Embed Size (px)
Citation preview
Financial SystemReviewJune 2008
Revue dusystème f inancierJuin 2008
www.bankofcanada.cawww.banqueducanada.ca
Finan
cial System R
eview—
June
Revu
e du systèm
e f inan
cier — Ju
in2008
2008
The Financial System Review and Financial Stability
The f inancial system makes an important contribution to the welfare of all Canadians. The ability of households and f irms to conf idently hold and transfer f inancial assets is one of the fundamental building blocks of the Canadian economy. As part of its commitment to promoting the economic and f inancial welfare of Canada, the Bank of Canada actively fosters a safe and eff icient f inancial system. The Bank’s contribution complements theefforts of other federal and provincial agencies, each of which brings unique expertise to this challenging area in the context of its own institutional responsibilities.
The f inancial system is large and increasingly complex. It includes f inancial institutions (e.g., banks, insurance companies, and securities dealers); f inancial markets in whichf inancial assets are priced and traded; and the clearing and settlement systems that underpin the fl ow of assets between f irms and individuals. Past episodes around the world have shown that serious disruptions to one or more of these three components (whether they originate from domestic or international sources) can create substantial problems for the entire f inancial system and, ultimately, for the economy as a whole. As well, ineff iciencies in the f inancial system may lead to signif icant economic costs over time and contribute to a system that is less able to successfully cope with periods of f inancial stress. It is therefore important that Canada’s public and private sector entities foster a f inancial system with solid underpinnings, thereby promoting its smooth and eff icient functioning.
The Financial System Review (FSR) is one avenue through which the Bank of Canadaseeks to contribute to the longer-term robustness of the Canadian f inancial system. It brings together the Bank’s ongoing work in monitoring developments in the system andanalyzing policy directions in the f inancial sector, as well as research designed to increase our knowledge. The strong linkages among the various components of the f inancial system are emphasized by taking a broad, systemwide perspective that includes markets, institutions, and clearing and settlement systems. It is in this context that the FSR aims to
• improve the understanding of current developments and trends in the Canadian and international f inancial systems and of the factors affecting them;
• summarize recent work by Bank of Canada staff on specif ic f inancial sector policies and on aspects of the f inancial system’s structure and functioning;
• promote informed public discussion on all aspects of the f inancial system, together with increased interaction on these issues between public and private sector entities.
The FSR contributes to a safe and eff icient f inancial system by highlighting relevantinformation that improves awareness and encourages discussion of issues concerningthe f inancial system. The Bank of Canada welcomes comments on the materialcontained in the FSR.
Bank of Canada234 Wellington Street
Ottawa, Ontario K1A 0G9
ISSN 1705-1304Printed in Canada on recycled paper La Revue du système f inancier et la stabilité f inancière
Le système f inancier contribue grandement au bien-être économique de tous les Cana-diens. La capacité des ménages et des entreprises de détenir et de transférer en touteconf iance des actifs f inanciers constitue en effet l’un des fondements de l’économiecanadienne. Conformément à l’engagement qu’elle a pris de favoriser la prospérité économique et f inancière du pays, la Banque du Canada s’attache à promouvoir active-ment la f iabilité et l’eff icience du système f inancier. Le rôle de la Banque dans cetimportant domaine vient compléter celui d’autres organismes fédéraux et provinciaux.
Le système f inancier est vaste et de plus en plus complexe. Il se compose des institutionsf inancières (p. ex., banques, compagnies d’assurance, maisons de courtage), des marchés f inanciers, sur lesquels les prix sont f ixés et les actifs sont négociés, et des systèmes de compensation et de règlement, qui permettent les échanges d’actifs entre les entreprises et les particuliers. L’expérience vécue de par le monde a montré que toute perturbation majeure d’au moins un de ces trois éléments (qu’elle trouve son origine au pays même ou à l’étranger) peut avoir de graves répercussions sur le système f inancier tout entier et, enf in de compte, sur l’ensemble de l’économie. En outre, des dysfonctionnements du système f inancier lui-même peuvent entraîner à la longue des coûts économiques substantiels et rendre ce système moins apte à résister aux périodes de diff icultés f inancières. Il est donc primordial que les organismes des secteurs public et privé du Canada s’emploient à étayer solidement le système f inancier af in d’en assurer l’eff icience et le bon fonctionnement.
La Revue du système f inancier est l’un des instruments par lesquels la Banque du Canada cherche à favoriser la solidité à long terme du système f inancier canadien. Ce document rassemble les travaux que la Banque effectue régulièrement pour suivre l’évolution de ce système et analyser les orientations politiques dans le secteur f inancier, ainsi que desrecherches visant à approfondir nos connaissances sur ce sujet. Les liens étroits qui unissent les diverses composantes de ce système sont mis en évidence par l’adoption d’uneperspective large, qui englobe les marchés, les institutions f inancières et les systèmes de compensation et de règlement. Dans cette optique, le but de la Revue est de :
• permettre de mieux comprendre la situation et les tendances actuelles des systèmes f inanciers canadien et international, ainsi que les facteurs qui infl uent sur ceux-ci;
• résumer les travaux de recherche récents effectués par des spécialistes de la Banque sur certaines politiques touchant le secteur f inancier et sur certains aspects de la structure et du fonctionnement du système f inancier;
• promouvoir un débat public éclairé sur tous les aspects du système f inancier et renforcer le dialogue entre les organismes publics et privés dans ce domaine.
La Revue du système f inancier contribue à la f iabilité et à l’eff icience du système f inancier, en s’attachant à mieux faire connaître les enjeux et à encourager les discussions. La Banque du Canada invite ses lecteurs à lui faire part de leurs commentaires au sujet de cette publication.
Banque du Canada234, rue Wellington
Ottawa (Ontario) K1A 0G9
ISSN 1705-1290Imprimé au Canada sur papier recyclé
FinancialSystem ReviewJune 2008
Members of the Editorial Committee
Pierre Duguay and David Longworth, Chairs
Lloyd BartonAllan Crawford
Paul FentonDonna HowardLouise HylandPaul Masson
Graydon PaulinGeorge Pickering
Jack SelodyRobert Turnbull
Mark Zelmer
Jill MoxleyMadeleine Renaud
Lea-Anne Solomonian(Editors)
The significant contribution of the working group mandated with the preparation andorganization of the Review is gratefully acknowledged.
Bank of CanadaJune 2008
The Bank of Canada’s Financial System Review is published semi-annually. Copies maybe obtained free of charge by contacting
Publications Distribution, Communications Department, Bank of Canada, Ottawa,Ontario, Canada K1A 0G9Telephone: 1 877 782-8248; email: [email protected]
Please forward any comments on the Financial System Review to
Public Information, Communications Department, Bank of Canada, Ottawa,Ontario, Canada K1A 0G9Telephone: 613 782-8111, 1 800 303-1282; email: [email protected]
Website: <http://www.bankofcanada.ca>
Contents
Developments and Trends ....................................................................... 1
Financial System Risk Assessment ...................................................................... 3Overview............................................................................................................. 3
Canadian financial situation .......................................................................... 5Risks ............................................................................................................ 5Mitigating risks to the financial system ............................................................ 6
The Financial System ........................................................................................... 8Financial markets .......................................................................................... 8Highlighted Issue: Decomposing Canadian corporate investment-
grade spreads: What are the drivers of the current widening? ........................ 15Highlighted Issue: Broad recommendations for reform in light of
the recent market turmoil ......................................................................... 17Financial institutions ..................................................................................... 20Highlighted Issue: The impact of the recent market turbulence on
credit growth in Canada............................................................................ 24The Macrofinancial Environment .......................................................................... 26
The international environment........................................................................ 26Canadian developments.................................................................................. 28
Important Financial System Developments.......................................................... 33Highlighted Issue: An introduction to covered bond issuance ............................. 33
Reports ................................................................................................. 39
Introduction .................................................................................................... 41
Bank of Canada Oversight Activities during 2007 underthe Payment Clearing and Settlement Act ...................................................... 43
Bank of Canada Participation in the 2007 FSAP Macro Stress-Testing Exercise ..... 51
The Role of Credit Ratings in Managing Credit Risk in Federal Treasury Activities . 61
Policy and Infrastructure Developments ................................................... 67
Introduction .................................................................................................... 69
Financial Market Turmoil and Central Bank Intervention ................................... 71
Research Summaries .............................................................................. 79
Introduction .................................................................................................... 81
A Model of Tiered Settlement Networks.............................................................. 83
The Effects of a Disruption in CDSX Settlement on Activity in the LVTS:A Simulation Study...................................................................................... 87
Family Values: Ownership Structure, Performance, andCapital Structure of Canadian Firms ............................................................. 91
The following people contributed to theDevelopments and Trends section:
Meyer AaronJim ArmstrongWilliam BarkerLloyd BartonDenise Côté
Jim DayAlejandro Garcia
Chris GrahamToni GravelleDylan Hogg
Guy MacKenzieKaren McGuinness
Chris ReidMicheline Roy
Patrick SabourinEric Santor
Gerald StuberVirginie Traclet
Elizabeth Woodman
Developments
and
Trends
Notes
The material in this document is based on information available to 22 May 2008 unlessotherwise indicated.
The phrase “major banks” in Canada refers to the six largest Canadian commercialbanks by asset size: the Bank of Montreal, CIBC, National Bank, RBC Financial Group,Scotiabank, and TD Bank Financial Group.
Assessing Risks to the Stability of theCanadian Financial System
The Financial System Review (FSR) is one vehicle that the Bank of Canada uses tocontribute to the strength of the Canadian financial system. The Developmentsand Trends section of the Review aims to provide analysis and discussion ofcurrent developments and trends in the Canadian financial sector.
The first part of this section presents an assessment of the risks, originating from bothinternational and domestic sources, that could affect the stability of the Canadianfinancial system. Key risk factors and vulnerabilities are discussed in terms ofpotential implications for the system’s overall soundness. The second part of theDevelopments and Trends section examines structural developments affecting theCanadian financial system and its safety and efficiency; for example, developmentsin legislation, regulation, or practices affecting the financial system.
The current infrastructure, which includes financial legislation, the legal system,financial practices, the framework of regulation and supervision, and the macro-economic policy framework, significantly influences the way in which shocks aretransmitted in the financial system and in the macroeconomy, and thus affectsour assessment of risks.
Our risk assessment is focused on the vulnerabilities of the overall financial sys-tem, and not on those of individual institutions, firms, or households. We there-fore concentrate on risk factors and vulnerabilities that could have systemicrepercussions—those that may lead to substantial problems for the entire finan-cial system and, ultimately, for the economy. In examining these risk factors andvulnerabilities, we consider both the likelihood that they will occur and theirpotential impact.
Particular attention is paid to the deposit-taking institutions sector because of itskey role in facilitating financial transactions, including payments, and its interac-tion with so many other participants in the financial system. For instance, theseinstitutions assume credit risks with respect to borrowers such as households andnon-financial firms. Thus, from time to time, we assess the potential impact thatchanges to the macrofinancial environment may have on the ability of householdsand non-financial firms to service their debts.
Risk factors and vulnerabilities related to market risks are also examined. We as-sess the potential for developments in financial markets to seriously affect the fi-nancial position of various sectors of the economy and, ultimately, to disrupt thestability of the Canadian financial system.
Financial System Review – June 2008
Financial System Risk Assessment
3
his section of the Review presentsan assessment of the risks arisingfrom both international anddomestic sources bearing on the
stability of the Canadian financial system.The objective is to highlight key risk fac-tors and vulnerabilities in the financialsystem and to discuss any potential impli-cations for the system’s overall soundness.
T
Key Points• Although there has been some improve-
ment in conditions over the past severalweeks, strains in global credit markets havebroadened since December.
• The process of financial system deleveragingappears to have led to wider spreads onfixed-income assets than would be justifiedby underlying credit risk.
• Enhanced disclosure of potential losses byseveral large global banks and efforts tostrengthen their balance sheets are begin-ning to reduce counterparty concerns.
• Tensions in Canadian credit markets havebeen somewhat less severe than those in theUnited States.
• The strong balance sheet positions of theCanadian financial, non-financial, andhousehold sectors have helped them toweather the turbulence.
• Weaknesses in the global financial systemare now better understood, and improve-ments are being developed.
• The key risk to the financial system is thatthe downturn in the U.S. economy may bedeeper than currently anticipated, increasinglosses and forcing additional deleveraging.
OverviewSince the publication of the December FinancialSystem Review (FSR), strains in credit marketshave broadened and deepened. Estimates ofpotential financial sector losses from the deteri-oration in credit markets have increased, anduncertainty about the distribution of thoselosses has fuelled counterparty concerns. Thisexacerbated tensions in money markets andincreased funding liquidity risks for financialinstitutions. Constraints on bank balance sheetsand heightened risk aversion have been the cat-alysts for a deleveraging process that has affectedeven some of the more traditional segments offinancial markets. The deterioration in financialconditions has also started to filter through tothe real economy. The underlying dynamics ofthe financial market turmoil are now better un-derstood, however, and work is under way inboth the public and private sectors to addressthe structural weaknesses that have been ex-posed in the global financial system. More re-cently, there have also been some tentativesigns that investors’ appetite for risk may bestarting to recover: yield spreads on fixed-income assets have retreated from their highs,demand has started to return to some of theworst-affected areas of the credit markets, andpressures in interbank funding markets haveeased somewhat.
The U.S. economy remains at the epicentre ofthe turmoil. The Federal Reserve has respondedto the recent deterioration in the macroeco-nomic outlook with aggressive interest rate cuts.Several other major central banks have also re-duced their key policy rates as their economieshave slowed.
Central banks have also employed less-con-ventional policy measures to address the supplyand distribution of liquidity in the financialsystem. When tensions in the interbank lendingmarkets intensified in December, both in Canadaand abroad, the Bank of Canada joined with
Developments and Trends
other major central banks in a coordinated ini-tiative aimed at improving market functioningthrough the provision of additional liquidity atmaturities longer than overnight. These actionsalleviated funding pressures at year-end, butspreads in term money markets remained abovetheir pre-August levels. Interbank markets havebeen particularly volatile in the United Statesand Europe this year, reflecting the acuteness ofcounterparty concerns in these regions. Re-newed pressures from financial market dislo-cations led to a further round of coordinatedmarket intervention by the world’s major centralbanks in March. Soon after these policy an-nouncements, the Federal Reserve extendedemergency funding to Bear Stearns, a large U.S.investment bank, facilitating its acquisition byanother firm. Against this background, globalcounterparty concerns have remained high, andmoney market stresses were only partially easedby policy actions.
The balance sheets of the major global bankshave come under increasing strain from thecombined impact of writedowns, the down-grading of monoline insurers, the diversion ofcapital to rescue off-balance-sheet vehicles, andthe reintermediation of loans into the bankingsystem. Being increasingly short of balancesheet capacity, global banks have reduced theircredit market exposures while also raising mar-gin requirements and cutting back on lines ofcredit to hedge funds and other geared invest-ment vehicles. The leverage that has supportedcredit market prices in recent years thus startedto unwind, as the forced sale of assets into an illiq-uid market by investors seeking to reduce riskresulted in price declines that prompted furtherrounds of margin calls and fire sales of assets.This process has been complicated by valua-tion difficulties stemming from the opaque na-ture of many structured products and the lack ofactive trading in secondary markets.
Against a background of weakness in the U.S.economy and deleveraging by financial institu-tions, market contagion cascaded from assetclasses with weaker fundamentals to those withstronger fundamentals. What started as a prob-lem with assets related to U.S. subprime mort-gages undermined the strength of the globalbanking sector, thus spilling over into thebroader financial markets and resulting in li-quidity-driven price declines that pushed up theyield spreads on a wide range of credit assets.Many valuation models based on underlying
4
economic fundamentals broke down as themarket continued to weaken; stabilization willrequire reaching a price that will rekindle in-vestor demand for these securities. The widen-ing of spreads contributed to a virtual shutdownof new issuance in several key areas of creditmarkets, including residential mortgage-backedsecurities (RMBS), commercial mortgage-backedsecurities (CMBS), collateralized loan obligations(CLOs), and high-yield bonds. Equity marketshave also been affected, as investors have begunto mark down their earnings expectations.
These broad trends have characterized the assetmarkets of industrialized nations around theglobe, including those in Canada. For example,spreads in Canadian corporate debt marketsrose to near all-time highs, despite the relative-ly favourable outlook for the domestic econo-my. As discussed in the Highlighted Issue onp. 15, these wide spreads are more likely areflection of the illiquidity of corporate debtmarkets than a measure of perceived underlyingcredit risk. Access to primary markets also re-mains limited in Canada; financial firms arestill raising funds in the primary debt markets,but often at larger concessions relative to sec-ondary-market pricing. While most markets forasset-backed securities have also shut down,mortgage securitization is still possible underthe government-insured National Housing ActMortgage-Backed Securities (NHA MBS) Pro-gram, and these securities can be sold throughthe Canada Mortgage Bonds (CMB) Program,albeit at a wide spread over Government ofCanada bonds.
More recently, there have been tentative signsthat the turbulence in global financial marketsmay be starting to subside. Several large globalbanks have enhanced the disclosure of potentiallosses and have succeeded in raising capital tostrengthen their balance sheets; this appears tobe helping to lower counterparty concerns, asreflected in the yield spreads on the banks’ creditdefault swaps (CDS), which have fallen backfrom their peaks. Demand for corporate bondshas also picked up moderately, and activity hasstarted to resume in some of the worst-affectedareas of the debt markets, such as the market forleveraged loans. Still, as detailed in the April2008 Monetary Policy Report, the Bank currentlyexpects global credit spreads to remain elevateduntil the end of 2008 and to recede only gradu-ally over 2009. Furthermore, spreads are not ex-pected to return to the unusually low levels thatprevailed before August 2007.
Financial System Review – June 2008
Canadian financial situation
Tensions in Canadian credit markets have beensomewhat less severe than those in the UnitedStates. The Canadian economy and financialsystem appear to be well placed to absorb the ef-fects of the recent turbulence.
On balance, Canadian banks remain healthy.Although earnings of the major Canadian bankshave been adversely affected by capital marketwritedowns, these writedowns have been rela-tively moderate compared with those of thelarge Wall Street banks. The Canadian bankingsystem remains well capitalized: core earningshave, so far, remained relatively intact, andbanks have been able to raise additional capital.Thus far, the degree of reintermediation of lend-ing into the banking system has not created sig-nificant difficulties for Canadian banks, whichhave generally been able to access sufficientwholesale and retail financing (albeit at higherspreads) to support their own lending activities.Persistence of current unfavourable financialconditions would, however, dampen loangrowth going forward. Bank profits will alsolikely come under further pressure. Revenuesfrom investment banking activities will fall assecuritization markets, a source of fee income,remain closed. Loan-loss provisions are begin-ning to rise from a low base, but the strong fi-nancial positions of the Canadian householdand non-financial corporate sectors should lim-it the deterioration in the credit quality of loans.Although the major Canadian banks have somesignificant exposures to various markets in theUnited States, these appear to be manageable(see Box 3).
The non-financial corporate sector remains ingood financial shape on the whole, reflectingstrong profit growth, solid margins, and largeholdings of liquid assets. This resilience in theface of a U.S. slowdown has been supported bystrong global demand for commodities. Non-energy sectors with high exposure to interna-tional trade, however, have generally seen prof-its decline over the past year, reflecting thefallout from the U.S. slowdown and the strongerCanadian dollar. With limited access to capitalmarkets, firms are increasingly drawing on theirexisting credit facilities with banks. Overall,growth in business credit has remained aboveits long-term average, despite some tightening
of bank lending conditions.1 This tightening islikely to translate into some modest slowing ofbusiness credit growth. Strong balance sheetsshould still allow firms room to continue toexpand their capital spending, but weakerdemand conditions and increased materialand fuel costs are likely to prove challengingfor some sectors, particularly forest productsand non-commodity, export-oriented industries.
The overall financial situation of the householdsector also appears sound, as evidenced by ag-gregate indicators such as mortgage loan arrearsand personal bankruptcies, which remain at lowlevels. With the cost of borrowing for house-holds having declined in spite of rising spreads,and with little evidence of any tightening in theterms and conditions of household credit, bor-rowing has continued to grow at a strong pace.This has contributed to an increase in the debt-service ratio, which nonetheless remains relativelylow by historical standards. Some modest deteri-oration in household balance sheets may beexpected, owing to declines in financial-assetprices and slower economic growth. Householdcredit growth should also decelerate. The paceof increase in house prices is likely to continueto moderate. While a widespread decline inhouse prices does not currently appear likelyin Canada, a weaker housing market representsthe main risk to household net worth.
Risks
Global credit markets remain very fragile, al-though there has been some general improve-ment in market conditions over the past monthor so. Investors remain wary of re-entering illiq-uid markets where further downside potentialexists, even though market participants largelyagree that credit assets have already undershotany conventional measure of fair value. Forcedselling by leveraged investors may again pushasset prices lower. Against this background, thesituation remains highly uncertain.
Further shocks could expose a number of vulner-abilities in the financial system. Such a shockcould emanate either from within the financial
1. See the Bank of Canada’s Business Outlook Survey, Spring2008, for indicators of credit conditions for businesses.The Highlighted Issue on p. 24 discusses broader trendsin credit growth within the economy.
5
Developments and Trends
system itself, such as the collapse of a hedgefund resulting in the sudden forced liquidationof a large portfolio of assets, or from a muchdeeper and more protracted downturn in theU.S. economy.
The key risk would appear to be the economicsituation in the United States. If the U.S. eco-nomic recovery proved slow to gain tractionamid low levels of business and consumer con-fidence, the “adverse feedback loop” betweenthe real economy and financial markets couldintensify. This would likely result in tighterlending criteria, leading to a more widespreaddeterioration in the underlying credit quality ofconsumer and corporate loan portfolios. This,in turn, could trigger additional liquidity prob-lems in financial markets. Associated losseswould further erode bank capital, leading torenewed concerns over counterparty risks andhigher funding costs in interbank markets. Fur-ther rounds of forced deleveraging by financialinstitutions and by leveraged investors coulddeepen the liquidity and credit crunch. Equitymarkets could also experience sharp falls in re-sponse to the deteriorating economic outlook,and the U.S. dollar could experience sharpdownward pressure.
A deeper-than-expected downturn in the U.S.economy would be transmitted to the rest ofthe world through trade channels, as well asthrough financial markets, particularly via thenegative impact on credit conditions. Emerging-market economies would be vulnerable to asudden reversal of capital inflows, with poten-tial adverse consequences for foreign investors.Commodity prices would also be vulnerable toa softening of global demand.
Such an outcome (weaker U.S. and global de-mand and a decline in commodity prices)would depress the profits of Canadian exporters,with knock-on effects for the broader economy.Canadian banks would likely experience furthersignificant writedowns, the profits of their corebusiness lines could contract, and loan-lossprovisions could rise sharply. This could havea significant negative impact on the capitalratios of the major Canadian banks.
Balance sheet constraints and higher fundingcosts for banks could translate into a reductionin aggregate lending to the economy. Creditconditions for businesses and consumers wouldtighten significantly, and default losses on
6
corporate and household lending would rise.Consumer wealth would also decrease as the val-ue of their financial investments declined andhouse prices weakened in some areas; housingmarkets in Western Canada would be particularlyvulnerable to a sharp fall in commodity prices.
While the probability of such outcomes materi-alizing is relatively low, they nonethelesswarrant careful consideration by financial insti-tutions because of the potentially large negativerepercussions. On the whole, major Canadianfinancial institutions appear to have the mar-gins for coping with such outcomes, and theCanadian financial system’s ability to weatherthese potential adverse developments remainssound.
Mitigating risks to the financialsystem
Recent events highlight the need for decisive ac-tion by both policy-makers and market partici-pants to address the shortcomings that havebeen exposed in the global financial system. Inthe public sector, the Financial Stability Forum(FSF) is playing an important international co-ordinating role; its report outlining wide-rang-ing recommendations for strengthening theglobal financial system was endorsed by the G-7finance ministers and central bank governorson 11 April.2
Immediate actions to ease the strains in finan-cial markets need to focus on restoring the con-fidence of investors. Central bank marketinterventions aimed at addressing elevated pres-sures in the short-term funding markets havehelped in this regard. The Federal Reserve’s ag-gressive reduction of its key policy interest ratehas also reduced the risk of a sharp downturn inthe U.S. economy. Many central banks havebeen working to strengthen their ability to re-spond to financial risks. In this respect, the Fed-eral Reserve has introduced new credit facilities,and the Government of Canada has proposed
2. The full report is available at <http://www.fsforum.org/publications/FSF_Report_to_G7_11_April.pdf>and the accompanying statement of the G-7 financeministers and central bank governors at <http://www.fin.gc.ca/activty/g7/g7110408e.html>; see theHighlighted Issue on p. 17 for a summary of therecommendations.
Financial System Review – June 2008
amendments to the Bank of Canada Act to mod-ernize the Bank’s powers, to allow it to supportthe stability of the financial system.3
The FSF report also identifies priority actions tobe undertaken in the private sector to facilitatethe adjustment of financial markets. It calls onfinancial institutions to enhance the disclosureof their risk exposures, writedowns, and fair-value estimates for complex and illiquid instru-ments; to strengthen risk-management practices;and to recapitalize bank balance sheets. It alsocalls for greater transparency regarding the com-position of structured products to improve thepricing of risk.
Policy-makers are also considering potential ac-tions aimed at enhancing the resilience of mar-kets and financial institutions over the mediumterm. The FSF report highlights several key areasfor reform, including strengthened supervisoryoversight of banks’ capital, liquidity, and risk-management practices, as well as measuresaimed at addressing shortfalls in the credit-rat-ing process for structured products.4 It indicatesthat, going forward, the FSF also intends to ex-amine more closely the forces that contribute toprocyclicality in the financial system and possi-ble options for mitigating them. The reportnotes that it is important that any policy re-sponse avoid exacerbating financial stress in theshort term. Overall, the report provides a usefultemplate for reform, but national policy-makersmust still be careful to ensure that any changesto the regulatory framework governing financialmarkets will promote an appropriate balancebetween the stability of the financial system andits efficiency.
3. See “Financial Market Turmoil and Central Bank Inter-vention” on p. 71 for a discussion of the economicrationale for these powers.
4. For a discussion of proposals for enhancing the credit-rating process, see “Reforming the Credit-Rating Pro-cess,” Financial System Review, December 2007.
7
Developments and Trends
Chart 1 iTraxx and CDX Spreads(Credit Default Swap Index Spreads)
Basis points
Note: iTraxx indexes composed of entities domiciled inEurope; CDX indexes composed of entities domiciledin North America
Source: Bloomberg
0
100
200
300
400
500
600
700
0
100
200
300
400
500
600
700
2007 2008
CDX investment-gradeCDX X-OveriTraxx investment-gradeiTraxx X-Over
The Financial System
Financial markets
Since the publication of the December FSR, con-ditions in global credit markets have deteriorat-ed, with strains spreading further beyondmortgage-related debt to the broader credit mar-kets. The broadening of the turbulence in finan-cial markets has developed against a backdropof mounting concerns over liquidity constraintsand systemic risk at financial institutions, and adeterioration in global (especially U.S.) economicprospects. Pressures on bank balance sheets re-lated to subprime-related writedowns, higherfunding costs, and the recapitalization of mono-line insurers (Box 1) have contributed to a rapiddeleveraging across financial market partici-pants. This process has become self-reinforcing,as fire sales of credit instruments by leveragedinvestors into illiquid markets (to meet highermargin calls, for example) have put further pres-sure on asset prices, leading to further rounds ofdeleveraging. Contagion has thus extendedthroughout the credit markets, resulting in a sig-nificant widening of spreads on a broad rangeof credit assets and a virtual shutdown of newissuance in several major segments of the globalfixed-income market (see Box 2 for details on theimportance of the affected markets in Canada).
Recently, however, there has been some encour-aging evidence that credit market tensions maybe starting to abate, as CDS spreads, includingthose for the major global banks, have startedto narrow (Chart 1). Demand for corporatebonds is also returning, and activity has re-sumed in some of the worst-affected areas of thecredit markets.
While volatility in equity and foreign exchangemarkets has been high compared with recenthistory, these markets have not experienced thedislocations occurring in the credit universe, andthey continue to function in an orderly manner.
With the process of deleveraging and recapital-ization across global financial intermediariescontinuing, some time will be required beforeregular liquidity conditions are restored to affectedmarkets and normal functioning resumes.
Money marketsThe elevation of spreads in funding markets forfinancial institutions reflects a sharp increase inthe size of their balance sheets and a correspond-ing increase in their demand for funds. This
8
Financial System Review – June 2008
Box 1
Why Monolines Matter
Traditionally, monoline insurance companies(monolines) have provided the issuers of U.S. mu-nicipal bonds with credit-default insurance byguaranteeing the timely payment of the interestand principal, thus allowing these issuers to obtaina AAA rating, which they otherwise would not havereceived.1 This innovation increased the market-ability of municipal debt and, thus, reduced the in-terest rates that municipal issuers had to offer. By2007, the six largest monolines insured approxi-mately half of the US$2.6 trillion U.S. municipaldebt market. Over the past decade, however, asstructured debt markets became increasingly im-portant, monolines began to change their businessmodel. They started to guarantee the timely repay-ment of the principal and interest on the underly-ing assets of structured debt instruments such asCDOs. For this service, they received insurance pre-miums, either from issuers of structured productsor from institutional investors who held specificproducts and wanted to hedge their exposure tocredit risk.2 Monolines devoted a growing share oftheir overall business to this area, and by late 2007,approximately 40 per cent of the US$3.3 trillion intotal outstanding debt insured by the monoline in-dustry was based on structured credit products.Insuring structured credit instruments impliedhigher levels of risk. Although monolines used un-derwriting criteria that were conservative by histor-ical standards, the very high level of leverage theytypically employed3 left very little margin for error.Given the significant losses in the market for struc-tured products, many financial market participantsbegan to question whether the monolines hadadequate capital to honour their obligations and,correspondingly, whether these insurers stillwarranted the credit ratings that allowed them tooffer AAA guarantees on insured products.The removal of the AAA credit rating from mono-lines would affect two groups of market partici-pants in particular: the U.S. municipal issuers andlarge global investment banks. Without the mono-lines’ guarantee, most U. S. municipal issuerswould likely face significantly higher costs of debt
1. “Monoline” insurance companies are so-calledbecause they focus on a single product line: insuringfinancial debt. Federally regulated Canadian insur-ance companies are prohibited from offering suchinsurance, and so monoline companies are non-Canadian firms.
2. Typically, monolines did this by writing creditdefault swaps with bank counterparties.
3. Their ratios of assets to equity ranged from the low90s to well over 200 at the end of 2006.
financing. The failure of the municipal “auction-rate securities” market in early 2008 clearly showedthe effects of the loss of investor confidence inmonolines: during this episode, even well-knownissuers of high-quality municipal debt saw theirfinancing costs briefly spike to interest rates over20 per cent.4
Large global investment banks, the primary coun-terparties to the monolines’ structured products,are exposed to monoline-related losses in two ar-eas. First, if monolines lose their AAA status, themarked-to-market value of their insurance is di-minished, and the underlying insured asset wouldhave to be marked down accordingly. Second, themarket value of any credit default swaps sold to thebanks by monolines would also decline. The in-ability of monolines to honour these insurancecontracts would cause any hedged credit exposuresto revert to unhedged status, against which thebanks would have to allocate capital reserves.It is difficult to estimate the total exposure of invest-ment banks to the downgrades of monoline creditratings. Accordingly, estimates of aggregate invest-ment bank exposure to monoline credit down-grades vary and tend to be scenario-specific. Mostestimates are large, however, with “worst-case” sce-narios of approximately US$125 billion.5 Theselosses—should they occur—would be in additionto all other losses suffered by investment banks ontheir subprime-mortgage and other credit-productinvestments, and would likely further constrain theability of banks to extend credit to the economy.One Canadian bank has had widely publicizedproblems with insurance purchased from ACA Fi-nancial Guaranty. Standard & Poor’s lowered thecredit rating on this monoline to CCC (the lowestjunk rating above default) late in 2007. This causedthe bank to write down the value of the contracts itheld with ACA Guaranty by $2.2 billion. This wasjust a partial writedown; the bank indicated thatfull writeoff would lead to further substantial loss-es. Other Canadian banks have also reported somemonoline exposure.
9
4. Auction-rate securities are a form of floating-rate,long-term municipal debt in which the coupon levelresets at regular dates according to an auction pro-cess. As the monoline guarantee for these securitiescame into question, investor willingness to bid forthem at the auctions declined, causing yields toincrease.
5. S. Glasser, “Monoline Downgrades: Understandingthe Impact,” Barclays Capital (January 2008): 1–19.
Developments and Trends
(a) All data are December 2007 at book value, except equities at market value. Sources: Bank of Canada Banking and Financial Statistics,National Balance Sheet Accounts (Statistics Canada), OSFI, and the TSX.
(b) Issued in Canada by Canadian and foreign corporations(c) Issued in Canada and abroad by Canadian corporations(d) Loans by Canadian financial institutions to Canadian and foreign borrowers for business purposes(e) These trusts were frozen under the Montreal Accord in August 2007.
Box 2
Structure of Canadian Capital Markets
Government
$836 billion
Bonds: $696 billion
Financial corporations
$3,078 billion
Residential mortgage-backedsecurities: $185 billion
Other asset-backedsecurities: $143 billion
• Government-guaranteed:$160 billion
• Non-government-guaranteed:$25 billion
Asset-backed securities
$328 billion—of which,
• Bank-sponsored ABCP:$76 billion
• Non-bank-sponsoredABCP: $32 billion(e)
Short-term paper: $140 billion• GoC t-bills: $116 billion• Other short-term paper:
$24 billion
• of which, CSBs: $13 billion• of which, other federal non-
marketable debt: $1 billion
Short-term paper: $103 billion
Non-financial corporations
$2,046 billion
• Commercial paper: $43 billion(b)
• Bankers’ acceptances: $60 billion
Commercial paper:$12 billion(b)
Bonds: $232 billion(c)Bonds: $220 billion(c)
Equities: $1,486 billionEquities: $607 billion
Loans: $316 billion(d)Loans: $191 billion(d)
Deposits at financial corporations:
$1,957 billion
Figure A Size of Canadian Capital Markets—2007(a)
This box outlines the size and structure of thecapital markets in Canada. In light of the ongoingfinancial market dislocations, this puts intoperspective the importance of the affected marketsin Canada.As in other mature capital markets, corporate equi-ties represent the largest source of corporate financ-ing in Canada, accounting for about one-third ofthe overall securities market.1 The relative impor-tance of corporate securities is somewhat higher inCanada and the United States than in the UnitedKingdom and the euro area, where firms have tradi-tionally relied more heavily on bank loans.2 Thissuggests that Canadian and U.S. corporations maybe somewhat more directly affected by a deteriora-tion in financial market conditions than firms inthe United Kingdom or the euro area. The balancesheet leverage of Canadian firms has improved
10
1. Global comparisons as of 2006. See McKinsey Glo-bal Institute, Mapping Global Capital Markets: FourthAnnual Report (2008) and the Bank of England’sFinancial Stability Report (October 2007). In Can-ada, the relative shares of the major componentschanged very little from 2006 to 2007.
2. For further details, see D. Côté and C. Graham,“Corporate Balance Sheets in Developed Econo-mies: Implications for Investment,” (Working PaperNo. 2007-24, Bank of Canada, 2007).
significantly in recent years relative to those in theUnited States and the United Kingdom, however,which would suggest that Canadian firms are betterpositioned to confront economic and financialshocks.In the United States and Europe, the issuance ofstructured financial products increased exponen-tially between 2000 and July 2007, when the finan-cial crisis began.3 Over the same period, the asset-backed security (ABS) segment in Canada nearly tri-pled in size; ABS represents a slightly smaller shareof the Canadian securities market (5 per cent) thanof global markets (7 per cent). As it has elsewhere,the erosion of investor confidence in structuredproducts has seriously impaired the functioning ofthe Canadian ABS market since July 2007. Mortgagesecuritization in Canada is still possible, however,under the NHA MBS Program, which benefits froman explicit government guarantee. This program ac-counts for close to 90 per cent of total outstandingresidential mortgage-backed securities in Canada,and recent growth has been strong (see HighlightedIssue on p. 24). It should also be noted, however,
3. The IMF 2008 Global Financial Stability Report (April)p. 56, notes that issuance of CDOs, ABS, and MBS inthe United States and Europe grew from US$500 bil-lion in 2000 to US$2.6 trillion in 2007.
Financial System Review – June 2008
Box 2
Structure of CanadianCapital Markets (continued)
increase is the result of the reintermediation ofoff-balance-sheet assets, constraints in accessingsecuritized funding markets, and the increasedreliance of corporations on bank lending (seeHighlighted Issue on p. 24).
The lack of transparency and secondary-marketpricing in many structured products has made itdifficult to gauge the potential loss exposure offinancial institutions with any accuracy. This hasincreased counterparty concerns, which has alsocontributed to significantly higher spreads in inter-bank funding markets (Chart 2) and raisedfunding liquidity risk for financial institutions.
Strains also continue to be evident in the marketfor asset-backed commercial paper (ABCP). How-ever, liquidity conditions have recently improvedfor bank-sponsored programs,5 and borrowingcosts have decreased. Nevertheless, costs remainelevated relative to short-term risk-free rates, andsecondary-market liquidity remains very limited.
Overall, credit spreads in money markets haveincreased significantly since the beginning of theturmoil and remain elevated. Liquidity remainsinversely related to the borrowing term, withlenders reluctant to provide funds for termsgreater than one month without significantcompensation. Given the strong interlinkagesbetween global financial markets, the persistenceof these pressures prompted a coordinated re-sponse from the Bank of Canada and other majorcentral banks to provide markets with additionalliquidity on two separate occasions.6 Althoughmarket liquidity has improved somewhat, withterm funding more readily available, anddespite the improvement in credit spreads sinceyear-end, signs of funding pressures remain.
5. As indicated in the December FSR, the market fornon-bank-sponsored ABCP is no longer active, withroughly $32 billion undergoing a restructuring underthe Montreal Accord, which is targeted to take effectthis month. The last hurdle was crossed when theproposed restructuring was approved by a majority ofcreditors.
6. Coordinated policy initiatives were announced on12 December 2007 and 11 March 2008. A third coor-dinated policy response was announced on 2 May,but the Bank of Canada did not participate. From20 March 2008 to 15 May, the Bank of Canada auc-tioned $2 billion of 28-day term purchase and resaleagreements every two weeks. (Later operations wererollovers of previous operations.) On 29 May 2008,the Bank reduced by $1 billion the outstandingamount of term financing.
that less than a quarter of total outstanding residen-tial mortgages in Canada are securitized, comparedwith almost 60 per cent in the United States.Nonetheless, exposure to U.S subprime mortgagesand CDOs has led to a crisis of investor confidencein Canada’s ABCP market, particularly in non-bank-sponsored ABCP, the segment of Canada’s se-curities market that has been most affected by themarket turmoil. Since last August, non-bank-spon-sored ABCP has been frozen under the terms of theMontreal Accord, as interested parties work toreach an agreement for restructuring the ABCP intolonger-term instruments. ABCP has been an impor-tant source of short-term financing for Canadianfirms over the past several years. Including non-bank-sponsored ABCP, it represents about 33 percent of the Canadian ABS market and nearly 50 percent of privately issued short-term paper.
11
Developments and Trends
Chart 4 Yield Spreads on Canadian Corporate Bonds
Basis points
1989–96: Scotia Capital corporate bond yields minus mid-termGoC yields
1997–2008: Merrill Lynch option-adjusted spreads1989–91: monthly averages; 1992–2008: end-of-month valuesSources: Scotia Capital, Bloomberg, Merrill Lynch
1990 1995 2000 20050
50
100
150
200
250
300
350
400
450
500
0
50
100
150
200
250
300
350
400
450
500
AAABBBInvestment-grade index
Chart 2 Spreads between 3-Month LIBOR andOvernight Index Swaps*
Basis points
200820070
20
40
60
80
100
120
0
20
40
60
80
100
120
United StatesEuro zoneCanada
* U.S. LIBOR, EU EURIBOR, and Canadian CDORSource: Bloomberg
Chart 3 Spreads on 5-Year Mortgage Bonds
Basis points
0
10
20
30
40
50
60
70
80
90
100
0
10
20
30
40
50
60
70
80
90
100
Source: Bloomberg
Fannie MaeCanada MortgageBonds
July Aug. Sept. Oct. Nov. Dec. Jan. Feb. Mar. Apr. May
2007 2008
Given persistent funding pressures in moneymarkets, the U.S. Federal Reserve has introducedadditional measures to address the need forliquidity in term money markets. These includethe introduction of a Primary Dealer Credit Fa-cility,7 a Term Securities Lending Facility,8 andmodifications to the Term Auction Facility9 thatwere introduced in December 2007. Evidence todate suggests that while these measures havebeen useful in containing concerns aboutconstraints on financial institution liquidity,they have not resulted in a material decline infunding pressures. In Canada, increased yieldspreads in term money markets have generallybeen less severe than in the United States andelsewhere (Chart 2). Since the end of April,there has been a general improvement inCanadian money market conditions: bankfunding costs have fallen markedly and are wellbelow those in a number of other currencies.
Mortgage debt marketsWhile the problems in the subprime sector ofthe market for residential mortgage debt arenow well known, concerns have spread to thebroader market for mortgage-backed debt. Forexample, the market for conforming mortgages,specifically those repackaged and sold by U.S.Government Sponsored Entities (GSEs), such asthe Federal Home Loan Mortgage Corporation(Freddie Mac) and the Federal National MortgageAssociation (Fannie Mae), experienced a dispro-portionately sharp rise in yield spreads relativeto the expected modest deterioration in the un-derlying assets (high-quality mortgage loans)(Chart 3). Although partly related to finan-cial concerns about the GSEs themselves,10 thiswidening bears witness to a strong investor
7. The Primary Dealer Credit Facility is an overnightfacility for primary dealers available on an ad hoc basisat the initiative of the borrower. Credit is rationedthrough prices, with loans granted at a rate equal tothat for primary credit at the Federal Reserve Bank ofNew York under the discount window.
8. The Term Securities Lending Facility provides primarydealers with the opportunity to borrow Treasuries fora term of 28 days in exchange for a broader list of secu-rities than the traditional securities-lending program.
9. The Term Auction Facility provides the depositoryinstitutions eligible for primary credit under the dis-count window with access to collateralized loans fora term of 28 days.
10. Both Fannie Mae and Freddie Mac have had to raisecapital on two separate occasions, as well as cut theirdividends, in order to repair their balance sheets inconjunction with reported mortgage-related losses.
12
Financial System Review – June 2008
aversion to any mortgage-related products, as wellas the deleveraging of market participants investedin these high-quality mortgage products.
Strains have also appeared in the U.S. marketfor commercial real estate debt. For instance,the AAA tranche of the CMBX index, a syntheticproxy for the value of commercial mortgage-backed securities, has fallen dramatically.The price declines in this market, while partlyreflecting the overall weakness in the U.S.economy, have also likely been driven bycontagion from the residential mortgage market.
While the Canadian subprime market is notcurrently a source of concern (refer to p. 30),the Canadian mortgage-debt market has beenaffected by recent developments. Specifically,issues of the Canada Mortgage Bond (CMB) Pro-gram, whose underlying assets are composed ofmortgage-backed securities insured under theNHA , have experienced a widening of their yieldspread relative to Government of Canada bonds(Chart 3). It is difficult to gauge the extent towhich this widening reflects investor aversion tomortgage-related products, rather than the sharpincrease in supply: Canadian financial institu-tions have increasingly turned to CMBs as a sourceof wholesale funding since the turmoil began(see Highlighted Issue, on p. 24 for more detail).
As an alternative source of funding, Canadianbanks have recently begun to issue coveredbonds in the European market (HighlightedIssue on p. 33). Covered bonds are secureddebt instruments issued by financial institu-tions and are traditionally seen by investorsas a safe, liquid alternative to governmentbonds. The European covered bond market hasnot been immune to the recent market disrup-tions, however, since many of the assets under-pinning covered bonds are tied to residentialmortgages. Investors have increasingly begunto differentiate among covered bonds on thebasis of issuer, country of origin, legal struc-ture, and quality of underlying assets, despitevery few rating downgrades. Demand for coveredbonds has fallen, particularly for bonds issuedby banks in those countries where housingmarkets have declined, such as Spain.11
11. As such, recent new issuance has been weighted moreheavily towards the private placement market, wheredealers can be more assured of demand conditionsbefore the actual sale of these bonds occurs.
Structured marketsDislocations have been particularly severe in themarket for structured products, such as collater-alized debt obligations (CDOs), where price de-clines have been dramatic. Since December, ascredit spreads continued to widen, many of thetriggers embedded in structured products havebeen activated, resulting in forced selling or re-hedging of credit exposure. This has served toreinforce and extend the repricing.
With secondary markets inactive, the valuationof many structured products remains difficult.When pricing is available, CDO tranches nowtrade at a significant discount relative to corpo-rate debt. This partly reflects the belated acknowl-edgement that the credit ratings of structuredproducts do not accurately reflect the probabilityor severity of downgrades. Overall, participantsin structured debt markets increasingly differen-tiate on the basis of superior transparency andsimplicity of structure. Nevertheless, the formthat structured markets will take going forwardremains unclear.
Corporate debt marketsGlobally, yield spreads on investment-gradecorporate bonds have increased. In Canada andthe United States, for example, spreads reached,or approached, all-time highs (Charts 4 and 5)before easing back in recent weeks.12 While atleast part of the increase in credit spreads lastsummer reflected a healthy correction of exces-sively low spreads, the recent widening has, inmany cases, gone beyond what current funda-mentals might dictate. For example, at its peakin April, the cost of default protection in manymarkets implied default rates significantlygreater than those experienced during the lasttwo recessions.
It appears that the process of financial systemdeleveraging has exaggerated the movement incorporate debt spreads relative to underlyingcredit risk. Indeed, as discussed in the HighlightedIssue on p. 15, the lack of market liquidity ap-pears to be a significant factor behind the wid-ening of spreads on corporate credit in Canada,
12. The pronounced movement in investment-grade creditspreads also reflects the high concentration of financialsector issuers in the index. The credit spreads of thesecompanies have been more affected by rising systemicconcerns.
13
Developments and Trends
Chart 5 Yield Spreads on U.S. Corporate Bonds
Basis points
1990 1995 2000 20050
500
1,000
1,500
2,000
2,500
3,000
0
500
1,000
1,500
2,000
2,500
3,000
Note: 1989–96: Merrill Lynch corporate index yieldsminus Merrill Lynch Treasury index yields1997–2008: Merrill Lynch option-adjusted spreads
Source: Bloomberg, Merrill Lynch
AAABBBBBBCCCInvestment-grade indexHigh-yield index
Chart 6 Yields on Canadian Corporate Bonds and10-Year Government Bond
%
Sources: Bloomberg, Merrill Lynch, Bank of Canada
2007 2008
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
AAABBBInvestment-grade index10-year Government of Canada
Chart 7 Issuance of U.S. and Canadian High-YieldBonds
US$ billions Number
0
10
20
30
40
50
60
70
0
25
50
75
100
125
150
175
2001 2002 2003 2004 2005 2006 2007 2008
Note: The data for 2008Q2 represent activity up to 22 May.Source: Thomson ONE Banker
Number (right scale)Proceeds (left scale)
particularly for debt issued by investment-gradefinancial institutions.
Despite the increase in credit spreads, the over-all impact on the cost of corporate borrowinghas been mitigated by the decreased yields onsovereign bonds, both in Canada and abroad(Chart 6). A significant tiering of credit has alsooccurred, however, with lower-rated borrowersexperiencing both higher costs and reducedaccess to markets.
In Canada, as elsewhere, access to primarymarkets for corporate debt has been a concern.Canadian firms, particularly those in the finan-cial sector, were still actively raising funds in theprimary debt market, but often at larger conces-sions relative to pricing in secondary markets.In recent weeks, however, access to markets hasincreased considerably, banks have fundedmore than they initially expected, and conces-sions to the secondary market have decreased.The strong balance sheet positions of non-finan-cial firms, as well as the economic outlook andvolatile market conditions, have likely causedmany of these firms to delay market issues and/or increase their reliance on bank financing.
While credit spreads in the U.S. market for high-yield bonds have also increased significantlysince July 2007, these spreads, unlike those forinvestment-grade bonds, remain well below thehighs witnessed earlier in the decade related tothe bursting of the technology sector bubble. Is-suance in the high-yield market has been mini-mal, however, since access to this market hasessentially been closed to many issuers (al-though activity has started to pick up again inrecent weeks). The lack of access to primarymarkets by non-investment-grade firms is animportant credit constraint, not only for U.S.firms but also for similarly rated Canadianfirms that rely on the U.S. market for funding(Chart 7).13
13. See S. Anderson, R. Parker, and A. Spence, “Develop-ment of the Canadian Corporate Debt Market: SomeStylized Facts and Issues,” Financial System Review(Ottawa: Bank of Canada, December 2003): 35–41.
14
Financial System Review – June 2008
Chart 8 Decomposition of Credit Spread forCanadian Investment-Grade Bonds
Basis points
Notes: Index composed of Canadian-dollar-denominatedinvestment-grade debt issued in the Canadian domesticbond marketSpreads are adjusted for embedded options (OAS).
Sources: Bloomberg, Merrill Lynch, authors' calculations
2001 2002 2003 2004 2005 2006 20070
50
100
150
200
0
50
100
150
200
SpreadExpected loss from default + credit-risk premiumExpected loss from default
2008
Highlighted Issue
Decomposing Canadian corporateinvestment-grade spreads: Whatare the drivers of the currentwidening?
Prepared by Alejandro Garcia and Toni Gravelle
Yield spreads on Canadian corporate bonds beganto widen last summer as the crisis in the subprime-mortgage market started to take hold (see topline in Chart 8). Spreads for investment-gradedebt are now far wider than one would expectbased on past experience with economic down-turns and given that Canada's economy is in ahealthier position than the U.S. economy. ThisHighlighted Issue examines the extent to whichrecent movements in the credit spread for in-vestment-grade corporate debt have been drivenby factors not related to credit risk, such as adrying up of market liquidity for debt issued bycorporations.
In general, two important components drivevariations in corporate yield spreads. One is theexpected loss from default, the other relates to riskpremiums. This latter component can be furtherdecomposed into two types: a credit-risk premi-um and an illiquidity premium. The expected lossfrom default generally reflects the fundamentalsof the firm, such as the degree of leverage and itsability to generate a stable stream of profits. Thecredit-risk premium is related to the variability of,or uncertainty about, potential loss from default.Both the credit-risk premium and the expectedloss from default are affected by changes in mac-roeconomic activity.14 When combined, thesetwo components comprise the part of the yieldspread attributed to default-related credit risks.
The illiquidity premium, a non-credit-risk factor,relates to a lack of general market liquidity. More-over, the credit-risk and illiquidity premiums, likeother risk premiums, can vary with any changein the risk appetite of investors and are thereforelikely to be positively correlated over time.
To identify the first two components, the expectedloss from default and the credit-risk premium,we use a structural credit-risk model. Thesemodels treat the firm’s equity and debt as
14. Credit-risk premiums rise during economic slowdowns,given the concurrent rise in macroeconomic (and, inturn, firm-level) uncertainty.
15
Developments and Trends
Chart 10 Yield Spreads on Canadian Corporate Bonds
Basis points
Chart 9 Risk Premiums on Canadian Investment-Grade Bonds
Basis points Basis points
Source: Authors’ calculations
2001 2002 2003 2004 2005 2006 2007-30
0
30
60
90
120
150
-10
0
10
20
30
40
50
Illiquidity-risk premium (left scale)Credit-risk premium (right scale)
2001 2002 2003 2004 2005 2006 20070
40
80
120
160
200
240
0
40
80
120
160
200
240
2008
2008
IndustrialsFinancials
Notes: Indexes composed of Canadian-dollar-denominated investment-grade debt issued in the Canadian domestic bond market Spreads are adjusted for embedded options (OAS).
Sources: Bloomberg, Merrill Lynch
contingent claims on the firm’s assets, with debthaving a priority claim.15 In particular, the totalvalue of the firm’s assets is modelled as a time-series variable. Default occurs when asset valuesfall below the value of the firm’s debt. The lowerthe value of assets relative to debt and thegreater the variability of asset values over time,the more likely a firm is to default and, as such,the greater should be the two components relatedto credit: the expected loss from default and thecredit-risk premium.
Using a structural model based on Churm andPanigirtzoglou (2005), we obtain the impliedexpected loss from default and credit-risk premiumfrom the level and variability of the asset valuesof the firms included in the credit-spread index.16
We derive the illiquidity premium from the differ-ence between the observed spread and the sumof the expected loss from default and the credit-riskpremium (i.e., the credit-risk component).17
Decomposition of spreads intorisk factors
The results from our model estimation, usingmonthly data, indicate that since July 2007,investment-grade firms have experienced anincrease in both the credit-risk component (sec-ond from the top line in Chart 8) and the illi-quidity premium (Chart 9).
As of 21 May 2008, while the actual spread was179 basis points, the expected loss, credit-risk pre-mium, and illiquidity premium were 20, 34, and125 basis points, respectively. Comparable figuresfor end-July 2007 were 85, 21, 5, and 59 basispoints, respectively. The increase in the invest-ment-grade credit spread can thus be attributedto an increase in the credit-risk and illiquiditypremiums above their recent historical norms.18
15. See Huang and Huang (2003) for a comprehensivereview of structural credit-risk models.
16. We thank the Bank of England for providing us withthe Matlab code for the model’s implementation.
17. Technically, this residual measures all non-credit-risk factors that can affect the credit spread. However,several empirical studies have documented that non-credit-risk factors tend to proxy the market illiquidityof corporate bonds. See Longstaff, Mithal, and Neis(2004) and Ericsson and Renault (2006).
18. The credit-risk component reached its peak level of89 basis points in March 2008, and the illiquiditypremium reached its peak level of 125 basis points inMay 2008.
16
Financial System Review – June 2008
The rise in the credit-risk premium and thesharper rise in the illiquidity premium of theinvestment-grade index can be partly explainedby the composition of the index. The Canadianindex for yield spreads on investment-gradebonds contains a high proportion of financialfirms (approximately 55 per cent of the index in2007). Globally, financial firms are at the epi-centre of the ongoing credit turmoil, and the risein the credit-risk and illiquidity premiums forthese firms, including those in Canada, is con-sistent with the rise in their funding liquidityrisks, as assets are brought back onto their bal-ance sheets and as access to certain segments ofthe securitized and/or structured markets ismore constrained for these firms than in thepast. This funding risk interacts with, anddrives, the rise in credit-risk factors for banksand the illiquidity premium for their debt in-struments.19 We observe the rise in the latterwhen investors indiscriminately reduce theirdemand for bank debt, securities become lessliquid, and yield spreads widen further thanbank fundamentals would indicate. Recent re-search on the components of corporate spreadsfor the United States and the United Kingdomalso shows that the illiquidity premium for in-vestment-grade firms increased during the lastquarter of 2007 (Webber and Churm 2007).
As illustrated in Chart 9, the rise in the illiquid-ity premium for investment-grade firms ismarkedly higher than recent historical norms.The behaviour of the illiquidity premium isconsistent with the nature of the crisis, becausethe repricing of credit risk is centred in financialinstitutions. Currently, as mentioned, financialsare perceived to be strained and dominate theCanadian investment-grade index. The strain infinancial institutions can be seen in Chart 10,which shows the yield spread for the two com-ponents making up the index: industrials andfinancial institutions. Chart 10 shows how thedifference between the yield spread on industrialsand that on financial institutions narrowed(and at times turned negative) since July 2007,indicating a rise in the perceived riskiness ofCanadian financial firms relative to industrial
19. See IMF (2008) for more on the interaction betweenmarket liquidity risk and the funding liquidity riskfaced by banks during the credit turmoil.
firms. This, in turn, was a reflection of globalconcerns regarding financial systemic risk.20
Caveat
Recent research that expands structural modelsby including them in a broader macroeconomicsetting has shown that credit-risk premiumsmay, in fact, account for a larger portion of theoverall spread than indicated by the “traditional”structural model (Chen 2008). This suggeststhat the results of “traditional” structural mod-els such as that used in this study should be in-terpreted with caution, and should focus on thedirection in which risk factors evolve, ratherthan on the specific values of the relative contri-butions of the factors.
Highlighted Issue
Broad recommendations forreform in light of the recentmarket turmoil
Prepared by Jim Armstrong
Committees of central banks, regulators, andinternational institutions have analyzed the or-igins of the recent turmoil in financial marketsand have proposed measures to strengthen theresilience of the financial system.21 There is abroad convergence of views among these pro-posals. Particular prominence has been given torecently published reports by the Financial Sta-bility Forum (FSF) and the Senior SupervisorsGroup (SSG). This Highlighted Issue reviewsthese reports, focusing on their main observa-tions and recommendations.
20. There is a rating differential between the industrialsand financials components. The mode rating forindustrials is generally lower than the correspondingrating for the financials. For example, in 2007 themode for industrials was A2, and for financials it wasAA3. In normal times, the ratings differentials areaccompanied by a spread differential.
21. For example, the President’s Working Group onFinancial Markets has issued a Policy Statementon Financial Market Developments available at<http://www.treas.gov/press/releases/reports/pwgpolicystatemktturmoil_03122008.pdf>.
17
Developments and Trends
Financial Stability Forum—April 2008In October 2007, the G-7 finance ministers andcentral bank governors asked the FSF to analyzethe causes and weaknesses that produced theturmoil, and to make recommendations for in-creasing the resilience of markets and institu-tions. The FSF submitted its final report on11 April. Its findings and recommendations arethe product of a collaborative effort between themain international bodies and national author-ities in key financial centres.22 The G-7 minis-ters and governors strongly endorsed the report,and committed to implementing its recommen-dations. They asked the FSF to actively monitorthe implementation of their recommendations.
The FSF proposes actions in five broad areas. Whatfollows is a high-level summary of the extensivelist of recommendations (67 in all) and somecomments from a Canadian perspective. Whilethe FSF regards it as essential that steps to en-hance the resilience of the global financial sys-tem be taken promptly to restore confidence inthe soundness of markets and institutions, italso recognizes the need to proceed in a waythat avoids exacerbating stress. Going forward,the FSF intends to examine the forces that con-tribute to procyclicality in the financial system.
a. Strengthened oversight of capital,liquidity, and risk management
The report concludes that Basel II provides theappropriate framework for addressing the weak-nesses that the turmoil has exposed, and that itsimplementation should proceed with priority.However, it also points to elements of Basel IIthat need to be strengthened to improve resil-ience. In particular, it proposes to enhance thecapital treatment of structured credit and off-balance-sheet activities, and recommends that
22. The FSF drew on a large body of coordinated work,comprising that of the Basel Committee on BankingSupervision (BCBS), the International Organizationof Securities Commissions (IOSCO), the InternationalAssociation of Insurance Supervisors (IAIS), the JointForum, the International Accounting Standards Board(IASB), the Committee on Payment and SettlementSystems (CPSS), the Committee on the Global Finan-cial System (CGFS), the International Monetary Fund(IMF), the Bank for International Settlements (BIS),and national authorities in key financial centres (theseinclude Canadian representatives from Finance Canada,OSFI, and the Bank of Canada). The report is availableat < http://www.fsforum.org >.
18
supervisors assess the cyclicality of the Basel IIframework and take additional measures asappropriate. It also calls on the Basel Committeeto issue for consultation a guide for the soundmanagement and supervision of liquidity, andto strengthen its guidance for risk managementand stress testing for capital-planning purposes.
In view of these recommendations, the BaselCommittee on Banking Supervision announceda series of steps on 16 April to help make thebanking system more resilient to financialshocks.23
b. Enhancing transparency and valuationTo restore market confidence, the FSF stronglyencourages financial institutions to use theleading disclosure practices summarized in itsreport to provide meaningful and consistentquantitative and qualitative information abouttheir risk exposures, valuations, and off-balance-sheet entities. It expects further guidance in thisarea from the Basel Committee on BankingSupervision and calls upon the InternationalAccounting Standards Board to accelerate itswork on accounting and disclosure standardsfor off-balance-sheet vehicles and to enhance itsguidance on valuing financial instrumentswhen markets are no longer active.
c. Changes in the role and use of creditratings
Credit-rating agencies play an important role inevaluating and disseminating information onstructured credit products. The FSF calls onthem to enhance their review of the informationprovided by originators, arrangers, and issuersof securitized products and to ensure that it is ofsufficient quality to support a credible rating. Italso asks agencies to differentiate the ratingsused for structured products from those used for
23. These steps include:(i) Enhancing various aspects of the Basel II frame-work, including the capital treatment of complexstructured credit products, liquidity facilities to sup-port asset-backed commercial paper (ABCP) con-duits, and credit exposures held in the trading book.(ii) Issuing a proposal for sound-practice standardsfor the management and supervision of liquidity risk.(iii) Initiating efforts to strengthen banks’ risk-man-agement practices and supervision related to stresstesting, off-balance-sheet management, and valua-tion practices, among others.(iv) Enhancing market discipline through better dis-closure and valuation practices.
Financial System Review – June 2008
corporate bonds. It warns investors that ratingsshould not replace their own risk analysis andadvises authorities to check that the role thatthey have assigned to ratings in regulations doesnot induce uncritical reliance on credit ratings.
d. Strengthening authorities’ responsivenessto risk
In its report, the FSF noted that some of theweaknesses that have come to light were knownor suspected within the community of financialauthorities before the turmoil began. Muchwork was under way at international levels that,if implemented, might have tempered the scaleof the problems that were experienced. However,international processes for agreeing upon andimplementing regulatory and supervisory re-sponses have, in some cases, been too slow, giventhe pace of innovation in financial markets. Thus,the FSF recommends that supervisors, regula-tors, and central banks, individually and collec-tively, take additional steps to more effectivelytranslate their risk analysis into actions that mit-igate those risks. This includes improving infor-mation exchange and co-operation amongauthorities. Of note is a recommendation thatlarge financial institutions share their contin-gency liquidity plans with their central bank.
e. Robust arrangements for dealing withstress in the financial system
The FSF report recommends that central bankoperational frameworks be sufficiently flexiblein terms of potential frequency and maturity ofoperations, available instruments, and the rangeof counterparties and collateral, to deal withextraordinary situations. Amendments to theBank of Canada Act proposed in the federalgovernment’s budget bill would increase theflexibility of the Bank’s operating frameworkby allowing the Bank to conduct open marketoperations with a wide range of assets asnecessary.
The FSF report also proposes that authoritiesat the national level need to review and, whereneeded, strengthen legal powers and clarify thedivision of responsibilities of different nationalauthorities for dealing with weak and failingbanks. In Canada, a structured interventionframework exists that enables federal agenciesto identify areas of concern at an early stage andintervene effectively, so as to minimize losses todepositors and the exposure of the CanadaDeposit Insurance Corporation (CDIC) to loss
(Engert 2005). A Guide to Intervention (OSFI2002) sets out the intervention measures thatan institution can normally expect from OSFIand the CDIC, summarizes the circumstancesunder which intervention measures may beexpected, and describes the coordination mech-anisms in place between OSFI and the CDIC.
The G-7 ministers identified the following rec-ommendations among the immediate prioritiesfor implementation within the first 100 days:
• Financial institutions were strongly encour-aged to make robust risk disclosures in theirupcoming mid-year reporting, consistentwith leading disclosure practices as set outin the SSG report and summarized in theFSF report.
• The International Accounting Standards Boardand other relevant standard setters wereurged to initiate action to improve theaccounting and disclosure standards for off-balance-sheet entities and enhance its guid-ance on fair-value accounting, particularlyon valuing financial instruments in periodsof stress.
• Firms were also urged to strengthen theirrisk-management practices, supported bysupervisors’ oversight, including rigorousstress testing, and to strengthen their capitalpositions as needed.
• The Basel Committee was asked to issuerevised guidelines for the management ofliquidity risk by July 2008, and IOSCO torevise its code of conduct fundamentals forcredit-rating agencies by the same date.
Senior Supervisors GroupEarly in this period of market turbulence, theSenior Supervisors Group—supervisors of majorfinancial services firms from France, Germany,Switzerland, the United Kingdom, and theUnited States—convened to assess whethershortcomings in risk management may havecontributed to the credit losses being registeredby major financial institutions.24
24. Seven supervisory agencies participated in this project:the French Banking Commission, the German FederalFinancial Supervisory Authority, the Swiss FederalBanking Commission, the U.K. Financial ServicesAuthority, the U.S. Office of the Comptroller of theCurrency, the U.S. Securities and Exchange Commis-sion, and the U.S. Federal Reserve.
19
Developments and Trends
More specifically, the group sought to identifyrisk-management practices that have tended towork well, and those that have not. They devel-oped an extensive questionnaire covering seniormanagement oversight and risk-managementperformance across key dimensions. The groupshared the questionnaire with 11 global bank-ing and securities firms that were major playersin key markets and had experienced a range ofoutcomes during this period. Its final report,“Observations on Risk Management Practicesduring the Recent Market Turbulence,” waspublished in March 2008.25
Importantly, the group found that firms thatdealt more successfully with the ongoing marketturmoil through year-end 2007 had adopted acomprehensive view of their exposures. Theyused information developed across the firm toadjust their business strategy, risk-managementpractices, and exposures promptly and proactive-ly in response to changing market conditions.This information was also centralized into onerisk-management unit reporting to the CEO. Thegroup noted that differences in risk appetite,business strategy, and risk-management ap-proaches in three particular business lines hadled to considerable variability in firms’ perfor-mance. The three business lines include: CDOstructuring, warehousing, and trading; syndicationof leveraged financing loans; and the conduit/struc-tured investment vehicle (SIV) business.
The group’s report discusses in detail key features ofthe broad functions of risk management—seniormanagement oversight, the management of liquid-ity risk, and the management of credit and marketrisk—needed to ensure the success of global finan-cial institutions through such challenging times.
Prior to the market turbulence, a series of inter-views conducted by the Bank of Canada in Janu-ary 2007 with the major Canadian banks ontheir risk-management practices found that theyhad improved over time and were broadly in linewith their global banking peers (see Aaron, Arm-strong, and Zelmer 2007). At that time, Canadi-an banks were striving, to varying degrees, toachieve a comprehensive and integrated view oftheir exposures, but they generally admitted thatthis was a “work in progress.”
25. The full report can be found at <http://www.fsa.gov.uk/pubs/other/SSG_risk_management.pdf>.
20
Financial institutions
Major Canadian banks reported weaker after-taxprofits of $4.6 billion through the first half offiscal 2008 (ending 30 April). Earnings were lowerby 53 per cent, compared with the same periodin 2007. The weaker performance can be largelyexplained by $8.1 billion of writedowns relatedto securities linked to U.S. subprime mortgages,and other credit market exposures experiencedover the period (in addition to the $2.1 billionin writedowns reported in 2007), and to a lesserextent, by rising loan-loss provisions. Averagereturn on equity in the first half of 2008 was8.9 per cent, compared with 21.0 per cent in 2007(Chart 11). It should be noted that these write-downs reduce “trading profits” (Chart 12),which, on average, have comprised about 25 percent of bank pre-tax income in recent years.While, to date, the writedowns for the majorCanadian banks have been rather moderatecompared with those for U.S. banks, the ongoingvolatility in credit and structured instrumentmarkets is expected to continue to hamper prof-it performance through 2008.26
Core earnings derived from banks’ lending andwealth-management operations have, to date,remained reasonably firm. But, given the unset-tled global economic and financial environment,there is likely to be some deterioration in thecredit quality of loans. Recently, loan-loss pro-visions have started to rise from a historicallylow base (Chart 13) and have been occurringboth in the household sector and the manufac-turing sector.
In the current volatile environment, Canadianbanks have provided enhanced disclosure oftheir exposures to key problem areas, includingsubprime mortgages, structured investmentvehicles, ABCP conduits, monoline insurers,hedge funds, and leveraged buyouts. While, onbalance, these exposures appear to be manage-able, some banks have announced certain con-centrated exposures that they are trying toreduce or unwind. Furthermore, Canadianbanks have significant exposures to variousmarkets in the United States (Box 3).
26. The Bank estimates that the writedowns realized byCanadian banks in 2007 and in the first quarter of2008 amounted to about 20 per cent of 2007 annualprofits (excluding writedowns). This compares withabout 42 per cent for U.S. banks and investmentdealers as a group.
Financial System Review – June 2008
Chart 11 Profits of Major Banks
% Can$ billions
Source: Banks’ quarterly financial statements
0
5
10
15
20
25
30
0
1
2
3
4
5
6
1998 2000 2002 2004 2008
Return on equity (left scale)Net income (right scale)
2006
Chart 12 Trading Profits: Level and Share ofPre-Tax Profits*
Major banks
Can$ billions % of pre-tax profits
* Includes gains and losses on instruments held for otherthan trading purposes
Source: Office of the Superintendent of FinancialInstitutions (OSFI)
-4
-2
0
2
4
6
8
-50
-25
0
25
50
75
100
1998 2000 2002 2004 2006 2008
Trading profits shareTrading profits level
Chart 13 Asset Quality
% % of income
Note: Blue bar indicates a period of recession.Source: OSFI
1990 1995 2000 20052
4
6
8
10
12
14
0
5
10
15
20
25
30Loan-loss provisions,4-quarter moving average(right scale)Net non-performingloans to Tier 1 capital(left scale)
In general, major Canadian banks, like theirglobal peers, are trying to cope with the follow-ing developments, which have put pressure ontheir capital ratios:
• Marked-to-market writedowns on securitiesrelated to U.S. subprime mortgages, as well asother credit markets exposures which,through the rules of fair-value accounting,have an immediate impact on retainedearnings and Tier 1 capital. As mentionedearlier, Canadian banks have recordedcumulative writedowns of $10.2 billion todate.
• A trend towards reintermediation as banksare pressed to bring assets from various off-balance-sheet structures onto their balancesheets, and to provide balance sheet financ-ing to borrowers who are no longer able toaccess capital markets. This has tended toput upward pressure on risk-weighted assets(RWA). The Highlighted Issue on page 24provides background on recent develop-ments in securitization in Canada and theimpact of the recent market turmoil on totalcredit growth and bank lending.
• The recent high-profile difficulties of U.S.monoline insurers (Box 1), which, in somecases, may call into question the reliabilityof the credit protection they have purchased,leading to the possibility of further increasesin RWA or reductions in capital.
Reflecting the pressures on their balancesheets—as well as high debt maturities in2008—the major banks have been active in rais-ing funds in a range of capital markets. While allof these instruments boost liquidity, for exam-ple, deposit notes (senior debt), NHA MBS, andcovered bonds (see Highlighted Issue on p. 33),other instruments also boost regulatory capital,for example, common shares (counts as Tier 1capital), subordinated debt (Tier 2), non-cumu-lative permanent preferred shares (Tier1), andinnovative instruments (Tier 1). In addition,growth in wholesale deposits has been strong,from both financial and non-financial sources.To help maintain capital ratios, some bankshave also suspended ongoing repurchase pro-grams for common shares and have postponedfurther dividend increases.
It is important to note that even though bankshave largely been successful in issuing debt atvarious terms, these instruments have been priced
21
Developments and Trends
Box 3
Exposure of Canadian Banks to the United States:An Aggregate View
Detailed public data on the U.S. exposures of Cana-dian banks are limited and often difficult to com-pare across banks. However, useful informationcan be obtained from the aggregate data collectedby the Bank of Canada. Overall, the direct exposureof Canadian banks to the United States representsabout 16 per cent of total bank assets, more thanthe combined exposure to any other group of for-eign countries.1 Although exposure to the UnitedStates as a proportion of total claims has remainedroughly stable since the early 1990s, its composi-tion has shifted as the activity of Canadian banks incapital markets has increased.As seen in Chart A, exposure to the United Stateshas increased mainly as a result of increased securi-ties holdings. In the mid-1990s, exposure to U.S.securities and loans represented roughly 3 per centand 10 per cent of total Canadian bank assets, re-spectively. In 2007, these figures had converged,each representing just under 8 per cent of totalCanadian bank assets.2 This shift, all else beingequal, implies that the return on U.S. exposures hasbecome more dependent on the performanceof financial markets.3 This, in turn, may increasethe volatility of bank earnings, and add downwardpressure to profitability during periods of poorfinancial market performance.
22
1. Data do not capture off-balance-sheet exposures(e.g., credit commitments) or indirect exposures(e.g., loans to a Canadian firm with extensive opera-tions in the United States). Data are expressed on animmediate borrower basis (i.e., claims are recordedunder the country where the immediate borrower islocated). A short dataset is also available on an ulti-mate-risk basis since 2005 (i.e., claims are recordedin the country of residence of the entity that willrepay the claim if the original borrower does not).The value of U.S. exposures differs only slightlybetween the two measures (the ultimate-risk mea-sure is about 2 per cent smaller), and the choice ofmeasure does not materially affect the figures usedhere.
2. Data do not account for TD Bank’s recent acquisition ofCommerce Bancorp in the United States. Data areunavailable regarding Commerce Bank’s U.S. versusnon-U.S. exposure; however, as of December 2007, thebank held roughly $18 billion in loans and $26 billionin securities. Combined, these figures represent justunder 9 per cent of TD Bank’s overall assets and lessthan 2 per cent of the overall assets of the Canadianbanking sector.
3. This fact is consistent with C. Calmès and Y. Liu,“Financial structure change and banking income: ACanada - U.S. comparison,” Journal of InternationalFinancial Markets, Institutions & Money (2007). Availableat < http://www.sciencedirect.com >.
Detail on the composition of Canadian bank hold-ings of foreign securities is limited. However, datashow that securities issued by U.S. non-bank pri-vate institutions make up the majority (about55 per cent) of U.S. securities exposure, havingovertaken government-issued securities (about35 per cent) in recent years. Securities issued byU.S. banks make up the remaining 10 per cent.In terms of profitability, available data from Cana-da’s five largest banks suggest that, on average, U.S.operations have underperformed relative to do-mestic operations. Although some improvementhas taken place in recent years, the prospectiveprofitability of U.S. operations remains uncertain,given the ongoing slowdown in the U.S. economy.Overall, exposure to the United States constitutes asignificant portion of the total assets of Canadianbanks, and given the expected future volatility inboth the U.S. real economy and financial markets,Canadian banks’ balance sheets may experiencefurther pressure.
Chart A Canadian Bank Exposure to theUnited States, by Type of Claim*
% of total assets
1996 1998 2000 2002 2004 2006
* Immediate borrower basisSource: Bank of Canada
0
5
10
15
20
25
0
5
10
15
20
25
TotalLoans toU.S. borrowers
Holdings of U.S. securitiesDeposits by Canadian banksheld at U.S. institutions
Financial System Review – June 2008
Chart 14 Total and Tier 1 BIS Capital Ratios
%
Note: Data reflect Basel II framework from 2008Q1 on.Source: OSFI
1998 2000 2002 2004 20064
5
6
7
8
9
10
11
12
13
14
4
5
6
7
8
9
10
11
12
13
14
Total capital ratioTier 1 capital ratioOSFI benchmarks
Chart 15 Distance to Default for Major Banks
Standard deviation
Note: Horizontal line is the average distance to default from December 1982 to present.
Sources: Bank of Canada calculations based on data from OSFIand Thomson Financial Datastream
Chart 16 Return on Equity of Life and HealthInsurance Companies
%
Note: Horizontal line is the average of return on equity from 1996Q1 to present.
Source: OSFI
0
2
4
6
8
10
12
14
16
18
0
2
4
6
8
10
12
14
16
18
2008
1990 1995 2000 20052
3
4
5
6
7
8
9
10
2
3
4
5
6
7
8
9
10
1998 2000 2002 2004 2006
MaximumAverageMinimum
at yield spreads over government bonds that aresignificantly higher than prior to the turmoil,and sometimes at significant concessions to themarket on the day of issue. Nonetheless, theiroverall success at funding in sizable amountshas permitted banks to maintain satisfactorylevels of liquidity through this difficult period,although liquidity-risk management continuesto be given enhanced attention as events un-fold.27
Thus, the Canadian banking system remainswell capitalized, with an average Tier 1 capitalratio of 10 per cent (well above OSFI’s thresholdof 7 per cent) and a total capital ratio of 12.7 percent (compared with the threshold of 10 percent) in the first quarter of 2008 (Chart 14).
The Basel II capital-adequacy framework forbanks came into effect in Canada in November2007. The initial effect for the first quarter end-ing 31 January 2008 was to raise the average re-ported Tier 1 capital ratio for the group by about0.4 percentage points and the overall total capi-tal ratio by about 0.3 percentage points overwhat it would have been under Basel I. On bal-ance, a new capital charge for operational riskwas more than offset by a lower charge forcredit risk.
Market-based metrics suggest that banks are stillin a sound financial position, although therehas been a noticeable weakening since August2007. For example, the distance to default formajor banks has declined sharply, reflecting theincreased volatility of bank share prices(Chart 15). While the impact has been greateron some banks than on others, the distance todefault of the six largest banks, on average, hasfallen moderately below its long-run mean,but remains above the lows reached duringthe technology-sector adjustment earlier thisdecade.28 These developments are similar tothe scenario presented in the December 2007Financial System Review in which heightenedequity volatility was assumed to continue oneyear into the future.
As noted earlier, provisions for loan losses atCanadian banks are also starting to rise, although
27. As of mid-May, the major banks had issued approxi-mately $45 billion of debt and equity in 2008.
28. This remains true when a 6-month, rather than thestandard 12-month, measure of volatility is used in thecalculation of distance to default.
23
Developments and Trends
they remain below their historical average. Evi-dence thus far suggests that the credit quality ofmajor Canadian banks remains relatively strongby historical standards, despite the negative im-pact of the U.S. subprime-mortgage crisis.
The threemajorCanadian lifeandhealth insurancecompanies reported generally strong earningsin the first quarter of 2008, with returns on eq-uity in the range of 15 to 20 per cent, as both theinsurance and wealth-management operationscontinue to do well (Chart 16). Recent favourableresults occurred despite ongoing headwindsfrom a strong Canadian dollar, which reducesreturns from their extensive foreign operations.The life and health companies continue to en-joy high credit quality in their fixed-incomeportfolios. They have reported minimal exposureto subprime instruments and some of the otherproblem areas of the capital markets. Further-more, they have no significant involvement inthe troubled bond insurance business. Con-tinued volatile markets would tend to have anunfavourable impact on the profitability oftheir wealth-management business, althoughthe insurance business should not be affected.
Highlighted Issue
The impact of the recent marketturbulence on credit growth inCanada
Prepared by Jim Armstrong
Since the start of the financial market turbu-lence in August 2007, there has been a majorloss of confidence in structured finance instru-ments, which has seriously impeded the financ-ing technique of securitization, both globallyand in Canada. Many securitization vehicles(often referred to as conduits) have had difficultyfunding themselves in the ABCP market, andthere has been pressure for the banks sponsoringthem to assume the securitized assets on theirbalance sheets. In addition, banks have had tocope with increased demand for credit fromborrowers with reduced access to the capitalmarkets. Here, we briefly review how importantsecuritization has been as a source of credit inCanada, as well as the impact of recent eventson growth in total and bank-originated creditin Canada.
24
The importance of securitization inCanadaChart 17 presents the trend in the share ofsecuritized credit (securitized loans residing inconduits) as a percentage of the major catego-ries of credit: residential, consumer, and busi-ness. The chart shows that securitization hasbeen highest in the residential mortgage market(particularly NHA-insured mortgages) but hasalso been fairly significant for consumer credit.It still accounts for a relatively minor portion ofbusiness credit.29
Impact of the recent developments onoverall credit growthThe financial market turmoil—with its associatedweakening effect on securitization activity andmarket-based finance—has not yet had a no-ticeable adverse impact on the overall growthof credit in Canada. Table 1 presents quarterlygrowth rates for total, securitized, and bankcredit for each major credit category since thebeginning of the turmoil in the middle of thethird quarter of 2007.
Growth in residential mortgage credit has beensustained at a very strong pace (about 13 per cent).The NHA MBS Program has played a crucialrole, exhibiting much higher growth (Table 1)than prior to the market turbulence. In contrast,securitization in non-NHA conduits declined.The major banks’ success in securitizing mort-gages through the NHA MBS Program has al-lowed them to reduce the rate of expansion ofresidential mortgages held on their balancesheets; growth was only 3.8 per cent per cent in2008Q1. It is important to note that much ofthis expanded use of NHA MBS was used to sup-port issuance of CMBs.
Table 1 also indicates that consumer credit hasremained firm, growing by about 10 per cent.While consumer loans held in securitizationconduits have contracted sharply, banks havestepped in to fund a large amount of consumercredit directly on their balance sheets, registeringvery strong growth of 16.5 per cent in 2007Q4and 13.5 per cent in 2008Q1.
29. This excludes “third-party” conduits sponsored byspecialty firms, which invested in non-traditionalstructured assets, such as CDOs (largely originatingfrom outside of Canada), and which are now thesubject of restructuring following successful negotia-tions by major participants in the Montreal Accord.
Financial System Review – June 2008
Table 1
Credit : Annualized Quarterly Growth
Pre-crisistrend a 2007Q3 2007Q4 2008Q1
Residential 10.1 13.5 12.9 13.4
Securitized 20.3 9.6 -14.8 -5.8
NHA MBS 20.5 39.4 65.4 39.6
Bank 9.2 14.8 5.7 3.8
Consumer 10.0 10.7 10.2 10.0
Securitized 16.9 6.1 -9.9 -7.1
Bank 9.0 12.8 16.5 13.5
Business 7.0 8.1 7.4 5.8
Securitized 24.2 12.4 -10.3 -8.4
Bank 12.3 20.5 19.6 15.2
Source: Bank of Canadaa. Average of the annualized quarterly growth rates from 2006Q3
through 2007Q2
Chart 17 Share of Securitized Credit in MajorCredit Aggregates
%
Source: Bank of Canada
1996 1998 2000 2002 2004 2006 20080
5
10
15
20
25
0
5
10
15
20
25
Residential mortgagesResidential mortgages (excluding CMHC MBS)Consumer creditBusiness credit
Chart 18 Domestic Bank Loans
Year-over-year growth rate
%
0
2
4
6
8
10
12
14
0
2
4
6
8
10
12
14
2004 2005 2006 2007 2008
Source: Bank of Canada
Total business credit (the sum of intermediatedcredit and market-based business credit) has re-mained reasonably firm since the beginning ofthe turmoil, with underlying growth of about6 to 8 per cent.
As noted earlier, securitization accounts for avery low share of total business credit in Cana-da. However, more broadly, in the difficultcredit environment, a contraction in traditionalcommercial paper (non-ABCP) for non-finan-cial corporations, combined with a slowing inthe issuance of bonds and debentures, equity,and trust units, resulted in banks assumingbusiness credit onto their balance sheets at avery strong pace of about 20 per cent in the sec-ond half of 2007.
Chart 18 shows that, against this background,there has been a marked increase in the rate ofgrowth in total domestic bank credit since Au-gust 2007.
ConclusionIn summary, the collapse of privately sponsoredsecuritization and associated reintermediationtrends has not seriously impeded the overallgrowth of credit in Canada. This outcome hasbeen facilitated by the banks’ willingness to as-sume more consumer credit and business crediton their balance sheets, while they themselveshave been able to sell substantial amounts ofresidential mortgages through the NHA MBSprogram, which experienced very high growthin this period. Besides issuing NHA MBS, bankshave also been successful in funding themselvesin various other markets (see Financial Institu-tions section on page 20).
25
Developments and Trends
Chart 21 U.S. House Prices and Inventory:Existing Homes
Year-over-year percentage change Months’ supply
Chart 19 Evolution of Consensus Estimates forAnnual World* Economic Growth in2008 and 2009
%
*46 of the world’s economies accounting for over 90% of global GDPSource: Consensus Economics Inc.
J F M A M J J A S O N D J F M
2008 2009
2.7
2.8
2.9
3.0
3.1
3.2
3.3
3.4
3.5
3.6
2.7
2.8
2.9
3.0
3.1
3.2
3.3
3.4
3.5
3.6
20082009
Sources: Office of Federal Housing Enterprise Oversight andNational Association of Realtors, and S&P Case-Shiller
2000 2001 2002 2003 2004 2005 2006 2007
OFHEO House Price Index (left scale)S&P Case-Shiller House Price Index (left scale)Inventory (right scale)
2008
A M
-20
-15
-10
-5
0
5
10
15
20
25
3
4
5
6
7
8
9
10
11
12
Chart 20 United StatesSenior Loan Officer Survey
% of banks reporting tighter lending standards
-40
-20
0
20
40
60
80
100
-40
-20
0
20
40
60
80
100
2000 2001 2002 2003 2004 2005 2006 2007 2008
Source: Federal Reserve
Commercial and industrial loansResidential mortgage loansPrime mortgage loansNon-traditional mortgage loansSubprime-mortgage loans
The MacrofinancialEnvironment
The international environment
The outlook for global economic growth in 2008has been revised downwards since December2007, led by a deceleration in growth in ad-vanced economies that was more pronouncedthan expected and growth that was more mod-erate than expected in emerging-market econo-mies, led by China and India (Chart 19).
In the United States, there are signs that the econ-omy is likely to experience a deeper and moreprolonged slowdown than had been projected.This slowdown stems from further weakening inthe residential housing market that is adverselyaffecting other sectors of the U.S. economy andcontributing to further tightening in credit condi-tions. The U.S. Federal Reserve’s April 2008 SeniorLoan Officer Opinion Survey showed a significanttightening of credit standards for business andconsumer loans (Chart 20). Sales of new and ex-isting homes, along with house prices, continueto decline, and the inventory of housing relativeto demand remains elevated (Chart 21). Theproblems in the U.S. subprime-mortgage marketappear to have spilled over into segments of thecorporate and commercial credit markets, nega-tively affecting bank balance sheets and posing afurther risk to the U.S. economy. The slowdownin consumer spending and tighter credit condi-tions have also led to lower business investment.U.S. GDP growth will be constrained by creditconditions that are unlikely to normalize untilearly 2010, but this should be somewhat mitigat-ed by the aggressive easing of U.S. monetary policyand the fiscal-stimulus package announced bythe U.S. government.
The deterioration in economic and financialconditions in the United States is expected to havesignificant spillover effects on the Canadian andglobal economies. For the advanced overseaseconomies, recent consensus forecasts predict amarked slowdown in growth in 2008 for theeuro area and the United Kingdom. Expecta-tions for growth in Japan have also been loweredslightly. Growth in Asia is expected to remainstrong, although lower than previously expected.Inflation risk appears to be increasing in the glo-bal economy, driven mainly by food and energyprices, which have pushed headline inflationabove targets in many countries. This may limitthe ability of some central banks to use monetary
26
Financial System Review – June 2008
Chart 23 Emerging-Markets Sovereign Bond Spread(EMBI+)
Basis points
Chart 22 United StatesExchange Rate and Current Account
March 1973 = 100 Per cent of GDP
2000 2001 2002 2003 2004 2005 2006 200780
85
90
95
100
105
110
115
120
-7.0
-6.5
-6.0
-5.5
-5.0
-4.5
-4.0
-3.5
-3.0
Source: JPMorgan Chase & Co.
1998 2000 2002 2004 20060
200
400
600
800
1,000
1,200
1,400
1,600
1,800
0
200
400
600
800
1,000
1,200
1,400
1,600
1,800
2008
2008
Real trade-weighted exchangerate (left scale)Current account (right scale)
Chart 24 Bank of Canada Commodity Price Index
1982–90=100
2000 2001 2002 2003 2004 2005 2006 2007Source: Bank of Canada
200850
100
150
200
250
300
350
400
450
500
50
100
150
200
250
300
350
400
450
500
Energy indexTotal indexNon-energy index
Sources: BEA, Federal Reserve
policy to counter the risks to economic growthfrom the ongoing financial turbulence.
The weakness in U.S. domestic demand and theassociated depreciation of the U.S. dollar arecontributing to an unwinding of global currentaccount imbalances (Chart 22). Trade surplus-es of the Asian and oil-exporting countriescontinue to expand, however, supported by of-ficial intervention to maintain fixed exchangeregimes, as well as by high commodity prices.
As suggested by the modest rise in the spreads inthe emerging-market bond index (EMBI), emerg-ing markets continue to hold up well in com-parison to previous episodes of financial turmoil(Chart 23). This resilience may be attributed tostructural reforms, stronger macroeconomic fun-damentals, and the accumulation of substantialforeign exchange reserves. There is, however,some underlying divergence between the con-tinuing strong growth in commodity-exportingLatin America and Russia, and slower growth inemerging Europe and India, where inflationarypressures have required a tighter monetary pol-icy response. As noted in the December 2007FSR, the main risk for these economies remainsthe potential for a slowdown in the global econ-omy with the associated reduced demand forexports and commodities.
For 2008, the important risks to the global outlookare expected to come from disruptions in creditmarkets, continued deterioration in house pricesin certain economies, the potential for a disor-derly unwinding of global imbalances, and a sud-den reversal of capital flows to emerging markets.
While subject to some additional volatility,commodity prices have continued to strengthenso far this year, despite the ongoing financialturmoil and the slowdown in U.S. economicactivity (Chart 24). In nominal terms, the pricesof several commodities (such as crude oil andpotash) have reached record highs. A number offactors may help to explain the strength of com-modity price gains in recent years, includingstrong demand from emerging-market econo-mies, supply constraints in the context of lowstocks, and the depreciation of the U.S. dollar.Recently, speculative activity may also haveplayed a role.30
30. This reflects increased demand from institutionalinvestors, hedge funds, and momentum traders, aswell as the growing use of derivative products linkedto commodities.
27
Developments and Trends
Chart 25 Real GDP Growth: Canada
%
Source: Statistics Canada
2000 2001 2002 2003 2004 2005 2006 2007-1
0
1
2
3
4
5
6
-1
0
1
2
3
4
5
6Year-over-year percentage changeQuarterly growth at annual rate
Chart 26 Financial Position of the CanadianNon-Financial Corporate Sector
% %
Sources: Statistics Canada and Bank of Canada
20072000 2001 2002 2003 2004 2005 2006-5
0
5
10
15
20
25
0.85
0.90
0.95
1.00
1.05
1.10
1.15
Return on equity (left scale)Debt-to-equity ratio (right scale)
Chart 27 Indicators of Credit Quality
Standard deviation %
Source: Bank of Canada calculation based on data from ThomsonFinancial Datastream, the Globe and Mail, the Financial Post, andMoody’s KMV. For the volatility indicator, all datapoints denotedJanuary 2005 and thereafter are provided by Moody's KMV;datapoints prior to January 2005 continue to be provided by theGlobe and Mail and Thomson Financial Datastream.
1994 1996 1998 2000 2002 2004 20060
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
0
2
4
6
8
10
12
14
16
18
20
2008
Volatility of return onCanadian corporateportfolio (left scale)Share of companieswith weak financialratios (right scale)
2008
Canadian developments
Canadian economyCanada’s economic growth slowed consider-ably towards the end of 2007 and into early2008, as manufacturers scaled back production(Chart 25). As detailed in the April 2008 Mone-tary Policy Report, Canada’s economy is expectedto be negatively affected by spillovers from thedeterioration in economic and financial condi-tions in the United States, both through lowerexports (in 2008) and through the dampeningeffect of tighter credit conditions and softeningsentiment on domestic demand. Nevertheless,domestic demand is still expected to remainstrong over the projection period. On the down-side for Canada’s economic outlook, greater-than-anticipated weakness in commodity prices(stemming from the projected slowdown in theUnited States and other industrialized econo-mies) could mean slower gains in real incomesand domestic demand in Canada.
Corporate sectorThe financial position of Canada’s non-financialcorporate sector as a whole continued to bequite robust in the first quarter of 2008, in spiteof the slowdown in economic growth. The over-all rate of return on equity, although easingsomewhat, remained at a high level, and the ratioof debt to equity stayed relatively low (Chart 26).Profitability was relatively high in commodity-producing sectors and in most sectors with alow exposure to international trade but consid-erably weaker in a number of sectors with highexposure to international trade, including forestproducts industries.
Although it remains low, the share of non-finan-cial corporations with weak financial ratios contin-ued a modest upward trend in 2007 (Chart 27).31
Backward-looking measures, such as bankrupt-cies and bond defaults, indicate that corporatecredit quality has declined slightly since the De-cember FSR. More forward-looking market-basedmeasures, such as the volatility of returns on theCanadian corporate portfolio (the Bank of
31. This microdata indicator represents the share of totalassets attributable to companies with a comparativelyhigh leverage ratio and weak current ratio and netprofit margin. The indicator is based on data up tothe end of 2007. Details on the indicator can befound in the December 2005 issue of the FSR,pp. 37–42.
28
Financial System Review – June 2008
Canada’s contingency claims approach (CCA)indicator), also suggest that corporate creditquality has deteriorated somewhat in recentmonths (Chart 27).32 This shift is a reflection ofthe increased volatility of stock prices acrossthe non-financial subindustries included in thecalculation. Nevertheless, this indicator remainswell below the peaks reached during the late1990s and early 2000s, suggesting that, over-all, the credit quality of non-financial corpora-tions remains relatively strong.
IndustryCanada’s forest products industry experiencedlosses in the fourth quarter of 2007 and the firstquarter of 2008, reflecting the adverse effects ofstill more production curtailments, the strongCanadian dollar, U.S. housing market devel-opments, and much higher fuel and wood fibrecosts. With the liquidity of a number of compa-nies under significant strain, and in the face ofdifficult credit conditions, restructuring of oper-ations is continuing, and some firms have re-structured their balance sheets.
Profitability in Canada’s auto manufacturing in-dustry was also quite weak during this period,with the softening of U.S. motor vehicle salesand the shift in the sales mix away from the moreprofitable larger vehicles. A prolonged drop inthe sale of U.S. vehicles would have a severe im-pact on activity in Canada’s auto manufacturingindustry, including the auto parts industry, sincearound 90 per cent of vehicle production is ex-ported to the United States.
Profitability in many other manufacturing in-dustries has also eased, partly owing to the fur-ther appreciation of the Canadian dollar. Withthe slowdown in the U.S. economy spreading tosectors beyond residential housing and motorvehicles, further adverse effects on the financialpositions of a broader range of Canadian man-ufacturers are likely over the near term. Indeed,some companies in the clothing and textile andprinting industries are currently experiencingsevere financial stress. The Canadian truckingindustry is also beginning to experience financialdifficulties as a result of the weakness in themanufacturing sector and sharply higher fuel costs.
32. The CCA indicator represents the volatility of market-valued assets in a portfolio consisting of nine broadnon-financial industries. It is currently based on dataup to the end of April 2008. Details of the CCA wereoutlined in the June 2006 issue of the FSR, pp. 43–51.
Although a number of manufacturing compa-nies are experiencing serious financial difficulties,their problems are unlikely to have significantadverse effects on the Canadian financial system,since the direct exposure of Canadian banks tothese industries remains limited.
House pricesIn contrast to the United States, conditions inCanada’s housing markets remain relativelyfavourable. Income growth, low unemploymentrates, and relatively good financing conditionshave continued to support rising house prices,although the pace of increase has slowed some-what (Chart 28). This deceleration has beenparticularly marked in markets that have postedvery steep price increases in the past two years,such as Alberta (Chart 29).33 Lower price growthis the result of increased housing supply com-bined with some softening of demand, as illus-trated by a decrease in home sales (Chart 30). Atleast part of this lowered demand for housingmay be attributed to a deterioration in housingaffordability.
Despite the increase in supply, the Canadianhousing market does not seem to be character-ized by excess supply at this time. The proportionof unoccupied newly built dwellings in most cit-ies remains below historical averages (Chart 31),suggesting that a major widespread reversal inhouse prices is unlikely.34 Moreover, the recentdecrease in building permits suggests that housingsupply is adjusting to the softening of demand.
The combination of an expected slowing in eco-nomic growth, more balanced housing supplyand demand, deteriorating consumer confi-dence, and lower home-buying intentions35
suggests that house prices should increase at aslower pace going forward. If a significant rever-sal in commodity prices were to occur, however,house prices could be severely affected, with
33. In contrast, there has been some pickup in house priceincreases outside of Alberta, reflecting recent migratorydevelopments (with people leaving Alberta for otherWestern provinces) and the deterioration in housingaffordability in the province.
34. This does not exclude the possibility of imbalances incertain segments of local markets.
35. CMHC’s Renovation and Home Purchase Report(available at <http://www.cmhc-schl.gc.ca/odpub/esub/65459/65459_2008_AO1.pdf>) indicates that6 per cent of households intend to buy a primaryresidence in 2008, down slightly from the 7 per centthat actually purchased in 2007.
29
Developments and Trends
Chart 30 Indicators of Housing Supply and Demand
Thousands Thousands
Sources: CMHC and MLS
Chart 28 Real Prices for Housing in Canada*
Year-over-year growth rate
%
* Deflated by CPISources: Royal LePage, Statistics Canada, and Bank of Canada
calculations
1985 1990 1995 2000 2005-20
-15
-10
-5
0
5
10
15
20
25
-20
-15
-10
-5
0
5
10
15
20
25
New housingExisting housing
Chart 29 Real Prices for Existing Houses by City*
Year-over-year growth rate
%
* Deflated by CPISources: Royal LePage, Statistics Canada, and Bank of Canada
calculations
1985 1990 1995 2000 2005-30
-20
-10
0
10
20
30
40
50
60
-30
-20
-10
0
10
20
30
40
50
60
CalgaryEdmontonVancouverMontréalToronto
1994 1996 1998 2000 2002 2004 2006100
120
140
160
180
200
220
240
260
280
50
60
70
80
90
100
110
120
130
140Housing starts (12-monthmoving average) (left scale)New listings of existinghomes (left scale)Sales of existing homes(annual rates) (right scale)
2008
price decreases in some local markets, where in-come and employment are particularly exposedto commodities. The direct impact on the bank-ing sector would be somewhat limited, since allbank mortgage lending with a loan-to-value ra-tio greater than 80 per cent must be insured,and since mortgage insurers benefit from an ex-plicit government guarantee. However, sincelending to households—notably through homeequity lines of credit—has been an importantsource of bank profits over the past few years, aslowdown in housing market activity wouldhave a negative impact on the banking sector.
Mortgage marketA number of product innovations in the mort-gage market and the mortgage-insurance marketsince 2006 (e.g., longer amortization periods,zero down payment options, subprime mort-gages) have boosted housing demand in Canadaand increased the vulnerability of the householdsector to changing circumstances. But a housingmarket collapse, similar to that in the UnitedStates (i.e., largely driven by subprime-mortgageinnovation) seems unlikely. The subprime-mortgage market remains small in Canada—ac-counting for less than 5 per cent of the residen-tial mortgage market, compared with 14 percent in the United States—and has not experi-enced the excesses of its U.S. counterpart.36 Thequality of the Canadian subprime-mortgagemarket remains good, as illustrated by stilllow (albeit rising) delinquency rates(Chart 32).37
Still, the ongoing financial market turmoil hashad some impact on the subprime-mortgagemarket in Canada. Subprime lenders that relyprimarily on securitization to fund their mort-gages have been significantly affected by the re-cent drying up of market liquidity: lendingconditions have been tightened, and somemortgage products considered “riskier” havebeen withdrawn; some posted losses at the endof the year, and a number of small players exit-ed the market altogether. Subprime lenders thatrely on a deposit base to fund their mortgageswere not affected as strongly by market-liquidi-ty problems, although their funding costshave risen. Overall, this suggests that, going for-
36. For details, see Box 1 in the December 2007 FSR, p. 8.37. By comparison, 14.4 per cent of U.S. subprime mort-
gages were in arrears over 90 days or in foreclosure in2007Q4 (up from 9.3 per cent in 2007Q2).
30
Financial System Review – June 2008
Chart 31 Recently Completed Unoccupied Dwellings,as a Percentage of Population
%
Sources: Statistics Canada and Bank of Canada calculations
1994 1996 1998 2000 2002 2004 20060
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
CanadaMontréalTorontoCalgaryVancouverEdmonton
Chart 32 Household Sector: Financial StressIndicators
% %
* As a percentage of population aged 20 and over** As a percentage of prime mortgage loans outstanding*** As a percentage of consumer loans outstanding**** As a percentage of subprime mortgage loans outstanding,
average of major subprime lendersSources: Statistics Canada and Bank of Canada
1992 1994 1996 1998 2000 2002 2004 20060.24
0.26
0.28
0.30
0.32
0.34
0.36
0.38
0.40
0.42
0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
Personal bankruptcies (left scale)*Prime residential mortgage loans in arrears 3 months or more(right scale)**Consumer loans in arrears 3 months or more (right scale)***Subprime residential mortgage loans in arrears 3 months or more(right scale)****
Chart 33 Household Sector: Indebtedness Indicators
% %
Sources: Statistics Canada, Ipsos Reid, and Bank of Canada calculations
1985 1990 1995 2000 2005
60
70
80
90
100
110
120
130
140
6
7
8
9
10
11
12
13
14
Debt-to-income ratio (left scale)Debt-service ratio (right scale)
2008
ward, the subprime-mortgage market shouldgrow at a slower pace.
The recent popularity of product innovations inthe mortgage market, such as low down pay-ments and longer amortization periods, suggeststhat a certain proportion of homeowners havelittle home equity and would be more sensitiveto adverse economic shocks. While the directimpact on the financial system would be limit-ed to the extent that those mortgages are in-sured, there could be secondary effects onfinancial institutions.
Household sectorDisposable income continued to increase at asolid pace (2.8 per cent increase in the secondhalf of 2007). The increase in household debt,however, outpaced that in income, leading to afurther rise in the debt-to-income ratio, whichstood at 131 per cent in 2007Q4 (Chart 33).Rising indebtedness was accompanied by high-er mortgage rates, which pushed up the debt-service ratio (DSR) to 7.7 per cent in 2007Q4from 7.3 per cent in 2007Q2.38
Although household liabilities increased morerapidly than household assets (at market value),causing the debt-to-asset ratio to rise to 17 percent in 2007Q4—up from 16.6 per cent in2007Q2—household net worth still increasedby 6.4 per cent in the second half of 2007.
Aggregate indicators of household financial stresscontinue to suggest that the Canadian house-hold sector is in good financial health (Chart 32).Mortgage loans and consumer loans in arrearshave remained at historically low levels, and thepersonal bankruptcy rate was unchanged at0.33 per cent in February 2008.
Notwithstanding the overall solid financial po-sition of the Canadian household sector, theproportion of debt owed by vulnerable house-holds (defined as households with a DSR above40 per cent)39 has increased slightly in the pastyear (Table 2). This suggests that some house-holds could become more sensitive to negativeeconomic shocks. Combined with possible ad-
38. It is estimated that the effective household borrowingrate increased by about 20 basis points over thisperiod.
39. This threshold is a rule of thumb used by financialinstitutions in Canada to assess whether a loanshould be granted. For more details about vulnerabil-ity thresholds, see the December 2007 FSR, p. 27.
31
Developments and Trends
Table 2
Vulnerable Householdsa
a. Sources: Ipsos Reid and Bank of Canada calculationsb. As a percentage of total households with debtc. We report data for 2001 because the share of debt owed by vulnerable
households was at its maximum during the sample period (1999–2007)in that year.
Proportion ofhouseholds
withDSR>40%b
Share of totaldebt owed byhouseholds
withDSR>40%
1999–2006 average 3.33 6.28
2001c 4.04 7.83
2006 3.13 6.17
2007 3.16 6.51
ditional decreases in financial asset prices andthe expected further slowdown in the growth ofhouse prices, this suggests that the financial po-sition of the Canadian household sector maydeteriorate going forward. This deteriorationwould be more significant in the event of asharp reversal in commodity prices, since such areversal could lead to a decrease in house prices—at least in some local markets—and would like-ly be accompanied by a tightening in credit condi-tions. At present, however, the financial situationof households does not pose a threat to the stabil-ity of the Canadian financial system.
32
Financial System Review – June 2008
Important Financial System Developments
his section of Developments and Trendsexamines the structural developmentsaffecting the Canadian financial systemand its safety and efficiency.
Highlighted Issue
An introduction to covered bondissuance
Prepared by Toni Gravelle andKaren McGuinness
The first covered bonds issued by a Canadianbank were offered in October 2007. This High-lighted Issue provides a review of the character-istics of and the market for these instruments.It also assesses the potential contribution ofcovered bonds to the efficiency of the Canadianfinancial system.
Covered bonds are marketable debt securitiesissued by banks and secured by dedicated col-lateral known as the “cover pool.” A large por-tion of the covered bonds issued, includingthose issued by Canadian banks, use mortgageloans as the underlying collateral.40 Coveredbonds have a long history in continental Europe,where a deep market has developed, reachingan outstanding amount of €1.7 trillion in mid-2007.41 It is only recently that Canadian andU.S. banks have started to issue covered bonds.
CharacteristicsThe defining feature of covered bonds is that,in the event of issuer insolvency, bondholdershave both a claim on the issuing bank and apriority claim (over unsecured creditors) on the
40. Although this Highlighted Issue focuses on coveredbonds that use mortgages as collateral, a large segmentof the covered bond market in Europe consists of bondsthat use loans from local authorities as collateral.
41. Packer, Stever, and Upper (2007).
T
bond’s dedicated collateral. That is, an investorin a covered bond is making a secured loan tothe bank and, as such, covered bonds garner ahigher credit rating (typically AAA) and requirea lower yield than unsecured bonds issued bythe same bank. Like asset-backed securities,such as mortgage-backed securities (MBS), acovered bond’s interest and principal paymentsare secured by bankruptcy-remote assets. But incontrast to MBS, the underlying assets for cov-ered bonds remain on the issuing bank’s bal-ance sheet.42 The bank must therefore continueto set aside capital (depending on the type andquality of the loans) for these assets, so the bankdoes not benefit from a reduction in capitalcharges related to the sale of loans off its bal-ance sheet. See Table 3 for a summary of thedifferences between MBS and covered bonds.A recent innovation in the covered bond marketis the advent of so-called “structured” coveredbonds. This innovation has enabled banks incountries that do not have specific financial leg-islation in place (typically those under commonlaw, such as Canada and the United States) toenter the covered bond market. Banks in mostEuropean countries issue “traditional” coveredbonds, which are regulated by country-specificlegislation. Like traditional covered bonds, struc-tured covered bonds provide for a claim on boththe collateral and the issuer. The key differenceis that the standards are defined in a contract,instead of in legislation.43
In most EU countries, traditional covered bondscarry lower risk-capital charges than MBS. In
42. The creation and issuance of MBS require banks tosell the loans that they originated to a special-pur-pose vehicle (SPV) which, in turn, issues bonds col-lateralized by these loans. The sale of these mortgageloan assets takes them off the bank’s balance sheet.See Kiff (2003) for more on asset-backed securities.
43. Standards relate, for example, to the assets used, the qual-ity of the assets, and the maintenance of asset quality.
33
Developments and Trends
Table 3
Traditional Covered Bonds versus Mortgage-Backed Securities
Covered bonds MBS
Balance sheet treatment ofcollateral
On the balance sheet of originating bank Off the balance sheet, assets sold to special-purposevehicle (SPV)
Legal treatment Collateral is bankruptcy remote Bankruptcy remote (given SPV structure)
Responsibility for principaland interest payments
Issuing bank (flows guaranteed by pledged collateral) SPV’s collateral cash flows
Bond structure Semi-annual payment of interest and principal at maturity Typically monthly payments, embeds prepayment risk
Rating Depends on quality of underlying assets and on the ratingof issuing bank (mainly AAA rated)
Depends on quality of underlying assets and creditenhancements to pool assets
Risk weightinga
a. This treatment is covered under the Consolidated Banking Directive of the European Union.
10 per cent for most EU countries 20 per cent for Fannie Mae and Freddie Mac MBS and50 per cent for non-agency MBS
addition, for many regulated investment fundsin the European Union,44 limits on asset con-centration are less restrictive for covered bondsthan for other corporate and securitized debt in-struments. Covered bonds take on the tradition-al structure of “bullet” bonds, with semi-annualinterest payments and principal payment at ma-turity. Holders of MBS, on the other hand, typi-cally receive monthly interest payments andface prepayment risks, since the prepayment ofany of the underlying mortgage loan is typicallypassed on to the MBS investor. These three factors,among others, increase the demand of institution-al investors for covered bonds relative to similardebt instruments. Since most of the preferentialtreatment offered by traditional covered bondsdoes not apply to the structured variety, however,the latter are typically issued at higher yields thantraditional covered bonds.
44. In Europe, regulated investment funds are subject tothe requirements of the undertakings for collectiveinvestment in transferable securities (UCITS) direc-tive. The directive lays down uniform requirementsfor the organization, management, and oversight ofinvestment funds across the European Union. Itimposes rules relating to fund diversification, liquid-ity, and the use of leverage, and defines eligible assetsin which the fund can invest. Funds covered by theUCITS are similar to Canadian mutual funds.
34
The dual nature of covered bonds (i.e., being asenior or priority claim on both the issuingbank and the underlying collateral) is highlight-ed by the differences in the way two of the larg-est credit-rating agencies, Standard & Poor’s andMoody’s, assess the credit risk of the individualbond issues. In the case of Moody’s, it first es-tablishes the senior unsecured credit rating ofthe issuing bank. Then, based on the legislativeor contractual framework of the bond (whichdefines the quality of the collateral), as well asthe degree of credit enhancement embedded inthe bond’s collateral pool, Moody’s will awardthe covered bond a rating several notches abovethe rating of the bank’s unsecured debt. Incontrast, Standard & Poor’s uses a structured-finance approach to assess the covered bond’screditworthiness that is essentially the same asthat used to assess asset-backed securities moregenerally. The only difference is that the issuingbank’s (unsecured) rating will affect the degreeof credit enhancements required for the coveredbond to be rated AAA. As such, the degree towhich the downgrade of an issuing bank willaffect the rating of the covered bond will varyacross bonds and will generally depend on thelegal or contractual structure of the bond, aswell as on the overall quality of the cover pool,including the pool’s credit enhancements.
The existence of an active secondary market is akey factor in the attractiveness of covered bonds
Financial System Review – June 2008
Chart 34 Royal Bank: Spread on Covered Bond vs.Senior Debt*
Basis points
* Swapped to floatingSource: Bloomberg
0
5
10
15
20
25
30
35
40
45
50
0
5
10
15
20
25
30
35
40
45
50
Nov. Dec. Jan. Feb. Mar. Apr. May
2007 2008
Senior debtCovered bond
for fixed-income investors, particularly for thebenchmark segment of the market. The bench-mark or “Jumbo” covered bond market consistsof those issues with an outstanding size of atleast €1 billion and represents approximatelyhalf of the total market. Although market li-quidity has recently declined as a result of themarket turmoil (see p. 13), the benchmark seg-ment of the covered bond market is typically thesecond most liquid bond market in Europe,after sovereign government bonds.45
Drivers of Canadian Bank Issuance ofCovered BondsThe interest of Canadian banks in issuing cov-ered bonds is driven by several factors. One ofthe most important is that covered bonds pro-vide for diversification of bank funding sources,as well as diversification of their investor base,since they tap into a largely European market. Inparticular, covered bonds help banks diversifytheir secured funding sources of mortgage lend-ing, such as National Housing Act (NHA) MBSfunding of the banks’ mortgage-loan portfolios(see the Highlighted Issue on p. 24 for more onthe issuance of mortgage securities under theNHA MBS Program). The banks’ access to theNHA MBS Canada Mortgage Bond Program canoccasionally be constrained, and covered bondscan provide Canadian banks with an alternativeavenue of mortgage funding in these situations.Moreover, the issuance of covered bonds allowsCanadian banks to use mortgages not eligiblefor NHA MBS issuance, thus better leveragingtheir broader pool of mortgage collateral togarner funding.
Covered bonds are particularly well suited tofunding a bank’s fixed-rate mortgage loan port-folio. As Packer, Stever, and Upper (2007) note,the issuance of covered bonds, like that of otherlonger-term fixed-rate debt, enhances a bank’sability to match the duration of its liabilities tothat of its mortgage loan portfolio, enabling
45. The total outstanding value of benchmark coveredbonds is approximately €700 billion as of 2006. SeeEuropean Covered Bond Council (2007).
35
Developments and Trends
better management of its exposure to interestrate risk.46,47
Another reason for Canadian banks to issuecovered bonds is that they are a potentially cost-effective alternative to the issuance of unsecuredsenior debt. This was particularly the case beforethe recent credit market strains spread to theEuropean covered bond market (Chart 34). Byreplacing the issuance of more expensiveunsecured debt, covered bonds may help lowera bank’s overall funding costs and, in turn, thecost of its capital. The lower overall cost of cap-ital generates, all other things being equal, high-er profits and/or lower mortgage loan rates forhomeowners. However, since covered bondsrepresent a priority claim on a pool of bank as-sets, their issuance results in a smaller pool ofassets available for the claims of depositors andother senior unsecured creditors in the case of abank default. Covered bond issuance thereforeposes additional risk to unsecured debt holdersand depositors. Thus, the issuance of a largeamount of covered bonds would affect a bank’screditworthiness and would likely lead to adowngrade of its unsecured debt and/or someform of remedial intervention by the bankingsupervisor.
Overall, the introduction of covered bonds inCanada should increase the effectiveness androbustness of the market-based funding pro-grams of banks and lower the banks’ overallcost of capital (if the amount of covered bondissuance is below some threshold). Coveredbonds thus represent a potential enhancementto the efficiency of the Canadian financialsystem.
46. The U.K. Treasury notes that the absence of a U.K.-basedmarket for covered bonds has impeded the develop-ment of longer-term fixed-rate mortgage lending. See<http://www.hm-treasury.gov.uk/media/F/D/consult_coveredbonds230707.pdf >.
47. An alternative source of secured funding for banks isrepo funding, in that the bank pledges collateral tosecure the loan (see Morrow 1994–95 for more onrepo lending). As with covered bonds, the lender has apriority claim on the pledged collateral. Unlike cov-ered bonds, however, repo funding is short term, typi-cally one day. Moreover, the collateral consists largelyof liquid government bonds, rather than (non-market-able) mortgage loans. As such, repo funding does nothave the asset-liability matching attributes offered bycovered bonds in funding mortgage loans.
36
Policy Developments in CanadaIn June 2007, OSFI completed an initial reviewof regulatory considerations regarding the issu-ance of covered bonds by Canadian banks andissued guidelines allowing for a limited issu-ance. In reaction to concerns that covered bondscreate a preferred class of claimants, OSFI im-plemented a limit on the level of covered bondissuance of 4 per cent of a bank’s total assets.48
The 4 per cent limit would amount to a maxi-mum issuance of roughly $95 billion by the bigsix Canadian banks, based on the 2007 level oftheir total assets.
In imposing limits on covered bond issuance,banking supervisors tend to weigh the previous-ly mentioned risks for depositors against theprudentially beneficial enhancements to abank’s creditworthiness offered by coveredbonds in the form of lower overall wholesalefunding costs and more robust access to liquid-ity. For example, the U.K. banking supervisor,the Financial Services Authority (FSA), envisag-es that an amount of covered bond issuancearound 20 per cent of total assets would likelypose a high enough risk to require an increase inbank capital for most U.K. banks.49 In other Eu-ropean jurisdictions, there is no limit on cov-ered bond issuance (e.g., France).50
So far, only a few Canadian banks have is-sued euro-denominated covered bonds in theEuropean market. While the maximum poten-tial issuance for major Canadian banks is roughly$95 billion in aggregate, the actual issuance todate by Canadian institutions is approximately$6.7 billion. Although the issuance of coveredbonds by these Canadian banks is seen as cost-effective compared with some of their othermarket-based sources of funds, these issueswere generally priced at concessions relativeto covered bonds of similar size issued by
48. See OSFI conditions at <http://www.osfi-bsif.gc.ca/app/DocRepository/1/eng/notices/osfi/cvbnds_e.pdf>.
49. See <http://www.fsa.gov.uk/pubs/international/cbsg_psletter.pdf> for more details on the FSA’sapproach to the supervision of banks in relation totheir covered bond issuance.
50. The volume of covered bonds issued by Spanish banksis limited to 90 per cent of “eligible” assets. The eligiblecollateral pool is constrained to first-lien mortgages,with the loan-to-value ratio capped at 80 per cent forresidential mortgages. An estimate based on a smallsample of Spanish banks indicates that the eligiblecollateral pool typically makes up roughly 50 per centof their total mortgage portfolio.
Financial System Review – June 2008
European banks. This partly reflects a lack of in-vestor familiarity, given that Canadian banksare just entering this market, as well as the yieldspread between “structured” and “traditional”covered bonds.
Compared with other forms of secured mort-gage funding,51 anecdotal evidence indicatesthat, at the time of issuance, the yields on Cana-dian covered bonds were roughly in line with,or were slightly below, those on NHA MBS. Sincelate January 2008, however, the yield spreadson Canadian covered bonds have widenedin relation to those for NHA MBS and senior
51. In most European jurisdictions, under typical marketconditions, the yield on MBS issued by Europeanbanks is typically higher (i.e., higher cost) than thatof covered bonds, owing to, among other things,their lower liquidity and lower investor appeal.
unsecured debt, implying that, at current yieldlevels, any new issuance of covered bonds bya Canadian bank would be a modestly morecostly choice of secured mortgage funding thanNHA MBS.
ConclusionCovered bonds provide financial institutionswith an additional funding vehicle and add tothe geographic diversification of their fundingsources. Since covered bonds also provide an al-ternative avenue of securitization, this shouldlead to the greater availability of lower-costfunding alternatives for financial institutionsand, ultimately, should contribute to the overallefficiency of the market.
37
References
Aaron, M., J. Armstrong, and M. Zelmer. 2007.“An Overview of Risk Management atCanadian Banks.” Bank of Canada Finan-cial System Review (June): 39–47.
Chen, H. 2008. “Macroeconomic Conditionsand the Puzzles of Credit Spreads andCapital Structure.” Available at: <http://www.moodyskmv.com/conf08/papers/macro_ec_cond.pdf>.
Churm, R. and N. Panigirtzoglou. 2005.“Decomposing Credit Spreads.” Bank ofEngland Working Paper No. 253.
Engert, W. 2005. “On the Evolution of theFinancial Safety Net.” Bank of CanadaFinancial System Review (June): 67–73.
Ericsson, J. and O. Renault. 2006. “Liquidityand Credit Risk.” Journal of Finance 61 (5):2219–50.
European Covered Bond Council. 2007. Euro-pean Covered Bond Fact Book. (September).<Available at: http://ecbc.hypo.org/Content/Default.asp?PageID=313>.
Huang, J. and M. Huang. 2003. “How Much ofthe Corporate-Treasury Yield Spread IsDue to Credit Risk? A New CalibrationApproach.” Proceedings of the 14th AnnualConference on Financial Economics andAccounting (FEA), May 2003. TexasFinance Festival.
International Monetary Fund (IMF). 2008.“Market and Funding Illiquidity: WhenPrivate Risk Becomes Public.” GlobalFinancial Stability Report (April): 86–117.
Kiff, J. 2003. “Recent Developments in Marketsfor Credit-Risk Transfer.” Bank of CanadaFinancial System Review (June): 33–41.
Longstaff, F., S. Mithal, and E. Neis. 2004. “Cor-porate Yield Spreads: Default Risk or Liquid-ity? New Evidence from the Credit-DefaultSwap Market.” NBER Working PaperNo. 10418.
Morrow, R. 1994–95. “Repo, Reverse Repoand Securities Lending Markets in Canada.”Bank of Canada Review (Winter): 61–70.
Office of the Superintendent of Financial Insti-tutions (OSFI). 2002. “Guide to Interven-tion for Federally Regulated Deposit-Taking Institutions.” Available at < http://www.osfi-bsif.gc.ca/app/DocRepository/1/eng/practices/supervisory/Guide_Int_e.pdf >.
Packer, F., R. Stever, and C. Upper. 2007.“The Covered Bond Market.” Bank forInternational Settlements Quarterly Review(September): 43–55.
Webber, L. and R. Churm. 2007. “DecomposingCorporate Bond Spreads.” Bank of EnglandQuarterly Bulletin (4): 533–41.
Reports
Financial System Review – June 2008
Introduction
eports address specific issues of rele-vance to the financial system (whetherinstitutions, markets, or clearing andsettlement systems) in greater depth.
The report on Bank of Canada OversightActivities during 2007 under the PaymentClearing and Settlement Act discusses theBank’s activities in 2007 with respect to thethree systems designated in accordance with theAct (the Large Value Transfer System, CDSX,and CLS Bank). This annual report by WalterEngert and Dinah Maclean also reviews otherBank activities related to that role.
In 2007, Canada’s financial system was thesubject of an FSAP (Financial Sector AssessmentProgram) update. This program is an Interna-tional Monetary Fund and World Bank initiativeaimed at helping countries to identify vulnera-bilities in their financial systems and determineneeded reforms. The report on Bank of CanadaParticipation in the 2007 FSAP Macro Stress-Testing Exercise, by Don Coletti, René Lalonde,Miroslav Misina, Dirk Muir, Pierre St-Amant,and David Tessier, contains a description of theBank’s role in the design and implementationof the macro scenario, as well as an indepen-dent assessment of the results provided by theparticipating financial institutions. The authorsreview the benefits of the exercise, as well as thechallenges, and outline directions for futurework.
The financial market turbulence experiencedover the past year has led to concerns over theexcessive reliance on credit ratings to assess thecredit quality of financial instruments and theirissuers. In the report, The Role of CreditRatings in Managing Credit Risk in FederalTreasury Activities, Nancy Harvey and MervinMerkowsky outline how ratings are used in the
R
Bank of Canada’s own activities and in the ac-tivities that the Bank conducts for the Govern-ment of Canada as its fiscal agent. Their keymessage is that the Bank and the governmentadhere to common marketplace practices intheir use of ratings, but are careful not to placeexcessive reliance on those ratings. Other mech-anisms employed to mitigate credit risk includejudgment and collateral requirements.41
Financial System Review – June 2008
Bank of Canada Oversight Activities during2007 under the Payment Clearing andSettlement ActWalter Engert and Dinah Maclean
his report reviews the Bank of Canada’soversight activities during 2007 pursu-ant to the Payment Clearing and Settle-ment Act (PCSA). This is the third in an
annual series of reports aimed at improving thetransparency and accountability of the Bank’sactivities in this area.1
The Bank of Canada has had formal responsibil-ity for the oversight of clearing and settlementsystems in Canada since 1996.2 The PCSA givesthe Bank of Canada this responsibility for thepurpose of controlling systemic risk. In this con-text, systemic risk is defined as the risk that thedefault of one participant in a clearing and set-tlement system could lead, through the activitiesof the system, to the default of other institutionsor systems.
A clearing and settlement system brings togethervarious financial system participants in a com-mon arrangement, such as a clearing house,where the participants are explicitly interlinkedso that the behaviour of one participant can haveimplications for others. In such an arrangement,each participant could face potentially signifi-cant risks and liabilities, depending on the be-haviour of other participants and on the designof the system. As a result, spillover or dominoeffects with broader economic consequencescan occur if the system is not properly designedand operated.
Under the PCSA, the Bank identifies clearingand settlement systems in Canada that could be
1. The authors thank N. Arjani, M. Bonazza, N. Chande,P. Higgins, A. Lai, P. Miller, R. Murray, S. O’Connor,and R. Turnbull for their assistance.
2. A clearing and settlement system is the set of instru-ments, procedures, rules, and technical infrastructurefor the transfer of funds or other assets among systemparticipants.
T
operated in a manner that could pose systemicrisk. Provided the Minister of Finance agreesthat it is in the public interest to do so, thesesystems are designated for oversight by the Bankof Canada, and must satisfy the Bank that theyhave appropriate risk controls in place to dealwith concerns related to systemic risk.3Three systems have been designated by the Bank:the Large Value Transfer System (LVTS), whichdeals with large-value Canadian payments;CDSX, which clears and settles securities trans-actions; and CLS Bank, a global system forforeign-currency transactions.
In the following sections, we discuss variousaspects of the Bank’s oversight work during thepast year. In 2007, the major payment, clearing,and settlement systems continued to evolve in away that supports the stability and efficiency ofthe financial system.
The Large Value TransferSystem
Owned and operated by the Canadian PaymentsAssociation (CPA), the Large Value TransferSystem began operation in February 1999. Dur-ing 2007, it processed, on average, 21,000 trans-actions per day worth approximately $185billion. In its nine years of operation, the LVTShas been a relatively stable system; that is,there have been few significant changes to itsdesign and risk controls. In 2007, the Bank re-viewed five rule changes concerning the LVTS,which were largely technical in nature, includingfurther improvements to contingency arrange-ments and an adjustment to the operation of
3. For a discussion of the Bank’s approach to the over-sight of designated systems, see Engert and Maclean(2006).
43
Reports
the so-called central queue, to improve liquiditymanagement.4
The past year, 2007, was the first full year in whichthe Bank of Canada and the CPA were operatingunder a Memorandum of Understanding (MOU),signed in November 2006. The MOU sets outthe roles and responsibilities of both partiesunder the PCSA and how they intend to worktogether to meet these responsibilities. TheMOU added clarity and structure to the rela-tionship between the Bank and the CPA, andimproved the oversight relationship betweenthe two organizations.
The gains from a well-functioning relationshipwere illustrated in late October and earlyNovember 2007. The CPA discovered a codingerror, which meant that, under certain circum-stances, it was possible, in principle, for a queuedpayment in the LVTS to be processed throughthe system, even though it may not have passedthe LVTS risk controls. Although the availabledata and anecdotal evidence suggest that thelikelihood of such an event was remote, theLVTS was operating without the benefit of allthe risk-control features working as intended.The implications of this could have been serious.
The CPA responded to the situation and workedwith the Bank of Canada and LVTS participantsto implement an interim solution acceptable toLVTS participants and the Bank, until a permanentsolution that corrected the system error was putinto place on 11 November.
Part of the oversight arrangement with regard tothe LVTS, as set out in the Bank’s MOU with theCPA, are regular meetings between the Bank andsenior CPA officials. This allows the Bank andthe CPA to discuss general payments systemdevelopments, as well as potential changes tothe LVTS, early in the process of developing thosechanges. During 2007, the Bank held four suchmeetings with senior CPA officials. The Bankalso met with the Board of Directors of the CPA.
CDSX
Owned and operated by CDS (Clearing andDepository Services Inc.), CDSX clears and settles
4. The central queue is an algorithm that offsets batchesof queued (delayed) payments against each other (ona multilateral basis) at specified intervals throughoutthe day.
44
securities transactions in Canada. On average,in 2007, CDSX processed about 520,000 tradesdaily worth about $250 billion.
In 2007, CDS established a new cross-borderclearing and settlement link with a major secu-rities settlement system in the United Kingdomoperated by Euroclear UK and Ireland Limited(formerly known as CrestCo Ltd.). Through thislink, called Euroclear UK Direct, CDS partici-pants can directly settle U.K. securities transac-tions, facilitating their trading in these securities.
As part of its review of this initiative, the Bankconducted a comprehensive assessment of thelink with the co-operation of CDS, to satisfy it-self that the creation and operation of the linkwould neither pose unacceptable risks to CDSnor compromise the risk controls of CDSX. Thislink became operational on 27 August.
An important development in 2007 was the Bankof Canada’s participation in the Financial SectorAssessment Program (FSAP) of the InternationalMonetary Fund (IMF) and World Bank.5 TheFSAP included a formal review of the complianceof CDSX with internationally accepted standardsfor the design, operation, and regulatory over-sight of securities settlement systems.
Accordingly, the IMF conducted an extensiveassessment of CDSX in collaboration with theBank, provincial regulators, and CDS. The assess-ment concluded that CDSX is
sound, efficient and reliable. The legalbasis for the system’s operation is solid,its functionality is well developed, its riskmanagement procedures to mitigatecredit, liquidity, and operational risks areappropriate, and its governance structureis effective and transparent.
The IMF also made several recommendationsto further enhance CDSX. Most important werethose related to two standards that the IMFassessed CDSX as not observing.6 More specifi-cally, the IMF advised that CDS diversify thesettlement of U.S.-dollar positions arising from
5. For more on this FSAP review, see the IMF website at<http://www.imf.org/external/pubs/ft/scr/2008/cr0859.pdf>.
6. These were the only standards that CDSX wasassessed by the IMF as not observing. The completeIMF assessment of CDSX is available at <http://www.imf.org/external/pubs/cat/longres.cfm?sk=21712.0>.
Financial System Review – June 2008
one of its services, currently concentrated on thebooks of a single, private bank. The Bank ofCanada has accommodated this particular set-tlement arrangement because of the relativelysmall size of potential losses, and the complexity,operational risk, and significant cost of apparentsolutions. The Bank will continue to monitorthe situation, however, and will seek solutionswith CDS if warranted.
The IMF also advised that CDS should notpermit the transfer into Canada of securitiesobtained through its link with the DepositoryTrust Corporation (DTC) in the United Statesuntil after settlement finality is obtained in thatsystem. DTC is a major securities settlementsystem in the United States and, under somecircumstances, DTC has the ability to reverse asecurity transaction until the end of the day ofthe original transaction. This can raise risks forparticipants who have already disposed of thesecurities they are asked to return.
Rather than delay securities transfers from DTCto the day after the original transaction, thusimporting inefficiencies into the Canadian mar-ketplace, CDS has established means to mitigatethe associated risks.7 The Bank’s view is that evenif such transaction reversals were to occur, anyresulting risk is sufficiently mitigated by the riskcontrols that CDS has in place. Moreover, oncea security is transferred into CDSX, the transac-tion is unlikely to be subject to such reversals.
The IMF assessment of CDSX provided a rigorousand useful complement to the Bank of Canada’soversight.8
7. This mechanism assigns the obligation to the CDSparticipant that had received the recalled security. Ifthat participant does not comply, CDS has in placecollateralized credit arrangements that would provideCDS with the funding needed to pay out the obliga-tion or replace the security. Ultimately, any associatedloss would be assigned to participants in a predeter-mined manner, protecting CDS.
8. In 2007, CDS was also assessed by a private custodyand risk-management-rating firm, Thomas Murray,and received a rating of AA, which is deemed “verylow risk overall.” No depository clearing organizationhas received a higher rating from Thomas Murray.(Thomas Murray has rated over 140 securities deposi-tories worldwide, including DTC, the U.S. FederalReserve, and all of the Euroclear group.) The outlookfor this rating is stable, suggesting that there are noforthcoming developments that would be expected toalter the assessment.
A valuable component of the Bank’s oversightof CDSX is the bilateral meetings between theBank and CDS that examine a range of topicsrelated to the operation of CDSX. These meetingsprovide the Bank and CDS with an opportunityto explore any concerns or questions related toproposed changes to CDSX on a timely andefficient basis. In this way, the Bank is alerted topossible significant changes early in the process,and can raise any concerns that it may have, sothat they can be dealt with efficiently by CDSin the process of developing system changes.During 2007, the Bank held two such meetingswith senior CDS officials. Among the issues dis-cussed, in addition to planned CDSX develop-ment, was the IMF assessment of CDSX.
The Bank of Canada approved 48 changes torules and procedures concerning CDSX in 2007.
CLS Bank
CLS Bank clears and settles foreign exchange(FX) transactions in 15 currencies, including theCanadian dollar. In 2007, CLS Bank settled anaverage daily value of US$3.6 trillion, whichincluded Canadian-dollar transactions with anaverage daily value of US$79 billion. Five of thesix major Canadian banks now use CLS Bank asone means of settling their FX transactions.
CLS Bank is overseen collaboratively by thecentral banks whose currencies are included inthe system, with the Federal Reserve Bank ofNew York acting as lead overseer. (CLS Bankis incorporated under U.S. laws, and the vastmajority of FX trades involve the U.S. dollar.)
In 2007, CLS Bank proposed a number of newservices and products that required regulatoryapproval. One of them is the provision of settle-ment services for over-the-counter (OTC) creditderivatives housed in the Depository Trust andClearing Corporation (DTCC) Deriv/SERVtrade information warehouse for which CLSwon a tender offer in December 2006.9 Thiswarehouse is a centralized database of recordsrelating to OTC credit-derivative transactions innine CLS-eligible currencies.
9. DTCC is a U.S. holding company that provides arange of clearing and settlement services through itssix subsidiaries, including clearing and settlement ofU.S. securities.
45
Reports
To this point, the focus of CLS Bank had beenon virtually eliminating FX settlement risk bysettling FX payments on a payment-versus-pay-ment basis.10 The proposal to settle these OTCcredit derivatives meant that CLS Bank wouldbe settling non-FX-related payments. Sincethese are one-way payments from one partici-pant to another (as opposed to two-way pay-ments characteristic of FX transactions), theycannot be settled on a payment-versus-paymentbasis.
In assessing this proposal, the overseeing centralbanks considered whether this would compro-mise the ability of CLS to comply with the CPSS(CommitteeonPaymentandSettlementSystems)core principles for systemically important pay-ments systems, which underpin the oversight ofCLS Bank.11 As well, the central banks also con-sidered broader risk issues associated with thefuture direction of CLS, and possible conflictswith the policies of some central banks if CLSwere to evolve beyond its focus on reducingFX-settlement risk. For example, there was con-cern that CLS Bank could become a general-purpose payments system and reduce the roleof major domestic systems, which could com-promise the ability of some central banks tooversee systemically important payments sys-tems involving their currencies.
The Bank of Canada continues to believe thatthe fundamental principle guiding the oversightgroup in considering such proposals should bethe continued compliance with the CPSS coreprinciples and, in particular, that any new serviceor product offered should not impair the con-trols put in place to manage FX-settlement risk.If these conditions are met, undue barriersshould not be placed on the evolution of CLSBank. It must also be recognized that, given thelarge fixed costs associated with CLS Bank’s coreFX-settlement business, complementary prod-ucts may be useful to amortize these costs andmay provide efficiencies.
10. FX transactions involve the sale of one currency foranother. In a payment-versus-payment arrangement,the currency payable is paid at the same time as fundsdue are received. In this way, risk arising from asyn-chronous exchange of payments (settlement risk) iseliminated.
11. These core principles are available at <http://www.bis.org/publ/cpss43.pdf?noframes=1>.
46
With regard to the proposal to settle these specif-ic OTC credit derivatives, approval was providedby the oversight group. Accordingly, CLS Bankbegan operation for transactions in five of thenine currencies in November 2007. Transac-tions in the other four currencies, including theCanadian dollar, will be phased in over time.
Approval was also given to CLS Bank to offertwo other products: the settlement of paymentsrelated to non-deliverable forward contracts;and the settlement of FX option premiums.12
Payments resulting from these instrumentsinvolve a single currency payment instead of atwo-way currency payment as in a standard FXtransaction.
The central bank oversight group also approvedchanges to the eligibility criteria for new curren-cies and settlement membership. The new crite-ria for currency eligibility will allow a currencywhose sovereign credit rating is below invest-ment grade to be settled in CLS Bank. However,such currencies cannot be used as collateral tosupport net debit positions within the system.
The minimum credit rating required for settle-ment membership was also lowered to belowinvestment grade. Settlement members ratedbelow investment grade cannot owe moneywithin the system, however, but are required tofund settlement of their payment instructionson a cash basis.
These latter two changes have the potential toreduce settlement risk among a broader rangeof currencies and trading institutions than thosepreviously eligible, without adversely affectingthe safe and efficient functioning of the CLSsettlement service.
Finally, an important outcome of the workamong the overseeing central banks regardingCLS Bank’s evolving business strategy has been
12. Non-deliverable forward contracts differ from nor-mal forward FX contracts in that there are no transfersof the principal currency sums between counterparties.Rather, on the contracted settlement date, profit/lossis calculated based on the difference between thecontracted exchange rate and the prevailing spotexchange rate, and a net cash payment is made by theparty suffering the loss to its counterparty, often inU.S. dollars. Such contracts are commonly used tohedge against currency risks in emerging markets. AnFX option premium is the price the buyer of anoption contract pays for the right to buy or sell a cer-tain currency for a specified price in the future.
Financial System Review – June 2008
a strengthening of the co-operative oversightarrangement that provides enhanced focus andresources with regard to CLS Bank develop-ments.
August Financial Turbulence
The turbulence in financial markets that beganin August 2007 led to increased volumes andactivity for each of the designated systems. Al-though there were notable volume increases inthe LVTS on some days in mid-August, therewere no capacity problems, and the systemfunctioned smoothly. Some system participantstemporarily increased their collateral pledged inthe system to support the higher value of LVTStransactions associated with intensified tradingin markets and increased transactions in otherclearing and settlement systems.13
The drying up of liquidity in the market for asset-backed commercial paper (ABCP) resulted in anumber of defaults and extensions of entitlementpayments related to securities held in CDSX.14
On 13 August 2007, approximately $2 billionin maturities of ABCP held at CDS was not paid.Issuers of this ABCP had either to extend thematurities of this paper, where this right existed,or to leave the maturities unpaid.
In the event, about $1.6 billion of ABCP wasleft unpaid, and another $500 million was ex-tended. CDS was not exposed to any financialrisks arising from the unpaid maturities, becauseCDS does not execute an entitlement paymentin CDSX unless it is prefunded by the issuers’paying agent. On 14 August, CDS took stepsto assist issuers and participants holding thedefaulted paper, including facilitating directinteraction between issuers and participants, toreach mutually agreeable solutions. It also en-abled procedures to process partial paymentson maturing ABCP. As well, CDS issued dailybulletins to system participants to provide in-formation on unpaid and extended maturities,and on the evolving value of unpaid ABCP heldat CDS.
13. Collateral is pledged by LVTS participants to the Bankof Canada to help manage risk in the system.
14. Entitlements include dividends, interest, paymentupon redemption or maturity, and other payments ordistributions to holders of securities. Entitlementsmay be distributed in the form of a money paymentor as a distribution of securities or other property.
Increased trading volumes in FX markets causedmany market participants, including CLS Bank,to face internal capacity pressure. CLS Bankrequired two settlement extensions on 17 August,in the Australian dollar and Korean won. Peak-input volumes received on 16 and 17 Augustgave rise to peak-settlement days on 20 and21 August, which were both record volume daysfor CLS Bank. Nevertheless, settlement and fund-ing occurred within specified timelines on bothof those days. CLS Bank had other peak-volumedays after August, notably on 13 November,with over 1.4 million currency sides processed.Settlement proceeded smoothly on all of thesepeak-volume days.
In sum, the core payment, clearing, and settle-ment infrastructure functioned well during theperiod of market turbulence in 2007.
Other Oversight-RelatedActivities
CPA and other payments-systemthemes
In 2007, the CPA launched a strategic review ofthe LVTS and, more generally, is working to de-velop a long-term strategy for the evolution ofthe payments system in Canada, taking accountof evolving needs and prospective developments.Such planning can provide helpful leadershipin addressing important issues affecting thestability and efficiency of the Canadian financialsystem. Through the CPA’s consultation withthe Bank on these matters, the Bank undertookto enhance the ongoing exchange of researchand analysis between the two organizations.
As part of the regulatory framework governingthe Automated Clearing Settlement System(ACSS), the federal Department of Financeregularly consults the Bank with regard to pro-posed rule changes and other developmentsaffecting the ACSS.15 In this regard, the Bankreviewed nine ACSS bylaw and rule changesduring the year.
15. The ACSS, which is owned and operated by the CPA,is used for payments not handled by the LVTS: gener-ally, small-value items, such as paper cheques, auto-mated bill payments, and debit card transactions. TheACSS is subject to oversight by the federal Depart-ment of Finance.
47
Reports
More generally, to facilitate the Bank’s interactionwith the Department of Finance on paymentsissues, including broader policy developmentsaffecting payment, clearing, and settlementsystems, senior and working-level officials meeton a quarterly basis to exchange informationand analysis. Among the issues considered in2007 were the CPA’s medium-term strategicplan, effects of the financial market turbulenceon core payments infrastructure, and work relat-ed to the IMF FSAP.
Central bank committees andworking groups
The Bank is an active member of the Committeeon Payment and Settlement Systems. The CPSSis a committee of central bankers that collabora-tively sets standards that guide oversight policiesaround the world. The CPSS also conducts anal-ysis and research on a range of issues relevantto clearing and settlement systems. (For moreon the CPSS, see <http://www.bis.org/cpss/index.htm>.)
In 2007, Bank staff were also actively involvedin two working groups established by the CPSS.One working group published a consultativereport in July on a global survey conducted in2006 on the management of FX-settlement riskat major international banks. The results showthat CLS Bank was used to settle 55 per cent ofthe value of the settlement obligations in thesurvey, contributing significantly to a reductionin FX-settlement risk.16 The analysis also suggeststhat further reductions in FX-settlement risk canbe made by making greater use of payment-versus-payment settlement methods (as providedby CLS Bank), increased reliance on bilateralnetting arrangements, and, in some cases, sim-ply through better measurement and control ofsettlement risk exposures. The CPSS is expectedto publish the final report in mid-2008.
A second CPSS working group has been analyz-ing interdependencies that exist between pay-ment and securities settlement systems in CPSScountries, as well as the potential role of theseinterdependencies in the transmission of riskacross systems and across countries. The group’swork highlights the fact that global paymentand settlement infrastructure is becoming moreinterdependent. Tighter direct relationships
16. The results for the Canadian banks included in thesurvey are summarized in Arjani (2007).
48
between systems, and more indirect relation-ships arising from the activities of large financialinstitutions and the use of common third-partyservice providers, have contributed to thistrend.
These relationships have strengthened the globalinfrastructure by reducing costs and diversifyingrisks, yet they also pose challenges by increasingthe potential for disruptions to spread widelyand quickly. Therefore, the increasing interde-pendence of payment and settlement systemscalls for broad risk-management perspectivesand coordination among stakeholders. As well,risk-management controls should be commen-surate with the roles of the system, institution,or service provider in the global infrastructure.The working group will be considering morespecific policy implications in 2008.
SWIFT
The Bank also continues to participate in theco-operative oversight of the Society for World-wide Interbank Financial Telecommunication(SWIFT). SWIFT is the principal payment-mes-saging service provider for financial institutionsaround the world and for critical systems, suchas the LVTS and CLS Bank. This co-operativeoversight group monitors and assesses the extentto which SWIFT maintains appropriate gover-nance arrangements, operational processes, riskmanagement, and controls to effectively addresspotential concerns that may arise for financialstability.
Business-continuity planning(BCP)
The Bank of Canada works with the operatorsand participants of systemically importantCanadian clearing and settlement systems toenhance arrangements for continuity of opera-tions. These systems are at the centre of Canada’sfinancial system, and serious economy-widerepercussions could arise if their operationswere not extremely reliable.
One of the key conclusions from the Joint BCPWorking Group in 2006 was the importance ofachieving a priority-recognition status for majorclearing and settlement systems from federaland provincial organizations that have respon-sibilities for emergency management (Goodlet2007). Recognition of priority access to essentialinputs such as electricity, diesel fuel, or municipal
Financial System Review – June 2008
services during a seriously disruptive event is animportant component of these systems and ofthe Bank’s BCP work.
In 2007, the Bank of Canada and other agenciescontinued to follow up on these recommenda-tions. In particular, the Bank, Finance Canada,and the system operators have been workingwith the Province of Ontario to achieve prioritystatus for critical clearing and settlement infra-structure.
The Bank was also active during the year toimprove its own business-continuity arrange-ments. The Bank conducted two integrated ITdisaster-recovery exercises under conditionsthat simulated a serious operational disruptionat the Bank. The test results were satisfactory,with the Bank able to meet its recovery-timetargets. In addition, a large-scale integratedbusiness-recovery exercise was conducted at theBank’s recovery site, which provided usefullessons for the Bank’s BCP planning.
High-availability banking system
As noted in last year’s oversight review (Goodlet2007), the Bank of Canada is committed toimproving its ability to deliver its unique bankingservices to critical clearing and settlement systemsand financial institutions on a high-availabilitybasis.
In 2007, significant progress was made in themulti-year redevelopment of a high-availabilitydelivery system for these banking services. Finalpreparation and testing are expected to continuein 2008, with implementation of the new systemtargeted for autumn 2008.
Publications in 2007
During 2007, the Bank published the followingstaff work related to clearing and settlementsystems:
• Arjani, N. and W. Engert. 2007. “The Large-Value Payments System: Insights fromSelected Bank of Canada Research.” Bank ofCanada Review (Spring): 31–40.
• Ball, D. and W. Engert. 2007. “UnanticipatedDefaults and Losses in Canada’s Large-ValuePayments System, Revisited.” Bank of CanadaDiscussion Paper No. 2007-5.
• Chapman, J. T. E. and A. Martin. 2007.“Rediscounting under Aggregate Risk withMoral Hazard.” Bank of Canada WorkingPaper No. 2007-51. This work was summa-rized in the Bank of Canada Financial SystemReview, December 2007, under the title, “TheProvision of Central Bank Liquidity underAsymmetric Information.”
• Chiu, J. and A. Lai. 2007. “Modelling Pay-ments Systems: A Review of the Literature.”Bank of Canada Working Paper No. 2007-28.This work is summarized in the Bank ofCanada Financial System Review, June 2007.
References
Arjani, N. 2007. “Management of ForeignExchange Settlement Risk at CanadianBanks.” Bank of Canada Financial SystemReview (December): 71–78.
Engert, W. and D. Maclean. 2006. “The Bank ofCanada’s Role in the Oversight of Clearingand Settlement Systems.” Bank of CanadaFinancial System Review (June): 57–63.
Goodlet, C. 2007. “Bank of Canada OversightActivities during 2006 under the PaymentClearing and Settlement Act.“ Bank ofCanada Financial System Review (June):33–37.
49
Financial System Review – June 2008
Bank of Canada Participation in the 2007FSAP Macro Stress-Testing ExerciseDon Coletti, René Lalonde, Miroslav Misina, Dirk Muir, Pierre St-Amant,and David Tessier
n 2007, Canada’s financial system was thesubject of an FSAP (Financial Sector As-sessment Program) update (Box 1). A keycomponent of this exercise was the stress
testing of the Canadian financial system with amacroeconomic scenario. The Bank of Canadatook the lead in developing, implementing, andassessing the results. This report describes therole played by the Bank and the main results of,and lessons derived from, this exercise.1
General Framework
The purpose of stress testing is to assess the re-silience of a segment of the financial system inthe face of “rare but plausible” events that haveeither resulted in vulnerabilities in the past orcould do so in the future. In macro stress test-ing, the events considered are macroeconomicshocks assembled (typically by means of a mac-roeconomic model) to form a macroeconomicscenario. The objective is to assess the impact ofthe scenario on a set of financial institutions.
There are two basic approaches to conductingmacro stress-testing exercises:
i) Bottom-up, in which the participating institu-tions assess the impact of a given scenario ontheir portfolio, and the authorities then aggre-gate and interpret the results.
ii) Top-down, in which the authorities assess theimpact of the scenario on financial institutionsat a more aggregate level and use the results as abasis for discussion with individual participants.
In the Canadian FSAP update, it was agreed thatmacro stress testing would be conducted usingthe bottom-up approach, and that the subject ofthe stress test would be the loans portfolio of theparticipating Canadian banks. The Bank of Can-ada played several roles in the exercise:
1. For a more detailed analysis, see Coletti et al.(forthcoming).
I
• designing the macroeconomic scenariousing an in-house model;• modelling the relationship between macrovariables and default rates in the businessand household sectors (the individual insti-tutions used these default rates to simulatethe losses in their loans portfolio);
• conducting an independent assessment oflosses that would arise under the scenario,to cross-check the results of individual banks.2
In the rest of this report, we elaborate on thesepoints by describing the general features of thechosen scenario, the method used to relate de-fault rates to macroeconomic variables, and theresults that were obtained. The last section con-tains a summary of the results, some thoughtson the lessons learned, and an outline of theareas where further work is needed.
The Macroeconomic Scenario
In 2006, during the preparation stage of the ex-ercise, it was agreed that the scenario to be con-sidered would be based on a disorderly adjust-ment of global imbalances brought about by adownward revision to expectations for produc-tivity growth in the United States. The impact ofsuch a scenario would be traced over a 2-yearhorizon, starting in 2007Q1.
Our scenario has its origins in the historicallyhigh level of trend labour productivity growthexperienced in the latter half of the 1990sand the early 2000s in the United States.3 As
2. The IMF has conducted its own independent assessment,the results of which are in their FSAP report at <http://www.imf.org/external/pubs/cat/longres.cfm?sk=21710.0>.
3. The macroeconomic scenario was developed using theBank of Canada’s version of the Global Economy Model(BoC-GEM) (Lalonde and Muir 2007). BoC-GEM is avariant of the GEM developed at the IMF (Faruqee et al.2007).
51
Reports
52
Box 1
FSAP Update: Highlights ofthe IMF Conclusions
The Financial Sector Assessment Program (FSAP),establishedby the InternationalMonetaryFund(IMF)and the World Bank in 1999, provides countries withcomprehensive assessments of the stability of theirnational financial systems. Canada undertook anFSAP assessment during the program’s pilot phasein 1999–2000. In September 2006, Canadian au-thorities formally requested that the IMF undertakean FSAP update for Canada.
This update included an assessment of Canada’s com-pliance with internationally accepted standards andcodes for financial sector regulation, as well as afocused review of Canada’s compliance with revisedBasel Core Principles for Effective Banking Supervi-sion. It also included a stress-testing exercise designedto assess the capacity of the Canadian financial systemto absorb various adverse economic and financialshocks.
The full FSAP report can be found at <www.imf.org/external/pubs/ft/scr/2008/cr0860.pdf>. The IMF’soverall conclusion is that Canada’s financial systemis mature, sophisticated, and well managed. The re-port states that financial stability is underpinned bysound macroeconomic policies and strong pruden-tial regulation and supervision, while deposit insur-ance and arrangements for crisis management andfailure resolution are well designed.
The IMF’s conclusion concerning the stress-testingexercise is that Canada’s major banks can withstandsizable shocks. Although capital drops below the reg-ulatory minimum in the stress scenario, it remainsadequate. The banking system thus appears sound,but it faces some challenges. In particular, the IMFnotes that global financial turmoil since mid-2007has highlighted the information and liquidity risks inthe structured financial products that Canadianbanks have embraced in recent years.
The report also concludes that CDSX, the securities-settlement system operated by Clearing and Deposi-tory Services Inc., is sound, efficient, and reliable andthat it complies with almost all recommendations forsecurities-settlementsystems.Aswell,Canadais foundto be compliant with the four revised Basel CorePrinciples for Effective Banking Supervision. Finally,concerning securities regulation, the IMF concludesthat the regulatory framework for the securitiesmarketin Canada implements the objectives and principlesof the International Organization of Securities Com-missions (IOSCO) in most respects, but that therewould be advantages in moving towards a singlesecurities regulator.
expectations of long-term growth in U.S. labourproductivity were gradually revised upwards to2 per cent and higher, perceived rates of returnon U.S. investments were boosted, leading toincreased investment demand, as well as in-creased capital inflows and a stronger U.S.dollar. In addition, expectations of higher per-manent income led to an increase in consump-tion and a drop in the savings rate. In turn, thesefactors led to a rise in imports and an expansionof the U.S. current account deficit (Ferguson 2005).
In the scenario, it is assumed that expectationsof a permanent rise in the growth of U.S. labourproductivity are revised sharply downwards to1.1 per cent per annum for the next 10 years from2 per cent. The resulting downward revision topermanent income growth and to expected rates ofreturn on investment lead to a retrenchment indemand that offsets the decline in the growth ofthe economy’s productive capacity. It is also as-sumed that increased economic uncertainty causesdeclines in the confidence of consumers and firms,leading to a reduction in consumption and in-vestment expenditures. Heightened uncertaintyis also assumed to motivate foreigners to sell offU.S.-dollar assets, resulting in a rapid depreci-ation of the U.S. dollar. The resulting deteriora-tion in the balance sheets of consumers and firmsleads to a significant rise in financial risk premi-ums, further magnifying the economic slowdown.4
Furthermore, it is assumed that Canadian trendlabour productivity growth slows to about0.8 per cent per annum over the next 10 yearsfrom 1.5 per cent. As in the United States, we as-sume similar but smaller declines in consumerand business confidence. Premiums on Canadi-an commercial interest rates also rise as a resultof the economic downturn, exacerbating theweakness in Canadian GDP growth.5
4. We assume that the risky spread (the differencebetween the rate on medium-term business loansand the rate on 5-year U.S. government bonds) widensto reach historic highs in early 2007 (the starting pointof the shock scenario). Our analysis assumes similarincreases in the spreads on consumer interest rates.
5. We assume that the risky spread between the rate onmedium-term business loans and the rate on 5-yearGovernment of Canada bonds widens to historicallyhigh levels. Risky spreads of this magnitude haveoccurred only about 2 per cent of the time in Canadaover the period 1980 to 2006. The rate on medium-term Canadian business loans is based on the yieldon bonds of A-rated firms, which represents themedian rating of Canadian corporate bond issuers.We assume similar increases in spreads on consumerinterest rates.
Financial System Review – June 2008
Table 1
Macro Stress-Testing Scenario (2007Q1–2009Q1)a
a. The values are year-over-year (Q1 over Q1).
2007Q1 2008Q1 2009Q1
United States
Trend labour productivity growth 1.9 0.9 1.1
GDP growth 2.2 -6.2 3.2
Unemployment rate (level) 4.5 7.7 8.0
Core CPI inflation 2.3 -2.0 -0.7
Real commodity price index(US$; 1997=100)
194 180 177
Real effective exchange rate(+=depreciation)
1.06 1.19 1.18
Federal funds rate 5.3 0.7 2.1
Rate on 5-year government bonds 5.1 3.2 4.1
Rate on medium-term business loans 7.6 8.4 7.7
Canada
Trend labour productivity growth 1.1 0.5 0.8
GDP growth 1.9 -4.3 2.0
Unemployment rate (level) 6.2 8.0 9.4
Core CPI inflation 2.2 -0.6 0.3
Real bilateral U.S. exchange rate(+=depreciation)
1.15 1.05 1.08
Target overnight rate 4.2 0.2 1.0
Rate on 5-year government bonds 4.4 1.9 2.6
Rate on medium-term business loans 5.0 4.0 3.8
Taken together, the shocks in our scenario areextremely large by historical standards. Asshown in Table 1, the scenario results in highreal interest rates for consumer and businessloans and an extremely severe economic con-traction in the United States. The recession em-bodied in the scenario is even more severe thanthat experienced in 1981–82, with year-over-year real GDP growth in the United Statestroughing at -6.2 per cent in 2008Q1. As a resultof the weakness in aggregate demand, U.S. labourmarket conditions deteriorate and the unem-ployment rate rises to a peak of about 8.5 percent in mid-2008.
The recession in the United States, a higherCanada/U.S. real exchange rate, falling worldcommodity prices, the downward revision toexpectations for the growth of domestic trendlabour productivity, losses in domestic consumerand business confidence, and the rise in domes-tic financial risk premiums lead to a significantrecession in Canada in 2007 (Table 1). In termsof cumulative output loss, the domestic recessionembodied in the scenario is about one-thirdlarger than that experienced in 1990–91, despitea significant easing in monetary policy.
As a result, core consumer prices fall throughout2008. Inflation picks up gradually in 2009 andreturns to the 2 per cent inflation target by theend of 2010. Canadian policy interest rates fallquickly to 0.25 per cent in early 2008 and re-main very low for several years. The aggressivedecline in policy rates, consistent with Canada’sinflation-targeting framework, plays a veryimportant role in mitigating the impact of theadverse shocks on the Canadian economy. Con-sequently, the rate on 5-year Canadian medium-term business loans actually declines in early2007, despite the large rise in the financial riskpremium. The relatively low level of domesticinterest rates at a time when GDP growth isquite weak distinguishes the macroeconomicoutlook from the events in the early 1990s.
53
Reports
Table 2
Peak Default Rates (Scenario and Historic Peaks)
History(peak date)
BoC model(peak date) Scenario
Scenario+0.25
standarddeviation
Accommodation 7.58(1992Q1)
6.26(1992Q1)
12.3 13.75
Agriculture 0.83(1992Q4)
0.78(1992Q1)
1.37 1.61
Construction 3.27(1992Q4)
3.61(1991Q2)
5.63 6.38
Manufacturing 8.28(1992Q1)
8.36(1992Q2)
11.1 12.22
Retail 5.31(1992Q1)
5.17(1992Q2)
3.76 4.31
Wholesale 4.63(1992Q1)
4.73(1992Q2)
6.58 7.42
Mortgage 0.63(1996Q4)
0.59(1996Q3)
0.55 0.57
54
Relating Default Rates toMacro Variables6
In modelling sectoral default rates, the objectivewas to identify systemic factors that affect defaultrates in all sectors. We assume that these factorsare related to the overall performance of theeconomy. The initial set of explanatory variablesincludes the Canadian GDP growth rate, un-employment rate, interest rate (medium-termbusiness loan rate), and the credit/GDP ratio.The paths of these variables under stress are ob-tained from the macro scenario.
The dependent variables are sectoral defaultrates. Because long time series of historical sectoraldefault rates with broad coverage are not availablefor Canada, proxies are constructed based onsectoral bankruptcy rates, supplemented byadditional information. The adjusted data spanthe period from 1988Q1 to 2005Q4, at aquarterly frequency.
The specification of the sectoral regressionsincludes non-linear terms. We find that non-linearities are the key to capturing the behav-iour of default rates around the historical extremes(MisinaandTessier2007).Withoutnon-linearities,even the extreme macroeconomic shocks have avery limited impact on default rates.
Simulation results are presented in Table 2, whichcontains information on historic peaks, as wellas fitted values from the non-linear specification.7
Fitted values are, on average, close to historicalpeaks. The values used in the stress-testing exercisecontain an ad hoc upward adjustment (equivalentto 0.25 standard deviation) reflecting the IMF’sopinion that the magnitudes of responses shouldbe larger than those generated by the default-rate models.8
6. Technical details related to this section can be found inMisina and Tessier (2007).
7. The sectors included were accommodation, agriculture,construction, manufacturing, retail, wholesale, anddefault rates on mortgages in the household sector. Sec-tors for which default rates were not provided couldeither be merged with the above or classified in a sepa-rate category, and an average of defaults in the above sec-tors used. Both approaches were used by individualbanks in implementing the scenario.
8. The performance of sectoral regressions under stress willdepend on the precise configuration of the macroeco-nomic variables in a particular sectoral regression. Duringthe recession in the early 1990s, interest rates were muchhigher than under the scenario. Thus, for interest-sensi-tive sectors, such as retail and mortgages, the situationunder the stress scenario is more favourable than in theearly 1990s, resulting in responses that are below his-toric peaks.
Financial System Review – June 2008
To compute losses under the stress scenario, weneed information on exposures and on loss-given-default, in addition to default rates. Data onloan exposures are taken at 2006Q4, the lastavailable point at the time of the exercise.
There is very little information on loss-given-default in Canada. A rough proxy can be obtainedby looking at the ratio of estimated assets toestimated liabilities at the time of bankruptcy.This information is available from the Officeof the Superintendent of Bankruptcy. For thecorporate sector, the average for the period1988–2006 is 0.35, which would suggest anexpected recovery rate of 35 per cent, or lossesgiven bankruptcy of 65 per cent. Since bankruptcyis the last stage of distress, and since most lossesoccur because of missed interest payments, webelieve that this recovery rate might be somewhatlow and, for the purpose of the FSAP exercise,have agreed with the IMF to set the recovery rateat 50 per cent.
Loss Assessment
Expected and unexpected losses
Each participating bank9 was asked to providean estimate of expected and unexpected losses(the mean of the loss distribution, and the99.9 per cent value-at-risk) arising from themacroeconomic scenario, for each quarter overa 2-year horizon. Individual results depend onhow the banks used the inputs provided tothem. While approaches vary across banks,some key commonalities had to be taken intoaccount in arriving at the estimates of lossesusing our internal model. In particular:
• The banks’ estimates of losses are based ontheir estimates of exposures at default(EAD), and these data were not publiclyavailable at the time of the exercise.10 Theseestimates are larger than the publicly avail-able balance sheet loan values, since thelatter are based on the drawn amounts,whereas the former take into accountundrawn commitments. Consequently, the
9. These are the “Big 5” banks: CIBC, RBC Financial Group,Bank of Montreal, TD Bank Financial Group, and Scotia-bank.
10. Under Basel II reporting rules, banks will be required toprovide this information.
use of the balance sheet exposures will resultin systematically lower estimates of losses.
• The banks’ results indicate that the loss dis-tribution has fat tails.
To deal with the first set of issues, the bankswere asked to provide the values of EAD thatthey used in their stress tests. This informationwas used to adjust the results of the Bank ofCanada’s internal model by the difference be-tween the exposures they used and those intheir balance sheets. The difference between thetwo varied from bank to bank, with the lowestbeing 4 per cent, and the largest being 45 percent.11 To deal with the second set of issues,the simulations are performed using a t-distri-bution (with four degrees of freedom), ratherthan the normal distribution.
Chart 1 is a summary of the impact of the sce-nario on the participating banks. It containsestimates of expected and unexpected lossesbased on the individual results provided by thebanks, Bank of Canada’s estimates based on ourinternal stress-testing model (labelled BoC),and the IMF’s estimates, which are based on theresults contained in the IMF’s FSAP report.12
The banks’ estimates of their expected losses arehigher than ours in the first year and lower inthe second year. Estimates of the unexpectedlosses are similar in terms of overall magnitude.
An examination of the results reveals that in-creases in losses are driven largely by develop-ments in the retail, manufacturing, and servicessectors. (In this stress test, the retail sector includesconsumer loans.)
The IMF’s estimates are somewhat different, be-ing either below (expected losses) or above (un-expected losses) the other two sets of estimates.
11. Scaling up the results by the difference between expo-sures is implicitly based on the assumption that the sec-toral distribution of exposures used by the bankscorresponds, in relative terms, to their balance sheetexposure. This assumption is difficult to verify, since thebanks provided only total exposures.
12. The IMF reports losses as a percentage of risk-weightedassets. These were converted into dollar amounts byusing publicly available information on the values ofrisk-weighted assets.
55
Reports
Chart 1 Losses under the Scenario
$ millions
Expected losses (reported)Expected losses (BoC)Expected losses (IMF)
2007 20080
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
10,000
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
10,000
Unexpected losses (reported)Unexpected losses (BoC)Unexpected losses (IMF)
Chart 2 Impact of the Scenario on the Capital-Adequacy Ratio
%
2007 20084
5
6
7
8
9
10
11
12
4
5
6
7
8
9
10
11
12
Banks’ resultsCross-check (BoC)Cross-check (IMF)
Basel IIrequirement
Impact on the capital-adequacyratio
The ability of the banks to absorb the losses thatarise under the scenario can be assessed basedon the impact on their capital-adequacy ratio(CAR). That assessment is based on the fact thattotal capital (Tier 1 + Tier 2) should be sufficientto cover the unexpected losses (at the 99.9 percent value-at-risk).
The results are presented in Chart 2, whichshows the average of the results reported bybanks, our estimates, and the IMF estimates.The horizontal line represents Basel II require-ment for total capital (8 per cent). The resultssuggest that if the unexpected losses material-ized at any point over the stress-testing horizon,the CAR would fall below the 10 per centthreshold for total CAR set by the Office of theSuperintendent of Financial Institutions. If theunexpected losses materialized after the fourthquarter of the scenario, the results indicate thatthe CAR would fall below the Basel II require-ments.13 The results do vary across banks, andin the best case the CAR remains above 8 percent throughout the exercise.
The above analysis is based on the assumptionof zero growth in regulatory capital over thestress-testing horizon (no mitigating action bythe management to reduce the impact of thescenario). This assumption results in very con-servative (worst-case) estimates. Some banksalso assessed the impact of the shock on theirCAR allowing for management action. Theseassessments were based on a variety of assump-tions, but in all cases, the estimated CAR understress remained above the regulatory require-ments.
There is no doubt that the banks would use avariety of measures to maintain their CAR abovethe stipulated threshold (e.g., reduce/halt divi-dend payments or change lending practices toinclude fewer risky borrowers). Nonetheless,banks’ estimates may be overly optimistic, sincesimultaneous actions by all banks to managetheir capital (e.g., tightening credit) may havea negative impact on the real economy andexacerbate the problem.
13. Since the default rates reach their peaks around2009Q1, we expect the biggest impact on the banks’CAR to occur around that time, with a possibleimprovement later on.
56
Financial System Review – June 2008
Conclusion
With a recession that is one-third larger thanthat experienced in the early 1990s, and increas-es in defaults exceeding predictions made byeconometric models, the stress-test scenario an-alyzed in the FSAP update is extreme. Nonethe-less, in terms of severity, the Canadian scenariois broadly in line with the scenarios used in otherdeveloped countries.
In the event of the materialization of the unex-pected losses implied by the scenario, the banks’capital is, on average, sufficient to absorb theselosses, although the average CAR falls belowthe regulatory requirements. In such circum-stances, banks would need to take action eitherto raise new capital, which could be difficultgiven the high financial stress assumed in thescenario, or to cut risk-weighted assets, whichcould be costly for the economy. Authorities,particularly monetary policy authorities, wouldneed to factor in these impacts on banks in de-termining the appropriate policy response.
While modelling and data differences make itdifficult to compare the outcomes of stress testsacross countries, generally speaking, the resultsindicate that the losses, while not negligible, didnot pose a systemic threat to the financial sys-tems of the countries tested. The results for Can-ada are in line with these general findings.
From the perspective of the Bank of Canada’songoing work to develop tools to assess theresilience of the financial sector and its impacton financial stability, the exercise was valuable,leading to:
• a deeper understanding of the complexitiesof the relationship between the micro andmacro aspects of the analysis of financialstability;
• improvements in the Bank’s internal stress-testing models and methodologies;
• increased awareness of data limitations anddata requirements;
• improved information sharing between thebanks and Canadian government agencies;and
• increased knowledge of risk-managementpractices at individual financial institutions.
The exercise has also revealed the limitations ofthe existing stress-testing tools and methodolo-gies, as well as the need for continued improve-ment, including:
• developing macro models that would bebetter suited to the analysis of extremeevents;
• further work on models that relate defaultrates to macroeconomic variables (betterintegration with the main macro model andexplicit modelling of economic behaviours);and
• gathering data on defaults beyond those oflarge publicly traded companies and gettingmore comprehensive data on the exposuresof financial institutions.
Difficult issues at the frontier of current researchefforts reflect broader problems: lack of the sec-ond-round/feedback effects that relate the ac-tions of financial institutions back to financialmarkets and the real economy; the interlinkagesamong financial institutions; and channels ofcontagion. Nonetheless, we think that, even atthe present stage, properly designed stress tests,based on scenarios that reflect rare but plausiblesources of stress, can be useful in identifyingvulnerabilities in the system. This informationcan help to guide official institutions in lookingmore deeply, in a risk-focused way, at possiblechannels through which vulnerabilities in onebank could be transmitted to others in the sys-tem. It can also form the basis for supervisory/macroprudential guidance to individual institu-tions or to the market as a whole.14
The FSAP has stimulated the Bank of Canada’sstress-testing work. We are planning to main-tain this accelerated momentum by investingin two priorities: (i) research efforts to addresssome of the shortcomings identified in the pre-vious paragraph, and (ii) regular updating, us-ing various scenarios of interest, of the top-down approach (Box 2). We are also consider-ing, together with financial institutions and oth-er government agencies, the possibility of
14. We thank Karl Habermeier and Mark Swinburne ofthe IMF for insightful comments and for providing abroader perspective regarding the nature of theseexercises.
57
Reports
Box 2
Macro Stress Testing: An Update of the FSAP Results
The FSAP exercise discussed in this report was 11.26 prior to stress to 7.69 in that quarter. If,
Chart A Capital-Adequacy Ratio under Stress
%
2
4
6
8
10
12
14
2
4
6
8
10
12
14
FSAP estimate2008Q1 estimate
2. The methodology used here does not allow usto infer the CAR values in the quarters followingthat in which the unexpected loss first materializes.
based on data for 2006Q4. Here we present anupdated top-down assessment of credit lossesin the loans portfolio based on data for2008Q1.
Highlights of the changes in the data since2006Q4
In the aggregate, there has been little change inthe composition of banks’ loans portfolios.Exposure to mortgage and retail sectors (thelatter includes consumer loans) continues toaccount for approximately 75 per cent of totallending.
• The banks’ loans portfolios have increasedby 12 per cent on average.
• On average, the capital-adequacy ratio(CAR) of total (Tier 1 + Tier 2) capital torisk-weighted assets has declined slightly.
• New data on exposures at default (EAD),which the banks started reporting in2008Q1, obviates the need to rely onadjustments to balance sheet exposures, aswas done in the FSAP exercise.
Results
Chart A summarizes the results of the stress-testing exercise that takes the above points intoaccount. Losses under the scenario are estimatedusing our internal model and are convertedinto an impact on the CAR using the samemethodology as in the FSAP exercise. The firstbar represents the median values of CAR for thefive banks prior to stress. The subsequent barsrepresent the level of the CAR in each quarter ifthe unexpected losses (i.e., 99.9 per cent valueat risk) associated with the scenario were tomaterialize.1 For example, if the scenario-induced unexpected losses were to materializein the fourth quarter after the initial shock, themedian value of the CAR would change from
58
1. Recall that the actual CAR is determined on thebasis of the estimates of the unexpected losses.
instead, the unexpected losses materialized inthe sixth quarter after the shock, the medianvalue of the CAR would change from 11.26prior to stress to 6.68 in that quarter.2
Higher exposures and a lower CAR startingpoint for values produce somewhat lower val-ues for CAR under the scenario than in the2006Q4 exercise. The levels of capital at thesebanks in 2008Q1 would still, in the aggregate,be sufficient to absorb the losses. Nonetheless,to continue to meet the regulatory require-ments (total CAR of 8 per cent under Basel II;OSFI threshold of 10 per cent), the bankswould have to take appropriate action: raisingtheir levels of capital, reducing their risk-weighted assets, or a combination of the two.The feasibility and broader implications ofthese adjustments would depend on the natureand timing of these actions, as well as the hori-zon over which the banks’ CAR values wouldreturn to their target levels.
No +2 +4 +6 +8+1 +3 +5 +7 +9
0 0
stress
Financial System Review – June 2008
conducting periodic bottom-up exercises, aswell as developing a more comprehensiveapproach to stress testing.
References
Coletti, D., R. Lalonde, M. Misina, D. Muir, andP. St-Amant. “Bank of Canada’s Participa-tion in the 2007 FSAP Macro Stress-TestingExercise.” Bank of Canada DiscussionPaper. Forthcoming.
Faruqee, H., D. Laxton, D. Muir, and P. Pesenti.2007. “Smooth Landing or Crash? Model-Based Scenarios of Global Current AccountRebalancing.” In G7 Current Account Imbal-ances: Sustainability and Adjustment, 377–451, edited by R. Clarida. National Bureauof Economic Research Conference Report.Chicago: The University of Chicago Press.
Ferguson, R. W. Jr. 2005. “U.S. Current AccountDeficit: Causes and Consequences.”Remarks to the Economics Club of theUniversity of North Carolina at ChapelHill, Chapel Hill, North Carolina,20 April.
Lalonde, R. and D. Muir. 2007. “The Bank ofCanada's Version of the Global EconomyModel (BoC-GEM).” Bank of CanadaTechnical Report No. 98.
Misina, M. and D. Tessier. 2007. “SectoralDefault Rates Under Stress: The Impor-tance of Non-Linearities.” Bank of CanadaFinancial System Review (June): 49–54.
59
Financial System Review – June 2008
The Role of Credit Ratings in ManagingCredit Risk in Federal Treasury ActivitiesNancy Harvey and Mervin Merkowsky
he ongoing turbulence in financial mar-kets has been accompanied by growingconcern that the use of credit ratings mayhave encouraged some investors to rely
too heavily on ratings as a summary statistic ofrisk. The Bank of Canada has raised the issueof overreliance on credit ratings through publicspeeches and in an article in the previous issueof the Financial System Review (Zelmer 2007).
Like many other central banks and market par-ticipants, the Bank of Canada (the Bank) uses avariety of tools, including credit ratings, in themanagement of credit risk in its own activitiesand in those that it carries out for the federalgovernment (the government) as its fiscal agent.This report provides a brief overview of thecredit risk management frameworks used by theBank and the government and how credit ratingsare used in these frameworks. As is outlined be-low, the Bank and the government are careful toavoid placing too much reliance on ratings.
Credit Risk
Credit risk can be defined as the risk that acounterparty may fail to meet its obligations asthey come due: that is, the risk of default. In itsbroadest sense, credit risk also includes the riskof a decline in the market value of investmentsthat may arise from a deterioration in the creditquality of a counterparty. This is known as credittransition risk.
The Bank is exposed to credit risk through itsroutine advances to members of the CanadianPayments Association (CPA) and through mar-ket transactions conducted in the form of pur-chase and resale agreements (PRAs) and loansof securities. The amount of credit risk borne bythe Bank is modest, however, because thesetransactions are fully collateralized with high-
T
quality securities denominated in Canadiandollars.1 In the unlikely event of a counterpartydefault, collateral can be liquidated to offset thecredit exposure. The credit quality of collateralis managed through a set of restrictions tied toasset type, credit ratings, and the term to matu-rity of the securities pledged as collateral.Credit risk is also evident in activities that theBank conducts for the government as its fiscalagent. Credit risk arises from the investment ofCanada’s foreign reserves, held in the ExchangeFund Account (EFA), in financial instrumentsissued by non-Canadian sovereign governments,their agencies, official international institutions,and major foreign financial institutions. Creditrisk is also engendered in the funding of reserveswhen swap transactions are conducted with ma-jor Canadian and foreign banks to transformdomestic-currency debt into foreign-currencyobligations. And it is present in the government’sCanadian-dollar cash balances, which are invest-ed in short-term deposits issued by the majorfinancial institutions operating in Canada.
Credit-risk policies have been established for allof these treasury activities to ensure that creditrisk is kept to a minimum. These policies specifythe types of transactions that can be conducted,the range of counterparties permitted, minimumcredit-quality thresholds, and how credit ratingsare used in the assessment of credit risk. Morebroadly, the policies seek to minimize creditrisk by promoting the use of a diversified poolof highly rated counterparties and, where ap-propriate, collateral frameworks.
1. In December 2007, the Bank of Canada announcedits intention to broaden eligible collateral for theStandard Liquidity Facility to include U.S. Treasuriesby mid-2008. Details available at <http://www.bankofcanada.ca/en/notices_fmd/2007/not121207a.html>.
61
Reports
Why and How Credit RatingsAre Used
Credit ratings published by the major ratingagencies offer important benefits to marketparticipants and public institutions. They pro-vide a commonly recognized source of indepen-dent opinions on creditworthiness, which canserve as a useful starting point for assessing thecredit quality of counterparties and their finan-cial instruments. The use of credit ratings is alsocost-effective, because rating agencies benefitfrom economies of scale in assessing credit risk.Indeed, agencies rate almost all of the counter-parties used in the treasury activities of the Bankand the government. But credit ratings are notflawless indicators of credit risk. Rating agencieshave been periodically criticized for, amongother things, overreliance on historical infor-mation and for being slow to react to newinformation.
Thus, the Bank and the government use creditratings in a prudent fashion. For any given creditrating, exposure limits and collateral haircutmargins vary across asset classes. For example,the investment limits and collateral haircutsfor AAA-rated government securities are moregenerous than those applicable to similarly ratedprivate sector instruments in recognition of thefact that the former are generally more liquidthan the latter. By the same token, the Bank andthe government have refrained from investingin, or accepting as collateral, some highly ratedstructured products, when the assets in questionwere judged to be incompatible with the objec-tives governing investment and collateral-management activities.
In selecting which rating agencies to use, the Bankand the government adhere to market practiceby using agencies that are widely accepted byprivate investors in the relevant markets andthat have been recognized by the regulatorsof those markets (e.g., the U.S. Securities andExchange Commission for markets operatingin the United States). Hence, the Bank and thegovernment have chosen to rely on credit ratingspublished by four rating agencies: DominionBond Rating Service (DBRS), Fitch Rating Service,Moody’s Investors Service, and Standard & Poor’s.
Ratings from these four agencies are used toassign a credit-quality grade to each counterparty,security issuer, or security issue. Thus, the credit-quality grade essentially represents the consensus
62
of the agency ratings. The number of externalcredit ratings used depends on rating availabilityand on the type of activity. Most activities requirea minimum of two ratings, and when the ratingagencies post different credit ratings, the credit-quality grade is usually based on the second-highest rating in accordance with the standard-ized credit-risk methodology proposed byBasel II.2 Moody’s and DBRS recently introducednew methodologies for rating commercial banks.These procedures explicitly consider the likeli-hood of external support (e.g., government orcentral bank support) in the determination oftheir ratings for commercial banks. This has ledthe Bank and the government to review theappropriateness of the official credit ratings forcommercial banks from these two agencies andto start using their “stand-alone” ratings forcommercial banks instead. (See Box 1 for addi-tional information on DBRS and Moody’s newrating methodologies and their implicationsfor the treasury activities of the Bank and thegovernment.) The next three sections describehow ratings are used in treasury activities. Furtherdetails can be found on the Bank of Canada andDepartment of Finance websites.
Management of the ExchangeFund Account
The Exchange Fund Account is the main reposi-tory for Canada’s foreign reserves. Assets heldin the EFA are managed by the Bank on behalfof the government in accordance with investmentpolicies approved by the Minister of Finance.Assets in the EFA are invested mainly in highlyrated securities issued by sovereigns, their agen-cies, and official international institutions. In-vestments in short-term securities, deposits,commercial paper, and certificates of depositissued by major foreign institutions are alsopermitted. Investments in more complex securi-ties, such as those that have embedded optionsand prepayment risk, structured products, andother asset classes not listed above are prohibited.
The assets in the EFA are managed against aportfolio of dedicated liabilities that are matchedin terms of duration and currency. Funding
2. See the treatment of multiple credit ratings in “CreditRisk—the Standardised Approach” in Part 2, SectionII.C.2 of: International Convergence of Capital Measure-ment and Capital Standards: A Revised Framework,available at <http://www.bis.org/publ/bcbs107.pdf>.
Financial System Review – June 2008
Box 1
Stand-Alone Ratings
In late 2006 and early 2007, DBRS and Moody’simplemented new methodologies for ratingcommercial banks. The new procedures arebased partly on their presumption that, in theevent of default, governments (and centralbanks) would likely stand behind the liabilitiesof major systemically important commercialbanks. As a result, when the new rating meth-ods were unveiled, many commercial bankswere given higher credit ratings by DBRS andMoody’s.
The methodologies highlight a fundamental is-sue for governments and central banks. That is,should they rely on credit ratings that are basedpartly on the presumption that they wouldcome to the aid of systemically important com-mercial banks? While other market participantsmay be willing to accept the new rating proce-dures, from a risk-management perspective, it isinconsistent for the Bank and the government touse ratings that are partly based on their owncredit strength and on their presumed willing-ness to provide support to the banking sector.
As a result, the Bank and the government havedecided to rely on the Bank Financial StrengthRatings published by Moody’s and on the In-trinsic Assessment ratings published by DBRSwhen assessing the credit quality of commercialbank counterparties in EFA investment andfunding activities.
Such stand-alone ratings are also used in theStanding Liquidity Facility to assess the spon-sors of asset-backed commercial paper pledgedas collateral. The latter must be sponsored by adeposit-taking institution that is federally orprovincially regulated and that has a minimumstand-alone credit rating equivalent to a creditrating of at least A- from at least two ratingagencies.
requirements are met primarily through anongoing program of cross-currency swaps, where-by domestic-currency liabilities are transformedinto foreign-currency liabilities in accordancewith the swap-management policies approvedby the Minister of Finance. The government isexposed to credit risk when swaps increase inmarket value, because it could experience a lossif swaps had to be replaced following the defaultof a counterparty.
Credit risk is mitigated in the foreign assets andliabilities by setting limits on credit exposurethat foster an appropriate diversification ofcounterparties and investments. Exposure limitsvary across asset classes and according to creditquality within each asset class. Credit ratingspublished by rating agencies are used to deter-mine: (i) the eligibility of a counterparty and(ii) exposure limits for individual counterpartieswithin each asset class. To be eligible for invest-ment, a counterparty or security issuer must havea minimum credit rating of A- from at least twoof the four rating agencies.3 In practice, however,almost all EFA investments are placed with sover-eigns, sovereign agencies, and official interna-tional institutions that are rated AAA, whilemost private sector counterparties are rated atleast AA-. Thus, the allowance of exposures rat-ed below AA- is meant to facilitate an orderlyreduction in exposures if a counterparty isdowngraded below that category. Within eachasset class, stronger-rated counterparties re-ceive larger exposure limits than those that havelower ratings. Since credit ratings change peri-odically, they are continuously monitored, andexposure limits are updated accordingly.
While credit ratings are used to determine coun-terparty eligibility and to set exposure limits,exposures vary within those limits. Investmentand swap transactions are executed based ontheir specific return and risk characteristics andon the credit outlook for the counterparties orthe security issuers. Exposure limits are not oftenused to their fullest. Moreover, exposures havebeen kept well below limits when credit assess-ments by the Bank and the government suggestedthat uncertainty surrounding the credit qualityof a counterparty was higher than normal andwas not fully reflected in public credit ratings.Thus, while credit ratings help set the parameters
3. Rating references in this document use the Standard& Poor’s ratings scale for illustrative purposes.
63
Reports
of the investment framework, they do not drivethe day-to-day investment and funding of theforeign reserves within those parameters.
Management of ReceiverGeneral Cash Balances
The management of the government’s Canadian-dollar cash balances differs from the investmentof the foreign reserves in that the former areplaced with counterparties on a short-term basisthrough a deposit-auction process rather thanon the basis of transactions initiated by the Bankon behalf of the government. Consequently,exposure levels are determined by the counter-parties themselves, subject to maximum biddinglimits. Hence, the determination of the eligibilityof participants and the setting of bidding limitsmust be thorough and transparent so that therules are understood by all auction participantsbefore the auctions take place. Thus, the optionof using internal credit assessments to gauge thecredit quality of counterparties and to set coun-terparty exposure limits is not practical. Instead,credit risk related to Receiver General depositauctions is mitigated mainly by (i) promotinga diversified set of counterparties through theuse of individual bidding limits that are partlylinked to credit ratings; (ii) typically limitingthe term of deposits to several business days;and (iii) where possible, taking collateral tolimit the amount of uncollateralized exposures.
Receiver General cash balances are investedthrough twice-daily auctions (morning andafternoon). Most of the government’s funds areusually auctioned in the morning, for terms thatcan range up to several business days, and arecarried out on a collateralized and uncollateral-ized basis. Eligible participants include a broadrange of counterparties whose bidding limits foruncollateralized funds are based in part on theircredit ratings. For example, they are required tohave minimum credit ratings of A- from at leasttwo rating agencies, and those with higher ratingsreceive larger uncollateralized bidding limits.The rules of the auction process are clearly for-mulated and are publicly available.4
4. The rules of the auction process can be found in“Terms and Conditions Governing the Morning Auc-tion of Receiver General Cash Balances” on the Bankof Canada’s website at <http://www.bankofcanada.ca/en/auction/rec_general.pdf>.
64
In contrast, the afternoon auction takes placelate in the day, after the government’s financialflows for the day have been finalized. Sincethe auction takes place after the close of thedelivery-versus-payment period of the auto-mated securities settlement system (CDSX) oper-ated by the Canadian Depository for SecuritiesLtd., it is not possible to conduct securitiestransfers at the same time as cash settlement. In-stead, credit risk in this auction is mitigated byrestricting access to direct participants in theLarge Value Transfer System (LVTS) and by lim-iting the term of these deposits to overnight.Bidding limits for this auction are based onthe size of the institution in the Canadian fi-nancial system based on Canadian PaymentsAssociation ratios, which represent an institu-tion’s share of total Canadian-dollar deposits.
Collateral Management
As mentioned, collateral is also used to protectthe Bank and the government against loss froma credit event. In the case of a counterparty de-fault, the proceeds from liquidating collateralcan be used to offset exposure from the under-lying transaction. The legal agreements in place,which must be signed by each counterparty (orparticipant) before any transaction occurs, areused to ensure that the Bank or the governmentobtains a valid, first-priority security interest inthe pledged collateral under the applicable law,while establishing, when applicable, thresholdswhere the Bank or the government have rightsto make margin calls for additional collateral asneeded. The collateral frameworks of the Bankand the government have been enhanced fromtime to time, in keeping with good market prac-tice and their own business requirements. Withthe broadening of eligible securities in recentyears, credit ratings have been used to help de-termine which securities can be pledged underthe various collateral frameworks. With the in-clusion of securities other than government-guaranteed securities in the eligible collateralpools, the need arose for a transparent mecha-nism to establish the creditworthiness of collat-eral so that pledgers understand ahead of timewhich securities can be pledged as collateral inthe Bank’s treasury activities and how they willbe valued and haircut.
In fiscal-agent activities, non-U.S. and non-Canadian government securities pledged as col-lateral for EFA securities lending or in support
Financial System Review – June 2008
of Receiver General deposits must adhere tominimum credit-rating thresholds of AA- and A-,respectively. In contrast, only U.S. Treasury, U.S.Agency, and Canadian government securities canbe pledged as collateral in support of EFA tri-party repo and swap transactions. In the case ofswaps, credit ratings are also used to set prede-termined thresholds for margin calls of addi-tional collateral. This mechanical approach isunavoidable since swaps are long-term contractsthat must contain explicit contingency plansfor credit migration.
For its own activities, the Bank can lend only ona secured basis and thus has collateral frame-works in place to support its operations underthe Standing Liquidity Facility (SLF) and in itsactivities involving securities-lending and pur-chase and resale agreements.5 The Bank usescredit ratings, in combination with other mech-anisms, to set eligibility requirements for securi-ties and applicable collateral haircuts or marginrequirements.
In its role as lender of last resort, the Bank rou-tinely provides liquidity to facilitate settlementin the payments system through the SLF byproviding collateralized overnight loans to par-ticipants in the LVTS. The Bank establishes thelist of assets acceptable for pledging as collateraland provides valuations of pledged securities.The latter are valued on a daily basis at currentmarket prices less an appropriate haircut toprotect the Bank against unexpected fluctuationsin their market value. The Bank determines theappropriate haircuts based on its own analysisof the market and the liquidity risks of the secu-rities in question.6 In particular, the Bank hasfound it useful to establish haircuts that varydepending on asset class, tenor, and credit qualityof the security.7 Credit ratings play a dual role inthis process. First, they are used to help determine
5. Terms and conditions of these programs are set out inthe document “Securities Eligible as Collateral underthe Bank of Canada Standing Liquidity Facility”<http://www.bankofcanada.ca/en/financial/securities.pdf> and “The Bank of Canada Securities-Lending Program: Terms and Conditions”<http://www.bankofcanada.ca/en/notices_fmd/2003/terms_en0403.pdf>.
6. A haircut is a percentage that is subtracted from themarket value of the assets that are being pledged ascollateral. The size of the haircut reflects the marketand liquidity risks associated with the assets.
7. Securities acceptable as collateral for SLF loans arealso eligible for intraday credit in the LVTS.
the minimum acceptable credit quality of a se-curity. Second, they are used, in combinationwith other indicators of market and liquidityrisk, to determine haircut levels for acceptablesecurities. In practice, haircuts are larger for lower-rated assets and for those with longer maturi-ties, since the prices of these securities tend toexhibit greater volatility, and their markets tendto be less liquid.8 There are, however, other safe-guards in place to mitigate collateral risk: pledg-ers may not pledge their own securities; and, inthe case of private sector securities pledged ascollateral under the SLF, pledgers cannot pledgemore than 20 per cent of the securities of re-lated issuers to promote a diversified pool ofprivate sector securities pledged as collateral.
The Bank also uses a collateral framework tomitigate credit risk in its own market operations.These are conducted in the form of PRAs andloans of its own holdings of Government ofCanada securities. Through its PRAs, the Bankoffers to temporarily purchase specific securitiesfrom designated counterparties with an agree-ment to sell them back at a predetermined priceand date. Under its securities-lending program,the Bank makes its Government of Canada se-curities available through a tender process whenthere are indications that those securities are un-available or trading at an unusually high premi-um in the market. Only primary dealers areeligible to participate in these activities, howev-er, since they are the main market makers in themarkets for Government of Canada securities,and thus have the strongest need for access tofunding and securities from the Bank to helppromote the liquidity of those markets. Thus,credit ratings are not used to determine who canaccess those facilities. Instead, they are usedonly to set eligibility and haircut requirementsfor securities pledged as collateral.
8. For example, a haircut of 1.5 per cent is applied to asecurity issued by the Government of Canada with a5-year term, while a haircut of 7.5 per cent is appliedto any asset-backed commercial paper (ABCP)pledged as collateral. In the case of ABCP pledged ascollateral, the pledger cannot be the sponsor or finan-cial services agent for the ABCP program, nor can thepledger be the provider of liquidity support to theprogram.
65
Reports
Conclusion
The need for a more sophisticated approach tomanaging credit risk in the treasury activities ofthe Bank and the government has grown overtime. Credit risk in these activities was tradi-tionally managed by restricting the list of eligi-ble counterparties to a small set of institutionsand by accepting only government-guaranteedsecurities as collateral. However, a more com-prehensive and transparent framework for man-aging credit risk became necessary as the list ofeligible counterparties and collateral expandedover time. This naturally led to the use of creditratings published by external rating agenciesto help assess the credit quality of counterpar-ties and of the securities pledged as collateral.
Credit-rating agencies provide a well-recognizedopinion on creditworthiness for a wide range ofcounterparties and financial instruments. Manyinvestors have found it cost-effective to rely ontheir opinions because rating agencies benefitfrom economies of scale in assessing credit risk.These benefits have led many central banks andmarket participants to use credit ratings to helpdetermine counterparty eligibility requirementsand to set credit-exposure limits.
The Bank and the government use a variety oftechniques to assess and manage credit risk,including rating-based frameworks in whichjudgment is applied. For example, they seek totransact with a wide range of counterparties andto minimize uncollateralized credit exposures.Furthermore, in the case of the Exchange FundAccount, exposures have been kept well belowlimit when the Bank and the government be-lieve that the uncertainty surrounding the creditopinion is higher than normal and not fully re-flected in public credit ratings. Thus, while ex-ternal credit ratings are embedded in manyfacets of the treasury activities of the Bank andthe government, they are not accorded undueweight as a summary statistic of risk. Creditratings are only one of many tools used tomanage credit risk in these activities.
66
References
Zelmer, M. 2007. “Reforming the Credit-RatingProcess.” Bank of Canada Financial SystemReview (December): 51–57.
Policy and
Infrastructure
Developments
Financial System Review – June 2008
Introduction
he financial system and all of itsvarious components (institutions,markets, and clearing and settle-ment systems) are supported by a
set of arrangements, including governmentpolicies, that influence its structure andfacilitate its operation. Taken together,these arrangements form the financialsystem’s infrastructure. Experience hasdemonstrated that a key determinant ofa robust financial system is the extent towhich it is underpinned by a solid, well-developed infrastructure. This section ofthe Review highlights work in this area,including that related to relevant policydevelopments.
Recent disruptions in financial markets have ledcentral banks around the world to re-examinetheir roles in providing liquidity to the financialsystem, and the Bank of Canada has been noexception. In the article, Financial MarketTurmoil and Central Bank Intervention,Walter Engert, Jack Selody, and Carolyn Wilkinsconsider the questions why, when, and how acentral bank might intervene when confrontedby financial market turmoil. They set out a policyframework and identify appropriate centralbank instruments, consistent with central bankpolicy goals and functions.
TT
69
Financial System Review – June 2008
Financial Market Turmoil and Central BankInterventionWalter Engert, Jack Selody, and Carolyn Wilkins1
ecent disruptions in financial marketshave led central banks around theworld to re-examine their roles in pro-viding“liquidity” to the financial system.
Analysts often refer to different types of li-quidity, that is, market liquidity, funding li-quidity, and central bank liquidity. Marketliquidity is an asset- or market-specific conceptthat refers to the ability to trade asset positions ofreasonable size with little price impact. Fundingliquidity is an institution-specific concept that re-fers to the ability of solvent counterparties to obtainimmediate means of payment to meet liabil-ities coming due. Central bank liquidity refersto access to money from the central bank.
At some risk of oversimplification, one mightconsider that liquidity generally refers to theavailability of assets that have predictable valueover time, and that can be transferred, bought,and sold with low transactions costs and with-out affecting the market value of the asset.
The fundamental concerns of a central bank re-late to two aspects of financial system liquidity.First, a central bank cares about aggregate systemliquidity because of its connection to futureinflation. Second, a central bank is concernedthat the financial system effectively distributesliquidity, because the system can become ineffi-cient and possibly unstable when liquidity isnot available where it is most needed.
In this article, we consider central bank inter-vention to address financial market turmoil with afocus on the questions of why, when, and how acentral bank might intervene. We set out a policyframework and identify appropriate central bankinstruments to respond to extraordinary financialmarket turmoil, consistent with central bankpolicy goals and functions.2
1. This work has benefited from comments provided bynumerous colleagues, for which we are grateful.
2. For a related perspective, see Carney (2008).
R
71
Why Intervene?
Endogenous liquidity creation
Central bank open market operations and lend-ing provide liquidity to the financial system.The central bank is not the only source ofliquidity in the financial system, however, nor isit the main source. In modern financial systems,liquidity is generated endogenously, that is,within the system through the normal interac-tion of private participants pursuing their owninterests. Central bank lending can be seen as anexogenous source of liquidity, determined bythe central bank to meet its policy objectives.This makes central bank lending especiallyimportant when the endogenous generation ofliquidity is impaired.
Considered in a highly stylized (or theoretical)framework, two functions are central to the pro-cess of endogenous liquidity creation: bankingand market making. Banks provide liquidity bytaking deposits that have a fixed value (at par)and that can be withdrawn on demand by theirowners. Banks expand liquidity by leveragingexisting deposits to issue new loans, which, inturn, can lead to new deposits. When a bankfinds itself short of liquidity, it can borrow fromother banks, sell assets in money markets, or goto the central bank for a loan. Markets provideliquidity by allowing assets to be readily sold atprices that correspond to the discounted streamof returns expected from the assets. Institutionsthat provide market-making services expandliquidity by leveraging their capital to buy and sellassets more frequently at such prices, which reflectfundamental value. When a market-makerfinds itself short of liquidity, it can borrow frombanks and other market-makers, or sell assets inmoney markets.3
3. Of course, banking and market-making services canbe provided by the same institution, and often are inCanada.
Policy and Infrastructure Developments
In a modern financial system, liquidity tends tobe generated from hubs (or tiers).4 A bankinghub is a group of major banks that are especiallyactive in providing loans to each other, to finan-cial institutions outside the hub, and to market-makers. A market-making hub is a group of in-stitutions that are especially active in makingmarkets and in buying or lending against theilliquid assets of other market-makers. Seen inthis way, the institutions active in these hubs,pursuing their own interests, create liquiditythat benefits other financial system participants.As a result, these banks and market-makers arecollectively important to the stability and effi-ciency of the financial system.
This process of endogenous liquidity generationand distribution will almost always create suffi-cient liquidity in the right places in the financialsystem. It is also generally accepted that themarket frictions and incentive misalignmentsthat exist in normal financial system conditionsare not sufficient to impair effective endoge-nous liquidity generation and distribution.It is also apparent, however, that in extraordi-nary circumstances, the process of endogenousliquidity creation can become impaired.
Endogenous liquidity creation canbreak down
When the endogenous generation of liquiditybreaks down, the central bank can improve thestability of the financial system by providingliquidity. For example, a small shock to thedemand for liquidity can ultimately lead to adisproportionate effect on a bank, given thatdeposits are redeemable at face value on demand.A resulting bank run can be associated with un-certainty about the solvency of the bank. And,as its creditworthiness becomes uncertain, thebank might not be able to obtain liquidity fromother banks or from markets. If the central bankhas access to information that indicates that thetroubled bank is solvent (through the supervi-sory authority, for example), the central bankcan improve the situation by lending to theaffected bank.5
4. For an analysis of tiering in the context of the pay-ments system, see Chapman, Chiu, and Molico(2008). See summary article, p. 83.
5. For discussion of the Bank of Canada’s lender-of-lastresort policies, see Bank of Canada (2004) or Daniel,Engert, and Maclean (2004–05).
72
The stability and efficiency of a modern finan-cial system rely on market participants generallybeing able to buy and sell assets at prices thatcorrespond to their fundamental values, espe-cially in money markets. Financial markets are“incomplete,” however, in the sense that allparticipants do not operate in all markets. Thiscan inhibit the flow of information and fundingin markets, which, in turn, can lead to pricinginefficiencies, including distorted liquidity pre-miums in important markets. Such pricing inef-ficiencies, which normally are minor transitoryfrictions, can become significant under certainconditions, such as a sudden widespread increasein uncertainty about counterparty solvency.This can discourage market participants fromfunding one another, and can cause importantparticipants to withdraw from the market, wors-ening market incompleteness. Pricing ineffi-ciencies can also be exacerbated by herdingbehaviour, where market participants follow thelead of others instead of relying on their ownanalysis.
In extreme circumstances, a lack of reliable in-formation in incomplete markets can lead tosignificant pricing inefficiencies and to a break-down of endogenous liquidity creation, furtherexacerbating pricing inefficiencies and decliningliquidity. Under such circumstances, banks maynot be in a position to respond by expandingtheir credit-intermediation services, either be-cause they face an increased need for liquiditythemselves or because they also lack reliable in-formation about the creditworthiness of institu-tions acting as market-makers and about othermarket participants.
For example, an anticipated rise in defaults for aparticular asset class could create uncertaintyabout the solvency of a market-maker for thatasset class, limiting its ability to obtain thefunding necessary (including through sales ofthe asset) to continue making the market, thusworsening pricing inefficiencies. Banks, havingalso been hit by the shock, might conserve li-quidity for their own needs, limiting their abilityto fund participants in such markets. As well,banks might not go to the central bank for addi-tional liquidity because they might lack the col-lateral needed to obtain a central bank loan, or,more likely, they might be reluctant to borrowfrom the central bank because of the potentialstigma (and possible supervisory intervention)associated with such borrowing. In addition,
Financial System Review – June 2008
banks might not have sufficient free capital (orbe able to raise sufficient capital) to replace thefinancing previously available through the issueof securities in the now dysfunctional market.As a result, banks could be ineffective in re-establishing credit intermediation at any rea-sonable level for the participants in the marketsuffering the shock. This could also lead toadverse effects in other asset markets.6
A central bank could help stabilize the systemby providing liquidity directly to the market. Itcould do so in this case by accepting as collateralfor central bank liquidity the securities thattraded in the now illiquid market, appropriatelydiscounted, which could also help re-establishefficient pricing.
Another example is analyzed by Allen and Gale(2007) who consider the collapse of an asset-price bubble. A bank with significant holdingsof this asset would become stressed because thevalue of many of its liabilities would be fixedwhile the asset would fall in value. This wouldforce the bank to conserve liquidity for its ownneeds, thus reducing the amount of liquidityavailable to others. The bank would also liqui-date assets, which would result in falling assetprices in illiquid markets, potentially under-shooting fundamental values, leading to an in-efficient allocation of resources.
A central bank can address this inefficiency byproviding liquidity to the illiquid market so thatasset prices can find their fundamental values,or it can lend to the affected banks so that theycan increase liquidity as needed.
The growing importance of marketliquidity
Since the events of August 2007, market condi-tions have solidified a growing realization thatan adequate supply and distribution of marketliquidity have become important to the stableand efficient functioning of the financial system.
Although economic theory suggests that thedistribution of liquidity matters for the soundfunctioning of the financial system, few practi-tioners have seen a need for the central bank toprovide direct liquidity support to individual
6. This scenario can be seen as a type of “market failure,”where decisions resulting from individual pursuit ofself-interest can lead to relatively poor collective oroverall results.
markets until recently. (See, for example,Banque de France 2008.) Altering liquiditythrough monetary policy, or in the core paymentsystems, or through a reallocation of liquidity tobanks was seen as sufficient action by a centralbank to maintain market and financial systemliquidity.
This view changed with the events of August2007 and the subprime-credit crisis. It is nowmore broadly accepted that the financial systemwill be more stable, and the effects of monetarypolicy actions more predictable, if the centralbank directly supports market liquidity in someextraordinary circumstances.
Behind this change in view is a realization thatthe financial system has become more depen-dent on market liquidity. One reason for thisincreased dependence is the greater use ofsecuritization to convert non-traded receivables(such as mortgages) into tradable securities(such as mortgage-backed securities), makingfinancial institutions increasingly reliant onmarket liquidity for funding their operations.
Another reason for the increased prominence ofmarket liquidity is the growing use of “mark-to-market” accounting that rapidly converts asset-price shocks into balance sheet shocks. Thismakes it more important that markets have suf-ficient liquidity to price assets efficiently so thatmarket prices adequately reflect economic value.Where securities are not sufficiently standardizedto be traded in a market, “mark-to-market”becomes “mark-to-model,” which creates addi-tional valuation uncertainty in times of financialstress. These phenomena also cause financialinstitutions to hoard liquidity in case they haveto restructure their balance sheets after a suddenchange in valuations. Liquidity is preserved, inturn, by cutting back on lending and tradingactivities.
Intervening in markets isconsistent with central bankpolicy objectives
Central bank provision of liquidity is governedby policies with a common objective. Suchpolicies mitigate potential financial systeminstabilities that can be addressed only by theexogenous provision of liquidity by the centralbank.
Monetary policy stabilizes the inflation rate. In amodern financial economy, the rate of inflation
73
Policy and Infrastructure Developments
is determined by the central bank setting a pathfor the riskless interest rate (i.e., the policy rate).This involves the central bank standing ready tolend to clearing banks in the payments systemand to conduct limited open market operationsto achieve the policy rate.7 The path for the pol-icy rate determines aggregate liquidity in thefinancial system. Central bank interventiondealing with the distribution of liquidity (dis-cussed below) may require an offsetting centralbank action to leave the policy rate at its target,thus keeping the setting of monetary policyunaffected.
Payment, clearing, and settlement policy protectsthe payments system against the destabilizingeffects of “gridlock,” which can occur if a partic-ipating bank does not have sufficient liquidityto meet its payment obligations. In a monetaryeconomy, banks are linked by a system that usescentral bank money to settle accounts, wherethe clearing banks (in the central hub) haveaccess to standing overdraft facilities from thecentral bank to facilitate settlement of payments.
Lender-of-last-resort policy (or, more specifically,emergency lending assistance) stabilizes banksin the face of a liquidity shock that could causea bank run because fixed-value deposits are re-deemable on demand. Such lending is providedonly when endogenous liquidity generationdoes not provide liquidity to a solvent bank,leaving the central bank as the only means ofobtaining liquidity.
Exceptional market intervention policy addressespotential instabilities arising from liquidity dis-tortions in money markets. These policies deter-mine the extent to which central banks lend tomarkets, as well as the means of such liquidityprovision, including choice of term to maturity,collateral, and counterparties.8
7. For more on how the Bank of Canada implementsmonetary policy, see Bank of Canada (2007) andEngert, Gravelle, and Howard (2008).
8. Fiscal agency policy complements these various centralbank lending policies. It contributes to financial sys-tem stability and efficiency by providing for the effi-cient pricing of government bonds, which are thebenchmark for many other securities prices in thefinancial system.
74
When to Intervene
In deciding whether to intervene in an episodeof financial market turbulence, a central bankshould address three basic questions, which areconsidered here.
Will central bank instruments beeffective?
In evaluating the potential effectiveness of itsinstruments, a central bank should focus onidentifying the nature of the market failurecausing the problem, and then judge whetherits instruments are well suited to addressing theproblem. Alternatively, a legislative, regulatory,supervisory, or market-practice change might,in some circumstances, be better suited toproviding the incentives needed to correct thepricing inefficiency.
Central bank instruments are likely to be effec-tive only when such intervention increases thewillingness to participate in markets, either byincreasing confidence that future prices willbe more predictable and will reflect reducedliquidity premiums, or by reducing the stock ofan illiquid asset held by the private sector.
What are the potential benefits ofintervention?
The central bank is a public institution thathelps manage the macroeconomy and shouldtherefore consider only benefits that are evidentat a macroeconomic level. In assessing thepossible benefits of intervention, the followingelements should be considered.
• The value of avoiding increasing financialsystem dysfunction that could occur frominaction.
• The avoided loss of selling assets at fire-saleprices, which could lead to insolvency andimplies dead-weight losses to the economy.
• The avoided cost of loss of confidence inthe financial system. For example, a majorbanking crisis specific to a country couldcause international investors to demand arisk premium, which would constrainnational growth.
• Benefits will be greater the more stronglyeconomic activity is linked to the marketunder stress.
Financial System Review – June 2008
What are the potential costs ofintervention?In evaluating the possible costs of intervention,a central bank should assiduously guard againstlosing focus on its primary responsibility of lowand stable inflation. A central bank should alsomitigate financial risks to itself that may arisefrom intervention. Another cost of interventionrelates to creating a sense of crisis when there isnone (a false negative signal), by interveningwhen there is no need.
Finally, “moral hazard” is a major consideration.Moral hazard is the prospect that a party pro-tected from risk will behave differently from theway it would behave if it were fully exposed tothe risk, and, in particular, with less regard forthe consequences of its actions, expectinganother party to bear the consequences of thoseactions.
If the central bank intervenes only in true liquiditycrises, then moral hazard would be limited to adistortion of the incentives to manage liquidityefficiently. Liquidity risk and solvency risk areoften confounded, however, making it difficultin practice to determine when to intervene. Thisalso raises the prospect that central bank inter-vention could discourage financial market par-ticipants from managing counterparty (credit)risk appropriately, with attendant adverse effectson the functioning of the financial system. Aswell, central bank intervention can create incen-tives for institutions to generate the conditionsthat would trigger such intervention, so thatthey can benefit.9
In sum, whenever a central bank intervenes,there are costs, and intervention creates thepotential for moral hazard. To the extent thatprivate agents expect a central bank to provideliquidity whenever financial markets encounterdifficulties, private agents will take less care inmanaging their liquidity and counterparty risks,which could make markets work less well in thefuture.
9. Explanations of financial crises often involve elements ofmoral hazard, usually excessive risk-taking behaviourencouraged by poorly designed safety nets. Similarly, theeconomic literature suggests that financial systems withmore conservative regulatory environments are betterable to withstand crises (Benston and Kaufman 1997;Caprio 1998; Dziobek and Pazarbasioglu 1997; andFurlong and Kwan 2006).
Mitigating moral hazardOne way of limiting the effects of moral hazardis to intervene only under very adverse circum-stances. The central bank could apply its toolsselectively so that private agents are unlikely toperceive such actions as a reason to change theirongoing behaviour.
In this regard, the application of a test would beuseful to determine when intervention wouldbe appropriate. The following test, consistentwith the questions posed in the preceding sec-tion, as well as the tests proposed by Summers(2007) and Buiter (2007), could be used toinform a decision on whether to intervene.
• Is there a significant common shock, sub-stantial contagion, or negative spillovereffects, with the prospect of significant realconsequences?
• Is the problem primarily a liquidity problem,where a contribution to stability can be pro-vided with high probability? (In contrast, ifthe problem is mainly one of solvency,central bank intervention is unlikely to besuccessful.)
• Is it reasonable to expect that interventionwill not impose costs on taxpayers?
• Is the intervention unlikely to have a materialimpact on the likelihood and severity offuture financial crises? (This would encom-pass, among other things, consideration ofthe nature of the intervention mechanism.)
• Will this action produce a net social benefit?
If the answers to these questions are “yes,” thenthere is likely a good case for the central bank tointervene. Importantly, this test suggests thatintervention would be infrequent and would beassociated with financial losses for market par-ticipants, which would provide an element ofcoinsurance to also help mitigate moral hazard.
Further, a penalty rate chosen at the discretionof the central bank could apply to the provisionof central bank funds to individual institutionsin this context.10 Finally, a central bank shouldpromote the sound supervision of liquidity and
10. Penalty rate here means a premium above the centralbank’s policy rate. (In Canada, the policy rate is the over-night interest rate, and the Bank of Canada’s minimumlending rate is the Bank Rate, that is, the overnight rateplus 25 basis points.)
75
Policy and Infrastructure Developments
related risks, and maintain some oversight ofthe management of liquidity risk by potentialborrowers to help mitigate the costs and risksof intervention.
How to Intervene
Principles
A central bank should intervene only whenthere is a market failure and when significantfinancial instability can be avoided or mitigatedwithout distorting the pricing of credit risk. Thepreceding discussion gives rise to five principlesthat should guide the use and design of centralbank intervention facilities.11
(i) Targeted intervention: Mitigate only thosemarket failures (liquidity distortions) of system-wide importance with macroeconomic conse-quences and which can be rectified by a centralbank. This principle acknowledges that the cen-tral bank cannot solve all problems, and indi-cates that the central bank should interveneonly when the problem is one that is likely tomaterially affect the macroeconomy and onethat could be reasonably addressed by centralbank intervention.
(ii) Graduated intervention: Intervention shouldbe commensurate with the severity of the prob-lem. This principle recognizes that there is a costassociated with the central bank doing too much.It suggests an escalated response that dependson the severity of the problem to guard againstcentral bank overreaction.
(iii) Well-designed intervention: Use the right toolsfor the job. Market-based transactions, providedthrough auction mechanisms, should be usedto alleviate marketwide liquidity problems, whileloans should be used to address liquidity short-ages affecting specific institutions.
(iv) Efficient, non-distortionary intervention: Centralbank transactions should be at market-deter-mined prices to minimize distortions. Inparticular, central bank intervention shouldnot distort credit-risk spreads, because thiswill create additional problems.
(v) Mitigation of moral hazard: The risk of creat-ing adverse incentives that could impair the
11. Any intervention by the Bank of Canada would be inaccordance with the terms of relevant statutes, mostimportantly, the Bank of Canada Act.
76
functioning of the financial system over timeshould be considered carefully, and measuresshould be taken to mitigate such risks. Suchmeasures include limited, selective intervention;an element of coinsurance; penalty rates asappropriate; and promoting the sound supervi-sion of liquidity-risk management.
Auction mechanisms
Central bank intervention in markets (as op-posed to loans to institutions under standingliquidity facilities) would likely be best achievedthrough auction mechanisms initiated at thediscretion of the central bank. An auction for-mat provides several benefits:
• Pricing is set competitively in an auction,and so generally should lead to the efficientpricing of the asset being auctioned.12
• The stigma that can be attached to centralbank lending could be mitigated or avoidedbecause an auction is a collective mecha-nism involving several borrowers simulta-neously.
• An auction can reveal information aboutmarket conditions useful to the central bankin managing the situation.
• An auction provides flexibility to vary thekey parameters of the transaction: that is,the term, eligible counterparties, and eligi-ble securities, depending on the situation.
• Appropriately designed, an auction can helpthe market find more efficient pricing andencourage the recovery of a troubled market.
Different facilities for differentcircumstances
Along with traditional central bank tools, suchas lender-of-last-resort arrangements, a rangeof facilities is likely necessary for the provisionof liquidity to the financial system, each withdistinct characteristics suitable for differentcircumstances.
12. An auction might not reveal the correct price of theasset being sold (for allocative efficiency) when thereis extreme uncertainty about the future market valueof the asset. Nevertheless, compared with othermechanisms, auctions appear to be a fairly robustand efficient means of allocating resources (Chap-man, McAdams, and Paarsch 2007).
Financial System Review – June 2008
Term purchase and resale agreements (or term repos)would be most useful for providing liquidity tomoney markets since they can be offered to anyfinancial market participants with marketablesecurities as the basis for the transaction. Termrepos would be most useful when liquidity pre-miums in money markets are distorted and areassociated with widespread liquidity problemsin an asset class or maturity.
Term securities lending would increase the supplyof high-quality securities that could be used forcollateral at times when there is a shortage ofsuch collateral needed for funding. This mecha-nism can also provide for a direct exchange ofless-liquid securities for more-liquid securities,thus reducing the incentive to hoard liquidityfor precautionary purposes.
Term loan facilities could be most useful when li-quidity premiums in money markets are distort-ed because specific financial institutions hadparticular liquidity shortages. Such an opera-tion could be conducted through an auction(subject to a minimum bid rate) when at leasttwo eligible institutions are facing pronouncedliquidity problems in this context, but do notyet need emergency lending assistance from thecentral bank.13
Concluding Remarks
Our conclusions can be summarized as follows.
First, central banks should provide liquidityto financial markets in extraordinary circum-stances because: markets require liquidity forefficient pricing, illiquidity can contribute tofinancial system instability with real economicconsequences, and a central bank’s unique char-acteristics make it well suited to be the ultimateprovider of liquidity to the financial system.
Second, a central bank should intervene to ad-dress financial market turbulence only whenthere is a significant market failure and signifi-cant financial instability and macroeconomicconsequences could be avoided or mitigated.
Third, a central bank should price the provisionof liquidity to financial markets competitivelythrough auctions.
13. According to the Bank of Canada Act, the Bank ofCanada can lend only to members of the CanadianPayments Association.
Fourth, a central bank should have a range offacilities with which to provide liquidity tofinancial markets, to better focus the provisionof liquidity as needed. These include term repos,term securities lending, and term lending.
Fifth, the provision of liquidity to financialmarkets should be guided by the followingprinciples.
• Targeted intervention
• Graduated intervention
• Well-designed intervention
• Efficient, non-distortionary intervention
• Mitigation of moral hazard
References
Allen, F. and D. Gale. 2007. “An Introductionto Financial Crises.” Wharton FinancialInstitutions Center Working PaperNo. 07-20.
Bank of Canada. 2004. “Bank of CanadaLender-of-Last-Resort Policies.” Bank ofCanada Financial System Review (Decem-ber): 49–56.
———. 2007. “A Primer on the Implementationof Monetary Policy in the LVTS Envi-ronment.” Available at <http://www.bankofcanada.ca/en/lvts/lvts_primer_2007.pdf >.
Banque de France. 2008. Financial StabilityReview. Special issue on liquidity (Febru-ary) (11).
Benston, G. and G. Kaufman. 1997. “FDICIAafter Five Years.” Journal of EconomicPerspectives 11 (3): 139–58.
Buiter, W. 2007. Comment on “Beware theMoral Hazard Fundamentalists,” byL. Summers, Financial Times EconomistsForum, 26 September. Available at<http://blogs.ft.com/wolfforum/2007/09/beware-the-morahtml/#comments>.
Caprio, G. 1998. “Banking on Crises: ExpensiveLessons from Recent Financial Crises.”World Bank Policy Research WorkingPaper No. 1979.
77
78
Policy and Infrastructure Developments
Carney, M. 2008. “Principles for Liquid Mar-kets.” Speech to the New York Associationfor Business Economics, New York, 22 May.
Chapman, J, J. Chiu, and M. Molico. 2008.“A Model of Tiered Settlement Networks.”Bank of Canada Working Paper No.2008-12.
Chapman, J., D. McAdams, and H. Paarsch.2007. “Bounding Revenue Comparisonsacross Multi-Unit Auction Formats underε-Best Response.” American EconomicReview 97 (2): 455–58.
Daniel, F., W. Engert, and D. Maclean. 2004–05.“The Bank of Canada as Lender of LastResort.” Bank of Canada Review (Winter):3–16.
Dziobek, C. and C. Pazarbasioglu. 1997. “Les-sons from Systemic Bank Restructuring:A Survey of 24 Countries.” IMF WorkingPaper No. 97/161.
Engert, W., T. Gravelle, and D. Howard. 2008.“The Implementation of Monetary Policyin Canada.” Bank of Canada DiscussionPaper No. 2008-9.
Furlong, F. and S. Kwan. 2007. “Safe and SoundBanking, Twenty Years Later: What WasProposed and What Has Been Adopted.”Prepared for the Conference on Safe andSound Banking: Past, Present, andFuture, sponsored by the Federal ReserveBanks of Atlanta and San Francisco andThe Journal of Financial Services Research,17 August 2006, Federal Reserve Bank ofAtlanta Economic Review 92 (1, 2).
Summers, L. 2007. “Beware the Moral HazardFundamentalists.” Financial Times Econo-mists Forum, 24 September. Available at<http://blogs.ft.com/wolfforum/2007/09/beware-the-morahtml/#comments>.
Research
Summaries
Financial System Review – June 2008
Introduction
ank of Canada staff undertakeresearch designed to improve overallknowledge and understanding of theCanadian and international finan-
cial systems. This work is often pursuedfrom a broad, systemwide perspective thatemphasizes linkages across the differentparts of the financial system (institutions,markets, and clearing and settlementsystems), linkages between the Canadianfinancial system and the rest of the econ-omy, and linkages to the internationalenvironment, including the internationalfinancial system. This section summarizessome of the Bank’s recent work.
The first two research summaries describe recentwork on payment and settlement systems. In thefirst summary, A Model of Tiered SettlementNetworks, James Chapman, Jonathan Chiu,and Miguel Molico describe their developmentof a dynamic equilibrium model of settlementnetworks, in which the settlement structure isdetermined endogenously, to study the degreeof tiering and the welfare effects of clearing-agentfailure. They show that, in the presence ofimperfect information and fixed costs of partici-pation, a tiered structure can be an efficientarrangement that supports cost saving and inter-bank monitoring. However, because settlementfailure may generate negative spillovers on otherparticipants, the market-determined concentra-tion and degree of tiering may not optimally di-versify the risk of a clearing-agent failure.
The second summary, The Effects of a Disrup-tion in CDSX Settlement on Activity in the LVTS:A Simulation Study, by Lana Embree and KirbyMillar, describes work that focuses on the inter-dependencies of two settlement systems: theLarge Value Transfer System (LVTS), which settlespayments, and CDSX, which settles debt and
B
equity trades. These are two of the main settle-ment systems in Canada, and they are closelylinked. At the end of each day, the final CDSXpositions must be settled through the LVTS, and,in planning their LVTS activity, participants takeinto consideration their expected CDSX settle-ment payment that day. Therefore, any eventaffecting CDSX settlement may have systemicimplications for the LVTS. In this work, theauthors quantitatively assess the impact on theLVTS of an operational event that would preventthe completion of CDSX settlement. The resultsindicate that this type of event could lead to aconsiderable amount of unsettled paymentsand payment delays. This study highlights theimportance of the contingency measures andmitigating actions that are available to safeguardthese payments systems.The final article, Family Values: OwnershipStructure, Performance, and Capital Structureof Canadian Firms, by Michael King andEric Santor, examines how family ownershipaffects the performance and capital structureof 613 Canadian firms from 1998 to 2005. Theauthors find that family ownership per se isnot negative for performance: it is the use ofcontrol-enhancing mechanisms that reduces afirm’s valuation. Whereas free-standing family-owned firms with single-class shares have amarket performance similar to that of otherfirms, family-owned firms that use dual-classshares have valuations that are lower by 17 percent, on average, relative to widely held firms,despite having similar returns on assets andfinancial leverage.
81
Financial System Review – June 2008
A Model of Tiered Settlement NetworksJames Chapman, Jonathan Chiu, and Miguel Molico
ettlement networks typically involve var- models of payments systems are developed to
ious tiers of intermediation. Some banksparticipate and clear directly in a “first-tier”network. A subset of these direct clearersthen act as clearing agents by operating a “second-tier” network and providing settlement accountsto indirect clearers downstream. In Canada, boththe Large Value Transfer System (LVTS) and theAutomated Clearing Settlement System (ACSS)exhibit a high degree of tiering. The efficiencyand risk associated with these tiered settlementnetworks are of particular interest to policy-makers. For example, what are the immediateimpact and long-term effects of the failure ofa clearing agent in a highly tiered settlementsystem? How do these effects differ from thosecaused by the failure of an ordinary direct clearer?
This article summarizes Chapman, Chiu, andMolico (2008), in which we develop a dynamicequilibrium model of settlement networks tostudy these questions. We demonstrate that, inthe presence of imperfect information and fixedcosts of settlement system participation, a tieredstructure can improve efficiency by supportinginterbank monitoring and cost saving.
Methodology
While policy-makers care about the efficiency andstability of settlement systems, guidance providedby economic theory has been limited. In partic-ular, there is little theoretical work on the tieredstructure in settlement systems. This is becausestandard economic models abstract from themechanism through which payments and set-tlement take place and thus are not suitable toolsfor modelling settlement systems. Our study isthe first to develop a dynamic equilibrium modelfor studying the degree of tiering and welfareeffects of clearing-agent failure.1 Economic
1. Related literature includes Kahn and Roberds (2002),Lai, Chande, and O’Connor (2006), and Chapmanand Martin (2007).
S
capture how the incentives and behaviour ofparticipants will adjust in response to changesin policy or in the economic environment.2Moreover, since we have limited historical dataon certain rare but highly significant events (e.g.,failure of clearing agents), using an economicmodel to conduct hypothetical experimentscan help us gain a better understanding of thepotential causes and consequences of suchextreme events.3
Model
Our analytically tractable model of the settle-ment system, in which the settlement structureis determined endogenously, is built on rational,strategic, and forward-looking agents. In themodel, the economy consists of two sectors:a trading sector and a settlement sector. In thetrading sector, agents meet bilaterally to tradeconsumption goods financed by private liabili-ties. In the settlement sector, agents interact toclear and settle these payment instruments. Un-derlying transactions in the trading sector gener-ate the bilateral payment flows in a settlementnetwork. In this environment, the mode of set-tlement (i.e., real-time vs. deferred) and thestructure of settlement networks (i.e., director indirect participation) are endogenously de-termined by agents, subject to the fundamentalcost structure and information structure.
The choice of settlement mode between real-time and deferred settlement involves the fun-damental trade-off between liquidity costs and
2. Much of the literature on payments system design isbased on payments system simulators such as thatdeveloped by the Bank of Finland (BoF-PSS2).Because they do not model the behaviour of systemparticipants, these tools are not appropriate forstudying the endogenous formation of tiered net-works.
3. See Chiu and Lai (2007) for a more detailed discus-sion of the microfoundations of payment economics.
83
Research Summaries
default risk. On the one hand, since real-timesettlement imposes a higher liquidity cost (forexample, the need to hold low-return liquidassets as collateral), payment senders (debtors)may prefer deferred settlement. On the otherhand, because of the settlement risk involved,payment recipients are willing to accept deferredsettlement only from creditworthy paymentsenders. Therefore, the choice of settlementmode depends critically on whether creditorspossess reliable information about debtors’credit history. This informational constraint isparticularly binding for trades involving debtorswhose creditworthiness is not well known toother agents and debtors who trade with otheragents only infrequently. We label these “small”agents. As a result, some of these small but safedebtors with no public history will be forcedto use real-time settlement. This is inefficient,because the unnecessary liquidity cost incurredby these safe debtors leads to a suboptimal allo-cation of resources.
This allocative inefficiency can be resolved byhaving some financial institutions act as clearingagents for these small agents. Typically, theseclearing agents are “large” in the sense that theyhave frequent transactions with a significant setof debtors and creditors. These large agents canimprove the efficiency of settlement by providinginformation and cost saving. Through their fre-quent dealings with creditors, they can establisha reputation and make their own creditworthinesspublic information. Through their frequentdealings with debtors, they can monitor debtors’credit history and choose the optimal settlementmode for each of them. This is their informationrole. When there are fixed costs associated withparticipation in the settlement system, clearingagents can also enjoy economies of scale andthus play a cost-saving role in a settlementnetwork.
Main Findings
Our main findings can be summarized as follows.First, we demonstrate that a tiered structure canimprove efficiency by supporting cost saving andinterbank monitoring. In a tiered settlementsystem, large agents work as clearing agents whoparticipate directly in a settlement system. Smallagents become indirect clearers who settle theirdebt through their clearing agent’s internalsecond-tier network.
84
This arrangement allows clearing agents to mon-itor the credit histories of the indirect clearersthat they serve and to then use this private infor-mation to choose the best mode of settlementfor their clients. Clearing agents have incentivesto appropriately monitor their clients becausethey will be held responsible if their clientsdefault. Furthermore, a tiered structure canimprove efficiency by economizing on the fixedcost of settlement system participation.
Second, the degree of tiering is decreasing in thefixed cost of operating the second-tier networkand the availability of public credit history. Asthe fixed cost of being a clearing agent increases,each clearing agent requires a larger number ofthe small agents as clients to be profitable.Therefore, there will be fewer larger clearingagents.
If a clearing agent’s provision of information isits primary motivation, then more public infor-mation regarding the creditworthiness of indirectclearers will lead to fewer clearing agents. Forexample, an increase in the number of agentswith credit ratings will reduce the equilibriumdegree of tiering.4 In the extreme case, in aworld where agents’ credit histories are perfectlyobservable, clearing agents have no informa-tional role.
Third, the failure of a clearing agent leads tosocial costs, which can be decomposed into:(i) default loss; (ii) participant loss; (iii) infor-mation loss; and (iv) operational inefficiency.The loss to default and the loss of participantsare transitory in nature and represent straight-forward losses as a result of the clearing agent’sfailure to perform its contracted role. The infor-mation loss and operational inefficiency are ofinterest, since they can have persistent welfareimplications and are closely related to the clear-ing agent’s unique role. The failure of a clearingagent leads to the loss of private informationregarding the trading history of its indirect clear-ers, which took time to accumulate. In addition,if there are economies of scale in the operationof the clearing agent’s second-tier network, then,unless that agent is immediately replaced, itsfailure will lead to operational inefficiencybecause the remaining clearing agents will need
4. In Canada, while all the direct clearers and clearingagents have credit ratings, many indirect clearers donot.
Financial System Review – June 2008
to serve too many indirect clearers and will beoperating above their efficient capacity.
Conclusion
Our study highlights that, in the presence ofimperfect information and fixed costs, thetiered structure can, indeed, improve efficiencyby supporting interbank monitoring and costsaving.5 One policy implication of this findingis that restricting the degree of tiering in paymentssystems such as the LVTS or ACSS may distortthe efficient monitoring structure of the system.6
Moreover, we identify the social costs resultingfrom the failure of a clearing agent. Since sucha failure may generate negative spillovers onother participants, the market-determinedconcentration and degree of tiering may notoptimally diversify the risk of such failure. Inconclusion, this framework can be expandedfor future analysis of specific payments systempolicies and their welfare implications.
5. Potentially, a tiered structure may also help to miti-gate the impact of systemic liquidity shocks (such asthe recent market events) on the indirect participants.
6. There is a volume restriction imposed on ACSS par-ticipation. There is no similar restriction to access inthe LVTS.
References
Chapman, J. T. E., J. Chiu, and M. Molico.2008. “A Model of Tiered SettlementNetworks.” Bank of Canada WorkingPaper No. 2008-12.
Chapman, J. T. E. and A. Martin. 2007. “TheProvision of Central Bank Liquidity underAsymmetric Information.” Bank ofCanada Financial System Review(December): 83–86.
Chiu, J. and A. Lai. 2007. “Modelling PaymentsSystems: A Review of the Literature.” Bankof Canada Financial System Review (June):63–66.
Kahn, C. M. and W. Roberds. 2002. “PaymentsSettlement under Limited Enforcement:Private versus Public Systems.” FederalReserve Bank of Atlanta Working PaperNo. 2002-33.
Lai, A., N. Chande, and S. O’Connor. 2006.“Credit in a Tiered Payments System.”Bank of Canada Financial System Review(December): 67–70.
85
Financial System Review – June 2008
The Effects of a Disruption in CDSXSettlement on Activity in the LVTS:A Simulation StudyLana Embree and Kirby Millar
safe and efficient payments system iscritical to the smooth functioning ofthe financial system. In Canada, theLarge Value Transfer System (LVTS),
for time-sensitive payments that are typicallylarge-value payments, and CDSX, for the clear-ing and settlement of debt and equity securities,are two of the systems that have been designatedas systemically important under the PaymentClearing and Settlement Act.1 The operationsof these systems are closely linked. For example,many LVTS participants are also CDSX partici-pants, and end-of-day funds exchange relatingto CDSX settlement occurs through the LVTS.
Given this link, a disruption to CDSX settlementcould potentially have a significant impact onLVTS activity. By monitoring their CDSX activity,LVTS participants can anticipate what their CDSXsettlement position will be and whether they willreceive a CDSX pay-out. Therefore, the partici-pants take into account their expected CDSXfunds when planning their LVTS activity through-out the day. An unexpected event that disruptsCDSX settlement could affect end-of-day activityin the LVTS.
The importance of this link between the twosettlement systems has long been recognized bysystem participants, system operators, and theBank of Canada. In our study (Embree and Mill-ar 2008), we try to quantify the potential impactof an operational event affecting CDSX settle-ment. Specifically, we simulate an event thatprevents CDSX settlement pay-outs from beingcompleted. Many possible events could disruptCDSX settlement in this way: for example, eventsaffecting the operator of CDSX, the system par-ticipants, or the Bank of Canada. We find that
1. For more information on CDSX, see McVanel (2003).For more information on the LVTS, see Arjani andMcVanel (2006).
A
such an event can have important potential im-pacts. There are, however, a number of mitigat-ing actions and contingency measures to preventsuch disruptions and reduce the impact shouldthey occur.CDSX and LVTS
Throughout the day, debt and equity trades andrelated entitlement payments (e.g., maturitiesand dividends) are settled in CDSX. At the endof the day, CDSX participants must settle theirnet funds position through the LVTS. The Bankof Canada is the settlement agent for the CDSClearing and Depository Services Inc. (CDS),the owner and operator of CDSX. The Bankreceives, through the LVTS, all the CDSX pay-ments from those participants in negative CDSXfunds positions. The Bank then makes the pay-outs to those in positive positions through theLVTS.2 CDSX settlement is usually completed by17:05.3 After CDSX settlement occurs, importantLVTS payment activity continues, as LVTS par-ticipants make payments for about an hour onbehalf of their clients or their own business. Thisis followed by a pre-settlement period, between18:00 and 18:30. LVTS settlement begins at18:30.
Data and Methodology
The LVTS has two payment streams, Tranche 1(T1) and Tranche 2 (T2). Each tranche is char-acterized by its own risk controls. Since CDSXsettlement takes place through T1, our studyfocuses on T1. The study makes use of payment-
2. Participants can draw on liquidity in CDSX prior toCDSX settlement through a CDSX–LVTS funds trans-fer. Anecdotal evidence suggests that this transfer isnot used frequently.
3. All times are in Eastern Standard Time.
87
Research Summaries
by-payment transactions data and intraday cred-it-limit data obtained from the Canadian Pay-ments Association, as well as paymentinstructions regarding the CDSX settlement ac-count held at the Bank of Canada. Our simulationspans the 65 business days from 1 June 2006 to31 August 2006.
Our data suggest that the value of the CDSXpay-outs and subsequent LVTS T1 activity is asignificant portion of the total daily T1 activity.Over the sample period, there were, on average,seven CDSX pay-outs each day worth $3.5 billion.This represents 2 per cent of the average daily T1volume and 16 per cent of the average daily T1value of $21 billion. The largest settlement pay-out to a single CDSX participant was $7 billion,while the daily maximum to all participants wasover $16 billion. In addition, after CDSX settles,there continues to be considerable LVTS activity,with a daily average of 17 payments of $2.6 billion.This represents approximately 12 per cent of thedaily average T1 value. During the sample period,up to 37 payments worth a total of $10.8 billionwere sent after CDSX settlement.
The main approach used is a simulation employ-ing the payments system simulator (BoF-PSS2)developed by the Bank of Finland and adaptedto replicate LVTS conditions.4 The simulatorallows us to use LVTS and CDSX data to recreateactual LVTS activity and to use this as a bench-mark. We then identify and remove the CDSXpay-outs, without removing the pay-ins, andsimulate the LVTS with these payments removed.In effect, this simulates a situation where theCDSX pay-outs are not completed. For example,an event affecting the Bank of Canada betweenpay-ins and pay-outs could result in the typeof event simulated. It is important to note thatthe simulation does not incorporate mitigatingstrategies, such as alternative payment methods,that can be used to circumvent, or at least reduce,the effect of such an event. The simulation pre-sents a possible worst-case scenario.
Results
We find that the simulated outage results in someLVTS payments being unable to settle duringthe day, and some payments being temporarily
4. For more information, see the Bank of Finland’s pay-ment simulator website at <http://www.bof.fi/en>.
88
delayed. In addition, we find that pre-settle-ment activity in the LVTS may be disrupted.
Our results indicate that a disruption to CDSXcan lead to payments that cannot settle. Unset-tled payments occur on 32 of the 65 days simu-lated. We examine how important the CDSXfunds are for end-of-day LVTS payment activityon the days with unsettled CDSX payments.First, we calculate the value of unsettled LVTSpayments as a share of the CDSX pay-outs. Wefind that, on average, the value of payments thatare unable to settle is equal to 27 per cent of theCDSX pay-outs. On some days, this value can beover 80 per cent. Second, we assess how muchof the LVTS payments that are sent after CDSXsettlement at 17:05 are affected by the simulat-ed disruption. We find that, on average, 36 percent of the T1 payments made after 17:05 areunable to settle.
The simulated outage also results in a substantialincrease in the delay of LVTS payments. Paymentsthat are unable to pass the risk controls whenthey are submitted may be entered into a queue.5
While some queued payments can subsequentlybe settled, they are delayed. We therefore exam-ine queue usage to understand the delay causedby the outage. In the base case, where CDSXsettlement occurs, the queue is used on 6 of the65 days, and in the simulated CDSX outage, thequeue is used on 39 days.
We also find that a disruption to CDSX settlementwill likely affect LVTS pre-settlement activity. Pre-settlement transfers allow participants to bringtheir end-of-day LVTS position close to zero,by making interbank payments. Pre-settlementpayments are made between 18:00 and 18:30.During the pre-settlement period, LVTS partici-pants with a positive position lend to thosein a negative position in order to bring theirpositions close to zero. Receipt of a CDSX pay-outmay cause some participants to have a positiveLVTS balance, allowing them to lend to partici-pants in negative positions. Some interbankactivity to bring positions close to zero can takeplace prior to the pre-settlement period. By ex-amining pre-settlement payments, we find thatinstitutions that did receive CDSX pay-outs are,in fact, making most of the payments during thepre-settlement period. For instance, on 35 of the65 days, over 70 per cent of the pre-settlement
5. For more information on the LVTS payment queue,see Arjani and McVanel (2006).
Financial System Review – June 2008
payments were made by recipients of CDSXpay-outs.
Contingency Measures andMitigating Actions
Contingency measures are in place that can beimplemented to prevent CDSX settlement frombeing delayed. While events affecting CDSXsettlement do happen, they are infrequent andof short duration. To ensure that importantpayments can be made during operational dis-ruptions, CDS, the LVTS, system participants,and the Bank of Canada have contingency mea-sures in place to safeguard their operations.These measures include making payments usingalternative payment methods, such as the LVTSdirect network, and moving operations to analternate site.6 These measures help to ensurethat any event that may prevent CDSX settlementfrom being completed is managed quickly, andthat the CDSX settlement payments can bemade with little or no delay.7
Furthermore, if an event does lead to a delay inCDSX settlement pay-outs, LVTS participants cantake actions to mitigate the impact of the delay.Our analysis of past operational events suggeststhat LVTS participants apportion additional col-lateral to T1 and move payments to the T2 pay-ment stream when CDSX settlement is delayed.8
Conclusions
The completion of CDSX settlement throughthe LVTS creates an important operational linkbetween these two systemically important sys-tems. Our analysis demonstrates that a disrup-tion to CDSX settlement can potentially haveimportant effects on end-of-day activity inthe LVTS. In the unlikely scenario that the CDSXsettlement funds are not available and mitigatingaction is not taken, a significant number of
6. For more information on the LVTS direct networkand other contingency measures, see the LVTS Rulesavailable on the Canadian Payments Association’swebsite.
7. For more information on the Bank of Canada’s con-tingency plans, see Allenby (2003).
8. Participants could implement other mitigatingactions that are difficult to analyze, such as changingthe order in which they submit payments so thatliquidity can be used more efficiently.
payments would be unable to settle or wouldbe delayed. In most cases, the participants wouldtake action to mitigate these impacts; for exam-ple, they may move payments to T2 or appor-tion additional collateral. Moreover, CDS, LVTSsystem participants, and the Bank of Canadacould employ contingency measures to ensurethe completion of CDSX settlement. The resultsof this study highlight the importance of well-designed systems and procedures, includingcontingency measures and mitigating actions,to safeguard the payments system.
References
Allenby, R. 2003. “Business-Continuity Plan-ning in Clearing and Settlement Systems:A Systemwide Approach.” Bank of CanadaFinancial System Review (June): 55–58.
Arjani, N. and D. McVanel. 2006. “A Primeron Canada’s Large Value Transfer System.”Available at <http://www.bankofcanada.ca/en/financial/lvts_neville.pdf>.
Canadian Payments Association. “LVTS Rules.”Available at <http://www.cdnpay.ca/rules/lvts_rules.asp>.
Embree, L. and K. Millar. 2008 “The Effects of aDisruption in CDSX Settlement on Activ-ity in the LVTS: A Simulation Study.” Bankof Canada Discussion Paper. No. 2008-7.
McVanel, D. 2003. “CDSX: Canada’s NewClearing and Settlement System forSecurities.” Bank of Canada FinancialSystem Review (June): 59–64.
89
Financial System Review – June 2008
Family Values: Ownership Structure,Performance, and Capital Structure ofCanadian FirmsMichael R. King and Eric Santor
heories of the relationship betweenconcentrated ownership and firm per-formance predict positive, negative, orno statistically significant relationship,
depending on the trade-offs between the align-ment and entrenchment effects.1 Likewise,empirical studies have produced mixed results,which may be due to two problems: one relatedto model specification, the other to model esti-mation. First, Demsetz and Villalonga (2001) andClaessens et al. (2002) argue that the relationshipbetween family ownership and performancecannot be identified without disentanglingownership (claims against the cash flow of thefirm) from control (the holding of voting rightsat the Board level).2 Studies that do not disen-tangle the alignment and entrenchment effectsof ownership and control may conflate theseeffects, leading to inconclusive results.
A second explanation for the mixed resultsrelates to unobserved firm heterogeneity: theremay be systematic differences between firmswith high and low ownership concentration.This generates an identification problem: whiletheory may suggest that causation runs fromfamily ownership to performance, an alterna-tive explanation is that causation is reversed.3
1. The alignment effect describes the positive incentiveof ownership on corporate governance. As the owner-ship stake increases, there are greater incentives forcontrolling shareholders to monitor firm performance.The entrenchment effect describes the negative conse-quences of greater ownership by managers, sincepoorly performing firms are insulated from the possi-bility of a takeover. Managers may also pursue theirprivate interests at the expense of other shareholders.
2. For the purpose of this study, we define control asholding 20 per cent or more of the firm’s voting shares.
3. Demsetz and Lehn (1985) argue that efficient mar-kets will lead to optimal ownership structures, sincefirms with inefficient ownership structures will fail tosurvive in the long run. Thus, the relationship betweenownership and performance may be endogenous.
T
One limitation of existing international studiesof ownership, performance, and capital structureis that most studies involve countries or regionswith legal, regulatory, and market institutionsthat differ markedly from those of the UnitedStates, making it difficult to disentangle firm-level effects (such as the choice of capital struc-ture, corporate governance, or managementquality) from country-level effects. Canadaprovides an ideal setting for studying this ques-tion. Canada and the United States share acommon legal ancestry, with Canadian corporateand securities laws adopted from Americanprecedents (Buckley 1997). Both countries havethe same English common-law legal system,require similar disclosure levels, and exhibitsimilar levels of shareholder protection (LaPorta et al. 1998, 2000). At the same time,Canada features more concentrated corporateownership than the United States and moreprevalent use of dual-class shares and pyramidalstructures that increase the risk of expropriationof minority shareholders (Attig 2005; Morck,Stangeland, and Yeung 2000).4 A study of Ca-nadian firms therefore provides a useful coun-terfactual assessment, since it featuresownership structures resembling those of Eu-ropean or Asian firms in an institutional settingsimilar to that of the United States.Theory
Increased ownership by insiders or the presenceof a large blockholder can sometimes lead tobetter performance. For example, greater equityownership by insiders improves corporate
4. We use the term “dual-class shares” to refer to threecategories of shares in Canada: non-voting shares,subordinate voting shares, and restricted votingshares. Pyramids occur when a blockholder controlsan apex firm or holding company that has controlstakes in a related group or chain of firms.
91
Research Summaries
performance because the monetary incentivesof the manager are better aligned with those ofother shareholders, thereby mitigating the stan-dard principal-agent problem. On the otherhand, many studies that document the preva-lence of family ownership around the worldhave expressed concerns that concentrated own-ership, particularly in the presence of control-enhancing mechanisms, may have negative im-plications for firm performance: it may contrib-ute to the entrenchment of poor managers, theexpropriation of resources from minority share-holders, capital misallocation, and reduced orinefficient investment. A high prevalence offamily ownership and control-enhancing mech-anisms may also lead to financial inefficiency,since investors would be unable to invest in aproperly diversified portfolio of widely held,and thus better-governed, firms (Morck, Stange-land, and Yeung 2000).5 Moreover, the Organi-sation for Economic Co-Operation andDevelopment (OECD 2007) notes that concen-trated ownership has led to cases of shareholderexpropriation and, subsequently, to large nega-tive externalities for financial markets. Taken to-gether, these issues have led some researchers toargue that the prevalence of family ownershipcan ultimately result in lower economic growth(Morck, Wolfenzon, and Yeung 2005).
This statement implies that if family ownershipdoes indeed have such negative effects, then pol-icy-makers may wish to consider implementingpolicies that discourage family ownership or,at the very least, discourage the use of control-enhancing mechanisms. As noted above, how-ever, empirical evidence regarding the effectsof concentrated ownership on firm performanceis mixed. It is therefore necessary to furtherexamine the relationship between family own-ership and performance to determine whethera policy response is warranted.
Methodology
Our study (King and Santor 2007) seeks to ad-dress these issues and makes four contributionsto the literature. First, we collect annual data for
5. For instance, in many countries, a large proportionof firms may be closely held and/or have control-enhancing mechanisms. Investors who wish to(or may be required to) hold a market index must,de facto, invest in such firms despite the greater riskof expropriation of minority shareholders.
92
613 Canadian firms that were members of theTSX 300 and the S&P TSX Composite Index from1998 to 2005 and identify the owner, the per-centage control of votes, the percentage cash-flowstakes, and the use of dual-class shares or pyra-midal structures in these firms. To our knowledge,this is the largest and most comprehensive data-base of Canadian ownership. Second, we distin-guish between the effects of family ownershipand control-enhancing mechanisms (specificallydual-class shares and pyramidal structures).Third, we examine the impact of ownershipstructure on both the market and accountingperformance of our full sample, using as proxiesTobin’s Q and return on assets (ROA), respec-tively.6 Fourth, we test different theories relat-ing ownership to capital structure. We are notaware of any other Canadian study that examinesthis issue.
To address the issues of endogeneity describedabove, we follow Claessens et al. (2002) Specif-ically, we use a random-effects specification toexamine the effect of ownership on firm perfor-mance and capital structure:
, (1)
where yit is either Tobin’s Q, ROA, or (for finan-cial structure) leverage (measured as the ratio oftotal debt to total assets); x is firm characteristics,namely firm size, sales growth, industry Tobin’sQ, ROA, financial leverage, firm age, member-ship in the composite index, and ratio of capitalexpenditures to sales (ROA and leverage areexcluded when they are the dependent variable);OWN is a measure of ownership, whether thesize of the control stake, dummy variables iden-tifying owner type, the use of control-enhancingmechanisms, or the size of wedge betweencontrol stakes from cash-flow stakes; it is themean-zero residual adjusted for firm-specificheterogeneity.
Results
The degree of family ownership and control-enhancing mechanisms exhibited by Canadianfirms is high relative to that in the United States:over 32 per cent of the firms in the sample arefamily owned at the 20 per cent threshold, and14 per cent have dual-class shares (compared
6. Tobin’s Q is (total assets + market value of equity-book value of equity)/total assets.
yit α β′ xit δ OWNit εit+ + +=
ε
Financial System Review – June 2008
with 20 per cent and 8 per cent, respectively,in the United States). We find that the marketperformance of free-standing, family-ownedfirms with a single-share class is similar to thatof other firms (based on Tobin’s Q ratios). Wealso find that these firms have superior ac-counting performance (based on ROA), andhigher financial leverage (based on the ratio ofdebt to total assets). These results are consistentwith the U.S. evidence in Anderson and Reeb(2003) and Villalonga and Amit (2006). In con-trast, family-owned firms with dual-class shareshave market valuations that are 17 per cent low-er, on average, than those of other firms, despitehaving similar ROA and financial leverage. Thisvaluation discount is consistent with evidencefrom U.S. and international studies that firmswith a separation between cash-flow rights andcontrol rights have lower valuations becausethey have a higher risk of expropriation of mi-nority shareholders (Claessens et al. 2002; Villa-longa and Amit 2006). This valuationdiscount is also robust when we control for Ca-nadian firms that are cross-listed on U.S. ex-changes. In summary, family ownership is notnegative for performance per se: rather, it is theuse of control-enhancing mechanisms that re-duces a firm’s valuation.
References
Anderson, R. C. and D. Reeb. 2003. “Founding-Family Ownership and Firm Performance:Evidence from the S&P 500.” Journal ofFinance 58 (3): 1301–28.
Attig, N. 2005. “Balance of Power.” CanadianInvestment Review 18 (3): 6–13.
Buckley, F. H. 1997. “The Canadian Keiretsu.”Journal of Applied Corporate Finance 9 (4):46–57.
Claessens, S., S. Djankov, J. P. H. Fan, andL. H. P. Lang. 2002. “Disentangling theIncentive and Entrenchment Effects ofLarge Shareholdings.” Journal of Finance57 (6): 2741–71.
Demsetz, H. and K. Lehn. 1985. “The Structureof Corporate Ownership: Causes andConsequences. ”Journal of Political Econ-omy 93 (6): 1155–77.
Demsetz, H. and B. Villalonga. 2001. “Owner-ship Structure and Corporate Perform-ance.” Journal of Corporate Finance 7 (3):209–33.
King, M. R. and E. Santor. 2007. “Family Values:Ownership Structure, Performance, andCapital Structure of Canadian Firms.”Bank of Canada Working PaperNo. 2007-40.
La Porta, R., F. Lopez-de-Silanes, A. Shleifer,and R. Vishny. 1998. “Law and Finance.”Journal of Political Economy 106 (6):1113–55.
———. 2000. “Investor Protection and Corpo-rate Governance.” Journal of FinancialEconomics 58 (1-2): 3–27.
Morck, R., D. Stangeland, and B. Yeung. 2000.“Inherited Wealth, Corporate Control,and Economic Growth: The CanadianDisease?” In Concentrated Corporate Own-ership, edited by R. Morck. Chicago: Uni-versity of Chicago Press.
Morck, R., D. Wolfenzon, and B. Yeung. 2005.“Corporate Governance, EconomicEntrenchment, and Growth.” Journal ofEconomic Literature 43 (3): 655–720.
Organisation for Economic Co-Operation andDevelopment (OECD). 2007. “Lack ofProportionality between Ownership andControl: Overview and Issues for Discus-sion.” Paris: OECD.
Villalonga, B. and R. Amit. 2006. “How DoFamily Ownership, Control and Manage-ment Affect Firm Value?” Journal of Finan-cial Economics 80 (2): 385–417.
93
