trying to keep prices fromfalling – but prices have to fall. Unless prices get back to some reasonablelevels, we are going to have the problem that buyers are not attracted back –and buyers are the only source of sustainability in this market.


Simkins: What market conditions discourage bubbles?

Smith: There is an old fashioned mortgage rule in capitalism – if you are going to borrow money, you have to put up asubstantial amount of equity, and that protects not only you but also the lender against the possibility that prices can decline.When we introduced margin buying in a laboratory stock market, it greatly exacerbated bubbles. And tight money and lessliquidity reduced the probability of bubbles. This conflicts with the idea of home ownership, where the normal pattern is that

you have family formation: people save for a period of time while they are renting. They save enough for a down paymentand they move into the low end of the housing market. They accumulate more wealth and then they move up into a higherpriced house. Anything you do to make it easier to buy a home (by using subsidies or creating an expectation that when yousell, your capital gains are tax free) means you can move people out of this period where they are accumulating and into thenext home earlier. What we had in this last housing bubble was a pretty dramatic movement of people out of this traditionalway people get started in home ownership. That increases prices and construction costs. Everyone (the buyers, the sellers,the lenders, the mortgage repackagers) believed that prices would go up. So who’s to blame?

Simkins: There has been a debate regarding the large increase in oil prices during the Summer 2008 and the subsequentdrop in prices, as to whether it is due to speculation or market fundamentals. What do you think?

Smith: I am puzzled by the increase in crude oil if it is not due to speculation. It is very hard to find any justification forthe price increase in terms of supply and demand. I think the price increase was very likely due to a lot of speculative capital.Hedge funds were going into crude oil and running the price up to $147 per barrel in a short time. It was a sharp peak and

There is an old fashioned mortgage rule in capitalism – if you are going toborrow money, you have to put up a substantial amount of equity, and thatprotects not only you but also the lender against the possibility that pricescan decline. When we introduced margin buying in a laboratory stock market,it greatly exacerbated bubbles. And tight money and less liquidity reducedthe probability of bubbles. This conflicts with the idea of home ownership,where the normal pattern is that you have family formation: people save fora period of time while they are renting. They save enough for a down paymentand they move into the low end of the housing market. They accumulatemore wealth and then they move up into a higher priced house. Anythingyou do to make it easier to buy a home (by using subsidies or creating anexpectation that when you sell, your capital gains are tax free) means you canmove people out of this period where they are accumulating and into the nexthome earlier. What we had in this last housing bubble was a pretty dramaticmovement of people out of this traditional way people get started in homeownership. That increases prices and construction costs. Everyone (thebuyers, the sellers, the lenders, the mortgage repackagers) believed that priceswould go up. So who’s to blame?


not sustainable. Now it has come back down to around $40.

With crude oil and gasoline prices back down, what’s not to stop people from buying bigger cars again? They did it beforein the 1970’s. We had a run up in oil prices that almost destroyed the motor home industry and the big gas guzzlers. Thenprices dropped back down. Bigger cars are what people wanted to buy.

Regarding how to prevent bubbles, I think that is as mysterious as ever. The sparks that actually lead to these bubbles, thatprecipitate the movement up or the current movement down. I think it is very hard to put your finger on what that is. We cantalk about things that affect the severity of the bubble but it doesn’t really give us an idea of what the root cause is. I’m notsure we ever know the root cause.


The 2008 Federal Intervention to StabilizeFannie Mae and Freddie Mac

W. Scott Frame

124

Fannie Mae and Freddie Mac are government-sponsoredenterprises that play a central role in US residentialmortgage markets. In recent years, policymakers becameincreasingly concerned about the size and risk-takingincentives of these two institutions. In September 2008,the federal government intervened to stabilize FannieMae and Freddie Mac in an effort to ensure the reliabilityof residential mortgage finance in the wake of thesubprime mortgage crisis. This paper describes thesources of financial distress at Fannie Mae and FreddieMac, outlines the measures taken by the federalgovernment, and presents some evidence about theeffectiveness of these actions. Looking ahead,policymakers will need to consider the future of FannieMae and Freddie Mac, as well as the appropriate scopeof public-sector activities in primary and secondarymortgage markets.

W. Scott Frame is a Financial Economist and Policy Advisor for theFederal Reserve Bank of Atlanta in Atlanta, GA.

The views expressed do not necessarily reflect those of the Federal ReserveBank of Atlanta, the Federal Reserve System, or their staffs. I would liketo thank Michael Hammill for research assistance and Mark Flannery,Diana Hancock, Wayne Passmore, Mario Ugoletti, Larry Wall, and LarryWhite for providing helpful comments on an earlier draft.

Fannie Mae and Freddie Mac are enormous government-sponsored enterprises, or GSEs, that play a central role inUS secondary mortgage markets.1 Together, as of mid-year2008, the two institutions held or guaranteed about $5.5

trillion in US residential mortgage debt – slightly more thanthe $5.3 trillion in publicly held US Treasury debt at thattime.

Both Fannie Mae and Freddie Mac have been the subjectof a great deal of attention and controversy in recent years.Each GSE has: faced accounting scandals, been criticizedfor not sufficiently targeting their activities toward low-and-moderate income communities and households, and hadpolicymakers voice concerns that they posed a systemic riskto the global financial system.2

At the heart of these (and other) issues is the GSEs’incentive structure. Fannie Mae and Freddie Mac are publiclytraded financial institutions that were created by Acts ofCongress in order to fulfill a public mission. These charterActs imbue the two GSEs with important competitiveadvantages (most notably, implied public-sector support fortheir obligations) and define the scope of their permissibleactivities.3 Over time, Fannie Mae and Freddie Mac becameexceptionally large, profitable, and politically powerful.

Recently, however, Fannie Mae’s and Freddie Mac’ssingular exposure to US residential mortgages – coupled witha thin capital base – resulted in both of these GSEs facingfinancial distress. US housing markets became increasinglystressed through 2007 and resulted in severe disruption tomortgage markets. Secondary market liquidity for mortgagesnot backed by Fannie Mae and Freddie Mac almost entirely

1“Fannie Mae” and “Freddie Mac” are widely used nicknames for theFederal National Mortgage Association and the Federal Home LoanMortgage Corporation, respectively.

2For a discussion of the accounting problems at Fannie Mae and FreddieMac, see the results of special regulatory examination reports at: <http://www.ofheo.gov/Regulations.aspx?Nav=199>. For an analysis of the GSEsfunding of mortgages for low-income borrowers and underserved areasearlier this decade see, for example, Brown (2001) and Bunce (2002).Former Federal Reserve Chairman Greenspan (2005), among others,described the systemic risks posed by the GSEs in testimony before theUS Congress.

3The charter acts may be found at 12 USC. § 1716 et seq. (Fannie Mae)and 12 USC. § 1451 et seq. (Freddie Mac).

125FRAME — THE 2008 FEDERAL INTERVENTION

dried-up, and GSE-backed mortgages saw liquidity pressureas evidenced by unusually wide yield spreads. Thesedevelopments resulted in a significant reduction in theavailability and cost of mortgage credit for homeowners.

As a result of these developments, the federal governmentwas compelled to intervene to stabilize both GSEs andmortgage markets more generally. On September 7, 2008,Fannie Mae and Freddie Mac were placed intoconservatorship by their federal regulator: the FederalHousing Finance Agency (FHFA). Concurrent with thisaction, the US Treasury entered into “senior preferred stockagreements” with each institution obligating the federalgovernment to inject up to as much as $100 billion each inFannie Mae and Freddie Mac. The Treasury also establisheda mortgage-backed securities purchase facility and a standingcredit facility in order to support the residential mortgagemarket.

The actions of the FHFA and the Treasury last Septemberstabilized Fannie Mae and Freddie Mac by effectivelyguaranteeing their debt and mortgage-backed obligations.4

A subsequent announcement by the Federal Reserve that itwould purchase substantial quantities of Fannie Mae andFreddie Mac debt and mortgage-backed securities during2009 has further acted to improve liquidity in those marketsand bring yield spreads back to historical norms.5

The remainder of this paper will proceed as follows.Section I provides some background information about FannieMae and Freddie Mac and Section II describes the sourcesof financial distress facing these two GSEs. Section IIIoutlines the steps taken by the federal government to stabilizethese systemically important institutions and also presentssome evidence relating to the effectiveness of these and otherrecent federal interventions into secondary mortgage markets.Some concluding remarks are offered in Section IV.

I. Who are Fannie Mae and Freddie Mac?

Fannie Mae’s roots stem from the Great Depression. TheNational Mortgage Association of Washington, as Fannie Mae

was first known, was created within the federal governmentin 1938. Its business was to purchase mortgages insured bythe Federal Housing Administration, or FHA, from financialinstitutions around the United States.6 Fannie Mae wassubsequently spun-off in 1968 as a publicly traded companyas a way to reduce the federal debt during the Vietnam War.7

By contrast, Congress in 1970 created Freddie Mac, whichwas owned by the 12 Federal Home Loan Banks and thesavings and loans that were members of these Banks.8 FreddieMac became publicly traded in 1989 as part of the thrift crisisresolution.9

Hence, today Fannie Mae and Freddie Mac are quasi-public/quasi-private financial institutions. On one hand, eachGSE was created by an Act of Congress and is broadly chargedwith providing liquidity and stability to the secondaryresidential mortgage market, with a particular emphasis onhousing for low- and moderate-income households and/or inareas viewed as historically underserved (central cities andrural areas).10 On the other hand, Fannie Mae and FreddieMac have been funded with private capital and their sharesare traded on the New York Stock Exchange. This unusualgovernance arrangement has resulted in two, sometimesopposing, corporate objectives: fulfilling certain social policygoals (and assisting related political constituencies) andmaximizing shareholder value.

By law, Fannie Mae and Freddie Mac are limited tooperating in the secondary conforming mortgage market and

4By law, the obligations of Fannie Mae and Freddie Mac must state thatthey are not guaranteed by the federal government. See 12 USC. §1719(b),(d)-(e) (Fannie Mae) and 12 USC. § 1455(h)(1) (Freddie Mac).Nevertheless, as discussed further below, financial markets have longviewed the GSEs’ obligations as carrying an “implicit” governmentguarantee. The federal government’s recent actions were intended to senda strong signal to financial markets that the US would protect the interestsof holders of Fannie Mae and Freddie Mac obligations on an ongoingbasis.

5The maturities of new debt issues by Fannie Mae and Freddie Mac alsoincreased as a result of these policy actions. Nevertheless, the GSE’s accessto long-term finance remains limited as it has been for all corporateborrowers.

6According to Frame and White (2005), by issuing debt and purchasing andholding FHA-insured residential mortgages, Fannie Mae was able to expandthe available pool of finance to support housing and also to provide a degreeof unification to mortgage markets. During this time, mortgage markets werelocalized for technological reasons as well as for reasons rooted in laws thatprohibited interstate banking and restricted intra-state bank branches in manystates during most of the twentieth century.

7Fannie Mae was replaced within the federal government by the GovernmentNational Mortgage Association, or “Ginnie Mae,” an agency within theDepartment of Housing and Urban Development, or HUD, that guaranteesmortgage-backed securities that have as their underlying assets residentialmortgages that are insured primarily by the FHA or by the Department ofVeterans Affairs (formerly the Veterans Administration, or VA).

8See Flannery and Frame (2006) for a history and overview of the FederalHome Loan Bank System, which those authors refer to as the “other”housing GSE.

9According to Frame and White (2005), a major motivation for the conversionof Freddie Mac to a publicly traded company was the belief that a widerpotential share-holding public would raise the price of the shares held by thethen ailing S&L industry and thus improve the balance sheets of the latter.

10Fannie Mae’s mission or “statement of purpose” can be found at 12 USC.§ 1716. A similar statement for Freddie Mac is located at 12 USC. § 1451[Note].


their activities take two broad forms.11 The GSEs’ “creditguarantee” businesses involve the creation and creditenhancement of mortgage-backed securities, or MBS. Thisis most often done through each institution’s “swapprograms,” whereby mortgage originators present pools ofqualifying mortgages and then exchange them for MBS thatrepresent an interest in the same pool. The GSEs agree toensure the timelypayment of principal andinterest on the MBS inexchange for a monthlypremium known as a“guarantee fee”. (Thisprocess is commonlyreferred to as“securitization” althoughthe credit enhancementstructure is much simplerthan that typically usedby investment banks forsimilar transformationsof loan pools intotradable securities.)GSE-backed MBS are very liquid (relative to other asset-backed securities and loan pools) and this liquidity facilitatesmore efficient balance sheet management for financialinstitutions.

Fannie Mae’s and Freddie Mac’s second line of businessis “portfolio investment”. This involves the two GSEs holdingMBS that they have purchased in the open market, wholemortgages (purchased from originators under their “cashprograms”), and liquid fixed-income investment securities.Fannie Mae and Freddie Mac largely fund these assets withso-called “Federal Agency” debt. The two GSEs havehistorically been highly leveraged with total accounting(book) equity equal to less than 4% of total assets.12

While Fannie Mae’s and Freddie Mac’s federal charters limitthe scope of their business activities to the secondary residentialmortgage market, they also provide them with a number ofadvantages that result in lower operating and funding costs.13

First, both GSEs are exempt from state and local income taxes.

Second, the Secretary of the Treasury has the authority topurchase up to $2.25 billion of Fannie Mae’s and Freddie Mac’ssecurities, which is often referred to as their federal line-of-credit.Third, the GSEs’ issue “government securities,” as classifiedunder the Securities Exchange Act of 1934, which in practicemeans that their securities are eligible for use as collateral forpublic deposits, for purchase by the Federal Reserve in open-

market operations, and forunlimited investment byfederally insureddepository institutions.14

Fourth, Fannie Mae andFreddie Mac use theFederal Reserve as theirfiscal agent, which meansthat their securities areissued and transferred usingthe same system as USTreasury borrowings.

The features of FannieMae’s and Freddie Mac’sfederal charters, coupledwith some past government

actions, has long served to create a perception in financial marketsthat the federal government “implicitly guarantees” the GSEs’financial obligations.15 This belief, in turn, allows Fannie Maeand Freddie Mac to issue debt at interest rates that are far morefavorable (better than AAA) than their stand-alone financialstrength ratings would warrant.16 This borrowing advantage hasbeen estimated empirically to be about 40 basis points, althoughsuch estimates vary depending upon the maturity and credit ratingof the comparison bonds and the sample period studied.17 The

Over time, Fannie Mae and FreddieMac became exceptionally large,profitable, and politically powerful.Recently, however, Fannie Mae’s andFreddie Mac’s singular exposure to USresidential mortgages – coupled with athin capital base – resulted in both ofthese GSEs facing financial distress.

11See 12 USC. 1719 (Fannie Mae) and 12 USC. 1454 (Freddie Mac).Conforming mortgages are those with balances below the legal limits onthe size of residential mortgages that Fannie Mae and Freddie Mac canbuy. For 2009, the conforming loan limit for single-family properties is$417,000, but can be as high as $625,500 in certain high-cost areas. See<http://www.ofheo.gov/Regulations.aspx?Nav=128>.

12By contrast, commercial banks (as a group) maintain a ratio of total equityto total assets of about 10 %.

13See, for example, US Congressional Budget Office (1996, 2001) forfurther discussion.

14A further implication is that they are exempt from the provisions of manystate investor protection laws and the registration and reporting requirementsand fees of the Securities and Exchange Commission (SEC). Notably, FannieMae voluntarily registered its stock with the SEC in March 2003 and FreddieMac did the same in July 2008.

15This perception arises despite explicit language on each GSEs’ securitiesthat they are not obligations of the federal government. US GeneralAccounting Office (1990, 90–91) discusses two past episodes during whichthe federal government assisted troubled GSEs. First, during the late 1970sand early 1980s, Fannie Mae was insolvent on a market value basis andbenefited from supervisory forbearance. Second, in the late 1980s, theFarm Credit System (another GSE serving the agricultural sector) requireda taxpayer bailout totaling $4 billion.

16Fannie Mae and Freddie Mac long received AA- ratings from Standardand Poor’s in terms of their “risk to the government”. However, theseratings incorporated whatever government support or intervention the entitytypically enjoyed during the normal course of business. See Frame andWall (2002) for a discussion.

17See Ambrose and Warga (1996, 2002), Nothaft, Pearce, and Stevanovic(2002), and Passmore, Sherlund, and Burgess (2005).


perceived implied guarantee also affects the interest rates onMBS that Fannie Mae and Freddie Mac issue, although theadvantage is difficult to estimate.18

The perception of an implied federal guarantee conveys asubsidy on Fannie Mae and Freddie Mac, part of which istranslated into lower mortgage rates for consumers. Inparticular, Fannie Mae’s and Freddie Mac’s activities resultin conforming mortgages’ carrying lower interest rates than“jumbo mortgages” with principal amounts above theconforming loan limit. Several econometric studies estimatedthe effect of GSEs on conforming mortgage rates, typicallyfinding the interest rate differential to be about 20-25 basispoints with variation in the estimates depending on theempirical specification, data sample, and time periodstudied.19

Fannie Mae and Freddie Mac have been largely free frommarket constraints on their size and risk because of the marketperception of an implied federal guarantee of their obligations.The GSEs have become enormous financial institutions – bothin absolute terms and relative to the mortgage market as awhole. As of June 30, 2008, Fannie Mae and Freddie Mactogether held almost $1.8 trillion in assets (almost entirelyMBS and whole mortgages) and had another $3.7 trillion innet credit guarantees outstanding – i.e., net of those held intheir own portfolios. This $5.5 trillion in obligationsrepresented almost half of all residential mortgage debtoutstanding at that time. The two GSEs have also grownmuch more rapidly than the residential mortgage market as awhole over the past three decades. For example, in 1980,Fannie Mae’s and Freddie Mac’s share of residential mortgagedebt outstanding was only 7% (Frame and White, 2005).

The perceived implied federal guarantee also distorts theGSEs’ risk-taking incentives in a way that may increase theprobability of financial distress. (A similar situation is well-understood in the context of federally insured depositoryinstitutions.) The idea is that a federal guarantee inducesdebt holders to accept artificially low interest ratesirrespective regardless of a GSE’s true default risk. A GSEcan then increase the riskiness of its activities – which promisehigh returns to equity holders if the risks turn out well –without needing to share those rewards with debt holders inthe form of higher coupon rates on their debt. The GSEs’

equity holders thus perceive a greater-than-normal benefitfrom risk-taking.

In order to maximize benefits to homebuyers and minimizetaxpayer risk, the federal government imposes a two-partregulatory structure on Fannie Mae and Freddie Mac. TheUS Department of Housing and Urban Development, orHUD, long regulated the GSEs for compliance with theirmission of enhancing the availability of mortgage credit bycreating and maintaining a secondary market for residentialmortgages. HUD was also responsible for establishing goals(and monitoring compliance with the goals) for Fannie Mae’sand Freddie Mac’s financing of housing for low- andmoderate-income families, housing in central cities, and other“underserved areas”. Congress formally established a safety-and-soundness regulatory and supervisory regime for Fannie Maeand Freddie Mac in 1992. The Office of Federal HousingEnterprise Oversight, or OFHEO, was authorized to set risk-based capital standards (subject to important statutorylimitations), conduct examinations, and take enforcementactions if unsafe or unsound financial or managementpractices were identified.20 Unfortunately, OFHEO’s structureand authorities proved deficient in many respects.21

GSE regulatory reform was an active legislative item thisdecade following the accounting scandals at both Fannie Mae(2004) and Freddie Mac (2003). However, it was not untilthe GSEs came under serious financial strain that reform waspassed as part of the Housing and Economic Recovery Actof 2008. The new law created the Federal Housing FinanceAgency (FHFA), which consolidated the mission and safetyand soundness oversight for Fannie Mae, Freddie Mac, andthe Federal Home Loan Bank System.22 The establishmentof the FHFA reflects an improvement in GSE safety-and-soundness supervision and regulation since the new regulator(among other things): (1) no longer requires Congressionalapproval for its budget, (2) has authority to set minimumleverage and risk-based capital requirements, and (3) hasreceivership powers.

18US Congressional Budget Office (1996, 2001) reported an MBS advantageof 30 basis points, but Passmore (2005, p. 9) critiques the approach thatgenerates this estimate and alternatively argues that the advantage is in therange of 0-6 basis points. See also Heuson, Passmore, and Sparks (2001) andPassmore, Sparks, and Ingpen (2002) for theoretical analyses of the relationshipbetween GSE securitization and mortgage interest rates.

19For an introduction to this literature, see US Congressional Budget Office(2001), McKenzie (2002), Passmore (2005), and Ambrose, LaCour-Littleand Sanders (2004), and the references in these papers.

20Prior to the Federal Housing Enterprises Financial Safety and SoundnessAct of 1992, HUD maintained exclusive regulatory oversightresponsibilities over Fannie Mae and (for 1989-1992) Freddie Mac. Priorto the passage of the Financial Institutions Reform, Recovery andEnforcement Act of 1989, Freddie Mac was the responsibility of the FederalHome Loan Bank Board.

21As discussed in Eisenbeis, Frame, and Wall (2007), OFHEO supervised onlytwo institutions making it prone to regulatory capture. The agency was alsoan independent arm of HUD, which is more focused on promoting housingthan contending with GSE safety-and-soundness. OFHEO was also subjectto Congress’ annual appropriations process and sometimes fell victim topolitical meddling. With respect to supervisory tools, OFHEO lacked theauthority both to adjust minimum capital standards and to resolve a failure ofeither Fannie Mae or Freddie Mac.

22By doing so, the FHFA succeeds the OFHEO, HUD’s GSE missionregulation, and the Federal Housing Finance Board.


II. Sources of Financial Distress

US housing and mortgage markets became increasinglystressed during 2007 and 2008, largely as a result of houseprice declines in many parts of the country. Between 2007:Q2and 2008:Q3 house prices declined 18.0% on a nationwidebasis based on the S&P/Case-Shiller composite index. Bycontrast, over the sameperiod, the OFHEOnationwide house priceindex fell 4.5%. While themagnitudes of decline inthese repeat-sales indicesdiffer – owing to coveragedifferences by geography,loan size, and loan quality –this national decline inhouse prices is unusual.23

House price declinesresulted in a large number of borrowers having mortgagebalances that exceeded the value of their homes — a conditionoften referred to as “negative equity”. Economic theory andevidence suggest that negative equity is a necessary conditionfor mortgage default.24 Borrowers may face incomedisruptions that temporarily limit their ability to pay and haveneither sufficient savings nor home equity to cover monthlyliving expenses. Other borrowers may default after findingthemselves in a situation where their expectations of futurehouse prices are such that they see little hope of attainingpositive equity in the foreseeable future. In any event, thehouse price declines witnessed in 2007 and 2008 have resultedin a tremendous wave of mortgage defaults and foreclosuresthat, in turn, has imperiled financial institutions withsignificant credit exposure to US residential real estate –particularly exposure to rapidly declining markets and/or toriskier subprime borrowers and investors. Fannie Mae andFreddie Mac certainly fit this bill, as did thrift institutionsoperating on a nationwide basis like Countrywide andWashington Mutual.

Fannie Mae and Freddie Mac were not only singularlyexposed to US residential mortgages, but also operated witha high degree of leverage, owing to a statutory minimumcapital requirement of only 2.5% for on-balance-sheet assetsand 0.45% for net off-balance sheet credit guarantees.Concerns about the GSEs’ concentration of residential

mortgage-related risk and leverage were a consistent themeraised by Federal Reserve officials throughout this decade(e.g., Greenspan 2005, Bernanke 2007).

As of mid-year 2007 (and prior to the beginning of thefinancial crisis), Fannie Mae and Freddie Mac maintainedbook equity values of $39.7 billion and $25.8 billion,respectively. This combined $65.5 billion in equity stoodagainst almost $1.7 trillion in combined assets (3.9% capital-

to-assets ratio) and another$3.2 trillion in net off-balancesheet credit guarantees. Oneyear later, the two GSEs hadexpanded to almost $1.8trillion in combined assets and$3.7 trillion in combined netoff-balance sheet creditguarantees, but their capitalcushions had begun to erode.During those four intervening

quarters, Fannie Mae posted $9.5 billion in losses (althoughit did raise $7.0 billion in new equity) and Freddie Mac lostanother $4.7 billion. Moreover, mark-to-market accountinglosses on ‘available-for-sale’ mortgage-backed securitiessubstantially reduced equity through negative entries to‘accumulated other comprehensive income’ on the GSE’sbalance sheets. As of June 30, 2008, Fannie Mae and FreddieMac reported book values of equity of $41.2 billion and $12.9billion, respectively. Perhaps more telling was that the GSEs’self-reported fair values of equity (i.e., the market value ofassets less the market value of liabilities) as of the same datewere $12.5 billion (Fannie Mae) and -$5.6 billion (FreddieMac).25

During 2008, significant problems emerged in both ofFannie Mae’s and Freddie Mac’s business lines – creditguarantees and portfolio investment. The credit guaranteebusinesses incurred rapidly increasing expenses, largelyowing to loan loss provisions. During 2006, Fannie Maeand Freddie Mac together incurred about $1.1 billion in credit-related expenses. These expenses rose to $1.6 billion duringthe first half of 2007 alone, and then jumped markedly to$6.5 billion during the second half of that year. For the firsthalf of 2008, Fannie Mae and Freddie Mac again saw credit-related expenses almost double to $12.8 billion, and for2008:Q3 alone they totaled $15.3 billion. Given the ongoingdecline in house prices, mounting foreclosures, and the

Given the ongoing decline in houseprices, mounting foreclosures, andthe weakening global economy, itis likely that the GSEs credit losseswill remain elevated for some time.

23Information about the S&P/Case-Shiller house price index can be foundat: <http://www2.standardandpoors.com/portal/site/sp/en/us/page.topic/indices_csmahp/0,0,0,0,0,0,0,0,0,1,1,0,0,0,0,0.html>. Information aboutthe OFHEO house price index can be found at: <www.ofheo.gov>. SeeLeventis (2007) for an analysis of the differences between the two indices.24See Foote, Gerardi, and Willen (2008) and references therein.

25According to Financial Accounting Statement (FAS) Number 157, “fairvalue” is the price that would be received to sell an asset or paid to transfera liability in an orderly transaction between market participants at themeasurement date. For the GSEs’ fair value balance sheets, see: <http://www.fanniemae.com/media/pdf/newsreleases/q22008_release.pdf> (Page17) and <http://www.freddiemac.com/investors/er/pdf/2008fin-tbls_080608.pdf> (Page 13).


weakening global economy, it is likely that the GSEs creditlosses will remain elevated for some time.

Fannie Mae’s and Freddie Mac’s portfolio investmentbusinesses also suffered from mark-to-market losses onmortgage-backed securities held either in trading accountsor as “available for sale”. (Under Generally AcceptedAccounting Principles, or GAAP, securities classified as“hold-to-maturity” are not marked-to-market unless there isan “other than temporary impairment” to value.) This wascaused by an unusual and unforeseen widening of the yieldspread between Fannie Mae and Freddie Mac-guaranteedMBS and 10-year Treasuries. Figure 1 presents the currentcoupon spreads on 30-year fixed-rate mortgages (to 10-yearTreasuries) for Fannie Mae and Freddie Mac between January2007 and July 2008. The observed widening is believed tobe primarily caused by the financial market turbulence, whichled to a heightened demand for US Treasury obligations thatwas reflected by lower Treasury yields. However, theaforementioned credit problems at Fannie Mae and FreddieMac also likely played a role by pushing-up required yieldson the GSEs’ MBS.

Mark-to-market losses also occurred in each GSEs’holdings of “private-label” mortgage securities backed bysubprime and Alt-A mortgages.26 As of mid-year 2007, thetwo GSEs held $252.7 billion in mortgage securities backedby subprime and Alt-A mortgages — virtually all of whichwere rated AAA.27 The GSEs’ holdings of such securitieslikely reflected at least two factors. One is the distorted risk-taking incentives faced by Fannie Mae and Freddie Macbecause of the perceived implied federal guarantee of theirobligations. Another factor was the HUD-regulatedaffordable housing goals that mandated a certain percent ofeach institution’s business devoted to affordable housing.28

Private-label MBS held by Fannie Mae and Freddie Mac weretypically backed by a greater concentration of affordablehousing goal-eligible loans than their own MBS.

During the summer of 2008, investors became increasinglyconcerned about the financial condition of both Fannie Mae

and Freddie Mac. Figure 2 illustrates how the GSEs’ shareprices fell during that time (and following an even moredramatic decline during the fall of 2007). Debt investorsalso sought clarity from the federal government about whetherbondholders would be shielded from any losses that mightultimately arise. Figure 3 shows prices for credit defaultswaps on Fannie Mae and Freddie Mac senior andsubordinated debt between January 2007 and July 2008. Ofparticular note are the spikes in March 2008 (just prior to theBear Stearns rescue) and then again during June and July2008. Holders of Fannie Mae and Freddie Mac senior debt(and MBS) appear to have only reacted modestly to thewidespread perception of GSE financial distress. However,one especially significant and risk-averse investorconstituency, foreign official institutions, began decreasingtheir holdings of Federal Agency obligations at that time.Figure 4 presents relevant weekly data based on holdings incustody accounts at the Federal Reserve Bank of New York.29

In response to increasing concerns that Fannie Mae andFreddie Mac would be unable to rollover their debt, formerUS Treasury Secretary Henry Paulson requested that thefederal government be given broad authority to invest in thetwo GSEs. That provision was included in the Housing andEconomic Recovery Act that passed in July 2008.

III. Federal Intervention

According to former Treasury Secretary Henry Paulson(2009), immediately following the passage of the new housinglegislation, the Treasury began a comprehensive financialreview of Fannie Mae and Freddie Mac in conjunction withthe FHFA, the Federal Reserve, and Morgan Stanley.30 TheGSEs believed in their solvency and thought that any capitaldeficiency below regulatory minimums could be rectified bysignificant asset sales. Given that US mortgage markets hadalready been disrupted for almost one year at that time, theprospect of Fannie Mae and Freddie Mac retrenching wasnot an appealing policy option.

In early August 2008, both Fannie Mae and Freddie Macreleased their second quarter earnings. As of June 30, 200826Private-label mortgage securities are those not guaranteed by Fannie Mae,

Freddie Mac, or Ginnie Mae. Subprime mortgages refer to those loansmade to borrowers with riskier credit characteristics, such as measured bycredit scores, loan-to-value ratios, or debt-to-income ratios. Alt-Amortgages refer to loans made with little or no documentation.

27Data provided in US Office of Federal Housing Enterprise Oversight(2008) indicate that, of this amount, Fannie Mae accounted for $81.4 billionand Freddie Mac $174.3 billion.

28The GSEs’ housing goals were established in the Federal HousingEnterprises Financial Safety and Soundness Act of 1992. The law requiredHUD to set annual housing goals for Fannie Mae and Freddie Mac and tomonitor the GSEs’ performance in meeting those goals. This responsibilitywas transferred to the FHFA as part of the Housing and Economic RecoveryAct of 2008.

29Fannie Mae and Freddie Mac account for about 70% of all Federal Agencyobligations outstanding. The other key issuer is the Federal Home LoanBank System. The data comes from a memorandum to the Federal Reserve’sH.4.1 release: Factors Affecting Reserve Balances. Quarterly Flow of Fundsdata (Table L.107) corroborates the trend illustrated by showing that foreignofficial holdings of Federal Agency obligations peaked in 2008:Q2.Interestingly, the same data indicates that holdings in foreign privateaccounts peaked sooner – as of 2007:Q4.

30Morgan Stanley was hired by the Treasury to provide market analysisand financial expertise in connection with its authorities to invest in FannieMae and Freddie Mac (e.g., Solomon and Paletta, 2008).


Figure 1: Federal Agency 30-year Current Coupon MBS Spread to 10-year Treasury

basis points

0

50

100

150

200

250

Jan-07 Mar-07 May-07 Jul-07 Sep-07 Nov-07 Jan-08 Mar-08 May-08 Jul-08

Fannie Mae

Freddie Mac

Source: Bloomberg

0

10

20

30

40

50

60

70

80


Fannie MaeFreddie Mac

Figure 2: Fannie Mae & Freddie Mac Stock Prices

Source: Bloomberg


Source: Bloomberg

pbasis points, 5-year (senior & subordinated debt)

0

50

100

150

200

250

300


Freddie Mac Subordinated Fannie Mae SubordinatedFreddie Mac Senior Fannie Mae Senior

Subordinated

Senior

Figure 3: Fannie Mae & Freddie Mac Credit Default Swaps

$ billions

400

500

600

700

800

900

1,000

1,100

Jan-07 Apr-07 Jul-07 Oct-07 Jan-08 Apr-08 Jul-08 Oct-08

Figure 4: Marketable Federal Agency Securities Held for Foreign Official & International Accounts

Source: Federal Reserve Board


both GSEs were both technically solvent insofar as the bookvalue of their equity capital was positive. (At that time, FannieMae had $41.2 billion in book equity and Freddie Mac $12.9billion.) However, there was a compelling case that – on aneconomic basis – both were actually insolvent. First, asmentioned previously, the GSEs’ reported fair values of equitywere much lower – and in Freddie Mac’s case fair value wasactually negative. Second, both institutions were carryingrelatively large “tax deferred assets” to allow them to reducefuture income taxes. These amounts were $20.6 billion forFannie Mae and $18.4 billion for Freddie Mac. If FannieMae and Freddie Mac were subject to the bank regulatorystandard for tax-deferred assets – and in light of theirextremely weak near-term earnings prospects – those assetswould have been written-off and taken total book equity downto $20.6 billion (Fannie Mae) and -$5.5 billion (FreddieMac).31

These facts, taken together with deteriorating mortgagemarket conditions and a view that the GSEs had beenespecially conservative in estimating expected future losses,made a compelling case for swift federal action.32 And onSeptember 7, 2008, FHFA Director James Lockhart, TreasurySecretary Henry Paulson, and Federal Reserve Chairman BenBernanke outlined a plan to stabilize the residential mortgagefinance market. This included: (1) placing both Fannie Maeand Freddie Mac into conservatorship, (2) having the Treasuryenter into senior preferred stock agreements with both GSEs,and (3) establishing two new Treasury-operated liquidityfacilities aimed at supporting the residential mortgage market— a mortgage-backed securities purchase facility and astanding credit facility.

The reasoning for the imposition of the conservatorshipswas that both Fannie Mae and Freddie Mac were financiallydistressed and could not perform their public missions – thatis, providing counter-cyclical support to mortgage marketsand financing affordable housing. By becoming aconservator, the FHFA assumed the responsibilities of thedirectors, officers, and shareholders of both Fannie Mae andFreddie Mac with the purpose of conserving each GSEs’ assetsand to rehabilitate them into safe-and-sound condition. NewCEOs were named to act as agents of the conservator.

Concurrent with the conservatorships, the Treasury entered

into a senior preferred stock agreement with each GSE.33 Thepurpose of the agreements is to ensure that Fannie Mae andFreddie Mac maintain positive net worth going forward. Ifthe regulator determines that either institution’s liabilitiesexceed assets under GAAP, the Treasury will contribute cashcapital equal to the difference in exchange for senior preferredstock. Each of these agreements is of an indefinite term andfor up to $100 billion. After its 2008:Q3 earnings release,Freddie Mac drew $13.8 billion. Both GSEs are expected torequire significant Treasury capital infusions after theannouncement of their respective year-end 2008 financials.Preliminary figures suggest that Fannie Mae will require asmuch as $16 billion and Freddie Mac as much as another$35 billion (Kopecki 2009).

The senior preferred stock accrues dividends at 10% peryear, a rate that steps up to 12% if in any quarter dividendsare not paid in cash. Also, in exchange for the senior preferredstock agreements, the Treasury received from each FannieMae and Freddie Mac: (1) $1 billion of senior preferredshares, (2) warrants for the purchase of common stockrepresenting 79.9% of each institution on a fully diluted basis,and (3) a quarterly commitment fee (starting March 31, 2010)to be determined by the Treasury and the FHFA (asconservator) in consultation with the Federal Reserve.

The senior preferred stock agreements require each GSEto begin shrinking their retained investment portfolios in 2010at a rate of 10% per year until they each fall below $250billion. This provision was intended to assuage policymakerconcerns about the GSEs’ investment portfolios, which hadbecome widely viewed as posing a systemic risk to thefinancial system and providing little social welfare benefit.34

The senior preferred stock agreements also included variouscovenants. Specifically, Treasury approval is required before:(1) purchasing, redeeming or issuing any capital stock orpaying dividends, (2) terminating conservatorship other thanin connection with receivership, (3) increasing debt to greaterthan 110% of that outstanding as of June 30, 2008, and (4)acquiring, consolidating, or merging into another entity.

The Treasury’s GSE credit facility is for Fannie Mae,Freddie Mac, and the Federal Home Loan Bank System andis operated by the Federal Reserve Bank of New York.35 Asof year-end 2008, no credit had been extended through thisprogram. The MBS purchase program, by contrast, had

31While acceptable under GAAP, bank regulators require institutions towrite-off all but the lesser of: (1) the amount of tax deferred assets theinstitution expects to realize in the next 12 months, or (2) 10% of Tier 1capital. For example, for state member banks, see: 12 C.F.R. 208 AppendixA, Section II(B)(4).

32 Morgenstern and Duhigg (2008) report that Morgan Stanley (workingon behalf of the Treasury) concluded that both GSEs had overstated theirfinancial condition by postponing various types of losses.

33See

34See Eisenbeis, Frame, and Wall (2007) for an overview of the policyconcerns and the related literature.

35Credit must be collateralized, can be extended for one-to-four weeks,and is priced at LIBOR plus 50 basis points. Eligible collateral is limitedto Fannie Mae and Freddie Mac mortgage-backed securities and FederalHome Loan Bank advances.

<http://ustreas.gov/press/releases/reportspspa_factsheet_090708%20hp1128.pdf.


accumulated $71.5 billion.36 Consistent with the GSEinvestment provisions in the Housing and Economic RecoveryAct of 2008, credit extensions and MBS purchases must bemade by year-end 2009 (although previously purchasedsecurities may be held beyond that time).

The intent of the senior preferred stock agreements andTreasury liquidity facilities was to provide comfort to FannieMae’s and Freddie Mac’s senior and subordinate creditorsand holders of mortgage-backed securities.37 By extension,these actions were expected to lower and stabilize the cost ofmortgage finance. Figures 5 and 6 illustrate the announcementeffect for Fannie Mae and Freddie Mac 5-year debt spreadsand current coupon MBS spreads, respectively. The tighterspreads on mortgage-backed securities, in turn, resulted inconforming mortgage rates falling by about 50 basis points.

Of course, the two agreements had significant negativeconsequences for the GSEs’ common and preferredstockholders. Fannie Mae and Freddie Mac common sharesquickly fell below $1, down from $60 just 12 months earlier.Indeed, as a result of trading at such low levels, the two GSEsnow face delisting.38 Preferred shares suffered a similar fate.Indeed, several community banks became financiallydistressed themselves as a result of having to write-down thevalue of their holdings of GSE preferred stock.39

The positive bond market reaction, coupled with a relativelysmooth operational transition, suggested that the impositionof conservatorships at Fannie Mae and Freddie Mac was, sofar, a success. However, by November 1, 2008, mortgagerates and yields on Fannie Mae and Freddie Mac obligationshad climbed back to pre-conservatorship levels because ofworsening financial market conditions. Policymakers thensearched for additional tools to lower and stabilize the costof mortgage finance. In response, the Federal Reserveannounced on November 25 that it was establishing new

36See <http://www.fms.treas.gov/mts/index.html>.

37On September 11, 2008, the Treasury issued a press release intended toclarify the status of the senior preferred stock agreements. See <http://www.ustreas.gov/press/releases/hp1131.htm>. The Treasury affirmed thatthe agreements are permanent and that legislative efforts to abrogate themwould give rise to government liability to parties suing to enforce theirrights under the agreements. The senior preferred stock agreements mayonly be terminated by either: 1) full funding by the Treasury ($100 billion),2) GSE liquidation, or 3) GSE satisfaction of all liabilities. In some sense,the senior stock purchase agreements have become an appendage to theGSE charters.

38The NYSE Listing Manual (Part 802.01C) notes that a company will bedeemed to be below compliance standards if the average closing price of asecurity is less than $1.00 over a consecutive 30-day trading period. Oncenotified, a company has six months to bring its share price and averageshare price above the $1.00 threshold. Fannie Mae and Freddie Mac wereeach notified in November 2008.

39See McGeer (2008) and Blackwell and Flitter (2008) for some discussion.

facilities to: 1) purchase up to $500 billion in mortgage-backed securities guaranteed by Fannie Mae, Freddie Mac,and Ginnie Mae, and 2) purchase up to $100 billion in debtobligations of Fannie Mae, Freddie Mac, and the FederalHome Loan Bank System. Figures 5 and 6 also show apositive market response to these announcements.

IV. Conclusion

Fannie Mae and Freddie Mac play a central role in the USresidential mortgage finance system. As real estate pricesfell and mortgage defaults and foreclosures mounted, the twohighly leveraged GSEs became financially distressed. Inresponse, Fannie Mae’s and Freddie Mac’s federal regulatorplaced both institutions into conservatorship and the USTreasury entered into senior stock purchase agreements witheach GSE and introduced new liquidity facilities aimed atsupporting the institutions and mortgage markets moregenerally.

The federal intervention into Fannie Mae and Freddie Machas been successful insofar as it improved the confidence ofcreditors and stabilized residential mortgage markets.However, the current arrangement of government ownershipand control over these two enormous financial institutionswill likely be revisited by the Congress in the months ahead.Today’s consensus appears to be that the previous public-private business model is inherently flawed and unstable.

Indeed it is unclear what role Fannie Mae and Freddie Macwill ultimately play in the US housing finance system, andthe reasons for this uncertainty do not solely rest with thetwo GSEs. The financial distress at Fannie Mae and FreddieMac has occurred along with significant and well-publicizedproblems at a host of mortgage originators, private mortgageinsurance companies, and monoline bond insurers. Hence,the federal government may need to redefine its role insupporting primary and secondary mortgage markets. FederalReserve Chairman Bernanke (2008) and former TreasurySecretary Paulson (2009) have offered some initial thoughtsabout various policy options. Nevertheless, additionalresearch and policy analysis should commence quickly aboutthe public-sector’s role in mortgage markets, the efficacy ofthe GSE model of financial intermediation, and the future ofFannie Mae and Freddie Mac.


basis points

0

30

60

90

120

150

180

Aug-08 Sep-08 Oct-08 Nov-08 Dec-08

Fannie Mae

Freddie Mac

Conservatorship

Fed Agency Purchase

Announcement

Figure 5: Federal Agency 5-year Debt Spread to 5-year Treasury

Source: Bloomberg

basis points

0

50

100

150

200

250

Aug-08 Sep-08 Oct-08 Nov-08 Dec-08

Fannie Mae

Freddie Mac

Conservatorship

Fed MBS Purchase

Announcement

Figure 6: Federal Agency 30-year Cuurent Coupon MBS SPread to 10-Year Treasury

Source: Bloomberg


Ambrose, B.W., M. LaCour-Little, and A.B. Sanders, 2004,“The Effect of Conforming Loan Status on Mortgage YieldSpreads: A Loan Level Analysis,” Journal of Real EstateEconomics 32 (No. 4), 541-569.

Ambrose, B.W. and A. Warga, 1996, “Implications ofPrivatization: The Costs to Fannie Mae and Freddie Mac,”In US Department of Housing and Urban Development,Studies on Privatizing Fannie Mae and Freddie MacWashington, D.C., HUD, 169-204.

Ambrose, B., and A. Warga, 2002, “Measuring Potential GSEFunding Advantages,” Journal of Real Estate Finance andEconomics 25 (No. 2-3), 129-150.

Bernanke, B.S., 2008, “The Future of Mortgage Finance in theUnited States,” Remarks to the UC Berkeley/UCLASymposium: The Mortgage Meltdown, the Economy, andPublic Policy (October 31).

Bernanke, B.S., 2007, “GSE Portfolios, Systemic Risk, andAffordable Housing,” Remarks to the IndependentCommunity Bankers of America, (March 6).

Blackwell, R. and E. Flitter, 2008, “Regulators and Bankers atOdds Over GSE Seizure,” American Banker (September 11).

Brown, J., 2001, “Reform of GSE Housing Goals,” In Peter J.Wallison, Editor, Serving Two Masters Yet out of Control,Washington D.C., AEI Press, 153-165.

Bunce, H.L., 2002, “The GSEs Funding of Affordable Loans: A2000 Update,” US Department of Housing and UrbanDevelopment Housing Finance Working Paper HF-013(April), http://www.huduser.org/publications/hsgfin/workpapr13.html.

Eisenbeis, R.A., W.S. Frame, and L.D. Wall, 2007, “An Analysisof the Systemic Risks Posed by Fannie Mae and Freddie Macand An Evaluation of the Policy Options for Reducing ThoseRisks,” Journal of Financial Services Research 31 (No. 2),75-99.

Flannery, M.J. and W.S. Frame, 2006, “The Federal HomeLoan Bank System: The ‘Other’ Housing GSE,” FederalReserve Bank of Atlanta Economic Review 91 (Q3), 33-54.

Foote, C.L., K. Gerardi, and P.S. Willen, 2008, “Negative Equityand Foreclosure: Theory and Evidence,” Journal of UrbanEconomics 64 (No. 2), 234-245.

Frame, W.S. and L.D. Wall, 2002, “Financing Housing throughGovernment-Sponsored Enterprises,” Federal Reserve Bankof Atlanta Economic Review 87 (Q1), 29-43.

Frame, W.S. and L.J. White, 2005, “Fussing and Fuming overFannie and Freddie: How Much Smoke, How Much Fire?”Journal of Economic Perspectives 19 (No. 2), 159-184.

Greenspan, A., 2005, “Testimony before the Committee onBanking, Housing, and Urban Affairs,” United States Senate(April 6), http://www.federalreserve.gov/boarddocs/testimony/2005/20050406/default.htm.

Heuson, A., W. Passmore, and R. Sparks, 2001, “Credit Scoringand Mortgage Securitization: Implications for Mortgage Ratesand Credit Availability,” Journal of Real Estate Finance andEconomics, 23 (No. 3), 337-363.

Kopecki, D., 2009, “Fannie, Freddie Funding Needs May Pass$200 Billion, FHFA Says,” Bloomberg (February 10).

Leventis, A., 2007, “A Note on the Differences between theOFHEO and S&P/Case-Shiller House Price Indexes,” http://www.ofheo.gov/media/research/notediff2.pdf.

McKenzie, J., 2002, “A Reconsideration of the Jumbo/Non-Jumbo Mortgage Rate Differential,” Journal of Real EstateFinance and Economics 25 (No. 2), 197-213.

McGeer, B., 2008, “Preferred Exposure Fallout,” AmericanBanker (September 9).

Nothaft, F.E., J.E. Pearce, and S. Stevanovic, 2002, “DebtSpreads Between GSEs and Other Corporations,” Journalof Real Estate Finance and Economics 25 (No. 2), 151-172.

Passmore, W., 2005, “The GSE Implicit Subsidy and the Valueof Government Ambiguity,” Real Estate Economics 33(No. 3), 465-486.

Passmore, W., S. Sherlund, and G. Burgess, 2005, “The Effectof Housing Government-Sponsored Enterprises onMortgage Rates,” Real Estate Economics 33 (No. 3), 427-463.

Passmore, W., R. Sparks, and J. Ingpen, 2002, “GSEs,Mortgage Rates, and the Long-Run Effects ofSecuritization,” Journal of Real Estate Finance andEconomics 25 (No. 2) 215-242.

Paulson, H., 2009, “The Role of the GSEs in Supporting theHousing Recovery,” Remarks to the Economic Club ofWashington (January 7).

Solomon, D. and D. Paletta, 2008, “Treasury Hires MorganStanley for Advice on Fannie, Freddie,” Wall Street Journal(August 6).


US Congressional Budget Office, 1996, Assessing the PublicCosts and Benefits of Fannie Mae and Freddie Mac,Washington, D.C., CBO.

US Congressional Budget Office, 2001, Federal Subsidiesand the Housing GSEs, Washington, D.C., CBO.

US General Accounting Office, 1990, Government-Sponsored Enterprises: The Government’s Exposure toRisks, Washington, D.C., GAO.

US Office of Federal Housing Enterprise Oversight, 2008,Report to Congress, Washington, D.C., OFHEO.

Book Review:Ending the Management Illusion: How to

Drive Business Results Using thePrinciples of Behavioral Finance

By Hersh Shefrin, McGraw Hill: 2008, vii + 317 pages

137

Someone whose business responsibilities or researchareas do not overlap with behavioral finance will findShefrin’s latest book an interesting and efficient way to learnabout how behavioral principles can be extended from therealm of investment decisions to the arena of corporatefinancial management. Most chapters, which can easily bedigested in separate sessions, are concise and interesting. Thetext is written at the middle-management level and will alsobe useful for academics that teach or consult for executives.

The first chapter of the text is one of the strongest. It usesexcellent real-world illustrations to explain the psychologicalarguments and traits that induce educated, powerfulindividuals to make biased decisions and introduces Shefrin’srecommendations for improving the way corporations aremanaged. Improving access to financial information andfinancial literacy for all of a firm’s employees are cornerstonesof Shefrin’s recommendations for ways to eliminatebehavioral biases, and he provides several examples of howfirms have successfully attained these laudable goals. Thesecond chapter discusses the financial metrics the authorwould include in each employee’s training, but the level ofdiscussion is a bit advanced, even for finance professionalsif they do not use valuation principles on a regular basis.

The middle chapters of the text are its weakest components.Chapter 3, “Narrow Financial Focus on Projects andFinancing: Traditional Approach”, contains a litany of storiesof financial mis-steps made by managers at a variety offamiliar firms that do not really serve to advance theeducational recommendations the author is making. It alsolavishes praise on firms that use managerial techniques ofwhich the author approves in a way that sounds self-servingand diverts attention from the author’s main points. Thediscussion of the difference between Free Cash Flow to theFirm and Free Cash Flow to Investors remains difficult toparse after several readings. Chapter 4’s reliance onaccounting measures of return and investment is puzzlinggiven earlier discussions of the importance of distinguishingbetween accounting profit and actual cash flow. Chapter 5,“Involving the Workforce in Financial Planning”, is quitedense. Corporate managers could probably convinced of theimportance of using simulated games and other experientialtechniques to teach employees about the crucial value driversof the business they are in without a long pretend“conversation” at an environmental firm.

Chapter 6, “Motivating the Workforce Through SmartCarrots and Sticks”, returns to the text’s strongest feature,which is the link between business decisions and behavioral


psychology via a discussion on constructing appropriatebonus and incentive plans. This issue is intriguing for anacademic, but the real-world focus here is quite narrow, withmost of the examples being drawn from a single local firm.The chapter could have easily been combined with chapter7, “Sharing Information Throughout the Organization”. Thischapter on information sharing has a fascinating explanationof how hard the management of the equity research practiceat Union Bank of Switzerland works to be sure that the analystreports they issue are not subject to typical psychologicalglitches that could bias the recommendations.

The text ends with chapter 8, “Integration: The Whole Ballof Wax”, that summarizes most of the author’s ideas andcontains an invaluable table (8.1) that lists the variouspsychological pitfalls and biases human beings could fall preyto when making financial or managerial decisions.Throughout the text, Shefrin discusses various types ofinformation that firms post on their walls to incentivizeemployees. Table 8.1 is a list of forces that should appear inevery conference room in corporate America wheremanagerial decisions are made.

In summary, Shefrin’s latest literary effort is a worthwhileattempt to educate practitioners and academics about howthe principles of behavioral psychology can be applied tocorporate decision-making. Its main messages are 1) thatsharing information, (especially financial information) withemployees at all levels improves productivity and efficiency,2) that all stakeholders should be educated about how financialinformation is presented and evaluated, and 3) that everyonemakes biased decisions but that these biases can be identifiedand corrected. The rationale behind these recommendationsis presented and defended using real-world examples todevelop a treatise that is worthwhlle reading for anyone whoneeds to motivate, manage or train individuals to battle the“gremlins”, (a favorite term the author uses to describe typicalbehavioral glitches that bias human decisions) that confrontus all every day.

Andrea HeusonProfessor of FinanceUniversity of Miami

Book Review

The Venturesome EconomyBy Amar Bhidé, Princeton University Press: 2008, ix + 499 pages

In The Venturesome Economy (2008, PrincetonUniversity Press), Amar Bhide offers an in-depth, refreshinglyoptimistic, and insightful counterpoint to the increasingly loudvoices decrying the trend of economic globalization. Dr.Bhide gives particular attention to what he terms the technonationalists – those who believe, “Leadership in science andtechnology gives the United States its comparative advantage”(p. 262).

He describes techno nationalists as people who promote anational agenda to divert scarce resources in order to facilitatean increase in the number of native born engineers,mathematicians, scientists, statisticians, and researchers. Thefundamental premise of the techno-nationalist argument isthat the competitive advantage of the US and other developedeconomies lies in their scientific and technological acumen– that they are the world’s preeminent innovators because oftheir mathematic and scientific excellence. The ability to“out innovate” the rest of the world drives the economy.Techno-nationalists argue that if that competitive advantageerodes, so will the economy. Former Chancellor of theExchequer in the UK, Gordon Brown, summed up theargument thusly (as Dr. Bhide reports on p. 264), “Everyadvanced industrial country knows that falling behind inscience means falling behind in commerce and prosperity.”The techno-nationalists believe this is already occurring.

The techno-nationalist position is supported by numerousstatistics, some of which Dr. Bhide outlines in his book. Hecites the work of respected scholars such as Clyde Prestowitzand Richard Freeman to outline the empirical foundation ofthe techno-nationalist movement. Freeman’s work isespecially representative of the cause. He notes that in 1970US educational institutions granted over half of all doctoratesawarded in science and engineering fields. By 2001,European Union countries were granting 54% more of these

doctorates than the US. Over that same period, China wentfrom granting almost no doctorates to granting roughly 1/3as many Ph.Ds as the US. Freeman also notes the decline inpublications and citations of US researchers. For instance,he quotes a New York Times article that states, “The share ofpapers (by US researchers) counted in the Chemical AbstractService fell from 73% in 1980 to 40% in 2003” (p. 262).Freeman and Prestowitz’s arguments Dr. Bhide outlines areconsistent with the much publicized recent studies showingthe US school children slipping in international academiccomparisons.

According to the techno-nationalists, the result of thisdecline in the US’s scientific and technological competitiveadvantage is a weakening of the US economy. Here too,they offer empirical support. On this point, Dr. Bhide devotesmore discussion to alarming forecasts than actual data. Forinstance, he cites a 2006 study by Alan S. Blinder in whichBlinder claims that 40 million jobs in the service sector wereat risk of being off-shored. Dr. Bhide writes of a similarreport from Forrester Research in 2002 that estimated 3.3million jobs in the service sector would be moved to othercountries by 2015. Surely, the recent troubles of the Detroitautomakers, along with the economy-wide meltdown of thesecond half of 2008, add fuel to the techno-nationalist fire.

Dr. Bhide, however, rejects the techno-nationalist paradigmand rebuts the fundamental thinking behind thistechnologically protectionist movement. While it’s aninjustice to the author to attempt to summarize his work in asingle paragraph, I believe the following excerpt from thebook captures the spirit (if only partly) of his message:

…the United States should welcomemore research from China and India,because an increase in the supply of high-level know-how helps mid-level innovators


based in the United States develop productsthat increase productivity and wages in theUnited States (p. 59).

His book is both timely and relevant. Regarding thecontinued prosperity of the United States, it is possible thatthere is no more important issue than economic globalization.The much publicized trend of outsourcing and off-shoring ofUS jobs – now extending far beyond the manufacturingindustry – has created a groundswell of support for increasedeconomic protectionism, further intensified by the collapseof the US auto industry. But the US’s attitude toward andresponse to economic integration on a global scale issomething we must get right; we cannot afford to take thewrong position on this issue. For that reason, I believe Dr.Bhide’s book should be required reading for anyone seekingto join the debate on what policies the US should embracerelated to free trade and international economic integration,regardless of whether the reader accepts or rejects Dr. Bhide’sthesis.

The Venturesome Economy is a very well-written andthoughtfully organized two-part analysis of the effects ofinnovation and globalization on the US economy. The firstpart (Book I) builds a foundation for the second part and isprimarily devoted to qualitative analysis stemming from anextensive study he performed involving 106 venture capital(VC)-backed businesses in the US. The first part issupplemented by numerous comparisons and contrasts of Dr.Bhide’s current work with his previous work focusing on firmsfrom the well-known Inc. 500 List. The second part of thebook (Book II) is devoted to a discussion of policy debatesrelated to the tide of globalization, especially the debateregarding the importance of developing high-level innovationon the US soil (and by the US citizens).

The book is written in easy to digest modules, which readerswill appreciate. Readers can choose to read the foundationin Book I and then proceed to the prescriptions of Book II,or they can simply jump directly to the policy debates ofBook II and use Book I as reference material. Similarly,each chapter is written to allow readers to spend as much (oras little) time as necessary. I found the chapter prefaces andconclusions to be concise and content rich, while the bodiesof the chapters were persuasive and intriguing. Readers willbenefit greatly from the bountiful mini-case studiesthroughout. Literally, almost every page in the book containsan example that relates the points Dr. Bhide is articulating tothe real-world companies he studied. And although his studyis intentionally short on econometrics (a point I return to atthe end of my review), what he lacks in numerical analysisDr. Bhide makes up for with enlightening qualitative analysisspringing from countless hours of interviews with CEOs ofimportant mid-level companies.

In Book I, Dr. Bhide uses his study to make the case the

innovation and economic prosperity related to innovation arenot as simple as many purport. He stratifies innovation intothree levels: high, mid, and ground-level innovation. Highlevel innovation is the kind that deals with molecular orphysical advances. Mid-level innovation is the applicationof a high-level innovation. It represents an innovation perse, but not to the same degree. Ground-level innovationrepresents the application of a mid-level innovation, and againrepresents an innovation per se, but again, not to the samedegree as the mid-level innovation. The example he gives isthat the invention of the silicon-based micro-processorrepresents a high-level innovation, the creation of themotherboard is a mid-level innovation, and the developmentof the personal computer represents a ground-levelinnovation.

Dr. Bhide argues in Book I, through his interviews andsurveys of CEOs of VC-backed businesses, that mid-levelinnovators play a key role. They take the somewhat difficultto understand and difficult to apply high-level innovation (andother mid-level innovations) and begin the critical processof transforming that innovation into a marketable andbeneficial product that helps to foster economic progress andprosperity. He goes on further to argue that mid-levelinnovators are relatively unconcerned with the location fromwhich the high-level innovation emanates, so long as theyare able to transform it into something marketable andbeneficial to their consumers. He also argues that thetransformation of high-level innovation into mid and ground-level innovation is critically influenced by the willingness ofconsumers to try new technologies and by the interaction ofmid and ground-level innovators with their customers. Thefirst point highlights the importance of an economy filledwith “venturesome” consumers who are willing to try newinnovation. Such venturesome consumers attract all levelsof innovation. On this point, Dr. Bhide believes the USrepresents an innovation Mecca. It is the country where allinnovators seek to land their products and services. On thelatter point, most mid-level innovators require a closerelationship with their beta customers that allows for aniterative development process to fine-tune products to meetcustomer demand. This iterative development processrequires, for the most part, geographic, language, and culturalproximity. In other words, Dr. Bhide is convinced that mid-level innovation is not currently and will not in the near futurebe something that can be moved off shore. These pointspresent critical ramifications: 1.) having consumers thatembrace new technology makes an economy the naturalmagnet for all levels of innovation, and 2.) mid-levelinnovation (the adaptation of high-level innovation intomarketable products) is something that has been and willcontinue to beneficial for domestic economies. Regardlessof where the high-level innovation comes from, the mid-level

141WRIGHT— BOOK REVIEW OF THE VENTURESOME ECONOMY

innovation that results will benefit the country where it occurs,both by creating jobs and by providing its citizens with abeneficial new product or service. Most of Dr. Bhide’sdiscussion on this topic is derived from his own work withthe CEOs of VC-backed businesses, which he describes asbeing almost universally mid-level innovators. Further, it isimportant to recognize that much of Dr. Bhide’s conclusionson the topic are derived from qualitative analysis and criticalthinking. There is little in the way of rigorous econometricanalysis. Nevertheless, his reasoning and qualitative evidenceis sound and convincing.

These keystone points provide support for his argument inBook II that the US should not fear high-level innovationfrom foreign countries or from foreign workers in the US.Instead, the US should embrace high-level innovation fromall sources without diverting scarce resources to attempt tosecure a position of superiority in the creation of those high-level innovations. The US is best served by pursuing a policyagenda that (a) invites high-level innovation from all localesand nationalities and (b) that facilitates mid-level innovationwithin the US itself.

Regarding the second point, facilitating mid-levelinnovation, Dr. Bhide provides evidence from his study thatPh.D.’s are not necessary for mid-level innovation. He evenquotes one CEO who states frankly that most Ph.D.’s at hiscompany consider the attainment of their Ph.D. to have beena career mistake (it was unnecessary in performing the rolesthey fill inside the company). Dr. Bhide ultimately concludesthat fostering an environment that invites high-levelinnovation into the country and encourages mid-levelinnovation within the United States likely does not requireany significant policy initiative to increase the number ofengineers, mathematicians, scientists, statisticians, andresearchers. In fact although I don’t see this explicitly statedin the book, I come away feeling as though Dr. Bhide issuggesting a relatively lassies faire rebuttal to the technonationalist push to address the shrinking number of Americanborn hard scientists in the United States.

The overarching message of the book is one of optimism.Perhaps the most important factor in enjoying economicprosperity spurred by technological innovation is to simplyhave a country full of venturesome consumers – people with

the means and gusto to buy new, high-tech products. Thiswillingness of consumers creates a magnet market for alllevels of innovation. As all levels of innovation come fromvarious sources, from Mumbai to Paris, mid-level innovatorsin the US will use the imported or domestically producedinnovations to develop even more innovations, which willnot be outsourced or off shored. (While he concedes groundlevel innovation and basic production can be and is off-shored,he is adamant that mid-level innovation is difficult to off-shore.) In short, innovation breeds innovation. Theinnovation cycle creates jobs. But equally important, theinnovation cycle creates products and services from whichconsumers, not innovators, reap the lion’s share of thebenefits. Dr. Bhide even provides convincing statisticalsupport for this rosy cycle of innovation and economicprosperity. For instance, he cites a study by Gene Epstein inwhich Epstein notes that from 2002 to 2006, employment inthe two sectors most vulnerable to off-shoring had increased7.7%, while employment in the overall service sectorincreased 4.5%. Dr. Bhide argues that pursuing a policy ofprotectionism and isolationism will do nothing but hinderthis process.

The subject of the US’s role in and response to globaleconomic integration is so critically important to the futureof the US’s prosperity that we cannot afford to get it wrong.Any decisions we make and any policies we enact must bedone only after careful and measured investigation. Giventhe importance of the subject, I believe Dr. Bhide’s workwould have benefited from a more rigorous econometricapproach to his study. He is very clear about why he chosenot to do that. But such an analysis used as a supplementcould not have hurt the veracity of his book and likely wouldadd increased credibility to his conclusions. In spite of thisone perceived deficiency, which was a conscious choice byDr. Bhide, I found the book well written and highlyinformative and believe it makes a substantial contributionto a critical topic that is currently being hotly debated in theUS.

Colby WrightAssistant Professor of Finance Central Michigan University


Financial Puzzles

Stewart C. Myers

142

14. This problem considers how corporate income taxes can affect IRRs for capital investment projects. Assume a constantfuture tax rate T.

(a) In what circumstances is the IRR of after-tax cash flows exactly equal to the IRR of pre-tax cash flows multipliedby 1 – T?

(b) What tax system always generates the same IRR for both pre-tax and after-tax cash flows?

(c) What tax system always generates the same IRR for any pattern of after-tax cash flows? Hint: the IRR is negative.

15. A convertible bond emerges from call protection with the conversion option well in the money. The issuing company cancall and force conversion now, but does not do so. Are the convertible bondholders necessarily better off because of thedecision not to call? Does the price of the bond necessarily increase? Assume that post-conversion dividends would equalthe coupon payments on the bond, so that there is no cash-flow advantage or disadvantage from conversion.

Proposed Answers for Problems 11 – 13In the Fall/Winter, 2007 issue of the Journal of Applied Finance, p. 131

11. Book return equals the true rate of return (IRR) in the following cases.

(a) Book depreciation always equals economic depreciation. Economic depreciation would be calculated as the changein the PV of project cash flows using the IRR as the discount rate.

(b) The firm settles into steady-state growth and the growth rate equals the IRR.

Condition (b) was first derived in Solomon and Laya (1967).1

12. Think of a project balance sheet, with PV(Revenue) on the asset side and PV(Fixed costs) and PV(Variable costs) on theliability side. NPV = PV(Revenue) – PV(Fixed costs) – PV(Variable costs). Project A’s costs are fixed. Project B’s costs arefixed, but with diversifiable noise added. Project C’s costs are variable, with = 0.5.

Net cash flows and NPV are safer (lower ) for project C than for A and B. C’s variable costs are a partial hedge againstC’s uncertain revenues. A’s and B’s costs do not provide such a hedge. A project amounts to a long position in PV(Revenue) and short positions in PV(Fixed costs) and PV(Variable costs).The overall position is safer (lower ) if the short position is in a positive- asset. Thus NPV(C) has a lower and a lower

1E. Solomon and J. Laya (1967), “Measurement of Company Profitability: Some Systematic Errors in Accounting Rates of Return,” in A. A. Robichek,ed., Financial Research and Management Decisions, John Wiley & Sons, Inc., New York.

With the previous issues of JAF, we began publishing Financial Puzzles by Stewart C. Myers. This set (problems 14-15) isthe fifth installment. The solutions for the previous set are also given below.

—The Editors

143MYERS — FINANCIAL PUZZLES

cost of capital than NPV(A) and NPV(B). C is the more valuable project.

13. (a) NPV > 0 when x is close to zero. If the institutional investor only has to put in $1 for a $100 million project, thenthe investor is almost certain to earn the return r = RCAP immediately. The return r includes a risk premium, but thepayoff is risk-free, so NPV > 0. NPV may be even larger if the first month’s cash flow is greater than 1 + RCAP andthe investor can keep all of that month’s cash flow. NPV < 0 when x = 1. When x = 1, the institutional investor contributes $100 million to a project with PV = $100million. However, the investor may not get all future cash flows. There is another residual claim that gets allsubsequent cash flows if and when IRR reaches RCAP. This residual claim has positive value, so the value remainingto the institutional investor must be less than $100 million.

(b) NPV at first increases as x increases and then declines. NPV > 0 for small values of x, for the reasons given in (a).But a larger x also means that the investor has to put in more money and wait longer to earn IRR = RCAP. The risk thatthe investor will not earn RCAP increases, and value finally declines. NPV < 0 at x = 1.

(c) Once the project is up and running, poor performance can actually benefit the institutional investor. Suppose thatIRR after 60 months is just shy of RCAP. A large project cash flow for month 61 could be just enough to achieve IRR> RCAP and put the investor out of the game. A slightly lower cash flow could leave IRR < RCAP, assuring the investorof cash in both months 61 and 62 and possibly in later months. Two or more cash flows can be better than one, evenif the first cash flow is not as high as it could have been. The ideal outcome for the investor would be a series of cashflows, each as large as possible without bringing IRR > RCAP. This would be a disappointing outcome ex ante, at thestart of month 0, but not ex post, when earlier cash flows are already money in the bank.

(d) Financing from the institutional investor could require RCAP > r, in order to give the investor additional upside tooffset the negative NPVs noted in (a). The value of the investor’s claim would have to be calculated from a MonteCarlo simulation of certainty-equivalent cash flows. The simulation is necessary because the test for IRR > RCAP inany given future period is path dependent. The residual claim that gets all cash flows after the institutionalinvestor is out of the game is also a path-dependent option.

Mission Statement: The mission of the Journal of Applied Finance is to disseminate information and to foster debate onthe practice and pedagogy of finance. The journal is devoted to the publication of original manuscripts that are accessible toa broad audience, including practitioners, academics as teachers, and students. Of particular interest are manuscripts with anapplied orientation falling into one of the following broad categories: traditional research articles (empirical, practical, survey,and synthesis), clinical studies (characterizations of real world situations using unique sources of data), and education (well-motivated, scientifically sound manuscripts that represent major contributions to the field of financial education).

Submission Procedure: Manuscripts should be submitted electronically as a PDF file at www.fma.org-Journal of AppliedFinance. Each paper must be accompanied by a submission fee for manuscript evaluation: $100 for FMA members, $200 fornon-FMA members, and $130 for doctoral students who are not FMA members. If paying by check, please make checkspayable to Financial Management Association. If you prefer to pay by credit card, please include in your submission emailthe following information: type of credit card, cardholder's name, credit card number, and expiration date. The non-membersubmission fees include a one-year membership in FMA for the submitting author.

Manuscripts are evaluated anonymously. Names of authors should not appear on the article itself. Attach a separate coverpage that includes the title, authors, and title and affiliation of each author to one copy of the article. Double space the textwith ample margins.

Authors should ensure that their identity does not appear under the “Properties” of the file they are submitting.

The cover page of the manuscript should show only the title of the paper and a short, one-paragraph (approximately 100word) abstract of the articlethat provides a brief overview of the paper.

Avoid tedious mathematical expressions. When algebraic terms do appear in the text, accompany them with a clearexplanation. Each equation should be numbered consecutively, with the number in parentheses and flush with the rightmargin. Place derivations and proofs in an appendix. When submitting a paper for review, please provide supplementalsheets showing all steps in algebraic derivations so that the reviewers do not have to re-create them.

Tables and figures should appear on separate pages labeled in numerical order and grouped at the end of the text. Label tablesat the top and follow the heading with a description of the table in sufficient detail so that it is capable of standing alone.Label figures at the bottom. Include marginal notation in the article for the approximate placement of all tables and figures.

Minimize extensive content footnotes. When preparing accepted papers, place all footnotes, double-spaced, at the end ofthe manuscript.

Place references in an unnumbered, alphabetical list at the end of the manuscript. Provide all relevant publication informationavailable (i.e., season/month, year, city and state, author(s) full names, etc.). Examples of references are provided below:

References

Baldwin, Carliss Y., 1991, “The Impact of Asset Stripping on the Cost of Deposit Insurance,” Harvard Business SchoolWorking Paper 92-053 (December).

Commerce Clearing House, 1993, 1994 U.S. Master Tax Guide, Chicago, IL.

Weston, J. Fred, 1994, “A (Relatively) Brief History of Finance Ideas,” Financial Practice and Education 4 (No. 1), 7-26.

Myers, Stuart C., 1993, “Finance Theory and Financial Strategy,” in D.H. Chew Jr., Ed., The New Corporate Finance,New York, NY, McGraw-Hill, 90-97.

Smith, Clifford W. Jr. and Charles W. Smithson, 1990, The Handbook of Financial Engineering, New York, NY, Harper Business.

Cite references in the text by citing the author(s) name(s) and then the year of publication in parentheses.

Authors of accepted papers must supply an MS Word copy of their article. Any questions about disk preparation shouldbe directed to the Managing Editor, Financial Management Association International, University of South Florida, Collegeof Business Administration, Tampa, FL 33620-5500, TEL 813-974-2084, FAX 813-974-3318,E-mail:[email protected]

JOURNAL OF APPLIED FINANCESTYLE NOTES FOR PROSPECTIVE AUTHORS

Documents

Behavioral Finance JAFFW2008