Embed Size (px)
Citation preview
Quantitative vs. Fundamental Analysis in Institutional Money Management: Where’s the Beef?
Russell B. Gregory-Allen∗Massey University
Hany A. Shawky University at Albany, SUNY
Jeffrey Stangl Massey University
September 2008
We thank Henk Berkman, Charles Corrado, Charles A. Trzcinka and seminar participants of the meetings of the Academy of Financial Services, FINSIA, Australian Banking and Finance Colloquium. All remaining errors are of course our own. ∗ Corresponding author
1
Quantitative vs. Fundamental Analysis in
Institutional Money Management: Where’s the Beef?
Abstract
In the money management industry, there is a “quiet” controversy over who does a better job, Traditional Managers (Fundamentalists), or Quantitative Managers. This issue has been examined by Gruber (1996), and Pastor and Stambaugh (2003) and more recently by Zhao (2006) and Wermers, Yao and Zhao (2007) using mutual fund portfolios. We reexamine this issue using the Plan Sponsor Network Database (PSN), a survivorship free database, which reports on how managers actually manage investment portfolios with respect to both style and types of stock selection methods used. Our empirical results indicate that when examining marginal performance that is purely attributable to the use of a distinct Primary Investment Process, only the Fundamental approach is shown to significantly add value.
2
I. Introduction There are two main approaches used in the selection of stocks in actively managed equity
portfolios. The first is the traditional approach which is based on fundamental analysis where
managers’ research and analyze the unique aspects of individual firms.1 The second approach is
the more quantitative approach, in which managers use pre-set mechanical models to identify
stocks. The popularity of the quantitative approach is attributed to the belief that it has the
potential to be less susceptible to cognitive errors and biases, and given the empirical power of
asset pricing theory, we expect quantitative management to do better than random chance. On the
other hand, fundamental analysis might also do better than random chance by simply using
accounting variables that incorporate beta, value/growth, and market value factors.
A few recent studies address this issue using mutual fund data. Zhao (2006) directly
examines how the stock selection approach affects mutual fund performance and economics of
scale. She characterizes the quantitative managers as “Quant Jocks” and the traditional managers
as “Tire Kickers”. She finds that there is no significant difference in their investment
performance. Moreover, she finds that although managers can cheaply screen a large universe of
stocks, the stocks that they invest in are smaller and less liquid, which results in higher
transaction costs and limited scalability of quantitative investment strategies.
Wermers, Yao and Zhao (2007) also examine the differences in the performance between
mutual fund managers that employ quantitative approaches and managers that use fundamental
analysis in their selection process. They find that employing quantitative models that are largely
designed to take advantage of known market anomalies does not produce above average
1 This approach is attributed to Graham and Dodd, and goes back about 75 years.
3
performance results.2 They then infer that any above average performance obtained by skilled
fund managers must have been generated by unique fundamental information on individual
stocks.
Several other recent studies address the issue of Quantitative vs. Fundamental using
mutual fund data. For instance, Casey and Quark (2004) examine a group of active Quantitative
funds, with which they have personal knowledge, and find that the Quants outperformed the
Fundamental managers. Similarly, Ahmad and Nanda (2005) examine a sample of Enhanced
Index funds, and use prospectuses to classify each fund as Quant or Fundamental. They find that
Quant managers outperform Fundamental managers in small-cap funds. Both Casey and Quark
(2005) and Ahmad and Nanda (2005) had fairly small samples, thus care must be exercised in
generalizing their findings.
What all of the above cited studies have in common is that they use their own
interpretation of the prospectuses to stratify funds into either Quantitative or Fundamental,
ignoring the subtleties of some of these funds being Bottom Up or Top Down.3 More
importantly, this stratification also obscures the potential for a particular manager to be using a
secondary technique. For example, while a given manager might be primarily Quantitative, and
uses a Top Down approach, they might also have in their stock selection model some
Fundamental ratios, meaning that they use a combination of Quantitative, Top-Down, and
Fundamental approaches. Fabozzi, Focardi, and Jonas (2008) specifically note this tendency,
finding that more than half of managers use what they call a “hybrid approach”, utilizing both
Quantitative and Fundamental techniques.
2 Among the market anomalies examined are; momentum investing, earnings revisions, accounting accruals, low PE, and other financial indicators. 3 "Bottom Up" simply refers to picking stocks starting with the individual stocks, while "Top Down" means picking stocks starting with the market.
4
This paper exploits a unique database to reexamine the investment performance
differences between quantitative and fundamental stock selection approaches.4 The Plan Sponsor
Network Database (PSN) is developed and marketed by INFORMA and reports on how
individual managers manage their portfolios with respect to both investment style and the types
of models used. We are able to explicitly distinguish between managers that employ a
quantitative approach and those that use fundamental methods in selecting stocks. Such
information can only be inferred for commonly used mutual fund data.5
Our empirical results indicate that when examining marginal performance that is
attributable to the use of a specific Primary Investment Process (Bottom-Up, Top-Down,
Fundamental, Quantitative, Computer Screening, and Technical Analysis), only the Fundamental
approach is found to significantly add value. Our results are consistent with the findings of
Wermers, Yao and Zhao (2007) that fundamental methods of stock selection are superior to
quantitative approaches. Moreover, when we examine the performance of individual strategies at
the margin and after controlling for the primary strategy, none of the secondary processes add
value.
This paper is organized in six sections. Section II reviews some of the related literature.
We describe some industry terminology in terms of the investment variables that are used in our
estimation models in section III, while section IV describes many aspects of the PSN data.
4The only exception that we are aware of is a study by Faugere, Shawky and Smith (2004) in which they use the PSN data to examine the issue of “sell discipline” in institutional portfolios. 5 A significant advantage of using this data is to examine a group of managers who are most likely to benefit from quantitative analysis. They manage portfolios much larger than mutual funds and can predict the inflows and outflows much better since they personally know their clients who are typically advised by quantitative consultants.
5
Section V presents the empirical models and discusses our findings. A summary and some
concluding remarks conclude the paper.
II. Related Literature
According to Fabozzi, et al., (2002), the use of Modern Portfolio Theory and fundamental
analysis is the most widely used methodology for portfolio selection. However, some authors
believe that an actuarial approach using a probabilistic method is more appropriate for building
efficient portfolios.6 While early studies have proved that technical analysis as a whole is not
effective in predicting rate of return and stock market movements, more recent studies by Brock,
Lakonishok and Lebaron (1992) have found that a relatively simple set of technical trading rules
have predicted with significant consistency changes in the Dow Jones Industrial Average over a
long sample period. Others have reported that this same set of rules can be applied for
movements in the Asian Stock Markets and foreign currency markets.
Berk and Green (2004) develop a rational equilibrium model of active fund management.
They assume decreasing returns to scale, the presence of competitive provision of capital by
investors and that there is differential ability to generate high returns across managers that is
learned through past performance. They conclude that active managers do not outperform
passive benchmarks and that manager performance does not persist. An important implication of
their model is that fund managers have skills and that the level of skill is not uniform. In
addition, their model implies that skill of a fund manger is not measured by the fund
performance but by the fund size.7
6An actuarial approach is one that is qualitative in nature relying on statistics and stock price movements rather on quantitative analysis of risk and return. 7 Berk and Green (2004) argue that in equilibrium, better managers manage larger funds and all managers who hold portfolios of the same risk are expected to earn the same return regardless of their skill level.
6
Much of the literature on portfolio performance, market anomalies and investment style
has been conducted using mutual fund portfolios. As a result, the literature is replete with
evidence to suggest that active mutual fund management underperforms passive benchmarks.
Very early on, Jensen (1968) documents that professionally managed funds do not beat an index
portfolio. Elton, Gruber, Das and Hlavka (1993), Malkiel (1995), Carhart (1997), Chang and
Lewellen (1985), Pastor and Stambaugh (2003), and Kothari and Warner (2001) reach similar
conclusions.
On the other hand, a growing number of studies seem to give more credit to professional
money managers. For instance, Ippolito (1989) finds that mutual fund risk-adjusted returns, net
of fees and expenses, are comparable to index funds, and Hendricks, Patel and Zeckhauser
(1993) document that good mutual fund managers have “hot hands”. Moreover, studies by
Grinblatt and Titman (1989, 1992), Wermers (1997), Daniel, Grinblatt, Titman, and Wermers
(1997) and Chen, Jagadeesh, and Wermers (2000) show that mutual fund mangers do actually
have some skills in selecting stocks.
III. Industry Terminology
Investment managers can be identified as Growth or Value, Quantitative or Fundamental,
and Passive or Active. Growth/Value refers to the target universe that a manager works within.
Quantitative/Fundamental describes the investment approach the manager applies in making
investment decisions, which is our main focus. Passive/Active describes how closely the
manager adheres to a benchmark, usually measured by estimating tracking error.8 Previous
8A fund manager who selects from stocks that are listed in the Russell 3000 Growth Index is said to be managing a Growth fund and following a Growth style. However, that says nothing about the process by which the manager arrives at the stock selection, and it gives no information about tracking error.
7
literature has often confounded these distinctions. To distinguish between these attributes, we
provide a detailed characterization of these investment terms that are used in our empirical
models and throughout the remainder of the paper.
Investment Policy
Investment funds are classified as Active, Enhanced Index, or Passive. We will refer to
this distinction as Investment Policy. Active funds have the least focus on tracking error and
sometimes are even without a clearly specified benchmark. Passive funds, at the other extreme,
strive for performance that closely matches a specific benchmark, reducing tracking error to as
little as 10-20 bps for a portfolio that attempts pure replication. The third type, Enhanced Index,
is a hybrid between the two extremes in which active risk is controlled by targeting a modest
tracking error, usually in the range of 50-100 bps, and various techniques are employed to
attempt to outperform the benchmark.
Investment Style
Investment Style refers to a manager’s choice of a specific universe. The majority of
funds are either Growth or Value, with nearly as many identified as blend, sometimes called
Balanced, Neutral or Core funds. Market Capitalization can also be considered part of this
categorization. Managers choose to focus on a specific market capitalization category such as
Large Cap, Mid Cap, or Small Cap. Whatever the investment style, this represents the target
universe the fund or manager has chosen, rather than the process used to make investment
decisions.
8
Investment Process
The approach a manager uses when selecting stocks is called the Investment Process.
This is the primary focus of this research. Two common investment processes are fundamental
and quantitative and our hypothesis mainly addresses the relative performance of these two
investment approaches. In addition to these two investment processes, the investment community
also focuses on the distinction between “Bottom Up” and “Top Down.” A manager who
emphasizes individual stocks is said to be Bottom Up, and a manager who focuses on
macroeconomic variables is said to be Top Down. While Fundamental managers tend to be
Bottom Up and Quantitative managers tend to be Top Down, there can be considerable overlap.
Moreover, Quantitative, Fundamental, Bottom Up or Top Down managers can be Growth or
Value, Large Cap or Small Cap.
IV. Description of Data
The PSN Database contains quantitative and qualitative information on over 11,000
independent equity and fixed income portfolios managed by 2000+ companies. The manager of
each of the portfolios fills out a rather lengthy questionnaire, and PSN collects this information
into a flexible and searchable database. This data is marketed to investment professionals,
primarily Pension Plan Sponsors, Endowments, Foundations and corporate and institutional
money mangers who use it as a tool to identify and select investment managers. Mangers have a
strong economic incentive to be complete and accurate in their reporting since PSN is the only
database widely used by institutions to identify managers.
To ensure that our sample is relatively homogenous with respect to active risk, we limit
our study to funds that have designated themselves to be "Active" funds. Passive funds and
9
especially Enhanced Index funds may employ some of the same techniques we discuss here, but
those funds have policy restrictions on the types of investments they can make, and may impose
restrictions on the weights of their holdings and their tracking error. This data cover exclusively
U.S. funds, and we further restrict our sample to equity funds, since the investment process of
fixed income managers can often be quite different.
One of the queries in the PSN questionnaire asks about the Investment Processes that
manager’s use. The top 10 Investment Processes are: Bottom Up, Fundamental Analysis,
Quantitative Research, Computer Screening/modeling, Top Down, Future Earnings Growth,
Theme Identification, Low Price/Earnings, Low Price/Book and Industry Sector Analysis, with
the majority of funds using one or more of the top 5. Table 1 lists all the Investment Processes
available in the PSN questionnaire, and the number of funds in our sample that use each
investment process.
Insert Table 1 about here
While we are particularly interested in the differences in the performance of
Quantitative/Research and the Fundamental Analysis approaches, we also consider other
investment methods that seem to be widely used in industry. For instance, Top-Down and
Bottom-Up, are thought of as being closely related to these two approaches, so we will examine
those as well. Since Quantitative managers typically employ computers, use modeling, and
might use screening, these managers could rank Computer Screening/ Modeling as very high,
hence we include that technique too. Finally, Technical Analysis is often thought of as being at
10
the other end of the spectrum from Fundamental Analysis and scores very low in use, but is
worthy of inclusion for comparative purposes.9
Managers are asked about their Investment Process in two ways: 1) what is the Primary
method they use, and 2) how important is each of these methods to their investment process.10
The difference between these two responses is that under the "importance ranking" managers
can, and frequently do, rank several styles equally as "Very Important", while with the Primary
method, they must choose only one method as being the most important. This double selection
gives us the flexibility to not just categorize funds by method, but also gain insight into what
other methods they use and to what degree. Table 2 shows the number of funds for each of our
examined Primary Processes, and the proportion of funds using each of the Ranked methods.
Insert Table 2 about here
With respect to the Primary method used by individual managers, "Bottom Up" is the
overwhelming favorite, but Fundamental and Quantitative are both well represented. What is
also apparent from Table 2, is that no manager uses one method exclusively - they all use at least
one other Ranked (secondary) method. This lends credence to our suspicion that other studies
may lose important information by classifying funds into mutually exclusive groups.
Surprisingly, some managers use multiple techniques that appear to be contradictory. For
example, 8% of Bottom-Up funds also use Top-Down and 52% of Top-Down funds use Bottom-
9 The remaining techniques are less representative of an identifiable investment approach and we will examine their impact using control variables. 10For each of the Ranked Investment Processes, the managers rank its importance on a scale of 1 denoting "Very Important" to 5 indicating "Not Utilized".
11
Up techniques. Similarly, 31% of Fundamental funds use Quantitative, and 40% of Quantitative
funds use Fundamental techniques.11
Table 3 shows the correlations between the Primary and the Ranked Process variables.
This demonstrates that the grouping of funds together, say, Bottom-Up and Fundamental, is
probably not representative of how managers work, even though one might be tempted to do so.
For example, at the Primary Process level, Bottom-Up and Fundamental funds are negatively
correlated (-0.57), and for Bottom-Up managers, the correlation with the ranked Fundamental
techniques is only 0.15, while for Fundamental managers, the use of Bottom-Up techniques has a
-0.04 correlation.
Insert Table 3 about here.
In addition to Investment Process, PSN also collects information on the Investment Style
of each fund. PSN identifies a total of 25 Investment Styles which managers can select from. As
with Investment Process, managers both 1) pick a Primary style, and 2) rank all 25 styles as to
their relative importance. While Investment Style is not the focus of this paper, using these
control variables allows us to distinguish between return performance that can be attributed to
the Investment Process and performance that is due to weighting on, say, Growth.
Individual fund returns are provided by PSN, as reported by the funds. To alleviate
concerns about using "self-reported" returns, we limit our sample to funds that are AIMR
compliant. Compliance gives some assurance that the return calculations are consistently
generated, and that they are accurate. In addition, these data are used widely by the industry and
11 One possible view of this is that the managers are confused - either about how they work, or about how to fill out the questionnaire. However, we think it reflects the way managers operate -- they use multiple techniques and strategies.
12
it very unlikely that errors or misrepresentations would persist undetected. Further, PSN
themselves run consistency checks on their data.
We have restricted our sample to firms (fund families) with a minimum of $100 million
in assets under management, and funds with a minimum of $1 million. We use monthly returns,
in excess of fees and the U.S. 90-day T-bill rate. We exclude funds with less than 2 years of
data, and set as “missing”, return observations that are identified as outliers.12 Our examination
period is January 2002 to December 2006.13 After implementing all screens, our sample had 947
funds.
V. Methodology and Results
In order to examine the marginal performance of individual fund managers that is due to
the use of a given investment process, we use a two stage procedure. First, we estimate a
performance alpha using the Carhart 4-factor model (1997) as follows:
(1) 1 500 2 2 3i i SPR R SMB HML MOM iα β β β β= + + + + +ε
where:
iR is the return for fund i, in excess of fees and the 90 day US Tbill rate
500SPR is the return on the S&P500, in excess of the 90 day US Tbill rate SMB, HML and MOM are the Fama-French and Carhart factors
And in a second stage, we use these alphas in a cross-sectional dummy variable regression to
estimate the marginal contribution of the various factors to performance.14 We define dummy
variables for "Bottom Up", "Top Down", "Quantitative", "Fundamental Analysis", "Technical
12 Monthly returns exceeding 5 standard deviations (of entire sample). 13 Although we have longer returns data available, funds do change styles over time, and we want to be reasonably certain that our categorizations are consistent throughout our examination period. 14 We also examined alphas from the Fama French 3-factor model (1993), with qualitatively the same results.
13
Analysis" and "Computer Screening." For each ranked variable, we assign the dummy a value of
1 if the managers rank it "Very Important" or "Important", and 0 otherwise. After controlling for
Investment Style, Size, and the Primary Processes, the contribution of each of the Ranked
Processes is estimated using the following model:
(2) 0 1 2 3 4 5
6 7 8 9
1 2 3 4 5 6
i
i i i i
i i i i i
Gro Val Cor Small MidPQ PF PB PTBotUp TopDwn Fndmntl Qnt Cmp Techi i
α γ β β β β ββ β β βγ γ γ γ γ γ
= + + + + + ++ + + +
+ + + + + ε+
where for fund i:
iα is the Carhart alpha, Dummies for the most common Investment Style variables, Growth, Value and Core ("other" omitted), and Fund Size, Dummies for the Primary processes: PQ = 1 if the manager identifies the fund's Primary process as Quantitative, PF = 1 if the manager identifies the fund's Primary process as Fundamental PB = 1 if the manager identifies the fund's Primary process as Bottom-Up PT = 1 if the manager identifies the fund's Primary process as Top-Down and dummies for the Ranked processes: BotUp, TopDwn, Qnt, Fndmntl, Cmp and Tech = 1 if the manager ranks it "Very Important" or "Important"
Insert Table 4 about here
Table 4 reports the estimated results of the contributions of each of the Primary processes
as well as the contribution of each of the ranked investment processes. For the Primary
Processes, only Fundamental is shown to add value. For the Secondary processes, measured by
the ranked variables, none of the Processes have a positive impact on return, and Computer
Screening demonstrates a negative impact.
14
Marginal Contribution of Additional Primary Investment Processes
While our results appear to suggest that only the Fundamental approach is shown to
contribute positively to portfolio returns, it is still possible that within a specific Primary Process,
using other additional techniques may be beneficial. For example, we saw in Table 2 that 42%
of the Bottom-Up funds also use Computer Screening techniques. Even though the Primary
Process did not help the funds’ relative performance, it could be that those funds do benefit from
using Computer Screening.
To examine the marginal impact of using more than one Investment Process by fund
managers, we add interaction terms to our model in equation (2) as follows:
(3)
0 1 2 3 4 5 6 7 8 9
1 2 3 4 5
6
* * * * *
*
i i i i i
i i i i
i
Gro Val Cor Small Mid PQ PF PB PT
PB TopDwn PB Fndmntl PB Qnt PB Cmp PB Tech
PT BotUp
α γ β β β β β β β β β
γ γ γ γ γ
γ
= + + + + + + + + + +
+ + + +
+ 7 8 9 10
11 12 13 14 15
16 17 18
* * * *
* * * * *
* * *
i i i
i i i i
i i i
PT Fndmntl PT Qnt PT Cmp PT Tech
PT BotUp PF TopDwn PF Qnt PF Cmp PF Tech
PT BotUp PQ TopDwn PQ Fndmntl
γ γ γ γ
γ γ γ γ γ
γ γ γ
i +
i
i
+ + +
+ + + +
+ + + 19 20 * *i iPQ Cmp PQ Tech
+
+
γ γ ε+ +
where each of the dummies are defined as in equation (2). The interaction terms are the cross
product terms. For example, the term PB*TopDwn equals 1 if the fund is both Primary Top-
Down and the fund manager ranks Top-Down as "Very Important" or "Important", and zero
otherwise.
Insert Table 5 about here
Table 5 reports the results of estimating equation (3). Similar to earlier results, none of
the individual or combination of techniques adds value at the margin. In addition, when
15
managers diverge from their main technique and use combinations of different techniques, their
performance suffers. For instance, the use of Fundamental techniques is shown to significantly
hurt Quant funds, and using Computer Screening impairs Bottom-Up funds.
Finally, we examine the issue of whether using multiple techniques improves
performance, in the sense of diversification benefits. This is accomplished by creating a simple
"process diversification" metric, which merely counts how many secondary processes a given
manager uses.15 We postulate that using multiple investment processes is likely to produce better
performance through enhanced diversification, as measured by:
( ), , , , ,Div sum Bot Top Fnd Qnt Cmp Tech=
To examine whether using multiple techniques adds value to any one specific Primary Process
(denoted as divPB, divPT, divPF and divPQ), we estimate, after controlling for other variables as
before:
(4)
0 1 2 3 4 5
6 7 8 9
1 2 3 4
i
i i i i
i i i i
Gro Val Cor Small MidPQ PF PB PTdivPB divPT divPF divPQ i
α γ β β β β ββ β β βγ γ γ γ
= + + + + + ++ + +
+ + +
+ε+
As shown in Table 6, we find that the impact from Process Diversification is negative and
statistically significant on Bottom-Up funds and not significant for Top-Down and Quantitative
managers. On the other hand, we find that Process Diversification has a significant positive
contribution on Fundamental managers.
Insert Table 6 about here
15 We thank Larry Rose for suggesting this metric.
16
VI. Summary and Conclusions
We examine the marginal contribution of the most commonly used Investment processes
(Bottom-Up, Top-Down, Fundamental, Quantitative, Computer Screening, and Technical
Analysis) on performance. We find that when examining the return performance that is
attributable to a distinct Primary Process, only the Fundamental approach significantly adds
value. When we examine the performance of these strategies at the margin and after controlling
for the primary process, none of the secondary processes add value. Furthermore, Quant
managers are hurt by using Fundamental techniques, and for Bottom-Up funds, the use of a
secondary process, especially Computer Screening, negatively impacts performance.
Our main finding in this paper is that the traditional approach which is based on
fundamental analysis where managers’ research and analyze the unique aspects of individual
firms performs better than any other stock selection approach. It is therefore possible to offer the
compelling argument that given the empirical power of asset pricing theory, fundamental
analysis might perform better than random chance by simply using accounting variables that
incorporate beta, value/growth, and market value factors.
Our finding that only Fundamental analysis adds value seems to be contrary to the results
presented in Zhao (2006). However, there are substantial differences in the classification of
funds in our study as compared with others. It is important to note that Zhao (2006) and others
categorized as Fundamental, funds that identify themselves as Quantitative but mention
fundamental techniques in the prospectus. This is done in order to capture the Fundamental
influence, assuming that the final stock-picking decisions are based on fundamental analysis.16
16 Zhao (2006) p. 14
17
However, that simplification in large part explains the differences between their findings
and our results. We find that the group of Quant funds using Fundamental techniques have
significant negative performance due to the use of Fundamental, and Zhao (2006) puts that
group into Fundamental. By doing so, Zhao’s approach biases upward the performance of the
Quant group and biases downward the performance of the Fundamental group.
With respect to the choice of a Primary Investment Process, after controlling for market
and investment style risk, only Fundamental managers adds value, however none of the others
detract. This is broadly consistent with market efficiency and with the Berk and Green (2004)
model that no fund manager should consistently outperform any other manager. However, after
controlling for market and investment style risk and the Primary Investment Process, no
technique or process adds value at the margin. Further, for Bottom-Up and Quantitative funds,
adding secondary processes actually hurts performance.
18
Table 1. Primary and SecondaryProcesses Complete list of "Decision Styles" as recorded by PSN, for all funds in our sample.
Investment Process Primary SecondaryNumber of Funds Number of Funds listing this process ranking this process as
as Primary "Very Important" or "Important"
* Bottom Up 684 899* Top Down 22 143* Fundamental Analysis 106 830* Quantitative Research 76 387** Computer Screening/modeling 20 461** Technical Analysis 1 104Cash Flow 0 561Dvd Discount 0 167Dvd Growth 0 185Dvd Yield 0 162Earnings Surprise 2 316Future Earnings Growth 11 659Industry Sectory Analysis 3 379Low Price/Book 5 353Low Price/Cash 0 397Low Price/Earnings 9 463Low Price/Earnings - Expected (PE) 0 403Low Price/Sales 0 265Momentum 1 191Quality 0 583ROA / ROI 0 558Theme Identification 6 203Yield Analysis 1 108
Total 947 N/A * These 4 are our focus for Primary Processes. Remaining form "Other Primary Processes". ** In addition to above 4, these are our focus for Secondary Processes. Remaining form "Other Secondary Processes".
19
Table 2. Percentage of Funds from each Primary Process, using each Secondary (Ranked) Process
TotalPrimary Proccess in sample Bot Top Fnd Qnt Cmp Tch
Bottom Up 692 100% 8% 92% 33% 42% 9%Top Down 23 52 100 78 57 52 39
Fundamental 100 93 33 100 31 43 14Quantitative 68 51 13 40 100 63 4
Other 64 92 21 82 52 50 18
Total 947
Ranked Proccesses
20
Table 3. Correlations of Primary and Ranked Processes PB PT PF PQ Bot Top Fnd Qnt Cmp Tch
PB 1.00 -0.25 -0.57 -0.47 0.33 -0.29 0.15 -0.24 -0.16 -0.09PT 1.00 -0.06 -0.05 -0.33 0.38 -0.05 0.06 0.01 0.14PF 1.00 -0.10 -0.04 0.16 0.13 -0.03 -0.01 0.03PQ 1.00 -0.28 0.05 -0.32 0.35 0.27 -0.02Bot 1.00 -0.20 0.28 -0.13 -0.04 -0.14Top 1.00 -0.03 0.05 0.01 0.12Fnd 1.00 -0.24 -0.17 -0.01Qnt 1.00 0.50 0.18Cmp 1.00 0.18Tch 1.00
21
Table 4. Marginal contribution of Secondary (Ranked) Processes across all Primary Processes Regressions (from Equation 1) of Carhart alphas estimated on Control variables (Gro, Val, Cor, Small, Mid) and Primary Processes dummies (Bottom Up, Top Down, Fundamental, Quantitative) and Secondary-Ranked Process dummies (Bottom-Up, Top-Down, Fundamental, Quantitative, Computer Screening, and Technical Analysis) using White standard errors.
Coef. t Coef. t
intercept -0.04 -0.96 0.00 0.08Control Log(Fund Size) 0.01 0.44 0.00 0.30
Variables Growth target -0.18 -7.31 -0.18 -7.31Value target 0.01 0.31 0.01 0.38
Primary Primary BottomUp 0.00 -0.01 0.00 -0.07Process Primary TopDown 0.05 0.82 0.05 0.73
Dummies Primary Fundamental 0.13 2.71 0.13 2.70Primary Quantitative -0.02 -0.52 0.00 -0.01
Secondary Secondary BottomUp -0.01 -0.15Process Secondary TopDown -0.01 -0.35
Dummies Secondary Fundamental -0.01 -0.35Secondary Quantitative 0.00 -0.13
Secondary Computer Screening -0.04 -2.02Secondary Technical Analysis 0.02 0.64
N 947 947R2 0.10 0.11
Estimation IIEstimation I
22
Table 5. Marginal contribution of Secondary (Ranked) Processes for each Primary Process Regressions (from Equation 2) of Carhart alphas estimated on Control variables (Gro, Val, Cor, Small, Mid, Primary Bottom Up, Primary Top Down, Primary Fundamental, Primary Quantitative) and interaction dummies (Primary Bottom-Up and Ranked Top-Down; Primary Bottom-Up and Ranked Fundamental, etc.) using White standard errors.
Coef. t NumDummy=1
intercept -0.04 -1.06Control Log(Fund Size) 0.01 0.64
Variables Growth target -0.18 -7.41 349Value target 0.00 0.20 299
Primary BottomUp 0.02 0.32 692Primary TopDown -0.06 -0.32 23
Primary Fundamental -0.04 -0.33 100Primary Quantitative 0.08 0.78 68
Interaction Primary BottomUp and Secondary: TopDown -0.03 -0.56 56Dummies Fundamental 0.02 0.31 635
Quantitative 0.00 0.01 230ComputerScreening -0.07 -2.71 294Technical Analysis -0.02 -0.62 60
Primary TopDown and Secondary: BottomUp 0.09 0.71 12Fundamental -0.01 -0.06 18Quantitative 0.08 0.54 13ComputerScreening 0.08 0.68 12Technical Analysis -0.06 -0.56 9
Primary Fundamental and Secondary: BottomUp 0.06 0.54 93TopDown 0.08 0.88 33Quantitative 0.05 0.60 31ComputerScreening 0.11 1.18 43Technical Analysis 0.15 0.99 14
Primary Quantitative and Secondary: BottomUp 0.06 1.46 51TopDown -0.04 -0.63 13Fundamental -0.11 -2.45 40ComputerScreening -0.09 -0.94 63Technical Analysis 0.10 1.47 4
N 947R2 0.14
23
Table 6. Marginal contribution of "Process Diversification" - the Number of Ranked Processes used by each Primary Process Regressions (from Equation 3) of Carhart alphas estimated on Control variables (Gro, Val, Cor, Small, Mid, Primary Bottom Up, Primary Top Down, Primary Fundamental, Primary Quantitative) using White standard errors, and includes the diversification metrics, counted for each process separately
Coef. t
intercept -0.05 -1.15Control Log(Fund Size) 0.01 0.72
Variables Growth target -0.18 -7.46Value target 0.00 0.22
Primary BottomUp 0.09 1.95Primary TopDown -0.08 -0.51
Primary Fundamental -0.17 -1.50Primary Quantitative 0.14 1.49
Diversification Total Secondary Processes for BottomUp -0.03 -2.61Metric Total Secondary Processes for TopDown 0.03 1.09
Total Secondary Processes for Fundamental 0.09 2.45Total Secondary Processes for Quantitative -0.04 -1.67
N 947R2 0.13
24
References Ahmed, P. and S. Nanda (2005). Performance of Enhanced Index and Quantitative Equity Funds.
Financial Review, 40, 459-479.
Bessembinder, Hendrick (1998). Market efficiency and the returns to technical analysis. Financial Management. 27, 5.
Berk, J. and R. Green, (2004), Mutual funds flow and performance in rational markets, Journal of Political Economics,112, 1269-1295.
Brock, William & Lakonishok, Josef & LeBaron, Blake, 1992. Simple Technical Trading Rules and the Stochastic Properties of Stock Returns, Journal of Finance, 47(5), 1731-64.
Casey, Quirk & Associates (2005) The Geeks Shall Inherit the Earth? CQA Research Insight. November, 1-4.
Carhart, M. (1997). On persistence in mutual fund performance, Journal of Finance 52, 57-82.
Chang, E. and Lewellen. (1985). An Arbitrage Pricing Approach to Evaluating Mutual Fund Performance, Journal of Financial Research, 8, 15-30.
Chen, H., N. Jagadeesh and R. Wermers. (2000). The value of active mutual fund management: An examination of the stock holdings and trades of fund managers, Journal of Financial and Quantitative Analysis 35, 343-368.
Cottle, Sidney., & Murray, F. Roger., & Block, E. Frank. (1988). Graham and Dodd's Security Analysis (5th ed.). New York: McGraw-Hill Book Company.
Daniel, K., M. Grinblatt, S. Titman, and R. Wermers. (1997). Measuring mutual fund performance with characteristic-based benchmarks, Journal of Finance 52, 1035-1058.
Elton, E., M. Gruber, S. Das and M. Hlavka. (1993). Efficiency with costly information: A reinterpretation of evidence for managed portfolios. Review of Financial Studies 6, 1-22.
Fabozzi, J.,Frank., Sergio M. Focardi, and Caroline Jonas. (2008) Research Foundation Publications, Challenges in Quantitative Equity Management: 1-110.
Fabozzi, J.,Frank., and Gupta, Francis., and Markowitz, M. Harry. (2002). The legacy of Modern Portfolio Theory. Journal of Investing, 11, 7(16).
Grinblatt, M. and S. Titman. (1989). Mutual fund performance: An analysis of quarterly portfolio holdings, Journal of Business 62, 393-416.
Grinblatt, M. and S. Titman. (1992). The persistence of mutual fund performance, Journal of Finance 47, 1977-1984.
Hendricks, D., J. Patel, and Zeckhauser. (1993). Hot Hands in Mutual Funds: Short-Run Persistence of Relative Performance, 1974-1988, Journal of Finance 48, 93-130.
Ippolito, R. (1989). Efficiency with costly information: A study of mutual fund performance 1965-1984, Quarterly Journal of Economics 104, 1-23.
Jegadeesh, N, and Titman, S. (2001). Profitability of Momentum Strategies: An Evaluation of Alternative Explanations. Journal of Finance 56 (2), 699-720.
25
Jensen, M. C. (1968). The performance of mutual funds in the period 1945-1964, Journal of Finance 23, 389-416.
Kothari, S. and J. Warner. (2001). Evaluating Mutual fund Performance, Journal of Finance 56, 1985-2010.
Malkiel, B. (1995). Returns from investing in Equity Mutual Funds 1971 to 1991, Journal of Finance 50, 549-572.
Melinkoff, Meyer. (1988). Investment performance analysis for investors. Journal of Portfolio Management, 25.
Pastor, L. and R. Stambaugh. (2003). Liquidity Risk and Expected Stock Returns, Journal of Political Economy 111, 642-685.
Neftci, N. Salih. (1991). Naive trading ndes in financial markets and Wiener-Komolgorov prediction theory: a study of "technical analysis." The Journal of Business, 549.
Smith, Tom. (2002, July 29). Investment strategies - Quantitative analysis - The art of investing with a clear head. Financial Times.
Wermers, R. (2000). Mutual fund performance: An empirical decomposition into stock-picking talent, style, transaction costs, and expenses, Journal of Finance 55, 1655-1695.
Wermers, R. T. Yao, and J Zhao. (2007). The Investment Value of Mutual Fund Portfolio disclosure, Working Paper, University of Maryland.
Zhao, J. (2006) Quant Jocks and Tire Kickers: Does the Stock Selection Process Matter?, working paper.
26
