This In Practice piece gives a practitioner’s perspective on the article “What Free Lunch? The Costs of Overdiversification,” by Shawn McKay, CFA, Robert Shapiro, CFA, and Ric Thomas, CFA, published in the First Quarter 2018 issue of the Financial Analysts Journal.
What’s the Investment Issue?
Although diversification is a primary duty for fiduciaries—and in the United States is codified in the Uniform Prudent Investor Act of 1994—fiduciaries often have difficulty knowing just how much diversification is sufficient. Fiduciaries who allocate capital between many already diversified asset managers risk exceeding what is prudent—as well as bearing the higher fees that come with hiring more managers.
Charged with outperforming a policy benchmark, institutional investors often allocate funds to active managers in order to hit their return objectives. Too much diversification can significantly reduce active risk—the risk a portfolio assumes as it attempts to beat the returns of its benchmark. The result is a plan with little scope to hit its targets. In the worst case, the outcome is indexation with high fees.This study examines the potential downsides of overdiversification by looking at pension funds that hold public equities. As new funds are added to an investment plan, the study explores, first, how quickly active risk declines and, second, how fees per unit of active risk are affected.
How Do the Authors Tackle the Issue?
To begin, the authors screen 88 of the 200 largest US defined benefit (DB) plans as of September 2015 (more than three-quarters of which have assets over $10 billion) and find that they allocate to an average of 17 equity funds each. This figure suggests that large DB plans do seem to favor manager diversification. This pattern applies even to smaller funds (with assets between $1 billion and $10 billion).
Using the concepts of active risk and active share (which measure how different a manager’s portfolio is from the benchmark index), the authors devise simple fee ratios: fees per unit of active risk (FAR) and fees per unit of active share (FAS). They estimate how quickly these ratios increase relative to the level of diversification and show that investors can use the ratios to help determine the optimal number of managers.
Next, the authors set out to examine empirically how active share and active risk are affected as more portfolios are added. They build a dataset of 663 active mutual funds with at least 50 holdings each and analyze the portfolios that are highest in three areas: active risk, active share, and percentage of active risk attributable to stock-specific risk. They then generate thousands of random combinations of active managers and conduct a holdings-based risk analysis of each combination. They extend their empirical analysis to examine the relationship between active risk and fees.Finally, the authors recognize that a potential decline in active risk could be beneficial if accompanied by an increase in the expected information ratio—a measure of how consistently managers generate better returns than the benchmark. To determine when the decline in active risk is beneficial, they use a framework called the “fundamental law of active management,” which estimates information ratios on the basis of the number of independent forecasts and the amount of active risk taken.
What Are the Findings?
Excessive diversification among portfolios of public equities can lower the overall level of active risk, causing the FAR and FAS ratios to rise. The authors show that the active risk opportunity set declined as the distribution of active risk narrowed moving from 1 to 10 equally weighted funds. As they increased the number of funds from 1 to 10, ex ante active risk fell from 3.3% to 1.2% while the FAR ratio almost trebled—from 19.4 bps to 53.2 bps—and the FAS ratio almost doubled, from 0.9 bps to 1.7 bps.Using their analytical framework, the authors show that the required information coefficient not only increases but also, in fact, increases exponentially at lower levels of active risk. In other words, the addition of active managers requires an exponential increase in forecast accuracy among the collective managers in order to achieve return objectives.
What Are the Implications for Investors and Investment Professionals?
The main takeaway from this analysis is that many pension funds have taken their model of picking multiple active managers and diversifying too far—leading to low active risk and high fees. The upshot is that unless the returns for beta exceed expectations, many of these funds will probably fail to hit their return objectives.
This study’s conclusions present a challenge to both investors in pensions plans and the investment professionals who oversee these plans. Although the authors make clear that they are not suggesting that all pension funds should avoid active management, they stress that any active manager program must ensure that the costs do not exceed the benefits. Many current programs are paying excessive costs and damaging their returns through overdiversification.
As for whether a decline in active risk may be mitigated by an increase in the expected information ratio, investors should ask themselves whether the exponential increase in forecast accuracy required is realistic.
Another implication of this study is that investors should scale their active manager programs appropriately. On the one hand, large pension plans may have more resources than small programs, but they also face capacity constraints above a certain level. On the other hand, small plans tend to face fewer constraints, but they may also have fewer resources to devote to due diligence. The consequence of this trade-off is that many plans need to restrict the number of external managers in their equity pool and consider boosting the allocation to their remaining managers. Ultimately, plans that tend toward simplification of their external equity manager programs are likely to reduce expenses—and thus optimize their returns.