Institutional investors need better research if they are going to enter the infrastructure space in a meaningful way.
Australia has pioneered institutional investment in infrastructure with pension fund allocations estimated at 6 per cent of assets. Specialised infrastructure investors argue that “for pension funds, the long duration, inflation hedging and steady cash flow nature of infrastructure investments holds considerable appeal”.
However, recent academic research on the performance of infrastructure investment is mixed. Using samples of Australian, US and global data, both listed and unlisted, recent papers conclude that infrastructure returns can be high but only because risk is also high, that diversification benefits are limited and may disappear with time, that inflation protection is seldom experienced except in the utilities sector, and that downside protection, while real, can be hampered by fund-level risk. Of particular concern when it comes to Australia is the possibility of a historical bias in the data since numerous assets were sold at a discount by distressed local governments in the early 1990s.
These results do not mean that infrastructure investment is not a good idea. Existing research relies on limited and biased data sets and we know that statistical estimates of returns are mostly unreliable. Instead, the inconsistency of research results suggests that infrastructure is an illdefined investment category lacking useful benchmarks.
Before churning more data, an effort is needed on the theoretical side: what should we expect from investment in infrastructure assets? The body of economic literature on the infrastructure sector has yet to be translated into financial economics. Likewise, project finance practitioners know a lot about how infrastructure project cash flows behave.
What is needed today is a concerted effort between academia and the industry to benchmark infrastructure assets as financial assets and work towards the most efficient building blocks that will allow institutional investors to integrate infrastructure in their asset allocation decisions.
I discuss two issues which should underpin future research efforts: infrastructure assets as real assets and infrastructure as an asset class.
Making infrastructure a real asset
No matter what kind of investment tool one has in mind, infrastructure investment always begins with discrete, real-world assets: large complex endeavours almost always characterised by a high up-front investment into a durable and immobile asset with little or no alternative use and a long economic life and repayment profile. But crucially, the value of privately financed infrastructure rests on contractual relationships.