A REVIEW OF STUDIES IN MUTUAL FUND
PERFORMANCE, TIMING, AND PERSISTENCE
Seth C. Anderson, Ph.D.
Professor of Finance
Department of Accounting and Finance
Coggin College of Business
The University of North Florida
4567 St. Johns Bluff Road
Jacksonville, FL 32224
Ph: (904) 620-2630
Fax: (904) 620-3861
Oliver Schnusenberg, Ph.D.*
Assistant Professor of Finance
Department of Accounting and Finance
Coggin College of Business
The University of North Florida
4567 St. Johns Bluff Road
Jacksonville, FL 32224
Ph: (904) 620-2630
Fax: (904) 620-3861
* Corresponding author
A REVIEW OF STUDIES IN MUTUAL FUND PERFORMANCE,
TIMING, AND PERSISTENCE
Concomitant with the development of the mutual fund industry over the last five decades
is the evolution of a rich academic literature that addresses a wide variety of mutual fund
issues. The purpose of this article is to review the more widely cited studies in the areas
of mutual fund performance, market timing, and persistence, and to offer some guidance
in these areas for potential future research. In the area of mutual fund performance, more
recent findings differ little from earlier results in that they too find that fund managers are
generally incapable of outperforming the market. Similarly, the papers on market timing
indicate that fund managers, by and large, are unable to time market movements. In the
area of persistence, a variety of studies have modified tests based on benchmarks,
models, time periods, and combinations of the same. No study to date has presented
convincing evidence that there is persistence in mutual fund performance. Currently,
mutual fund research appears to be evolving in the following directions: (1) the
assessment of mutual fund performance at the fund family level; (2) the investigation of
fund managers’ market timing ability using conditional models of performance, which
have not yet been fully exploited; and (3) the evaluation of performance persistence
relative to both fund flows and to fund families. Other extant factors in the mutual fund
industry, such as increased regulation, will also likely affect the direction of future
mutual fund research.
Over the past fifty years, mutual fund assets have grown from approximately $2
billion to more than $7 trillion presently; the number of funds has risen from 140 to more
than 8,300, with 4,800 of these funds now investing primarily in equities. These funds’
portfolio objectives and compositions are highly disparate, both domestically and
internationally, and range from index funds to specialty funds. As the industry has
developed, there has concomitantly evolved a rich academic literature addressing a wide
range of topics, including issues of performance, persistence, market timing, security
selection, fund flows, and expenses.
Since the earliest research of pioneering authors like Close (1952) and Jensen (1968),
one of the most frequently addressed topics in the literature is that of performance. The
investigation of performance has evolved from the examination of benchmarking and
modeling issues to analyses of other factors that may impact performance. For example,
institutional changes within the industry have spawned an investigation into performance
based on the structure of funds. In the area of portfolio management, for instance,
investment decisions in the mid-1900s were usually implemented by investment
committees. Over the years, the importance of committees waned, and today the norm
for management is the individual portfolio manager. Is there a difference in the efficacy
of management structures? Do they affect mutual fund performance and/or lead to
persistence in mutual fund performance? Additionally, a different, more recent vein of
inquiry into performance investigates whether a Bayesian model, which incorporates
prior beliefs about future returns, may be more effective than earlier models in predicting
performance. These two foci and resultant questions are among many that have fomented
numerous interesting studies that contribute to the literature on mutual funds.
Because of the vastness of the mutual fund literature it is not reasonable to attempt a
review of the entire body in the constraints of one article. Hence, in this article we
attempt to review the more widely cited works that focus primarily on the issue of fund
performance. While early articles focus on straightforward evaluation, more recent
articles also focus on tangentially pertinent issues, including persistence and market
timing. Therefore, the structure of this article is as follows: Section 2 chronologically
presents articles directly assessing performance. The next section is concerned with fund
managers’ ability to time the market. Section 4 reviews the works concerned with
performance persistence. The final section presents a brief summary of the prior sections
and a discussion of potential future research topics. Some of the research addressing
multiple topics is reviewed in more than one section.
Prior to 1965 a mutual fund’s performance was often rated by comparison to other
funds’ returns or by averaging returns over a number of periods. The shortcomings of
such methods are briefly addressed by Treynor (1965) in “How to Rate Management of
Investment Funds,” a watershed work wherein he presents a new way of viewing
performance results. Drawing from modern portfolio theory, Treynor discusses both
market influence on portfolio returns and investors’ aversion to risk. The article has three
parts: (1) a development of the characteristic line, which relates the expected return of a
fund to the return of a suitable market average; (2) a development of the portfolio-
possibility line, which relates the expected value of a portfolio containing a fund to the
owner’s risk preferences; and (3) a development of a measure for rating management
performance using the graphical technique developed in (1) above.
Shortly thereafter, Sharpe (1966), in “Mutual Fund Performance,” explains in a
modern portfolio theory context that the expected return on an efficient portfolio, E(Rp) and
its associated risk (σp) are linearly related:
p) = RF + βσp, (1)
wherein RF is the risk-free rate and β is a risk premium. The optimal portfolio with the risky
portfolio and a risk-free asset is the one with the greatest reward-to-variability ratio:
The author examines 34 open-end mutual funds (period 1954-1963) and finds
considerable variability in the Sharpe ratio, ranging from 0.78 to 0.43. He provides two
potential explanations for the results: (1) the cross-sectional variation is either random or
due to high fund expenses, or (2) the difference is due to management skills.
Two years later Jensen (1968), in “The Performance of Mutual Funds in the Period
1945-1964,” investigates the performance of mutual funds with a model that statistically
measures a fund’s performance relative to a benchmark. The estimating equation is:
where the α is termed Jensen’s alpha and the error term ujt is expected to be serially
independent. A positive α indicates superior security price forecasting. A negative α
indicates either poor security selection or the existence of high expenses.
Jensen investigates 115 mutual fund returns (1945-64) relative to the S&P 500 index,
and finds that the funds on average earned 1.1% less than expected given their level of
systematic risk. On average, mutual funds do not produce returns to offset their internal
expenses and fees.
Over the next half-decade, the papers of Carlson (1970) “Aggregate Performance of
Mutual Funds, 1948-1967,” and McDonald (1974) “Objectives and Performance of
Mutual Funds, 1960-1969,” address performance relative to fund type and fund
objectives, respectively. Carlson shows that regressions of fund returns on the S&P
index returns have a high unexplained variance which is significantly reduced when a
mutual fund index (diversified, balanced, or income) is used as the market proxy. In a
related vein, McDonald reports that more aggressive portfolios appear to outperform less
aggressive ones. As a reward-to-variability ratio, the author uses mean excess return
divided by standard deviation and finds that a majority of the estimated ratios fall below
the ratio for the market index.
A later paper, “Mutual Fund Performance Evaluation: A Comparison of Benchmarks
and Benchmark Comparisons,” by Lehmann and Modest (1987), provides empirical
evidence on whether the choice of alternative benchmarks effects the measurement of
performance. Among their findings are results showing that the Jensen measures (α) are
sensitive to the choice of APT benchmarks. However, both mean Jensen measures and
the rankings of funds are insensitive to the choice of the number of common factors (5,
10, or 15). The authors conclude that the choice of a benchmark portfolio is the first
crucial step in measuring the performance of a mutual fund.
In contrast to earlier studies which examine the actual returns realized by mutual fund
investors, Grinblatt and Titman (1989), in “Mutual Fund Performance: An Analysis of
Quarterly Portfolio Holdings,” employ both actual and gross portfolio returns in this
study. The authors report that superior performance may exist among growth funds,
aggressive growth funds, and smaller funds, but these funds have the highest expenses,
thus eliminating abnormal investor returns. In a 1993 study these authors introduce a
new measure of portfolio performance, the “Portfolio Change Measure” and conclude
with essentially the same findings.
In a comprehensive study, “Returns from Investing in Equity Mutual Funds: 1971 to
1991,” Malkiel (1995) employs all diversified equity mutual funds sold to the public for
the period 1971-1991 to investigate performance, survivorship bias, expenses, and
performance persistence. To consider performance he calculates the funds’ alpha
measure of excess performance using the CAPM model and finds the average alpha to be
indistinguishable from zero. Using the Wilshire 5,000 Index as a benchmark, he finds
negative alphas with net returns and positive alphas with gross returns, but neither alpha
to be significantly different from zero. He also finds no relationship between betas and
total returns. He concludes that his findings do not provide any reason to abandon the
efficient market hypothesis. Following Malkiel, Gruber (1996), in “Another Puzzle: The
Growth in Actively Managed Mutual Funds,” offers four reasons for mutual funds’
(1) customer service, including record-keeping, (2) low trading costs, (3) diversification
benefits, and (4) professional portfolio management. Using a sample of 270 funds (1985-
1994) Gruber finds that mutual funds underperform the market by 1.94% per year. With a
single index model the underperformance is 1.56%, and with a four-index model the
underperformance is 0.65% per year. Non-surviving funds underperform the market by
2.75% per year, and the average fund’s expense is 1.13%. Gruber also tests index funds
and finds that they have an average annualized alpha of -20.2 basis points, with average
expenses of 22 basis points.
In a study focusing on non-surviving funds Lunde, Timmermann, and Blake (1999),
in “The Hazards of Mutual Fund Underperformance: A Cox Regression Analysis,”
investigate the relationship between funds’ conditional probability of closure and their
return performance. The authors explain that the process of fund attrition rates is
important because: (1) survivorship bias is impacted by the funds’ lives and their relative
performance; (2) duration profiles of funds is important for understanding fund
managers’ incentive environments; and (3) termination processes may provide
information about investor reaction to poor performance. The paper measures the
importance of various factors influencing the process and rate by which funds are
terminated. After examining a data set of dead and surviving funds (973 and 1402,
respectively), the authors present some reasons why funds are terminated: (1) not
reaching critical mass in capitalization, (2) merging a poorly performing fund with a
similar, more successful fund, and (3) merging or closing a poorly performing fund to
improve family group performance overall. All of these are related to fund performance,
which the authors use to explain fund deaths.
In a different vein Indro, Jiang, Hu, and Lee (1999), in “Mutual fund Performance:
Does Size Matter?,” explore the question, “Does size of fund have any adverse impact
on the performance of a fund?” The authors explain that added economic value via
economies of scale can result from having the optimal amount of assets under
management. However, high growth may create some cost disadvantages such as higher
impact costs, difficulty in exploiting information asymmetries, and more administrative
complexities. Researching data for 683 funds (1993-1995), the authors show that three-
year returns increase as fund size increases and conclude that the optimal fund size for
growth, value, and blend funds is approximately $1.4 billion, $0.5 billion, and $1.9
In a widely-cited paper by Wermers and Moskowitz (2000), “Mutual Fund
Performance: An Empirical Decomposition into Stock-Picking Talent, Style,
Transactions Costs, and Expenses,” the authors decompose mutual fund returns by
attributable factors such as stock holdings, expense ratios, and transaction costs. Findings
indicate that annual trading costs were lower and expense ratios were higher in 1994 than
in 1975. Furthermore, mutual funds on average hold stocks that outperform a market
index by roughly their combined expenses and transaction costs, but the funds’ net
returns is about one percent lower than the CRSP index. The difference between the
funds’ stock returns and net returns is attributable to non-stock portfolio components,
expense ratios, and transaction costs. High turnover funds seem to have managers that
pick high-return stocks.
Some studies shift to more detailed considerations of fund performance rather than
overall modeling. For example, Dickson, Shoven, and Sialm (2000), in “Tax
Externalities of Equity Mutual Funds,” investigate how the after-tax performance of a
mutual fund is affected by sales and redemptions, and by the accounting cost method
used. Results indicate that funds with positive net sales perform better than funds with net
redemptions on an after-tax basis. Funds with net sales dilute capital gains, while those
with net redemptions may result in additional liquidations by the fund. Furthermore, a
costing method that costs out the stocks with the highest price first (HIFO) appears to be
the most tax-efficient costing method. Tangentially, Jain and Wu (2000), in “Truth in
Mutual Fund Advertising: Evidence on Future Performance and Fund Flows,”
investigate actions by fund managers and how they relate to performance by considering
funds that advertised in either Barron’s or Money Magazine between 1994 and 1996.
Carhart’s (1997) four-factor alpha for those funds is -3.45% in the year following the
month in which the ad appeared. Moreover, relative to a control group, these funds attract
more assets subsequent to the advertisement, indicating that funds advertise for the
purpose of attracting more money rather than to signal superior skill.
A popular question in recent years is whether socially responsible investing results in
superior returns. Statman (2000), in “Socially Responsible Mutual Funds,” finds that the
Domini Social Index (DSI) beat the S&P 500 index by a small margin between 1990 and
1998. However, the DSI index exhibits more risk over the sample period and has a
negative Jensen’s alpha relative to the S&P 500. Nonetheless, the mean performance of
socially responsible funds is better than that of a size-matched sample of conventional
funds. Statman concludes that socially responsible investing is not necessarily inferior to
conventional mutual fund investing.
Simultaneously, newer models to assess mutual fund performance have emerged. For
example, Baks, Metrick, and Wachter (2001) take a novel approach to mutual fund
performance in “Should Investors Avoid all Actively Managed Mutual Funds? A Study
in Bayesian Performance Evaluation.” They focus on an investor’s perspective using
Bayesian performance evaluation wherein an investor chooses to invest in an active fund
when the prior point estimate of alpha is positive. For 1,437 domestic equity funds in
1996, the authors calculate the posterior expectation of alpha over a range of prior beliefs.
They conclude that a mean-variance investor would require extremely skeptical beliefs
about the possibility of managerial skill to be induced not to invest in an actively
Kothari and Warner (2001), in “Evaluating Mutual Fund Performance,” also
expand existing methodologies in evaluating fund performance. Using simulation, they
find that the power of tests based on Jensen’s alpha, the Fama-French three-factor model,
and the Carhart four-factor model, is less than optimal and often results in incorrect
conclusions. The authors confirm that an event study methodology based on mutual fund
holdings may be a superior way to analyze mutual fund performance.
Research using existing performance measures to investigate different factors of fund
performance continues in Bliss and Potter (2002) with “Mutual Fund Managers: Does
Gender Matter?” The authors expect female fund managers to be more risk-averse and
less overconfident than men, but they find that female managers take more risk and
outperform men (based on Sharpe ratios and alphas) over the 1991 to 2000 period.
However, after controlling for fund size, P/E ratios, market capitalization, manager
tenure, the turnover ratio, and beta, there is no significant difference between male and
female fund manager performance in a cross-sectional regression. Another paper
investigating fund performance differences by characteristics is Chan, Chen, and
Lakonishok’s (2002) “On Mutual Fund Investment Styles.” The authors examine
whether mutual fund performance differs by the style of the fund. Using Carhart’s four-
factor model, they find that the alpha for growth managers is 1.2% larger than that of
value managers over the period from 1976 to 1997.
The issue of survivorship and performance is addressed in earlier works and again in
“Mutual Fund Survivorship” by Carhart, Carpenter, Lynch, and Musto (2002).
Although survivorship-free databases such as CRSP are available, many mutual fund
databases are still subject to survivorship bias. The authors investigate the issue of
survivorship bias by examining the effect of survivor conditioning on estimates of
average performance using data from 1962 to 1995 for 1,346 surviving funds and 725
nonsurviving funds. For a one-year sample period, the average survivor bias is only about
0.07% annually. However, for sample periods of greater than 15 years, the average bias is
about 1% annually. Survivorship bias increases with sample length, but at a decreasing
rate. Also, in a pooled time-series, cross-sectional regression, a negative survivorship
bias exists in the slope coefficient for fund size, and a positive survivorship bias exists in
the slope coefficient of the net expense ratio.
In a set of two papers, Pastor and Stambaugh (2002), in “Investing in Equity
Mutual Funds” and in “Mutual Fund Performance and Seemingly Unrelated Assets”
modify existing performance methodology by incorporating non-benchmark (or
“seemingly unrelated”) assets. In the first paper, the authors develop a framework in
which prior views about pricing models and managerial skill are incorporated into the
investment decision through the use of benchmarks prescribed by the Jensen, Fama-
French, and Carhart models and by several non-benchmark assets. Findings indicate that
actively managed funds can be better substitutes for the benchmarks than existing passive
funds, indicating that active funds can be selected even by investors who admit no
possibility of selection skill. In their second paper, the authors attempt to show that an
estimate of either alpha or the Sharpe ratio can be improved with the use of non-
benchmark assets, including a book-to-market factor and Carhart’s momentum factor.
Results indicate that, when including the nonbenchmark assets, new Sharpe ratio
estimates are typically four to five times more precise than usual estimates. Furthermore,
30% of funds that rank in the top Sharpe ratio deciles based on the benchmark estimates
fall into the bottom two-thirds of the rankings based on the new estimates. Also, the
difference between Fama-French and CAPM alphas is substantially reduced when the
non-benchmark assets are included. Empirically, however, estimated alphas for the
majority of equity funds are negative when the non-benchmark assets are either included
or excluded, confirming previous findings in the literature.
Do performance outcomes of a team of mutual fund managers differ from that of an
individual mutual fund manager? Prather and Middleton (2002) ask this question in
“Are N + 1 Heads Better than One? The Case of Mutual Fund Managers.” Using Jensen’s
alpha and four alternative benchmarks, the results indicate little evidence that team-
managed funds outperform individually-managed funds over the sample period from
1981 to 1994 for 147 individually-managed funds and 15 team-managed funds. In a
related vein, the use of incentive fees for managers is the primary topic in Elton, Gruber,
and Blake’s (2003) “Incentive Fees and Mutual Funds.” The authors investigate whether
incentive fees affect mutual fund performance over the 1990 to 1999 period. To
investigate whether funds with incentive fees have superior security selection ability, the
excess risk-adjusted return is obtained from a multi-index model using a variety of
benchmarks. Results indicate that the multi-index alpha across all incentive-fee funds is
0.048% per month, higher than the alpha for non-incentive fee funds by 0.084% per
month. Also, internal managers have significantly higher alphas than external managers,
and incentive-fee funds with a beta less than one give back a substantial portion of the
average alpha. Incentive-fee funds take on more risk than non-incentive-fee funds, and
they tend to increase risk after a period of poor performance. Lastly, incentive-fee funds
attract more new cash flows, consistent with the argument that investors are aware of the
better performance of incentive-fee funds relative to their conventional counterparts. The
issue of fees and fund performance is also addressed by Massa (2003) in “How Do
Family Strategies Affect Fund Performance? When Performance-Maximization Is Not
the Only Game in Town.” The author investigates whether mutual fund companies
attempt to attract investors through both performance and diversification of the mutual
fund family. Between 1962 and 2000, the author finds that there is a statistically strong
and negative relationship between performance and the degree of product differentiation,
which exists at both the family and the category level. Furthermore, those fund families
with high product differentiation are also more likely to set up new funds or to enter into
a new category.
Reminiscent of Indro, et al. (1999), the issue of economies of scale is addressed by
Chen, Hong, Huang, and Kubik (2004) in “Does Fund Size Erode Mutual Fund
Performance? The Role of Liquidity and Organization.” The authors use four benchmarks
to adjust their fund returns over the 1963 to 1999 sample period, including the Fama-
French and Carhart models. Results indicate a negative relationship between lagged fund
size and mutual fund performance. The results are explained using the “liquidity
hypothesis,” wherein the negative relationship is more pronounced in smaller funds. The
authors also investigate whether “hierarchy costs,” which may prevent bottom managers
from uncovering investment ideas, contribute to the negative relationship. They report
that funds belonging to large families, where decisions are decentralized, do better than
other funds, and that among small cap funds, small funds do better in local stock
investments than do large funds.
Cohen, Coval, and Pastor (2004), in “Judging Fund Managers by the Company
They Keep,” develop two new performance evaluation measures based on the idea that a
manager’s skill depends on how closely his holdings resemble those of other successful
managers. The first measure is essentially the covariance between the weights of a given
mutual fund manager and those of a skilled mutual fund manager. The second measure
entails managers making similar decisions if their trades are similar, and incorporates
changes in holdings.
The empirical results indicate that the new measure based on changes in holdings is
more powerful than either the CAPM, Fama-French, or Carhart alphas, over the 1980 to
2002 period. Furthermore, both new measures contain information not contained in
traditional alphas. Investors buying funds in the top quintiles according to both alpha and
the holdings-based measure, while concurrently selling funds in the bottom quintiles,
earn the highest risk-adjusted returns. Although the new measures appear to contain
information, investors do not seem to be aware of the information since fund flows do not
appear to vary across quintiles.
Nanda, Wang, and Zheng (2004), in “Family Values and the Star Phenomenon:
Strategies of Mutual Fund Families,” investigate mutual fund performance in the context
of mutual fund families. Using over 16,000 equity funds from CRSP (1992 to 1998), the
authors find that “star” funds (those that rank among the top 5% of performance in the
previous twelve months) result in a strong positive spillover effect to other funds in the
family compared to a stand-alone star. For example, a star performer in a family with
seven member funds delivers an aggregate cash flow increase that is more than three
times larger. Also, fund families with more risk underperform their low risk counterparts,
suggesting that increasing return variability to increase the probability of generating a star
results in poorer performance.
Closely related to the study above is the one by Gaspar, Massa, and Matos (2005),
“Favoritism in Mutual Fund Families? Evidence on Strategic Cross-Fund Subsidization,”
wherein they investigate whether mutual fund families strategically transfer performance
across member funds to favor those more likely to increase overall family profits.
Examining the top 50 mutual fund families in the U.S. from 1991 to 2001, the authors
investigate whether the families transfer performance from low value (defined as low fee,
low-performing funds) to high-value (defined as high fee, high past performers) funds.
The results indicate that the performance gap between low and high value funds
inside fund families is greater than for collections of equivalent funds by 6 to 28 basis
points of extra net-of-style performance per month (0.7% to 3% per year). This finding is
particularly pronounced when the styles of low value funds are doing relatively well, and
cross-subsidization appears to be common in families that are large, that manage many
funds, and that are heterogeneous in terms of size of the funds they offer. The results also
indicate that fund families allocate relatively more underpriced IPOs to high fee and high
past performance funds.
Reminiscent of Statman (2000), Bauer, Koedijk, and Otten (2005), in “International
Evidence on Ethical Mutual Fund Performance and Investment Style,” investigate the
performance of 103 ethical mutual funds from Germany, the U.K., and the U.S. over the
period 1990 to 2001. Using both the CAPM and the Carhart four-factor model, their
results indicate that these funds do not outperform a matched sample of conventional
funds over the full sample period. However, when dividing the entire period into
subperiods, several ethical funds that perform worse in the first period from 1990 to 1993
perform better in the 1998 to 2001 period, indicating the ethical funds “caught up” to
conventional funds in terms of performance. The results also indicate that ethical funds
tend to be less exposed to market return variability and that U.K. and German funds are
heavily exposed to small caps, whereas U.S. ethical funds invest more in large caps.
3. MARKET TIMING
In the earliest paper to directly address market timing Treynor and Mazuy (1966), in
“Can Mutual Funds Outguess the Market?”, discuss how investors frequently expect
managers to be able to anticipate market moves, and the dilemma of whether or not
managers should try to time the market. In addressing the issue, they explain that the
only way a fund can translate ability to outguess the market into higher shareholders’
returns is to vary the fund’s systematic volatility in a manner that results in an upwardly
concave characteristic line. Returns for 57 funds (1953-1962) are employed to determine
if the volatility of a fund is higher in up-years than in years when the market does poorly.
They compute a characteristic line wherein a managed fund’s return is plotted against the
rate of return for a suitable market index. They find no evidence of curvature in any
fund’s characteristic lines and conclude that none of the managers outguesses the market.
The specific model used by Treynor and Mazuy to test mutual fund managers’ market
timing ability is stated below:
tptmptmpptp rrr ,
+++= , (4)
where is the excess return on a portfolio at time t, is the excess return on the
measures timing ability (if a mutual fund manager increases the portfolio’s
market exposure prior to a market increase, then the portfolio will be a convex function
of the market’s returns, and gamma will be positive).
A model similar to the one used by Treynor and Mazuy was developed by
Henriksson and Merton (1981). Their model is stated below and is also frequently used
in tests of market timing ability, either as an alternative to the Treynor and Mazuy model
or in addition to it.
tptmptmpptp rrr ,
+++= , (5)
where is an indicator function that equals one if is positive and
zero otherwise, and
tmtmtm rrIr ,,
measures the difference between the target betas and is positive
for a manager that successfully times the market.
A decade after Mazuy and Treynor, Grant (1977), in “Portfolio Performance and the
‘Cost’ of Timing Decisions,” and Miller and Gressis (1980), in “Nonstationarity and
Evaluation of Mutual Fund Performance,” address the issue of market timing. Grant
provides a context for investigating the implications of treating the systematic relative
risk of an investment portfolio as a random variable. The author compares the
performance of a managed portfolio and that of the relative benchmark under the
assumption that beta and market return are not independent variables. He concludes by
noting that the relationships investigated are significant both in theory and in application.
Miller and Gressis explain that estimates of fund alpha and beta may provide misleading
information if nonstationarity is present in the risk-return relationship and is ignored.
They examine 28 no-load funds and find that only one fund has stationary betas, and that
the number of betas for any given fund varies considerably over periods. Their findings
indicate both weak, positive relationships and weak, negative relationships between betas
and the market return.
During the following years there appear several widely cited articles addressing the
issues of timing and securities selection. Kon and Jen (1979), in “The Investment
Performance of Mutual Funds: An Empirical Investigation of Timing, Selectivity and
Market Efficiency,” employ several models of market equilibrium to evaluate stock
selectivity performance when managers are also engaged in market timing. Using a
sample of 49 mutual funds with different investment objectives, the null hypotheses of
risk-level stationarity and of constant selectivity performance are rejected for many
individual funds. They report that some funds generate superior selectivity performance
but that fund managers are unable to select securities well enough to recoup research
expenses, management fees, and commissions. These finding are supported by Kon
(1983) in “The Market-Timing Performance of Mutual Fund Managers,” who reports that
a sample of funds produces better selectivity than timing performance. Similar results are
reported by Chang and Lewellen (1984) in “Market Timing and Mutual Fund Investment
Performance.” These authors jointly test for either superior market-timing or security-
selection skills for a sample of 67 mutual funds during the 1970s. Using both quarterly and
monthly returns series, they find that managers’ security selection abilities are significant in
magnitude in only five instances out of 67, and three of these five have negative values.
Like findings are reported for managers’ market-timing abilities.
In “Assessing the Market Timing Performance of Managed Portfolios,” Jagannathan
and Korajczyk (1986) discuss earlier reported puzzling evidence that funds exhibiting
significant timing characteristics show negative performance more frequently than
positive performance. The authors demonstrate both theoretically and empirically that
portfolios can be constructed to show artificial timing ability when no real ability exists.
They offer that certain parametric techniques for determining timing and selectivity
performance can yield spurious performance (of the opposite sign) when applied to
option-like securities. They propose two methods of testing market-timing models: (1)
testing linearity by examining the difference between OLS and WLS parameter estimates,
and (2) testing restrictions on the coefficients of additional regression independent
variables. The tests generally reject linearity when spurious timing is statistically
significant. They call for an extension of this analysis involving different mutual fund
Ten years later Ferson and Schadt (1996), in “Measuring Fund Strategy and
Performance in Changing Economic Conditions,” address the effects of incorporating
informational variables in an attempt to better capture the performance of managed
portfolios such as mutual funds. They modify the traditional Jensen (1968) model as well
as the market timing models of Treynor and Mazuy (1966) and Henriksson and Merton
(1981) to incorporate conditioning information. The conditional models allow estimation
of time-varying conditional betas, as managers are likely to shift their bets on the market
to incorporate information about market conditions. Using 67 mutual funds over the
period 1968-1990, they find that the use of conditioning information is significant.
Traditional measures of performance show that more funds have negative Jensen’s alpha
than positive. In contrast, conditional models produce alphas that have a mean value of
zero thus there is no evidence of perverse market timing. In a related vein Ferson and
Warther (1996), in “Evaluating Fund Performance in a Dynamic Market” present a
conditional or dynamic model which utilizes three factors: the S&P 500 Index, the lagged
value of the market dividend yield, and the lagged value of the short-term Treasury yield
in order to account for the dynamic strategies followed by many fund managers. Using
data for 63 funds, the authors show that, unlike the unconditional models, funds do not
usually underperform the S&P 500 Index on a risk-adjusted basis. Soon afterward
Becker, Ferson, Myers, and Schill (1999), in “Conditional Market Timing with
Benchmark Investors,” investigate the market-timing ability of mutual funds by employing
models that: (1) allow the manager’s payoff function to depend on excess returns over a
benchmark, and (2) distinguish timing based on public information from timing based on
superior information. Employing two fund samples, they initially report “negative” market
timing, which makes no economic sense. However, their conditional market-timing model
yields no evidence of timing ability, which is more reasonable than that reported in the prior
literature on market timing.
In another study Volkman (1999), in “Market Volatility and Perverse Timing
Performance of Mutual Fund Managers,” investigates fund managers’ security-selection
and market-timing abilities over the 1980s, as well as performance persistence prior to
and after the 1987 crash. Three measures of abnormal fund performance are utilized:
Jensen’s alpha, Bhattacharya and Pfleiderer’s selectivity measure, and an adjusted timing
model. Using data for 332 funds (1980-1990), he finds negative correlation between a
fund’s timing and selectivity performance, which suggests that managers focus on one
source of performance to the detriment of the other source. He concludes that during
periods of high volatility, few funds correctly anticipate market movements, although
many funds outperform the market via security selection.
Bollen, and Busse (2001), in “On the Timing Ability of Mutual Fund Managers,”
extend the previous literature on market timing by examining daily (instead of monthly)
data and by comparing the timing ability of mutual fund managers to a synthetic sample
that exhibits no timing ability by construction. The authors focus on 230 domestic equity
growth funds from 1985 to 1999 and use the Treynor and Mazuy (1966) and Henriksson
and Merton (1981) models to test market timing ability.
Findings indicate that regression-based tests are more powerful for daily than for
monthly data. Furthermore, daily returns increase the number of significant estimates of
timing ability; 34.2% of the funds exhibit significantly more timing ability than the
synthetic funds using daily data, whereas only 11.9% exhibit timing ability using monthly
Jiang (2003) develops a new, nonparametric measure of timing ability in “A
Nonparametric Test of Market Timing.” The new measure, theta: (1) requires only the ex
post returns of funds and their benchmarks; (2) is not affected by the manager’s degree of
risk aversion; and (3) is more robust to different information and incentive structures and
is independent of the underlying return distribution. Statistically, the measure results in
the following properties: (1) the risk level taken on by the manager is a non-decreasing
function of the expected market return; (2) it measures how often a manager correctly
reacts to a market movement; (3) it can be extended to a conditional market timing
context based on the manager having superior market timing information; and (4) it
offers a timing measure that has little correlation with the estimation error in the standard
An empirical test is conducted from 1980 to 1999 using data from both Morningstar
and CRSP and 1,827 surviving funds and 110 dead funds. Results indicate: (1) the
value is -1.33%, indicating that the probability that a manager of an average
live fund moves the fund’s market exposure correctly direction is 1.33 percentage points
lower than the probability of a incorrect move; (2) except for the technology sector, theta
values are negative across different subcategories of funds; (3) the top 5% of market
timers have theta values above 8.47%; (4) the average timing performance bears a
positive relation with fund age or management tenure and fund load; and (5) theta is
negatively related to fund size. Overall, the relationship between market timing ability
and fund characteristics in the study is weak.
Sapp and Tiwari (2004), in “Does Stock Return Momentum Explain the ‘Smart
Money’ Effect?,” investigate whether the “smart money” effect can be explained by
momentum, since neither Gruber (1996) nor Zheng (1999) control for Carhart’s (1997)
momentum factor when they document the smart money effect. Essentially, they suggest
that investors have selection ability by showing that risk-adjusted returns on new cash
flows to funds are higher than average returns for all fund investors.
The authors include all U.S. equity funds (1970 to 2000) in the CRSP Survivor-Bias
Free Database and form two new-money portfolios at the beginning of each quarter. The
first portfolio includes all funds realizing positive net cash flows during the prior quarter,
and the second portfolio includes all funds realizing negative net cash flow. Subsequent
performance is examined using alternatively the Fama-French three-factor model and the
Carhart four-factor model. Results indicate that a strategy mimicking investor fund flows
by going long in the positive cash-flow portfolio and short in the negative cash-flow
portfolio earns a significant Fama-French alpha of 2.09% annually. However, the Carhart
alpha on the flow of money is essentially zero. It does not appear that investors
deliberately invest in funds following momentum strategies, as cash flows to funds are
not strongly correlated with momentum factor loadings.
Although earlier works such as Sharpe (1966) and Grinblatt and Titman (1993) report
evidence of performance persistence, Hendricks, Patel, and Zeckhauser (1993) focus
primarily on the persistence issue in “Hot Hands in Mutual Funds: Short-run Persistence
of Relative Performance, 1974-1988.” The authors employ an initial sample of 165 funds
to test for short-run persistence. Returns are computed using three benchmarks: (1)
single portfolio benchmarks, (2) an eight-portfolio benchmark, and (3) an equally-
weighted index of sample mutual funds. They find positive performance persistence for
four quarters and a reversal thereafter and attribute this pattern of returns to an incorrect
model specification or to several other reasons. The authors rank portfolios into octiles
on the basis of the most recent four quarters’ returns and find that mean excess returns,
Sharpe’s, and Jensen’s alpha, rise monotonically with octile rank. The authors confirm
their findings of short-term persistence via additional simulations and tests on another
sample of funds for 1989-1990. They report that poor performance persists over time and
that this performance is more inferior than “hot hands” performance is superior. In
another study, “Do Winners Repeat? Patterns in Mutual Fund Return Behavior,”
Goetzmann and Ibbotson (1994) employ data for 728 mutual funds (1976-1988) and
consider two-year, one-year, and monthly gross and Jensen risk-adjusted returns. They find
support for the winner-repeat question with both type returns for funds overall, and with
growth funds separately. Top-quartile and lower-quartile funds experience the greatest
In the following year Malkiel (1995), in “Returns from Investing in Equity Mutual
Funds: 1971 to 1991,” finds that there is some fund return persistence during the 1970s,
but that this persistence does not hold during the 1980s. From this he suggests that
persistence may have existed earlier, but has since disappeared. However, even when
persistence existed during the 1970s, many investors would not have benefited from
buying funds with a “hot hand” because of the load charges (up to 8% of asset value)
entailed with their purchase. Similar findings are reported by Brown and Goetzmann
(1995) in “Performance Persistence..” The authors’ analysis of fund data (1976-1988)
shows that 1,304 past winners are repeat winners; 1,237 past losers are repeat losers; and
1,936 funds reverse roles. However, persistence is not found to be a result of a winning
management style each year. It is seen that performance persistence is more likely due to
repeat-losers than to repeat-winners. They conclude that past patterns yield clues about
which funds to avoid but do not provide strong indications about which funds will
outperform their benchmarks in the future.
Using a different tact, Kahn and Rudd (1995), in “Does Historical Performance
Predict Future Performance?,” analyze funds’ performance relative to a set of style
indices. This contrasts with a single index model used in many earlier works. The
authors employ 300 equity funds and a large sample of taxable bond funds (1983-1993)
for analysis. Thirty-six month in-sample data are used to classify the funds’ style, and
performance is calculated with out-of-sample data. Out-of-sample period performance is
regressed against the in-sample performance, and results show no evidence of persistence
among equity funds but some evidence of persistence among fixed-income funds.
In a more comprehensive work Carhart (1997), in “On Persistence in Mutual Fund
Performance,” investigates the persistence issue using a sample of 1,892 equity funds (free
of survivorship bias) from 1962-1993. He employs two models: (1) the Capital Asset
Pricing Model (CAPM), and (2) his four-factor model involving excess returns on a market
proxy and returns on factor-mimicking portfolios for size, book-to-market equity, and one-
year return momentum. Carhart’s model is reproduced below:
where is the return on a portfolio in excess of the one-month T-bill return; is
the excess return on a value-weighted aggregate market proxy; and SMB
are returns on value-weighted, zero-investment, factor mimicking portfolios for size,
book-to-market equity, and one-year momentum in stock returns.
tYRPR1, tt HML,
Portfolios of funds are formed on lagged one-year returns, and performance is estimated.
With the CAPM model, post-formation excess returns on the decile portfolios decrease
monotonically in rank, exhibiting an annualized spread of approximately 8%, versus 24% in
the ranking year. In contrast, the four-factor model explains much of the spread among
portfolios, with size and momentum factors accounting for most of the explanation. The
author repeats the analyses using two-to-five-year returns and finds that only top and bottom
decile funds maintain their rankings more than would be expected randomly. Following
Carhart, Hendricks Patel, and Zeckhauser (1997), in “The J-shape of Performance
Persistence Given Survivorship Bias,” discuss that social scientists must generally base
their inferences on observations of non-experimental information, thereby presenting a
challenge to unbiased robust inference from this data. They explain that for groups with
performances above the population mean, relative ranks will be positively correlated
across sub-periods. Thus, considering all survivorship-biased sample groups, they
contend that a spurious j-shaped relation exists between first- and second-period
performances. The authors employ a simple regression-based approach to discriminate
between a j-shaped pattern of persistence performance and a monotonic persistence in
performance. The method appears to be effective in the simulations conducted. They
conclude that mutual funds exhibit a monotonic increasing pattern effected by true
ter Horst and Verbeek (2000), in “Estimating Short-Run Persistence in Mutual Fund
Performance,” recognize that previous techniques to estimate performance persistence
may result in spurious confirmations of persistence because they regress a sample of
funds’ current returns upon a series of lagged returns. If cross-sectional variation is
controlled for by subtracting some measure of “expected returns” from the dependent
variable, then a mechanical relation between risk-adjusted returns and lagged returns
The authors investigate the nature of these spurious findings using Monte Carlo
simulation and find that, when expected returns are constructed as the mean return over
the sample period or as predicted returns from a factor model, the induced biases are
negative and decreasing in the number of time periods. The implication of this result is
that biases do not generate spurious findings of hot hands in mutual funds but that
persistence in mutual fund performance is more pronounced than previous studies show.
Additional empirical results confirm findings of persistence over the 1987 to 1994 period.
Another potential bias in estimating mutual fund persistence is investigated by ter
Horst, Nijman, and Verbeek (2001) in “Eliminating Look-Ahead Bias in Evaluating
Persistence in Mutual Fund Performance,” which focuses on look-ahead bias. This bias
occurs because persistence studies typically utilize a ranking period and an evaluation
period. Funds tend to disappear in a nonrandom way during the ranking and/or evaluation
period, which causes a look-ahead bias that may result in spurious persistence. The
authors model the survival probability as a function of 12 lagged quarterly returns, fund
age, and aggregate time effects.
Monte Carlo simulation reveals a spurious persistence pattern that exists as a result of
look-ahead bias that disappears once a corrected model is used. Furthermore, an
empirical investigation over the 1989 to 1994 period reveals no evidence of performance
persistence in a sample of U.S. growth and income funds using Carhart’s four-factor
The Carhart, Carpenter, Lynch, and Musto (2002) study in mutual fund
survivorship is discussed in the above performance section, but this study also contains
an investigation of how survivorship bias affects performance persistence. To investigate
how survivorship bias affects mutual fund performance persistence, the authors form
deciles for each year, sorted on either lagged returns or lagged four-factor alphas. After
one year, portfolios are re-formed. This process is repeated for one-year returns, five-year
returns, and three-year estimates of alpha from the four-factor model and for both end-of-
sample conditioning and look-ahead conditioning. Results reveal that the full sample
exhibits strong performance persistence in the subsequent period. However, even the
survivor-only sample shows some performance persistence in the spread between deciles
one and ten, indicating that performance persistence cannot be completely explained by
survivorship bias. This contradicts the findings by ter Horst, Nijman, and Verbeek
As stated previously, the present synthesis does not focus on the relationship between
fund flow and performance. However, in the study by Berk and Green (2004) “Mutual
Fund Flows and Performance in Rational Markets,” the authors develop a model showing
that investments with active managers do not outperform passive benchmarks because
investors competitively supply funds to managers and there are decreasing returns for
managers in deploying their superior ability. Specifically, successful managers increase
both the size of their funds by investing more in passive benchmarks and thereby their
own compensation to the point at which expected returns to investors are competitive
going forward. Overall, the model implies that investors competitively supply funds to
successful managers, which decreases the expected returns of the fund. This, in turn,
explains the absence of performance persistence across the mutual fund universe.
Bollen and Busse (2005), in “Short-Term Persistence in Mutual Fund Performance,”
use an argument similar to that of Berk and Green (2004) and ask, “What if the
documented absence of long-term persistence in mutual fund performance is due to the
fact that investors increase their capital investment to the best performing funds?”
In their investigation of 230 domestic equity growth funds, the authors utilize both the
stock selection four-factor model of Carhart (1997) and the market timing models
developed by Treynor and Mazuy (1966) and Henriksson and Merton (1981). The
authors combine stock selection and timing models by including the three additional
Carhart factors (size, book-to-market, and momentum) in the two market timing models.
Moreover, the authors investigate the stock selection model, the market timing models,
and a mixed model to identify abnormal returns.
Depending on the model used, the authors find that the top decile of funds generates a
statistically significant abnormal return in the post-ranking quarter of 25 to 39 basis
points. When modifying the tests, the short-term persistence phenomenon disappears. For
example, when the funds are placed into deciles according to raw returns, no statistically
significant difference between the top and bottom deciles exist, which the authors
attribute to the fact that their procedure identifies different funds. Although the results are
robust to a barrage of robustness tests, the authors acknowledge that the findings may not
be economically significant once transaction costs and taxes are taken into account.
The purpose of this article is to review the more widely cited works in the areas of
mutual fund performance, the market timing ability of mutual fund managers, and mutual
fund performance persistence. Although early studies assessing mutual fund performance
determine abnormal performance by using Jensen’s (1968) alpha, any study attempting to
assess mutual fund performance today would be remiss in not correcting performance for
the size and book-to-market factors identified by Fama and French (1993), and for the
momentum factor incorporated into the assessment of fund performance by Carhart
(1997). Yet, perhaps not surprisingly, more recent findings in the area of mutual fund
performance differ little from early results in that they too find that fund managers are
generally incapable of outperforming the market, even if the “market” is proxied for
using a variety of benchmarks.
It is too early to tell whether other factors, such as those based on other managers’
performance developed by Cohen, Coval, and Pastor (2004) will render more widely
accepted measures of abnormal mutual fund performance. However, it appears that there
has lately been a shift in the focus of performance from the individual fund level to the
fund family level. This shift indicates that mutual fund companies cross-subsidize funds
in their families to maximize their profit. Consequently, a focus on the aggregate
performance of fund families may be a fecund area for future investigation.
In the area of market timing, a vast majority of studies use the models of Treynor
and Mazuy (1966) and Henriksson and Merton (1981) to determine whether fund
managers have the ability to time the market successfully. One sub-area of investigation
emerging recently is that of conditional market timing. This area investigates whether
mutual fund managers use “dynamic” market timing models conditional on the
performance of the market. Overall, the papers to date on market timing indicate that
mutual fund managers, by and large, are unable to time market movements.
In the last section above we focus on mutual fund performance persistence which
addresses how those funds that perform well (poorly) previously continue to perform well
(poorly) in the subsequent period. Early papers such as Carhart (1997) indicate that there
is little performance persistence for mutual funds. While a variety of studies have
modified tests of persistence based on benchmarks, models, time periods, and
combinations of the same, no study to date has presented convincing evidence that there
is persistence in mutual fund performance.
Over the last five decades the literature on fund performance, market timing
ability, and performance persistence, has evolved around various attributes such as
models and benchmarks used and time period investigated. The basic results have not
changed; it appears that: (1) mutual funds underperform the “market;” (2) fund managers
in aggregate are incapable of timing the market; and (3) mutual fund investors are ill-
advised to invest based on prior fund performance.
In addition to the topics addressed in the extant literature, there are also certain
issues that have recently crystallized in the mutual fund industry that have not, by and
large, found their way into the academic literature. For instance, there is a rapidly
evolving body of literature focusing on exchange-traded funds, and many of the issues
addressed here may pertain to various types of mutual funds. Thus, one potential venue
for future research is how the emergence of exchange-traded funds affects the
performance, timing, and persistence of mutual funds. For instance, competition from
exchange-traded funds may impact the performance of index funds if investors in the
latter shift funds to exchange-traded funds.
Another topic that has not yet been widely addressed in the mutual fund literature
is that of increased regulation and poor publicity as a result of mutual fund scandals.
How do increased scrutiny and regulation issues affect mutual funds? Does such an
environment affect return persistence and the ability of fund managers to time the
market? Since these scandals are relatively recent, empirical analyses of these issues may
be forthcoming over the next few years.
Lastly, the evolution of financial conglomerates as a result of the Financial
Services Modernization Act of 1999 may ultimately affect fund performance, as financial
conglomerates that operate in various areas try to maximize their overall profits. A
conglomerate involved in investment banking, brokerage, and mutual funds, may transfer
funds and subsidize certain areas at the expense of others, similar to the cross-
subsidization of funds within fund families mentioned in this article. Will the different
businesses serve their stakeholders as well as they would as stand-alone entities?
How all of the above issues may affect fund performance and related aspects of
mutual funds will likely be seen in the years to come, as sufficient data to analyze these
issues becomes available. Nonetheless, research continues apace, as is evidenced by the
33 mutual fund papers presented at the FMA 2005 Annual Meeting in Chicago.
Baks, K., A. Metrick, and J. Wachter, 2001, “Should Investors Avoid all Actively
Managed Mutual Funds? A Study in Bayesian Performance Evaluation,” The Journal
of Finance, 56, 45-85.
Bauer, R., K. Koedijk, and R. Otten, 2005, “International Evidence on Ethical Mutual
Fund Performance and Investment Style,” Journal of Banking & Finance, 29, 1751-
Becker C., W. Ferson, D. Myers, and M. Schill, 1999, “Conditional Market Timing with
Benchmark Investors,” Journal of Financial Economics, 52, 119-148.
Berk, J. and R. Green, 2004, “Mutual Fund Flows and Performance in Rational Markets,”
Journal of Political Economy, 112, 1269-1295.
Bliss, R. and M. Potter, 2002, “Mutual Fund Managers: Does Gender Matter?” Journal of
Business and Economic Studies, 8, 1-17.
Bollen, N. and J. Busse, 2001, “On the Timing Ability of Mutual Fund Managers,”
Journal of Finance, 56, 1075-1094.
Bollen, N. and J. Busse, 2005, “Short-Term Persistence in Mutual Fund Performance,”
Review of Financial Studies, 18, 569-597.
Brown, S. and W. Goetzmann, 1995, “Performance Persistence,” The Journal of Finance,
Carhart, M., 1997, “On Persistence in Mutual Fund Performance,” The Journal of
Finance, 52, 57-82.
Carhart, M., J. Carpenter, A. Lynch, and D. Musto, 2002, “Mutual Fund Survivorship,”
Review of Financial Studies, 15, 1439-1463.
Carlson, R., “Aggregate Performance of Mutual Funds, 1948-1967,” Journal of Financial
and Quantitative Analysis, 1-32.
Chan, L., H. Chen, and J. Lakonishok, 2002, “On Mutual Fund Investment Styles,”
Review of Financial Studies, 15, 1407-1437.
Chang, E. and W. Lewellen, 1984, “Market Timing and Mutual Fund Investment
Performance,” The Journal of Business, 57, 57-72.
Chen, J., H. Hong, M. Huang, and J. Kubik, 2004, “Does Fund Size Erode Mutual Fund
Performance? The Role of Liquidity and Organization,” American Economic Review,
Close, J., 1952, "Investment Companies: Closed-End versus Open-End," Harvard
Business Review, 29, 79-88.
Cohen, R., J. Coval, and L. Pastor, 2004, “Judging Fund Managers by the Company They
Keep,” Journal of Finance, forthcoming.
Dickson, J., J. Shoven, and C. Sialm, 2000, “Tax Externalities of Equity Mutual Funds,”
National Tax Journal, 53, 607-628.
Elton, E., M. Gruber, and C. Blake, 2003, “Incentive Fees and Mutual Funds,” Journal of
Finance, 58, 779-804.
Fama, E. and K. French, 1993, “Common Risk Factors in the Returns on Stocks and
Bonds,” Journal of Financial Economics, 33, 3-56.
Ferson, W. and R. Schadt, 1996, “Measuring Fund Strategy and Performance in Changing
Economic Conditions,” The Journal of Finance, 51, 425-461.
Ferson, W. and V. Warther, 1996, “Evaluating Fund Performance in a Dynamic Market,”
Financial Analysts Journal, 52, 20-28.
Gaspar, J., M. Massa, and P. Matos, 2005, “Favoritism in Mutual Fund Families?
Evidence on Strategic Cross-Fund Subsidization,” Journal of Finance, forthcoming.
Goetzmann, W. and R. Ibbotson, 1994, “Do Winners Repeat? Patterns in Mutual Fund
Return Behavior,” The Journal of Portfolio Management, Winter, 9-18.
Grant, D., 1977, “Portfolio Performance and the ‘Cost’ of Timing Decisions,” The
Journal of Finance, 32, 837-846.
Grinblatt, M. and S. Titman, 1989, “Mutual Fund Performance: an Analysis of Quarterly
Portfolio Holdings,” The Journal of Business, 62, 393-416.
Grinblatt, M. and S. Titman, 1993, “Performance Measurement without Benchmarks: An
Examination of Mutual Fund Returns,” The Journal of Business, 66, 47-68.
Gruber, M., 1996, “Another Puzzle: The Growth in Actively Managed Mutual Funds,” The
Journal of Finance, 51, 783-810.
Hendricks, D., J. Patel, and R. Zeckhauser, 1993, “Hot Hands in Mutual Funds: Short-
run Persistence of Relative Performance, 1974-1988,” The Journal of Finance, 43,
Hendricks, D., J. Patel, and R. Zeckhauser, 1997, “The J-shape of Performance
Persistence Given Survivorship Bias,” Review of Economics and Statistics, 79, 161-
Henriksson, R. and R. Merton, 1981, “On Market Timing and Investment Performance,”
The Journal of Business, 54, 513-533.
Indro, D., C. Jiang, M. Hu, and W. Lee, 1999, “Mutual Fund Performance: Does Size
Matter?” Financial Analysts Journal, 55, 74-87.
Jagannathan, R. and R. Korajczyk, 1986, “Assessing the Market Timing Performance of
Managed Portfolios,” The Journal of Business, 59, 217-235.
Jain, P. and J. Wu, 2000, “Truth in Mutual Fund Advertising: Evidence on Future
Performance and Fund Flows,” Journal of Finance, 55, 937-958.
Jensen, M., 1968, “The Performance of Mutual Funds in the Period 1945-1964,” The
Journal of Finance, 23, 389-416.
Jiang, W., 2003, “A Nonparametric Test of Market Timing,” Journal of Empirical
Finance, 10, 399-425.
Kahn, R. and A. Rudd, 1995, “Does Historical Performance Predict Future
Performance?” Financial Analysts Journal, 51, 43-52.
Kon, S. and F. Jen, 1979, “The Investment Performance of Mutual Funds: An Empirical
Investigation of Timing, Selectivity and Market Efficiency,” The Journal of Business,
Kon, S., 1983, “The Market-Timing Performance of Mutual Fund Managers,” The
Journal of Business, 56, 323-347.
Kothari, S. and J. Warner, 2001, “Evaluating Mutual Fund Performance,” Journal of
Finance, 56, 1985-2010.
Lehmann, B. and D. Modest, 1987, “Mutual Fund Performance Evaluation: A
Comparison of Benchmarks and Benchmark Comparisons,” The Journal of Finance,
Lunde, A., A. Timmermann, and D. Blake, 1999, “The Hazards of Mutual Fund
Underperformance: A Cox Regression Analysis,” Journal of Empirical Finance, 6,
Malkiel, B., 1995, “Returns from Investing in Equity Mutual Funds: 1971-1991,” The
Journal of Finance, 50, 549-572.
Massa, M., 2003, “How Do Family Strategies Affect Fund Performance? When
Performance-Maximization Is Not the Only Game in Town,” Journal of Financial
Economics, 67, 249-304.
McDonald, J., 1974, “Objectives and Performance of Mutual Funds, 1960-1969,” Journal
of Financial and Quantitative Analysis, 311-333.
Miller, T. and N. Gressis, 1980, “Nonstationarity and Evaluation of Mutual Fund
Performance,” Journal of Financial and Quantitative Analysis, 15, 639-654.
Nanda, V., Z. Wang, and L. Zheng, 2004, “Family Values and the Star Phenomenon:
Strategies of Mutual Fund Families,” Review of Financial Studies, 17, 667-698.
Pastor, L. and R. Stambaugh, 2002, “Investing in Equity Mutual Funds,” Journal of
Financial Economics, 63, 351-380.
Pastor, L. and R. Stambaugh, 2002, “Mutual Fund Performance and Seemingly Unrelated
Assets,” Journal of Financial Economics, 63, 315-349.
Prather, L. and K. Middleton, 2002, “Are N + 1 Heads Better than One? The Case of
Mutual Fund Managers,” Journal of Economic Behavior & Organization, 47, 103-
Sapp, T. and A. Tiwari, 2004, “Does Stock Return Momentum Explain the ‘Smart
Money’ Effect?” Journal of Finance, 59, 2605-2622.
Sharpe, W., 1966, “Mutual Fund Performance,” The Journal of Business, 39, 119-138.
Statman, M., 2000, “Socially Responsible Mutual Funds,” Financial Analysts Journal, 3,
ter Horst, J. and M. Verbeek, 2000, “Estimating Short-Run Persistence in Mutual Fund
Performance, Review of Economics and Statistics, 82, 646-655.
ter Horst, J., T. Nijman, and M. Verbeek, 2001, “Eliminating Look-Ahead Bias in
Evaluating Persistence in Mutual Fund Performance,” Journal of Empirical Finance,
Treynor, J., 1965, “How to Rate Management of Investment Funds,” Harvard Business
Review, 43, 63-75.
Treynor, J. and K. Mazuy, 1966, “Can Mutual Funds Outguess the Market?” Harvard
Business Review, July, 131-136.
Volkman, D., 1999, “Market Volatility and Perverse Timing Performance of Mutual
Fund Managers,” The Journal of Financial Research, 22, 449-470.
Wermers, R. and T. Moskowitz, 2000, “Mutual Fund Performance: An Empirical
Decomposition into Stock-Picking Talent, Style, Transactions Costs, and Expenses,”
The Journal of Finance, 55, 1655-1703.