How Should Investors Benchmark Infrastructure?

Since no standard exists for benchmarking the performance of unlisted infrastructure investments, institutional investors must choose a benchmark with “tolerable imperfections.”

(March 23, 2012) — As pension funds become increasingly interested in infrastructure as a long-term investment, a new paper asks: How should we benchmark and value unlisted asset classes?

“Benchmarks for Unlisted Infrastructure” asserts that institutional investors must choose a benchmark with “tolerable imperfections.” The paper was written by Jagdeep Singh Bachher, deputy CIO at the Alberta Investment Management Corporation (AIMCo) who recently received an aiCIO Industry Innovation Award, Ryan J. Orr, executive director at Stanford University’s Collaboratory for Research on Global Projects, and Daniel Settel, co-founder and vice president of operations at Zanbato Group — says that institutional investors must choose a benchmark with “tolerable imperfections.”

“Different benchmark selection choices have different inherent strengths and weaknesses. The Bailey criteria provide a generally accepted framework for assessing benchmark quality. When we apply the six Bailey criteria to assess the quality of the available families of infrastructure benchmarks, we find that all of the benchmark options are flawed—that is, none satisfy all six criteria,” the report, originally published by the CFA Institute, asserts. 

Bailey’s criteria include 1) unambiguous, 2) investable, 3) measurable, 4) appropriate, 5) Reflective of current investment options, 6) Specified in advance. 

The authors outline a number of potential benchmarks used for institutional investors. While the hybrid benchmarks may combine both equity and debt components in a composite index, the custom portfolio benchmark satisfies the largest number of the Bailey criteria, but still suffers from being composed of listed companies. Meanwhile, the peer group benchmark is the only benchmark that is fully appropriate and reflective of current investment options, the paper notes, but it fails on most of the other scores. 

The paper continues to note that within certain limits, chief investment officers will allocate assets to infrastructure if expected returns exceed the opportunity cost of not investing in a portfolio of public debt and equity. “As one CIO describes, ‘In the absence of attractive opportunities for active management, our partners would simply buy and hold broad-based equity and bond indices.'”

In other words, according to the authors, investing in illiquid infrastructure assets with a buy-and- hold strategy over long time horizons is a lot like managing a multicrop farm. “Once capital has been committed, it is difficult to redeploy. Returns cannot be known precisely until after the harvest, when the actual performance is tallied.”

Therefore, the authors outline describe several general principles for selecting and implementing benchmarks for unlisted assets. 

1) Separate the “opportunity cost” and “manager value-add problems” problems, 2) Take the long view, 3) Create the right incentives, 4) Match benchmark and vintage period, 5) Match benchmark and valuation frequency, 6) Contemplate cost–benefit trade-offs, 7) Track the business plan, 8) Revalue conservatively.

The paper is the second part of an earlier report by the same authors that note that different investors have different goals for their infrastructure portfolios, leaving no single “right” way to benchmark the asset class. To conduct the analysis, AIMCo conducted a three-week brainstorming, critiquing, and idea refining session, with the paper’s analysis coming from interviews with nine institutions that maintain dedicated infrastructure investment allocations, as well as a review of scholarly literature.

According to the paper, the diversity of benchmark approaches reflects a number of factors — namely the newness and heterogeneity of infrastructure, variations in risk–return expectations across institutions, and a lack of widely cited performance data for infrastructure. “Several institutions noted a desire for greater benchmark stability when selecting real return benchmarks,” the paper asserts. “Some hybrid approaches may represent an attempt to integrate more of the desired features of infrastructure investing (low volatility, inflation-linkage, cash yield, etc.) into a single composite. The diversity also reflects the fact that infrastructure investments play different roles in different investors’ portfolios.”

Furthermore, the paper maintains that despite the different allocations, institutional investors generally conclude that on the risk–return continuum, infrastructure investments fit somewhere between regular equities and fixed income.