In this third and final article on the EDHECinfra/G20 survey of infrastructure benchmarking practices the role of infrastructure investment benchmarks for the purpose of risk management is discussed. More than 300 respondents took part in the survey, including representatives of 130 asset owners accounting for $10 trillion in assets under management, or more than 10 per cent of global AUM. Most respondents declared needing to change their current benchmarks when it comes to the risk management of their infrastructure investments.
Infrastructure investment as a collection of risk factor exposures
In a portfolio context, risk management aims to control and optimise the amount of risk taken by investors per unit of expected reward (excess return or spread). As such, it focuses on the sources of remunerated risk found in various securities i.e. the factors that explain and predict the price and therefore the returns of these securities.
Priced risk factors are the result of fundamental economic and financial mechanisms but are usually proxied using the characteristics of investments that systematically explain or drive asset values.
For instance, most asset values are impacted by movements in interest rates. But not all assets are equally exposed to interest rate risk, depending on their expected life and payouts. An important aspect of risk management for infrastructure investors then is to understand the underlying risk exposure created by a given infrastructure investment strategy or mandate.
Moreover, as discussed in the second article in this series, the construction of an infrastructure portfolio can be a lengthy and uncertain process and investors’ risk exposures can be expected to evolve significantly overtime.
Infrastructure investors also face evolving risk exposure at the universe level: the underlying investible universe keeps changing as new countries embrace infrastructure privatisation, or others turn their back on certain types of concession contracts or revise existing regulation or subsidies.
This is reminiscent of the sub-optimality issues found in cap-weighted market indices: standard stock indices exhibit both sector and style biases (concentrations) that make them either relatively inefficient or relative unstable in terms of risk exposures.
Moreover, these biases tend to change over time, making standard cap-weighted indices unsuitable as benchmarks since their implicit risk exposures drift in the long-run in a manner that investors cannot control. The solution to this issue is to build benchmarks that have constant sector and geographic weights or, even better, target a constant exposure to certain risk factors.
The absence of infrastructure investment risk management
None of these issues are currently taken into consideration in the risk management process of infrastructure investors.
In the EDHECinfra survey, most investors declared using the same benchmarks for risk management as they do for strategic asset allocation and performance monitoring.
That is, nearly 70 per cent of investors in unlisted infrastructure continue to use absolute-return benchmarks for the purpose of risk management i.e. benchmarks that do not represent risks.
This suggests that the infrastructure portfolio risk management function is very limited amongst these investors.
Unsurprisingly, only 10 per cent of respondents said that their choice of benchmark is adequate for risk-management purposes.
Among the vast majority of respondents there is a consensus that current practices present a number of challenges.
More than 50 per cent of respondents are concerned that their choice of benchmark do not allow for measurement of diversification indicators such as effective number of factors/constituents.
Half worry that these benchmarks do not measure exposure to traditional risk factors such as size and momentum, which are likely to be found in multiple asset classes involving equity investment.
Likewise, around 40 per cent of equity investors said that current benchmarks do not allow for stress testing or default risk mapping, nor do they measure contributions to asset-liability-management (ALM) objectives.
Investors in infrastructure will benefit from understanding underlying risk factor exposures
As also discussed earlier, for infrastructure investors the choice of strategic benchmark effectively embodies two challenges: 1) the creation of the portfolio to which the benchmark refers, which is a lengthy and potentially costly endeavor, and 2) for this portfolio to also out-perform the benchmark.
Hence, from a risk management perspective, it is not enough for investors in infrastructure to simply know that they added another bridge or another airport to their portfolio. They need to know which risk factors they are becoming exposed to through each new investment.
A decomposition of risk exposures by individual factors would create more control in the way an infrastructure portfolio is built over time (since factor exposures are present in all investments) and also would also allow optimising the portfolio in order to achieve the desired risk exposure determined at the strategic level.
Moreover, to the extent that risk factors are found within multiple asset classes, investors’ total portfolio risk is also partly determined by the dependencies between assets classes created by common risk factor exposures.
For instance, interest rate or credit risk can be expected to be present in multiple asset classes like fixed income and also infrastructure, including infrastructure equity, since leverage is typically high in infrastructure companies and repayment period very long. As a result, the current value of any stream of future dividends to unlisted infrastructure equity investors is partly driven by the movement of interest rates (discount rates) and the possibility of being “wiped out” by a default.
Eventually, understanding how each asset-class component of the portfolio loads on various cross-asset-class risk factors is the key to a successful the risk-measurement and management process.
Because investments in infrastructure are illiquid, they are not easily or rapidly changed and optimizing the portfolio is more easily done though more liquid asset classes. Hence, if the infrastructure portfolio creates a certain exposure to interest rate risk, along with, say, bonds, because the former is the most illiquid, an investor owning both types of assets can optimize their total exposure to interest rate risk by buying or selling bonds, while taking the interest rate risk exposure of the infrastructure part of the portfolio into account.
While such approaches are not common in the unlisted asset space, recent evolutions and the need to better measure the sources of performance are becoming increasingly apparent amongst investors and regulators. This requires a robust statistical model of expected returns to be calibrated using observable and predictable inputs.
For instance, the technology developed by EDHECinfra to value unlisted infrastructure assets is based on a multifactor model of expected returns that allows measuring over time the impact of various priced risk factors such as interest rate risk, size (illiquidity), credit risk, investment (capex intensity), etc.
In due course, it will become possible to approach the risk management of unlisted infrastructure as a series of choice to gain exposure to a combination of risk factors. This evolution will occur as better benchmarks that represent the risks taken by investors gradually become used for strategic asset allocation and performance monitoring, as the respondents of the 2019 EDHEC/G20 Survey have strongly suggested in their responses.
The other two stories in this three part series are