A new paper from EDHEC Infrastructure Institute decides that there is no such thing as a listed infrastructure asset class.
What is the practical significance of that inference? It means that for investors (individual or institutional) looking to diversify their portfolio properly, a dedicated index focused on a listed infrastructure theme may not add any value to what one can get from a combination of capital market instruments and alternatives, or from a factor-based asset allocation.
Likewise, although listed infrastructure stocks are often used to proxy investments in unlisted equity, the findings of the new paper throw some suspicion on the adequacy of such proxies.
Three Definitions and 22 Tests
For their own empirical tests of the asset-class-ness of listed infrastructure, the authors focused on three different possible definitions of that class:
- One using industrial sector classifications and the percentage of income generated from predefined infrastructure sectors;
- One accepting existing listed infrastructure indexes, offered for example by Dow Jones, FTSE, MSCI, and S&P, or buy financial institutions such as Brookfield, Macquarie, or UBS, and;
- One consisting of a basket of stocks that offer exposure to underlying projects defined in terms of long-term public-private contracts.
The authors ran 22 statistical tests: nine based on the first (or “naïve”) definition, twelve using the index providers, and one using the contract-based definition.
They tested for persistence by dividing the period of observation into two parts: from 2000 to 2008, and then from 2008 to 2014; that is, before and since the global financial crisis. Part of the motivation for this is that the excessive leverage of the former period may have played a distorting role.
The majority of test portfolios that improved the mean-variance efficiency frontier before the GFC failed to repeat this feat after the GFC. Further, of the 22 test portfolios used, “only four manage to improve on a typical asset allocation defined either by traditional asset class or factor exposure after the GFC….”
A Critique of Existing Literature
The EDHEC study has three authors: Frédéric Blanc-Brude, Director of the EDHEC Infrastructure Institute Singapore; Tim Whittaker, associate research director there and head of data collection; and Simon Wilde, a Ph.D. candidate at the University of Bath.
The existing literature includes studies that have used each of the first two definitions presented above. Each one of those that targets the naïve first definition proceeds somewhat differently from the others. In general, though, when only a small group of stocks gets through the filters, the result is anecdotal. In studies where a larger group of stocks has gotten through, the results have been heterogeneous. Neither situation supports the notion of that it is a true asset class that is being tested.
What of those scholars who have made ad hoc use of listed infrastructure indexes? Some of them have found that the supposed asset class “exhibits similar returns, correlations, and tail-risks” as the market portfolio. An example is a paper published in 2014 by Ron Bird et al. in European Financial Management.
Blanc-Brude et al also acknowledge that there have been some studies on the performance of infrastructure indexes that have reported potential diversification benefits, but they call shenanigans. That is, they say that these authors claim that those earlier studies failed to test these supposed benefits for statistical or economic significance.
Not Beating the Baseline
In their own statistical tests Blanc-Brude et al employ what they call a “relatively unambitious baseline” to represent the broad equity markets. They use the MSCI World Index, a “free float adjusted market capitalization weighted index” that draws on the equity markets of 23 developed nations. This “should not be difficult to beat,” yet adding a new infrastructure component as per the accepted indexes doesn’t help to beat it in terms of a better risk-adjusted performance.
But what about those four proxies managed above that do manage to improve on reference asset allocation after the GFC? The secret sauce for those four stand-outs was the ownership of assets with regular and high dividend payouts. What makes those possible is “the contractual and governance structure of the underlying businesses, not their belonging to a given industrial sector….”
The notion of “infrastructure” as a locus of investing can be useful, then, but only as a heuristic drawing an investor’s attention to contracts and governance structures, and neither as “a thing nor an end in itself.”