Abhishek Gupta, Senior Research Engineer and Product Specialist at EDHECinfra, answers 4 frequent questions following our latest webinar on Strategic Asset Allocation.
In the October 27th webinar, you emphasised the importance of taking unlisted infrastructure into account in an investors’ strategic asset allocation (SAA). What was the primary motivation behind this choice of topic?
AG: Investors have a number of motivations to invest in infrastructure including yield pickup and portfolio diversification but also asset-liability management, which can benefit from assets with significant duration like infrastructure. But to really receive these benefits they need to ensure that their long-term or strategic asset allocation to infrastructure makes sense.
Indeed, well-established research has shown that strategic asset allocation is central to investment management and explains most of investment performance. This being said, benchmark selection is equally important and this is where using EDHECinfra data can really make all the difference.
The primary goal of EDHECinfra institute is to increase transparency in the infrastructure asset class and help investors to make more informed decisions. Along with a highly-experienced group of researchers, we also have a dedicated team of financial analysts who have studied more than 600 infrastructure companies extensively going back to their creation up to 20 years ago. Hence, we are well-placed to perform this level of research in infrastructure and provide meaningful insights to our audience.
This webinar is the first in a research series exploring the role of infrastructure in a multi-asset class portfolio. Next year, we also intend to extend this research and look at the role of infrastructure in an asset-liability management setting.
While there is an increasing interest in the asset class, the allocation to infrastructure in institutional portfolios remains low. Why do you think that’s the case?
AG: It is true that infrastructure allocation remains conservative in institutional portfolios but it is certainly not due to the unwillingness of the investors. In the poll conducted during the webinar, the majority of our audience reported their intention to increase their investment in infrastructure from less than 2% currently to close to 10% or more.
The current low allocations to infrastructure are in part the result of the investors’ lack of confidence in the available benchmarks. Indeed, more than 90% of the respondents in our 2019 survey considered their current benchmarks to be inadequate.
Until now, investors have traditionally relied on three types of benchmarks to represent unlisted infrastructure: The first is an absolute-return benchmarks typically designed as a fixed spread over the prevailing CPI or risk-free rate. Such benchmarks offer absolutely no information on risk and are useless for the purposes of strategic asset allocation. These issues have been discussed in detail in our recent paper arguing that “infrastructure investors should abandon absolute return benchmarks”.
A second type of benchmark used by investors is a listed infrastructure index which, as we highlighted in the webinar, is merely a segment of the stock market and does not represent the characteristics of unlisted infrastructure equity.
A third type comes from appraisals: they suffer from several data biases such as survivorship bias, reporting lags, etc. More importantly, these indices are based on smooth valuations leading to vastly underestimating risk and covariance with other asset classes. Without an appropriate measure of risk, this data is wrong and leads to extremely high and unrealistic allocations to infrastructure, typically more than 50% of the portfolio!
It is not surprising that without a reliable benchmark, investors have been very cautious in allocating to infrastructure, but we believe that this can change with an appropriate benchmark which can truly represent the risks of unlisted infrastructure.
EDHECinfra indices have been in the market for over an year now How do these indices can overcome the challenges of the ‘fake’ benchmarks used so far?
AG: EDHECinfra indices were launched in 2019 and include more than 400 sub-indices covering the various segments of both unlisted infrastructure equity and private infrastructure debt markets.
These indices are built in a bottom-up fashion: each constituent is carefully selected to represent the overall market and thoroughly studied by our team of financial analysts. Hence, the data is free from the selection bias typically observed in the appraisal-based indices which have to rely on the contributed data. These indices also retain the historical returns of bankrupt companies, thus avoiding survivorship bias.
The indices are based on a sophisticated and consistent asset pricing methodology grounded in modern finance theory. We decompose the risk premia into several risk factors and update the prices of these risk factors every quarter as we observe new secondary market transactions. This is a parsimonious and statistically robust process, and allows us to estimate mark-to-market or fair valuations for each constituent in the universe. One of the important characteristics of this approach is the absence of smoothness in the returns that we report which provide a realistic assessment of the risks of the asset class.
In fact, the calculation methodology follows the IFRS approach to measuring Fair Value and is based on contemporaneous market inputs, such as interest rates and secondary market transactions. It follows that the EDHECinfra indices exhibit significant correlations with other asset classes, in particular fixed income and to a lesser extent public equity but also property.
Not only do these indices tick almost all the boxes of a ‘good’ benchmark, they also confirm that infrastructure has an attractive risk-adjusted return with high cash yield and should play an important role in any portfolio.
How does this research help infrastructure investors?
AG: In the first part of the webinar, we conclusively show that bad proxies can only lead to bad allocation results. We encourage investors to abandon the use of listed and appraisal-based indices since they do not capture the risks of the infrastructure asset class.
We run a simple substitution exercise in a 60/40 portfolio, and find that, unlike the EDHECinfra indices, appraisal-based index data leads to an inconsistent allocation behaviour. We also explain that the common practice of unsmoothing appraisal-based returns does not resolve these issues and, in fact, asset allocation outcomes become a direct function of the choice of unsmoothing techniques and not of the data. Instead, the main EDHECinfra index (infra300) consistently substitutes with equities in the portfolio.
We also show, using EDHECinfra sub-indices, that granular segments of infrastructure have different characteristics as they represent quite different risk factor exposures. We demonstrate that a precise infrastructure benchmark tailored to the portfolio can improve the Sharpe Ratio of the total portfolio.
The second part of this webinar presents a forward-looking view of SAA in a multi-asset class portfolio. Acknowledging the differences in the objectives of various investors, we implement three optimisation methods, namely return-targeting, risk-targeting, and equal risk contribution. We show that infrastructure asset class can always play a role in an institutional portfolio with a potential allocation of around 10% in most scenarios, significantly higher than the average of 2% reported year after year in the OECD surveys of pension fund allocations to infrastructure.
We also show that adding infrastructure equity and debt to the other asset classes, improves the Sharpe Ratio of the portfolio by up to 6% under reasonable constraints.
This webinar is a precursor to a forthcoming research paper. Following this work, investors will gain a better understanding of the risks in unlisted infrastructure asset class and can confidently use EDHECinfra indices to represent it in their own SAA exercise.