Abnormal performance of US global equity funds

Abnormal performance of US global equity funds

Bachelor Thesis

Abstract

This research empirically tests US global equity funds on abnormal performance between 2009 and 2014. This study examines if US global equity funds underperform relatively to the market. The dataset contains 123 funds and the sample is tested on a significant negative alpha in three asset-pricing models (Capital Asset Pricing Model, Fama-French Three-Factor Model and Carhart’s Four-Factor Model). The empirical results show that there is sufficient evidence to reject the hypothesis that alpha is equal to zero, based on the significance level of 1%. This indicates that US global equity funds underperform the market. Further, the Carhart Four-Factor model shows the highest explanatory power with a R-Squared of 0.81.

Student: Maarten Alblas Student number: 10152032 Thesis supervisor: T. Homar

Specialization: Finance & Organization Research Field: Asset Pricing

Statement of originality

This document is written by Student Maarten Alblas who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

1.Introduction

In the book “A random walk down Wall-Street” written by Burton Gordon Malkiel (1973), the situation is sketched where a blindfolded monkey throwing darts at the financial pages of a newspaper selects a portfolio just as well as a mutual fund manager. In line with this analogy Jensen (1969), Grinblatt and Titman (1989), and Malkiel (1995) stressed that equity funds do not outperform the market. Additionally when the returns net of management fees are computed and the dataset is corrected by survivorship bias there is general consensus that equity funds underperform the market.

Despite that there is significant evidence in past academic research that active management of investments do not benefit the investor, the mutual funds industry is still a huge market. According to the Investment Company Fact Book (2014) the US mutual funds industry holds 15 trillion in assets. When the major force in long-term investments is still active managed mutual funds, the question can be asked: Is there still significant evidence

that mutual funds underperform the market when management fees are taken in consideration?

The remainder of this study is used to examine the research question. In the next section the literature review is established in which the relevant concepts are discussed. This is followed by the third section, in which the research design and methods are explained, followed by an overview of the results in the fourth section. These results are discussed in the next section after which a conclusion is made

2. Literature

The theory underlying the research question is examined in this section. First, the efficient market hypothesis is discussed. Second, the three asset pricing models that are used in the analysis are introduced. And finally, previous empirical research on the performance of mutual funds is discussed.

2.1. Efficient Market Hypothesis

The Efficient Market Hypothesis (EMH) is the central theory in finance and asset pricing. Fama (1965) defined an “efficient” market as a market where security prices reflect all available information. Because security prices reflect all available information, it’s not possible for active traders to outperform the market. The EMH can be separated in three forms: the weak, the semi-strong and the strong form. These forms are separated in the interpretation of “all available information”. The weak EMH assumes that all historical data is reflected in security prices (Fama, 1970). The semi-strong form assumes that historical data and fundamental analysis are reflected in security prices. And the strong form assumes that all relevant information, including insider information, is reflected in prices.

Although it’s central role in financial theory, the EMH has received a lot of critique on the underlying assumptions. For example, Grossman and Stiglitz (1976) disagree with the fact that information relevant to security prices is free. They say that searching for information about securities requires resources. The EMH is also criticized by the assumptions that investors are fully rational and always try to maximize their wealth. Advocates of behavioral finance claim that this is not a realistic assumption. They argue that markets are inefficient, because of behavioral biases of investors. Multiple studies, like De Bondt and Thaler (1990),

investigated these biases and found that irrational investor behavior creates inefficient markets.

Fama (1991) recognized that financial markets are not fully efficient, but approximate it. In his new EMH he made some improvements to the model. It is still a lively debate in academic research if markets are efficient. The claim that markets are perfectly efficient is not realistic, neither is the claim that markets are completely inefficient. Remaining undecided if the EMH is true or not, it’s a good starting point in the development of asset pricing models.

2.2 Asset Pricing Models

In the following part three broadly used asset-pricing models are discussed, which are useful for the analysis of mutual fund performance. First, the Capital Asset Pricing Model is described. Second, the Fama-French Three-Factor model is introduced, and third the Carhart Four-Factor model is discussed.

2.2.1 Capital Asset Pricing Model

The Capital Asset Pricing Model (CAPM) is the asset-pricing model that is based on the EMH. The model was introduced in the research of Markowitz (1952), and refined in the works of Sharpe (1964), Lintner (1965) and Mossin (1966). The idea behind the CAPM is to divide risk into systematic and idiosyncratic risk. Idiosyncratic risk can be diminished by diversification, but systematic risk is market risk that cannot be diversified away. The CAPM is expressed in the following formula:

E(Ri – Rf) is the expected excess return of the asset, the systematic risk of the asset is βi, and E(Rm – Rf) the expected market premium. When the excess return and the market premium are known, it is possible to include the abnormal return (αi), as the excess return that is not explained by systematic risk (βi). By rewriting the CAPM formula in 1967 Jensen developed a new performance measure for securities. Because it is based on the CAPM the same assumptions are applied for Jensen’s Alpha. The performance measure can be expressed in the following equation:

α

= (

–

) – β

(

–

) + ε

2.2.2 Fama-French Three-Factor Model

Basu proved in 1977 that stocks with low price/earning ratios outperformed stocks with high price/earnings ratios. And in 1981 Reinganum showed that small firm stocks outperformed big firm stocks. These two researches were not in line with the CAPM, and were introduced as anomalies to the traditional theory.

In reaction to these findings, Fama and French (1993) countered the CAPM, and introduced the Fama-French Three-Factor model. The two factors that were added were respectively the Small Minus Big (SMB) and High Minus Low (HML). SMB is the return of a small cap stock portfolio minus the return of a big cap stock portfolio. HML is the return of a high book-to-market stocks portfolio minus a low book-to-market stock portfolio.

When calculating the abnormal return (

α

Rf) is the excess return of the portfolio, (Rm – Rf) is the market risk premium and β1, β2, β3

are the coefficients of the risk factors.

At the first sight, SMB and HML do not seem to fit as risk factors, but Fama-French (1993) argue that the two factors are serving as proxies for the risks that are not included in the original CAPM. They note that it is empirically proven that small company stocks and high book-to-market stocks are getting higher returns than is estimated by the CAPM.

Critics of this type of reasoning like Black (1993), stress the possibility of data snooping biasedness. Which is the belief that it is always possible to find a variable that is by chance correlating with the error term. Still Fama & French (1993) proved that the SMB and HML factors are beneficial to the explanatory power of the model.

2.2.3 Carhart Four-Factor Model

In 1993 Jagadeesh and Titmann exposed a fourth anomaly, in their research they discovered a pattern for good and bad performing stocks to carry on over time. Carhart (1997) discovered that the alpha of many mutual funds was related to this tendency factor. Therefore he created the Carhart Four-Factor Model by adding a momentum factor (MOM) to the Fama-French Three-Factor Model. The Carhart Four-Factor Model can be expressed in the following equation:

When calculating the abnormal return (

α

excess return of the portfolio, (Rm – Rf) is the market risk premium and β1, β2, β3, β4 are the

coefficients of the risk factors.

Similar to the Fama-French Model, the Carhart model is also criticized by the fact that the MOM factor doesn’t intuitively fit as a risk factor. But modern research brought a lot of sympathy to the idea that the momentum effect might be related to market liquidity in the determination of asset prices. For example, Pastór and Stambaugh (2003) researched if high liquidities have an effect on the alphas computed by the Carhart model and conclude that to some extent, liquidity risk accounts for the momentum effect.

2.3 Performance of mutual funds

In the timespan from 1945 to 1964 Jensen (1967) did not find confirmation that equity mutual funds can outperform the market. Additionally, after extracting the management fees he found evidence that the funds were actually underperforming the market. In 1989 Grinblatt and Titman examined the timeframe 1975-1984 and reached the same outcome for the average mutual fund, but saw that some individual funds outperformed the market. Still the alphas of the individual funds were not significant when management fees were extracted.

Survivorship bias is an important concept in the research of mutual funds. It is the idea that a methodological issue arises when some mutual funds are vanished from the dataset during the research period. In 1995 Malkiel stated that survivorship bias has a bigger effect than was assumed in the past. To prevent the research from survivorship bias, he stressed that it is necessary to include every mutual fund that has existed in the research period. Also the mutual funds that died out, or vanished for another reason. So for the period 1971-1991, he corrected the dataset for survivorship bias and also found evidence that mutual funds underperform the market.

In the research of Jensen (1969), Grinblatt and Titman (1989), and Malkiel (1995) it is stressed that equity mutual funds do not outperform the market. Additionally when the returns are deducted with management fees and the dataset is corrected by survivorship bias there is general consensus that equity mutual funds underperform the market. This statement is in line with the EMH, where security prices reflect all available information and it is not possible for individuals to outperform the market.

3.Methodology and Data

The research design and methodology are explained in this section. First, the estimation models used are presented, followed by the hypothesis of this thesis. Second, the methodology and regression specifications are explained. Third, the process of data collection is described, followed by the explanation of survivorship bias in this particular dataset. And finally, the descriptive statistics of the data are presented.

This study examines the performance of US global equity funds in the period 2009-2014. Performance of a security is at first measured by its return. Though this return has to be adjusted for risk, because risk and return are positively correlated. To examine the risk-adjusted returns of the funds, the three different asset-pricing models explained in the literature section are used. The coefficients in the CAPM, Fama-French Three-Factor Model and Carhart Four-Factor Model are estimated with the following regressions:

CAPM: (Rp– Rf) = αp + β1 (RM – Rf)

+ ε

Fama-French Three-Factor Model: (Rp– Rf) = αp + β1 (RM – Rf)

+

ε

Carhart Four-Factor Model: (Rp– Rf) = αp + β1 (RM – Rf)

+

ε

In line with the research of Jensen (1969), Grinblatt and Titman (1989), and Malkiel (1995) the following hypothesis is stated: The alpha of US global equity funds in the period

2009-2014 is significantly different from zero in the CAPM, Fama-French Three-Factor Model, and Carhart Four-Factor Model.

To test this hypothesis, OLS regressions are conducted, and robust standard errors applied to correct for potential serial correlation and heteroskedaticity.

To test the research question, a sample of 123 US global equity funds are examined. This sample is selected through the advanced search option in Datastream. The search query used for selection was specified with the keywords “global” and “equity”, the market “United States”, and the currency “United States Dollar”. After the selection the monthly prices in between 2009 and 2014 (60 months) of the 123 funds were downloaded. In excel the prices were prepared for regression and computed to monthly returns. The monthly MKT, SMB, HML and MOM factors (60 months) were downloaded from the Kenneth R. French Data Library.

Following the research of Jensen (1969), Grinblatt and Titman (1989), and Malkiel (1995), the first aim was to conduct this study with a dataset that is survivorship bias free. Survivorship bias is the idea that a methodological issue arises when some mutual funds are vanished from the dataset during the research period. Unfortunately the University of Amsterdam did not have access to the survivorship bias free mutual fund database of CRSP. Therefore the dataset was selected with the non-survivorship bias free mutual fund database of Datastream.

When the dataset is non-survivorship bias free, there is a chance that some underperforming funds that started in 2009 died and were excluded from the dataset. Because of that there is a chance that the abnormal performance is less (or more negative) than presented in the empirical results of this study. In case of a negative alpha, there is a chance that the survivorship bias free alpha could be even more negative. In case of a positive alpha, there is a chance that the survivorship bias free alpha could be neutral or negative. But the precise effect of this bias cannot be specified.

The regression of the three asset-pricing models is done in the statistical program Stata. The monthly returns and MKT, SMB, HML and MOM factors are imported to the

data-editor. The 123 funds are all assigned an id, and the returns and factors are reshaped in long form. These variables are included in a regression analysis, in order to compute the coefficients and statistics of the three asset-pricing models. Table I presents the summary statistics for the dependent variable and the independent variables

Table I. Descriptive statistics.

Variable Mean SD Min Max

Return 1.17 5.85 -16.29 19.97

Rmkt – Rf 1.55 4.62 -10.10 11.35

SMB 0.37 2.18 -4.22 5.78

HML -0.02 2.69 -9.87 7.57

4.Results:

The empirical results of the three asset pricing models are presented in this section. First, in

Table II the estimations of the CAPM, Fama-French Three-factor Model and Carhart

Four-Factor Model are presented. Followed by the description of the results in three separate paragraphs.

As shows below, Table II further presents the coefficients of alpha, the market beta, SMB beta, HML beta and MOM beta together with the p-value of those coefficients. The lowest section of the table shows the R-Squared, number of funds and number of observations.

Table II. Estimation results of regression analysis

Model CAPM Fama-French Carhart's

Alpha -0.57 *** (0.00) -0.66 *** (0.00) -0.70 *** (0.00) Rmkt-Rf 1.11*** (0.00) 1.27*** (0.00) 1.25 *** (0.00) SMB -0.42*** (0.00) -0.42 *** (0.00) HML -0.19*** (0.00) -0.24*** (0.00) MOM -0.097*** (0.00) R-Squared 0.78 0.80 0.81 Funds 123 123 123 Observations 7380 7380 7380 Note. ***p < .001

The coefficient of the alpha in the CAPM regression is negative, and with a p-value of 0.00 also significantly different from zero. At the significance level of 1% there is sufficient evidence to reject the hypothesis that alpha is equal to zero. Further the market beta is significantly positive, and the R-Squared is relatively high with a value of 0.78. Which indicates that the model has relatively high explanatory power.

In the Fama-French Three Factor Model regression the coefficient of the alpha is negative, and has a p-value of 0.00, therefore it is also significantly different from zero. Based on the significance level of 1% there is sufficient evidence to reject the hypothesis that alpha is equal to zero. Further, the market beta, the SMB beta and the HML beta are also significantly different from zero. The Squared is with a value of 0.80 higher then the R-Squared of the CAPM regression. Which suggests that the Fama-French Three Factor Model has a higher explanatory power than the CAPM for this particularly dataset

The coefficient of the alpha in the Carhart Four Factor Model regression is negative, and with a p-value of 0.00 also significantly different from zero. At the significance level of 1% there is sufficient evidence to reject the hypothesis that alpha is equal to zero. Again, the market beta, the SMB beta, the HML beta and the MOM beta are all significantly different from zero. The R-Squared is with a value of 0.81 higher then the R-Squared of the CAPM and Fama-French Three Factor Model regression. Therefore, the results show that the Carhart Four Factor Model has a higher explanatory power than the CAPM and Fama-French Three Factor Model for this particularly dataset.

5. Discussion and conclusion

In this section the results of this study are discussed and after which a conclusion is made. First, the results are linked to the relevant theory. Followed by the limitations of this study and suggestions for further research.

The aim of this study was to investigate whether US global equity funds underperform relatively to the market in the period 2009-2014. Therefore, this study examines if the alpha is significantly different from zero in the CAPM, Fama-French Three-Factor Model, and Carhart Four-Factor Model. This study indeed rejects the hypothesis that alpha is equal to zero, which relates to the work of Jensen (1969), Grinblatt and Titman (1989), and Malkiel (1995). Therefore, it implies that as expected US global equity funds underperform relatively to the market.

A further contribution of this study relates to the explanatory value of the asset-pricing model as a tool for performance evaluation, as was explained in the literature review. The CAPM model was introduced by Markowitz (1952). Second, Fama and French (1993) added the SMB and HML factor to the CAPM. And third, Carhart (1997) added the MOM factor to his model. The suggestion that the asset-pricing models explanatory power improved by adding these anomalies is supported by this research. The R-Squared metric describes the explanatory power of the models. As presented in Table II the Carhart Four-Factor Model has the highest explanatory power. The Fama-French Three-Factor Model has the second highest explanatory power and the CAPM the lowest explanatory power.

Overall, this study contributes to the existing literature that US global equity funds in the period 2009-2014 as well did not benefit the investor. In the introduction of this study the analogy is sketched that a blindfolded monkey throwing darts at the financial pages of a newspaper selects a portfolio just as well as a mutual fund manager. According to the empirical results of this study there is the possibility to stretch this analogy, and make the

non-academic claim that a blindfolded monkey throwing darts at the financial pages of a newspaper selects a portfolio just as well or maybe even better than a US global equity fund manager in the period 2009-2014. From a more practical perspective, this study would suggest investors to focus on market trackers rather than actively managed mutual funds.

This study is limited by the fact that the dataset is non-survivorship bias free. Based on the research of Jensen (1969), Grinblatt and Titman (1989), and Malkiel (1995), the first aim was to examine this research with a dataset that is survivorship bias free. As was described above, the University of Amsterdam does not have access to the survivorship bias free mutual fund database of CRSP. Therefore the dataset was selected with the non-survivorship bias free mutual fund database of Datastream. Because of a non-survivorship bias free dataset there is the chance that the alphas of this study are actually more negative than presented in the empirical results. Therefore suggestions for further research are the usage of a survivorship bias free dataset.

References

Basu, S. (1977). Investment Performance of Common Stocks in Relation to Their Price Earnings Ratios: A Test of Market Efficiency. Journal of Finance, 32(3), 663-682 Black, F. (1993). Beta and Return. The Journal of Portfolio Management, 20(1), 8–18.

De Bondt, W.F.M., & Thaler, R. (1990). Do Security Analysts Overreact? American

Economic Review, 80(2), 52-57.

Carhart, M.M. (1997). On Persistence in Mutual Fund Performance, Journal of Finance, 52 (1), 57-82.

Fama, E.F. (1965). The Behavior of Stock-Market Prices. Journal of Business, 38(1), 34-105. Fama, E.F. (1970). Efficient Capital Markets: A Review of Theory and Empirical Work.

Journal of Finance, 25(2), 383-441.

Fama, E.F. (1991). Efficient Capital Markets: II. The Journal of Finance, 46(5), 1575-1617. Fama, E.F., & French, K.R. (1993). Common Risk Factors in the Returns on

Stocks and Bonds, Journal of Financial Economics, 33(1), 3-56.

Grinblatt, M., & Titman, S. (1989). Mutual Fund Performance: An Analysis of Quarterly Portfolio Holding. Journal of Business, 62(3), 393-416.

Grossman, S.J., & Stiglitz, J.E. (1976). Information and Competitive Price Systems. The

American Economic Review, 66(2), 246-253.

Jegadeesh, N., & Titman, S. (1993). Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency. Journal of Finance, 48(1), 65-91.

Jensen, M.C. (1967). The Performance of Mutual Funds in the Period 1945-1964. Journal of

Finance, 23(2), 389-416.

Lintner, J. (1965). The Valuation of Risk Assets and the Selection of Risky Investments in Stock Portfolios and Capital Budgets. Review of Economics and Statistics, 47(1), 13– 37.

Malkiel, B.G. (1973). A Random Walk Down Wall Street (W.W. Norton & Co., New York). Malkiel, B. G. (1995). Returns from Investing in Equity Mutual Funds 1971 to 1991. Journal

of Finance, 50(2), 549-572.

Mossin, J. (1966). Equilibrium in a Capital Asset Market. Econometrica, 34(4), 768– 783. Pastór, L., & Stambaugh, R.F. (2003). Liquidity Risk and Expected Stock Returns. Journal of

Political Economy, 11(3), 642–685.

Reinganum, M.R. (1981). Abnormal Returns in Small Firm Portfolios. Financial Analysts

Journal, 37(4), 52-56.

Sharpe, W.F. (1964). Capital Asset Prices: A Theory of Market Equilibrium. Journal of