The Historical Record on Active vs. Passive Mutual Fund ...

The Historical Record on Active vs. Passive Mutual Fund Performance

David Nanigian* Mihaylo College of Business and Economics at California State University, Fullerton

This Version: March 1, 2019

Comments are enormously welcome!

* Associate Professor of Finance, Mihaylo College of Business and Economics at California State University, Fullerton, 800 N. State College Blvd., SGMH 5113, Fullerton, CA, 92831, Telephone: (657) 278-4690, Email: dnanigian@fullerton.edu. I am grateful for the insightful comments received from Mehmet Akbulut and Larry Swedroe. JEL Classification Codes: G11, G23

Keywords: Passive investing, Index funds, Actively managed funds, Mutual fund performance, Mutual fund fees, Mutual fund industry competition ? 2019 David Nanigian

Abstract This study examines the risk-adjusted performance of actively managed mutual funds vs. passively managed mutual funds between 1991 and 2018 and finds that the statistical significance of the difference in performance between the two types of funds disappears when the passively managed funds are compared to competitively priced actively managed funds. The practical implication of this study is that, setting tax considerations aside, as long as investors are cost conscious in their fund selection process, investing in passively managed funds does not meaningfully improve investor outcomes.

1

The Historical Record on Active vs. Passive Mutual Fund Performance

1. Introduction

A plethora of studies have documented that mutual funds with higher expenses have lower performance.1 This has led to a belief that a passive investment strategy beats an active strategy and the resultant rise in the popularity of index fund investing (French (2008)). The trend towards index fund investing is illustrated in Figure 1. The purpose of this study is to perform an empirical analysis into the risk-adjusted performance of actively managed mutual funds vs. passively managed mutual funds between 1991 and 2018. The central finding of this study is that the statistical significance of the difference in performance between the two types of funds disappears when the passively managed funds are compared to competitively priced actively managed funds. The practical implication of this study is that, setting tax considerations aside, as long as investors are cost conscious in their fund selection process, investing in passively managed funds does not meaningfully impact household portfolio performance.

2. Data and Methodology

To evaluate the historical performance of actively and passively managed mutual funds, monthly net-of-expense returns and share-class level total net assets (TNA), along with annual report net expense ratios are gathered from Morningstar Direct's survivor-bias-free United States Mutual Funds database. In order to measure risk-adjusted performance using the Carhart (1997) FourFactor Model, the sample is restricted to funds with a Morningstar Category that belongs to the U.S. Category Group of U.S. Equity. The April 2018 Morningstar Methodology Paper on Morningstar category classifications for portfolios available for sale in the United States is used to assign historical Morningstar Categories to broad U.S. Category Groups. Returns and expense ratios are aggregated to the fund portfolio level by weighting them by their TNA at the end of the previous month.

1 See, for example, Carhart (1997), Chevalier and Ellison (1999), Cremers, Ferreira, Matos, and Starks (2016), Elton, Gruber, Das, and Hlavka (1993), Gil-Bazo and Ruiz-Verdu (2009), Gruber (1996), and Harless and Peterson (1998) for evidence of this.

2

Funds are sorted into two groups: passive and active. The passive group consists of funds that are classified by Morningstar as index funds. The active group consists of those that are not classified by Morningstar as index funds. The performance of the two groups of funds over the time period from January 1991 to December 2018 is then evaluated based on the alpha from the Carhart Four-Factor Model. Factor data, along with a one-month Treasury bill rate proxy for the risk-free rate, is gathered from Wharton Research Data Services.

3. Main Results

Table 1 displays the annualized Carhart Four-Factor Model alpha on an equally-weighted "passive" portfolio and on an equally-weighted "active" portfolio. The table also displays the annualized Carhart Four-Factor Model alpha on value-weighted portfolios, in which the weights to each fund are based on the value of each fund's TNA at the end of the prior month.

The results involving equally-weighted portfolios indicate that the actively managed funds underperform the passively managed funds by 0.59% per year. The results involving the valueweighted portfolios indicate that the actively managed funds underperform the passively managed funds by 0.90% per year. To determine whether these differences were statistically significant, paired t-tests were performed. Interestingly, the results involving the value-weighted portfolios (tvalue of -2.59) were statistically significant at the conventional 5-percent level yet the results involving the equally-weighted portfolios (t-value of -1.36) were not significant at even the 10percent level.

Table 2 displays the average annualized Carhart Four-Factor Model alpha on passively managed funds and on actively managed funds in each of the 28 years of this study. This year-byyear analysis illuminates how well anchored mutual fund performance is to its Grossman and Stiglitz (1980) equilibrium. There is no discernable trend in the performance of passively managed funds relative to actively managed funds. In the equally-weighted analysis, the passively managed funds outperformed the actively managed funds in 16 of the 28 years of the study. In the valueweighted analysis, the passively managed funds outperformed their actively managed counterparts in 18 of the 28 years of the study. Taken together, this year-by-year analysis indicates that mutual fund performance is strongly anchored to its Grossman and Stiglitz equilibrium.

3

Numerous studies have documented that mutual funds with higher expenses have lower performance. The very reason for the existence of passively managed funds is to minimize expenses and thereby theoretically improve household wealth accumulation. Table 3 shows that, over the time period of this study, between 67% and 90% of passively managed funds ranked in the bottom quintile of expense ratio amongst the broader set of U.S. Equity Mutual Funds. Table 4 shows that between 93% and 100% of the dollars invested in passively managed funds were invested in funds that ranked in the bottom quintile of expense ratio amongst the broader set of U.S. Equity Mutual Funds. Table 4 also shows that between 30% and 48% of the dollars invested in actively managed funds were also invested in funds that ranked in the bottom quintile of expense ratio amongst the broader set of U.S. Equity Mutual Funds. These observations of the distribution of investor capital across funds indicate that investors are cost conscious in their fund selection process. The prior research on the impact of expenses on fund performance, combined with investors' revealed preference for inexpensive funds, motivates a more "apples to apples comparison" involving only funds that ranked in the bottom quintile of expense ratio.

Table 5 displays the results for the portfolios of funds that ranked in the bottom quintile of expenses. None of these portfolios generated performance that was statistically significant at the conventional 5-percent level. The equally-weighted portfolio of actively managed funds underperformed the equally-weighted portfolio of passively managed funds by a statistically insignificant (t-value of -0.67) 0.24% per year, which is less than half the magnitude of that which was observed in the full sample without the expense restriction (0.59%). The value-weighted portfolio of actively managed funds slightly underperformed the value-weighted portfolio of passively managed funds by a statistically insignificant (t-value of -0.39) 0.13% per year. This performance differential is merely one-seventh the magnitude of that which was observed in the full sample (0.90%).

Table 6 displays the yearly average annualized Carhart Four-Factor Model alpha of passively managed funds and actively managed funds that ranked in the bottom quintile of expense ratio. Consistent with the year-by-year analysis without the expense restriction, there was no discernable trend in the performance of passively managed funds relative to actively managed funds. However, it is interesting to note that there was somewhat greater equality in performance between the two groups of funds. In the equally-weighted analysis, the passively managed funds

4

outperformed the actively managed funds in 15 of the 28 years of the study. In the value-weighted analysis, the passively managed funds outperformed their actively managed counterparts in exactly half (14) of the 28 years of the study.

Taken together, the analysis involving the sample of funds that ranked in the bottom quintile of expense ratio suggests the difference in performance between the passively managed funds and the actively managed funds in the full sample is largely attributable to the presence of actively managed funds with excessive fees. To more closely examine this possibility, I pair each passively managed fund with an actively managed "partner". To identify the "partner", for each passively managed fund, I first restrict the sample of actively managed funds to only those that belong to the same Morningstar Category and then choose the one fund that has the expense ratio that is closest to that of the passively managed fund at hand. In the event that multiple actively managed funds are identified as eligible partners, then the passively managed fund is paired with a hypothetical "fund of funds" that allocates capital equally among each of the eligible partners. The mean and median differences in the expense ratios across the fund-months were three basis points and one basis point respectively. This indicates that the passively managed funds are paired with actively managed funds that have very similar expenses.

Table 7 displays the results for the portfolios of passively managed funds and actively managed "partners". The equally-weighted portfolio of passively managed funds underperformed by 0.47% per year and the equally-weighted portfolio of actively managed "partners" underperformed by 0.64% per year. The statistically insignificant (t-value of -0.39) 0.17% difference in performance between the two equally-weighted portfolios was less than one-third the magnitude of that which was observed in the full sample (0.59%).

The value-weighted portfolios displayed in Table 7 compare the performance of a valueweighted portfolio of passively managed funds to the performance of a portfolio of actively managed "partners" where the weight given to each "partner" is proportionate to the that of its passively managed counterpart. The purpose of this alternative weighting methodology is to discover what the performance of a representative investor in passively managed funds would have been if he or she invested in comparably priced actively managed funds instead, while leaving his or her asset allocation unchanged. Although the difference in performance between the two valueweighted portfolios was not statistically significant (t-value of 0.24), the value-weighted portfolio

5

of actively managed "partners" outperformed the value-weighted portfolio of passively managed funds by 0.11% per year (-0.03% vs. -0.14%).

Table 8 displays the average annualized alphas of passively managed funds and their actively managed "partners" in each of the 28 years of this study. Consistent with the year-by-year analysis involving the funds that ranked in the bottom quintile of expense ratio, there was no discernable trend in the performance of passively managed funds relative to actively managed funds and great equality in performance between the two groups of funds. In the equally-weighted analysis, the passively managed funds outperformed their actively managed "partners" in 13 of the 28 years of the study. In the value-weighted analysis, where weights are based on those of the passively managed funds, the passively managed funds outperformed their "partners" in 16 of the 28 years of the study. Table 8 also displays the percentage of passive funds that outperformed their "active partners" in each year. This ranged from 22% in 1992 to 80% in 1998. The average annual percentage of passive funds that outperformed their "active partners" was 52%. This further indicates that, when passively managed funds are compared to similarly priced actively managed funds, there is great equality in performance between the two groups of funds.

4. Additional Results

Cremers and Petajisto (2009) and Amihud and Goyenko (2013) illuminate the prevalence of "closet indexing" by U.S. Equity Mutual Funds. "Closet Indexing" is when a fund claims to be actively managed yet fails to demonstrate a distinctive portfolio management strategy. Therefore, when evaluating the historical record on the performance of actively managed vs. passively managed funds, it may be more fair to compare passively managed funds to truly actively managed funds. To perform this analysis, the 2 value from the Carhart Four-Factor Model that is run over a twelve-month estimation period is used to gauge each fund's level of strategy distinctiveness. Then, in each month, the sample of actively managed funds is truncated to those with an 2 value that was below the median 2 value for the month among U.S. Equity Mutual Funds. This subset of funds is referred to as "truly actively managed" funds. It should be noted that, due to sporadic TNA data prior to 1991, the portfolio performance evaluation period for this analysis begins in 1992.

6

Table 9 displays the annualized alpha on the portfolio of passively managed funds and the portfolio of truly actively managed funds. In the equally-weighted analysis, the truly actively managed funds underperformed the passively managed funds by one basis point (-0.45% vs. 0.43%). In the value-weighted analysis, the truly actively managed funds underperformed the passively managed funds by 50 basis points (-0.57% vs. -0.06%). Table 10 displays the yearly average annualized Carhart Four-Factor Model alpha of passively managed funds and truly actively managed funds. The results corroborated the year-by-year analysis involving the full sample of actively managed U.S. Equity Mutual Funds, demonstrating that mutual fund performance is strongly anchored to its Grossman and Stiglitz equilibrium.

Table 11 displays the annualized alpha on the portfolio of passively managed funds that ranked in the bottom quintile of expenses and the portfolio of truly actively managed funds that ranked in the bottom quintile of expenses. Most strikingly, the equally-weighted portfolio of truly actively managed funds outperformed the equally-weighted portfolio of passively managed funds and the value-weighted portfolio of truly actively managed funds outperformed the value-weighted portfolio of passively managed funds. In the equally-weighted analysis, the truly actively managed funds outperformed the passively managed funds by 40 basis points (0.17% vs. -0.23%). In contrast, in the equally-weighted analysis involving the sample of passively managed and truly actively managed funds without the expense restriction, the truly actively managed funds underperformed the passively managed funds by one basis point (-0.45% vs. -0.43%). In the valueweighted analysis with the expense restriction, the truly actively managed funds outperformed the passively managed funds by 5 basis points (-0.02% vs. -0.07%). In contrast, in the value-weighted analysis involving the sample of passively managed and truly actively managed funds without the expense restriction, the truly actively managed funds underperformed the passively managed funds by 50 basis points (-0.57% vs. -0.06%).

Table 12 displays the yearly average annualized Carhart Four-Factor Model alpha of passively managed funds and truly actively managed funds that ranked in the bottom quintile of expense ratio. Consistent with the year-by-year analysis without the expense restriction, there was no discernable trend in the performance of passively managed funds relative to actively managed funds. However, it is interesting to note that there was somewhat greater equality in performance between the two groups of funds. In the equally-weighted analysis, the passively managed funds

7

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download