Quantitative vs. Fundamental Analysis in Institutional Money Management: Wheres the Beef?

Russell B. Gregory-Allen Massey University r.gregory-allen@massey.ac.nz Hany A. Shawky University at Albany, SUNY h.shawky@albany.edu Jeffrey Stangl Massey University j.stangl@massey.ac.nz

April 2009

We thank Henk Berkman, Stephen Brown, Charles Corrado, Oguzhan Dincer, Larry Rose, Charles Trzcinka and seminar participants of the meetings of the Academy of Financial Services, FINSIA, Australian Banking and Finance Colloquium, and Financial Management Association. All remaining errors are of course our own.

Corresponding author

Electronic copy available at: http://ssrn.com/abstract=1269715

Quantitative vs. Fundamental Analysis in Institutional Money Management: Wheres 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 recently been examined by Zhao (2006), Wermers, Yao and Zhao (2007) and others. We reexamine this issue over the period January 2002 December 2006 using the Plan Sponsor Network Database (PSN), which reports how managers actually manage investment portfolios regarding the types of stock selection methods used. This allows us to examine not just the Primary method used (quantitative, fundamental, or other), but also the impact of Secondary methods. Our empirical results indicate that when examining performance attributable to the use of a distinct Primary Investment Process, only the Fundamental approach significantly adds value. However, when examining marginal performance of a Secondary Process, over and above a Primary approach, no process adds value, and in fact some detract.

JEL classification: G23 Keywords: Mutual Funds, Quantitative, Fundamental

2Electronic copy available at: http://ssrn.com/abstract=1269715

I. IntroductionThere are two main approaches used in the selection of stocks in actively managed equity portfolios. The first is the traditional approach based on fundamental analysis, where managers research and analyze the unique aspects of individual firms.1 The second is the quantitative approach, in which managers research models, and then mechanically use those 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.2 In a recent study on the challenges facing the investment community, Fabozzi, Focardi, and Jonas (2008) identify an investment process as fundamental if it is performed by human asset managers using information and judgment, while a quantitative process is one in which the value-added decisions are based on quantitative outputs generated by computer-driven models using fixed rules. Further, they refer to a process as being hybrid if a manager uses a combination of the two approaches. Several recent studies examine the relative performance of these investment processes using mutual fund data. Zhao (2006) directly examines how the stock selection approach affects mutual fund performance and economies 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 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 average1 2

This approach is attributed to Graham and Dodd, and goes back about 75 years. 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.

3

performance results.3 They then infer that any above average performance obtained by skilled fund managers must have been generated by unique fundamental information on individual stocks. Casey and Quark (2004) examine a group of active Quantitative funds, and find that Quantitative managers outperformed Fundamental managers. Similarly, Ahmad and Nanda (2005) examine a sample of Enhanced Index funds and use prospectuses to classify each fund as Quantitative or Fundamental. They find that Quantitative managers outperform Fundamental managers in small-cap funds. However, both Casey and Quark (2005) and Ahmad and Nanda (2005) had fairly small samples and thus care must be exercised in generalizing their findings. What all of these 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 managers being Bottom Up or Top Down. Further, this stratification also obscures the potential for a particular manager being able to use 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.4 Much of the related literature on portfolio performance, market anomalies and investment style has been conducted using mutual fund data. 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 Wermers3

Among the market anomalies examined are; momentum investing, earnings revisions, accounting accruals, low PE, and other financial indicators. 4 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.

4

(1997) and Chen, Jagadeesh, and Wermers (2000) have shown that mutual fund mangers do have some skills in selecting stocks. More recently, 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 non-uniform skills, and that skill of a fund manager is not measured by the fund performance but by the fund size.5 In this study, we use a unique database to reexamine the investment performance between the quantitative and the fundamental stock selection approaches. The Plan Sponsor Network Database (PSN), developed and marketed by INFORMA, 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 by other studies that commonly use mutual fund data.6 Our empirical results indicate that when examining the marginal performance attributable to the use of a specific Primary Investment Process for the period January 2002 to December 2006 (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. However, when we examine the performance of investment processes at the margin and after controlling for the primary processes, none of the secondary processes add value. This paper is organized in five sections. In the first section, we identify the quantitative and fundamental approaches and provide a brief discussion of the related literature. In section II we describe commonly used industry terminology in terms of the investment variables that are5

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 A significant advantage of using this data is to examine a group of managers who are most likely to benefit from quantitative analysis. They can predict the inflows and outflows better than mutual funds since they often personally know their clients, who are typically advised by quantitative consultants.

5

used in our estimation models. In section III we describe some of the salient features of the PSN database. In section IV we present the empirical models used to assess portfolio performance and discusses our findings with respect to the differences between quantitative and fundamental approaches. A final section provides a summary and some concluding remarks.

II. Industry Terminology and StyleInvestment managers can be identified as Growth or Value, Quantitative or Fundamental, and Passive or Active. Previous literature has variously used the term style for all three of these distinctions, sometimes even within the same paper. 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 the main focus of this paper. Passive/Active describes how closely the manager adheres to a benchmark, usually measured by estimating tracking error.7 To distinguish between these attributes, we provide a detailed characterization of these investment terms that are used not only in industry but are extensively used and referred to in our models and empirical estimates 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.

7

A 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.

6

Investment Style Investment Style refers to a managers choice of a specific universe from which to select stocks. 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 often 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. Investment Process The approach a manager uses when selecting stocks is called the Investment Process. This is the primary focus of our research. Two common investment processes are fundamental and quantitative, and this study 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 the characteristics of 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.

III. Description of DataThe PSN Database contains quantitative and qualitative information on over 11,000 independent equity and fixed income portfolios privately 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. Managers 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 funds for their clients. 7

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 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. These data cover exclusively U.S. funds, and we further restrict our sample to equity funds, since the investment process of fixed income managers can be quite different. One of the queries in the PSN questionnaire asks about the Investment Processes that managers 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 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. 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 as well. Finally, Technical Analysis is often thought of as being at the other end of the spectrum from Fundamental Analysis; though it scores very low in use, it is worthy of inclusion for comparative purposes.8 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.9 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 Primary8

The remaining techniques are less representative of an identifiable investment approach than an investment strategy and we examine their impact using control variables. 9 For each of the Ranked Investment Processes, the managers rank its importance on a scale of 1 denoting "Very Important" to 5 indicating "Not Utilized".

8

method, they must choose exactly one method as most important. This double selection gives us the flexibility not just to 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 secondary 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 BottomUp techniques. Similarly, 31% of Fundamental funds use Quantitative, and 40% of Quantitative funds use Fundamental techniques.10 Table 3 shows the correlations between the Primary and the Ranked Process variables. This demonstrates that grouping 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 from which managers can select. 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

10

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.

9

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 it is 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 in assets under management. 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.11 Our examination period is January 2002 to December 2006.12 After implementing all screens, our sample includes 947 managed funds.

IV. Methodology and ResultsIn order to examine the marginal performance of individual fund managers 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) Ri = i + 1 RSP 500 + 2 SMB + 3 HML + 4 MOM + i

where: i is the Carhart alpha, Ri is the return for fund i, in excess of fees and the 90 day US Tbill rate RSP 500 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

11 12

Monthly returns exceeding 5 standard deviations (of entire sample). 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.

10

In a second stage, we use these alphas in a cross-sectional dummy variable regression to estimate the marginal contribution of the various factors on fund performance.13 We define dummy variables for "Bottom Up", "Top Down", "Quantitative", "Fundamental Analysis", "Technical 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. Also controlling for Investment Style, Size, and the Primary Processes, the contribution of each of the Ranked Processes is estimated using the following model:

i = 0 + 1 LogSizei + 2Groi + 3Vali + 4Cori +(2)

5 PQi + 6 PFi + 7 PBi + 8 PTi + 1 BotUpi + 2TopDwni + 3 Fndmntli + 4Qnti + 5Cmpi + 6Techi + i

where for fund i:

i is the Carhart alpha, LogSize is the natural log of Fund is average size over the examination period, Dummies for the most common Investment Style variables, Growth, Value and Core ("other" omitted),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, secondary 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.

13

We also examined alphas from the Fama French 3-factor model (1993), with qualitatively the same results.

11

Marginal Contribution of Additional Investment Processes While our results suggest that only the Fundamental approach is likely to contribute positively to portfolio returns and no secondary process adds value across all Primary Processes, it is still possible that within a specific Primary Process, the use of 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 might benefit from using Computer Screening as a secondary investment process. To estimate 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) i = 0 + 1 LogSize + 2 Gro + 3Val + 4Cor + 5 PQi + 6 PFi + 7 PBi + 8 PTi + 1 PB * TopDwni + 2 PB * Fndmntli + 3 PB * Qnti + 4 PB * Cmpi + 5 PB * Techi + 6 PT * BotUpi + 11 PF * BotUpi + 12 PF * TopDwni + 16 PQ * BotUpi + 17 PQ * TopDwni + 18 PQ * Fndmntli + 7 PT * Fndmntli + 8 PT * Qnti + 9 PT * Cmpi + 10 PT * Techi + 13 PF * Qnti + 14 PF * Cmpi + 15 PF * Techi + 19 PQ * Cmpi + 20 PQ * Techi +

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 Bottom-Up 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 previous results, none of the individual techniques or combination of techniques adds value at the margin. In addition, when 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. Consistent with our findings, Fabozzi, Focardi, and Jonas (2008) present survey results on what they called a hybrid investment process. Specifically, when survey participants were asked if they thought a fundamental overlay would add value to their quantitative process, two12

thirds of the participants disagreed. They further report that most of the investment consultants and fund-rating firms surveyed shared the same view that adding a fundamental overlay to a quantitative investment process does not add value. 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.14 We postulate that using multiple investment processes may produce better performance through enhanced diversification, as measured by: Div = sum ( Bot , Top, Fnd , Qnt , Cmp, Tech ) To test whether using multiple techniques adds value to any one specific Primary Process (denoted as divPB, divPT, divPF and divPQ), we estimate the following regression, while again controlling for other variables as before:

i = 0 + 1Gro + 2Val + 3Cor + 4 Small + 5 Mid +(4)

6 PQi + 7 PFi + 8 PBi + 9 PTi + 1divPBi + 2 divPTi + 3 divPFi + 4 divPQi + i

As shown in Table 6, we find that the impact from Process Diversification is negative and statistically significant for 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 for Fundamental managers. Insert Table 6 about here

V. Summary and ConclusionsWe examine the marginal contribution of the most commonly used Investment processes (Bottom-Up, Top-Down, Fundamental, Quantitative, Computer Screening, and Technical Analysis) on portfolio performance over the period January 2002-December 2006. We find that when examining the return performance attributable to a distinct Primary Process, only the Fundamental approach significantly adds value. When we examine the performance of these14

We thank Larry Rose for suggesting this metric.

13

strategies at the margin and after controlling for the primary process, none of the secondary processes add value. However, 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 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. Zhao (2006) categorized as Fundamental, funds that identify themselves as Quantitative but use fundamental techniques15. This is done in order to capture the Fundamental influence, assuming that the final stock-picking decisions are based on fundamental analysis.16 However, we find it is precisely this sub-group that underperforms. This simplification simultaneously biased upwards the Quant group and biased downwards the Fundamental group, explaining the difference between her findings and our results. With respect to the choice of a Primary Investment Process, after controlling for market and investment style risk, only Fundamental managers add 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. While our main finding in this paper is that during the period January 2002 through December 2006, the traditional approach, 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 reasonable to suggest that in times of economic turmoil, it is possible that fundamental techniques are better able to respond to change. However, in more stable economic times, quantitative techniques may be better able to stay the course and not respond to cyclical distractions. Future research into the impact of economic cycles on the relative performance of these two investment approaches should shed some light on that continuing debate.

15 16

Identified in the prospectus Zhao (2006) p. 14.

14

Table 1. Primary and SecondaryProcessesComplete list of "Decision Styles" as recorded by PSN, for all funds in our sample.

Investment Process

PrimaryNumber of Funds listing this process as Primary

SecondaryNumber of Funds ranking this process as "Very Important" or "Important"

* Bottom Up * Top Down * Fundamental Analysis * Quantitative Research ** Computer Screening/modeling ** Technical Analysis Cash Flow Dvd Discount Dvd Growth Dvd Yield Earnings Surprise Future Earnings Growth Industry Sectory Analysis Low Price/Book Low Price/Cash Low Price/Earnings Low Price/Earnings - Expected (PE) Low Price/Sales Momentum Quality ROA / ROI Theme Identification Yield Analysis Total

684 22 106 76 20 1 0 0 0 0 2 11 3 5 0 9 0 0 1 0 0 6 1 947

899 143 830 387 461 104 561 167 185 162 316 659 379 353 397 463 403 265 191 583 558 203 108 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".

15

Table 2. Percentage of Funds from each Primary Process and their use of a Ranked Secondary Process Primary Proccess Bottom Up Top Down Fundamental Quantitative Other Total Total in sample692 23 100 68 64 947

Bot100% 52 93 51 92

Top8% 100 33 13 21

Ranked Proccesses Fnd Qnt92% 78 100 40 82 33% 57 31 100 52

Cmp42% 52 43 63 50

Tch9% 39 14 4 18

Table 3. Correlations of Primary and Ranked Secondary ProcessesPB PB PT PF PQ Bot Top Fnd Qnt Cmp Tch1.00

PT-0.25 1.00

PF-0.57 -0.06 1.00

PQ-0.47 -0.05 -0.10 1.00

Bot0.33 -0.33 -0.04 -0.28 1.00

Top-0.29 0.38 0.16 0.05 -0.20 1.00

Fnd0.15 -0.05 0.13 -0.32 0.28 -0.03 1.00

Qnt-0.24 0.06 -0.03 0.35 -0.13 0.05 -0.24 1.00

Cmp-0.16 0.01 -0.01 0.27 -0.04 0.01 -0.17 0.50 1.00

Tch-0.09 0.14 0.03 -0.02 -0.14 0.12 -0.01 0.18 0.18 1.00

16

Table 4. Marginal contribution of Secondary (Ranked) Processes across all Primary ProcessesRegressions (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.Estimation I Coef. t intercept Log(Fund Size) Growth target Value target Primary BottomUp Primary TopDown Primary Fundamental Primary Quantitative Secondary BottomUp Secondary TopDown Secondary Fundamental Secondary Quantitative Secondary Computer Screening Secondary Technical Analysis N R2 947 0.10 -0.04 0.01 -0.18 0.01 0.00 0.05 0.13 -0.02 -0.96 0.44 -7.31 0.31 -0.01 0.82 2.71 -0.52 Estimation II Coef. t 0.00 0.00 -0.18 0.01 0.00 0.05 0.13 0.00 -0.01 -0.01 -0.01 0.00 -0.04 0.02 947 0.11 0.08 0.30 -7.31 0.38 -0.07 0.73 2.70 -0.01 -0.15 -0.35 -0.35 -0.13 -2.02 0.64

Control Variables

Primary Process Dummies

Secondary Process Dummies

17

Table 5. Marginal contribution of Secondary (Ranked) Processes for each Primary ProcessRegressions (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

Control Variables

intercept Log(Fund Size) Growth target Value target Primary BottomUp Primary TopDown Primary Fundamental Primary Quantitative

-0.04 0.01 -0.18 0.00 0.02 -0.06 -0.04 0.08

-1.06 0.64 -7.41 0.20 0.32 -0.32 -0.33 0.78

349 299 692 23 100 68

Interaction Dummies

Primary BottomUp and Secondary: TopDown Fundamental Quantitative ComputerScreening Technical Analysis Primary TopDown and Secondary: BottomUp Fundamental Quantitative ComputerScreening Technical Analysis Primary Fundamental and Secondary: BottomUp TopDown Quantitative ComputerScreening Technical Analysis Primary Quantitative and Secondary: BottomUp TopDown Fundamental ComputerScreening Technical Analysis N R2

-0.03 0.02 0.00 -0.07 -0.02 0.09 -0.01 0.08 0.08 -0.06 0.06 0.08 0.05 0.11 0.15 0.06 -0.04 -0.11 -0.09 0.10 947 0.14

-0.56 0.31 0.01 -2.71 -0.62 0.71 -0.06 0.54 0.68 -0.56 0.54 0.88 0.60 1.18 0.99 1.46 -0.63 -2.45 -0.94 1.47

56 635 230 294 60 12 18 13 12 9 93 33 31 43 14 51 13 40 63 4

18

Table 6. Marginal contribution of "Process Diversification" - the Number of Ranked Processes used by each Primary ProcessRegressions (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 separatelyCoef. intercept Log(Fund Size) Growth target Value target Primary BottomUp Primary TopDown Primary Fundamental Primary Quantitative Total Secondary Processes for BottomUp Total Secondary Processes for TopDown Total Secondary Processes for Fundamental Total Secondary Processes for Quantitative N R2 -0.05 0.01 -0.18 0.00 0.09 -0.08 -0.17 0.14 -0.03 0.03 0.09 -0.04 947 0.13 t -1.15 0.72 -7.46 0.22 1.95 -0.51 -1.50 1.49 -2.61 1.09 2.45 -1.67

Control Variables

Diversification Metric

19

ReferencesAhmed, 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. 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 of CFA Institute Publications, Challenges in Quantitative Equity Management: 1-110. 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. Jensen, M. C. (1968). The performance of mutual funds in the period 1945-1964, Journal of Finance 23, 389-416.

20

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. 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.

21