New Research Shows Opportunities and Perils of Pension Funds Investing in Infrastructure

Infrastructure investments can improve risk-adjusted pension portfolio performance by adding diversification and the potential for relatively high and predictable returns. On the other hand, poor asset selection and high expenses can wipe out the benefits offered by such investments.  These are some of the findings in a recent working paper by Clive Lipshitz and Ingo Walter of New York University.

The authors analyzed asset allocation and return data for the nation’s 25 largest pension funds over a 10-year period. They found no correlation between allocations to infrastructure (and other alternative investments) and net investment returns. They hypothesize that higher expenses for these exotic investments offset their higher gross returns, but cannot verify this contention due to inadequate and inconsistent expense data reported by the funds.

Looking at Costs

Examining what little data is available on this subject could be extremely valuable. One system that reports investment expenses in a relatively detailed manner is the California State Teachers Retirement System (CalSTRS), which supplements the relatively limited expense information in its Comprehensive Annual Financial Report with an annual investment cost report.

The CalSTRS cost report shows investment expenses by major asset class. CalSTRS infrastructure assets make up most of an asset class that CalSTRS calls “inflation sensitive.”  At the end of 2017, infrastructure assets accounted for about two-thirds of the CalSTRS inflation sensitive portfolio, so it is likely that the system’s Inflation Sensitive cost ratio of 128 basis points (or 1.28 percent annually) is reasonably close to the cost ratio for its infrastructure investments. This expense ratio is much higher than those for global equity, 14 basis points (bp), and fixed income (8 bp), which compose the bulk of CalSTRS’ portfolio. But inflation sensitive investments had lower costs than real estate (202 bp) and private equity (521 bp). These expenses include “carried interest” —an additional share of profits reaped by a fund manager when returns exceed a certain threshold.

Since the end of 2017, which is when the most recent expense information is available, CalSTRS has increased its exposure to infrastructure investments. Total holdings grew from $2.4 billion in 2017 to $3.0 billion in 2018, with a major new investment announced in May 2019.

While details on CalSTRS infrastructure investments are incomplete, much of its exposure appears to involve clean energy and other environmentally-oriented projects.  For example, one of its infrastructure asset managers, Capital Dynamics, funded a large solar power array in Monterrey County, California.  Another infrastructure investment manager hired by CalSTRS, Argo Infrastructure Partners, cites its green credentials on its public web site. This comports with CalSTRS investment principles, which include the following:

CalSTRS believes that, in addition to traditional financial metrics, timely consideration of material environmental, social, and governance (ESG), factors in the investment process for every asset class, has the potential, over the long-term, to positively impact investment returns and help to better manage risks.

While it would be nice if green infrastructure projects consistently produced superior returns, we do not know this to be the case. Socially responsible investments can certainly produce solid returns. But if CalSTRS places significant weight on the environmental impact of prospective investments, the system may be receiving lower gross returns and paying higher fees than necessary. Infrastructure projects that are not necessarily green, such as toll roads, may be overlooked despite the prospect of offering higher net returns.

That said, the CalSTRS infrastructure portfolio is doing quite well. As of December 31, 2018, the system reported one-year and three-year net returns of 10.00 percent and 11.32 percent respectively on these assets. That compared to total portfolio net returns of -3.22 percent and 6.91 percent over one year and three years, respectively.

Nonetheless, it would be useful to compare the returns on CalSTRS infrastructure funds with others that put less emphasis on ESG factors, a difficult task given the relative opacity of the space. As the study’s authors note, we are going to need better, more standardized data to fully assess the effectiveness of infrastructure investing. Infrastructure investment advocates will need more comprehensive data if they hope to rebut the recent finding in a May 2019 report by Joshua Rauh and his co-authors that this investment class performs no better than private equity funds that buyout corporations or invest in real estate.

Investment Duration

Lipshitz and Walter find that pension funds usually invest in infrastructure through closed-end funds, which liquidate on a specified date, typically 10-15 years after fund inception. Given the long duration of some pension liabilities, the authors argue that pension systems should increase their exposure to open-end funds. They note that open-end funds generate more of their returns from operating cash flows, as opposed to capital appreciation, making them a good match for pension funds which are called upon to make periodic payments over long periods of time.

Aside from investing open-end funds, pension funds can invest directly in infrastructure projects. Canadian pension systems often build and operate their own projects. For example, Caisse de Depot et Placement du Quebec, the province’s second largest fund, is building and plans to operate a new rapid transit system in Montreal.

Lipshitz and Walter note that direct investments require building in-house infrastructure expertise. Although this involves substantial start-up costs and would require developing expertise (whether via in-house staff or experienced consultancies under contract) to navigate the development of complex deals, large US pension systems may be able to justify such investment by the greater flexibility they could then achieve when making infrastructure investments. Direct investments also do not involve fund management fees.

Other Thoughts and Conclusion

Aside from advocating better investment expense reporting and longer duration infrastructure investments, Lipshitz and Walter offer several other suggestions. They suggest increasing portfolio allocations to infrastructure. While large American funds usually invest 0-2 percent of their assets in infrastructure, many of their Canadian counterparts allocate 10 percent or more. These larger allocations would encourage funds to regard infrastructure as a discrete category rather than as a subset of “inflation sensitive”, “private equity” or “alternatives.” This would facilitate better tracking of infrastructure asset returns and costs.

The authors also suggest that US pension funds consider the asset recycling model employed in Australia and make judicious use of public-private partnerships while noting that labor unions and other traditional privatization opponents may be more accepting of innovative infrastructure financing arrangements when public sector employees can share in the benefits. Further, they explore the benefits of greenfield investing—the creation of new infrastructure as opposed to buying existing “brownfield” facilities—observing that it creates the opportunity for value capture, under which the investor and/or government sponsor shares in the appreciation of real estate near newly built facilities.

Pension funds in Australia, Canada and elsewhere show the potential of infrastructure investing for US pension systems. Greater infrastructure allocations offer opportunities for improved risk-adjusted returns if lessons from other countries are applied judiciously and expenses are carefully measured and managed.