What Free Lunch? The Costs of Overdiversification

Abstract

Institutional investors, charged with outperforming a policy benchmark, often allocate to external active managers in order to hit their return objective. The challenge is to do so without overdiversifying the plan. Hiring too many managers can significantly reduce active risk, leaving the plan with high fees and limited ability to outperform a policy benchmark. We review the number of external investment strategies held by the largest US public and corporate pension funds. Our analysis shows that most large pension funds are overdiversified, allowing us to suggest a simpler framework for moving forward.

Disclosure: The authors report no conflicts of interest.

Editor’s Note

This article was externally reviewed using our double-blind peer-review process. When the article was accepted for publication, the authors thanked the reviewers in their acknowledgments. Rajna Gibson Brandon and Maria Vassalou were the reviewers for this article.

Submitted 20 October 2016

Accepted 1 June 2017 by Stephen J. Brown

Acknowledgment

We thank the staff at FactSet for providing invaluable research support for this article. We also thank Jennifer Bender, Rajna Gibson Brandon, Stephen J. Brown, Steven Thorley, CFA, and Maria Vassalou for helpful commentary and guidance after viewing initial drafts of the article.

Notes

¹ The Uniform Prudent Investor Act states in Section 3 that “a trustee shall diversify the investments of the trust unless the trustee reasonably determines that, because of special circumstances, the purposes of the trust are better served without diversifying.” See www.uniformlaws.org/shared/docs/prudent%20investor/upia_final_94.pdf.

² In an email to the authors, P&I offered the following details about its data collection: “Questionnaires were sent to more than 1,200 fund sponsors in P&I’s database. The largest 1,000 were identified from completed questionnaires. Data for funds that did not respond were culled from published annual reports and the Form 5500s filed with the Department of Labor. P&I’s survey generally covers the 12 months ended September 30, 2015. In cases where no information was available from the fund, or the data were older than that date, P&I calculated estimates to September 30.”

³ According to Axioma, as of 2013, the models covered more than 8,700 securities (more than 23,500 historically) listed on various US stock exchanges, including American depositary receipts (ADRs). The models also covered more than 250 exchange-traded funds (ETFs) and more than 300 European Investment Fund (EIF) contracts. Dynamic selection criteria are used to identify stocks on the NYSE and NASDAQ with sufficient size and market liquidity. Typically, only common stocks are eligible (ADRs and foreign issuers excluded), but a few exceptions may arise from time to time (e.g., REITs). Grandfathering logic is applied to ensure stability and robustness. Throughout the model history, the estimation universe amounts to roughly 3,300 stocks, on average. See www.axioma.com/media/uploads/document/axus3-equityrm-201309.pdf.

⁴ Kahn and Lemmon (2016) provided an instructive summary of the relationship between factor (smart beta) risk and specific (orthogonal) risk, with several helpful examples. Expanding on some of these ideas, Garvey et al. (2017) included factor correlations (greater than zero) in their examples.

⁵ The risk model decays observations in such a way that the half-life parameters are 125 days for variances and 250 days for correlations.