The Effect of Management Design on the Portfolio Concentration and Performance of Mutual Funds

Abstract

T he economic role of actively managed equity mutual funds is to delegate the stock selection decisions of individual investors to professional fund managers. The general belief is that these fund managers can generate abnormal returns relative to passive investment strategies. There is an ongoing debate among academics regarding whether actively managed funds outperform passive mutual funds or index funds. Many studies have found that active fund managers do not outperform the market-for example, Wermers (2000) showed that, on average, actively managed funds do not outperform the market after fees and expenses.The finding that the average actively managed fund does not generate abnormal returns has led researchers to investigate what limits the ability of fund managers to generate performance. For example, Ackermann and Ravenscraft (1999) focused on regulatory restrictions faced by funds; Ellis (2014) argued that the lack of private information is a hindrance. A recent strand of research has examined fund or manager characteristics that affect mutual fund performance in order to identify which mutual funds perform better. 1 Our study builds on the literature that has identified portfolio concentration as a key dimension that affects performance (e.g., Brands, Brown, and Gallagher 2005; Kacperczyk, Sialm, and Zheng 2005; Ivković, Sialm, and Weisbenner 2008; Cohen, Polk, and Silli 2010; Huij and Derwall 2011) and extends it in several directions. First, many of these studies measured the concentration of portfolios on the basis of concentration across industry sectors, but we found an additional, marginal impact on performance by within-sector concentration. This finding strengthens the conclusion that concentration has more to do with mutual fund ability than with picking up abnormal returns associated with industry momentum strategies, which also lead to industryconcentrated portfolios (see Wu 2015). Second and perhaps more important, we identified the organizational design behind the loss of abnormal returns associated with less concentrated portfolios. In particular, we found that mutual funds run by a single manager tend to have a much higher portfolio concentration, both across and within industries, than funds run by multiple managers. We further found that when funds' management designs are changed from single manager to multiple managers (or from multiple to single), portfolio concentration decreases (increases) and performance deteriorates (improves).To construct measures of mutual fund portfolio concentration, we followed Kacperczyk et al. (2005), who first showed that mutual funds with portfolios concentrated in a few industry sectors tend to outperform. They argued that this cross-sector concentration is an indication of fund managers' self-assurance and ability. Similarly, we constructed an industry concentration index (ICI) on the basis of 10 industry sectors. 2 In addition, we developed a new within-sector concentration index (WCI) to capture the degree of mutual fund portfolio concentration withi...

What Free Lunch? The Costs of Overdiversification

“…Cremers and Petajisto (2009) showed that funds with the highest active share tend to outperform their benchmarks significantly. Goldman, Sun, and Zhou (2016) found that mutual funds with higher withinsector concentration perform better than those with more diversification and less active risk. Cremers and Curtis (2015) made a strong implicit case for high-active-risk strategies by showing that the fees charged by "closet index funds" are unjustified.…”

Hunting High and Low: Visualising Shifting Correlations in Financial Markets

“…Investment portfolios often contain high numbers of holdings or securities, with US mutual funds typically holding between 40 and 120 stocks [SS05]. More recent work suggests that these numbers have increased, with a mean of 144 and a standard deviation of 271 stocks held [GSZ16]. The ability to visualise a substantial number of pairwise correlations concurrently is therefore of particular interest.…”

“…Using pre- calculated layouts (to gauge the underlying rendering engine's performance), the largest test dataset was animated at an average frame rate of 65.7fps. Extrapolating from the performance on the largest test data set, the frame rates reported should be achievable for approximately 85% of investment portfolios, assuming that portfolio sizes are normally distributed and that estimates from Goldman et al [GSZ16] generalise.…”

“…The study of dynamically varying covariances and correlations is thus of great importance in many areas of the securities industry, including, inter alia, investment selection, portfolio construction, derivatives pricing, structuring and trading, financial index construction and the measurement and management of financial risk [SF12,GSZ16]. In current industry practice, however, correlations are often represented in a manner which requires considerable analytical effort from the user (see section 3.2).…”

Trading Concentration and Industry-Specific Information: An Analysis of Auto Complaints

“…Despite longstanding conventional financial advice to minimize idiosyncratic risk by holding well-diversified portfolios (c.f., McKay et al, 2018), several studies document that both individuals and investment funds concentrate their trades and holdings in certain firms or industries (Brands et al, 2005;Baks et al, 2006;Goldman et al, 2016). Researchers have offered several reasons, including reducing investment information overload (Simon 1972;Agnew andSzykman, 2004: Garvey et al, 2017), possessing an information advantage (Lee and Rahman, 1991;Kacperczyk et al, 2005;Hiraki et al, 2015), and behavioral biases like familiarity (Pool et al, 2012) or overconfidence (Goetzmann and Kumar, 2008), for why individuals and funds would concentrate investment activity and holdings.…”