Scientific Beta: Smart infrastructure portfolio construction

3 september 2024

By Frédéric Blanc-Brude (Director), Abhishek Gupta (Associate Director), and Moataz Farid, (Quantitative Analyst), EDHEC Infrastructure & Private Assets Research Institute

Unlisted infrastructure investments are notably illiquid due to their substantial size and their few secondary market transactions, creating challenges for portfolio construction. Despite these limitations, large institutional investors pursue infrastructure investments for diversification purposes.

Achieving broad diversification in this asset class is difficult because the lack of liquidity restricts entry and exit points, complicating portfolio rebalancing. Additionally, the inability to short sell in private markets also limits the scope of infrastructure investing to a long-only strategy.

Improving total portfolio diversification with unlisted infrastructure

Previous research has suggested that unlisted infrastructure equity offers significant potential to improve total portfolio diversification. However, this is on the assumption that the asset class is accessed on a well-diversified basis. But what does this mean for an investor in infrastructure in practice? Is there a minimum number of assets or sectors that justifies using the phrase ‘well-diversified’ in the annual report? Is a portfolio of just 10 infrastructure assets necessarily under-diversified? We show that answering these questions is not as simple as totting up the assets, sectors, or countries in which individual investments have been made.

In recent research[1], we aimed to answer two questions:

1) Is it feasible to build a diversified infrastructure portfolio?

We construct several portfolios using a ‘brute force approach’. These are 100 equally weighted random portfolios for each target number of constituents, from 5-100, irrespective of any other sector or geography criteria. These ‘brute force’ strategies show that on average portfolio diversification can be achieved by increasing the number of assets in the portfolio.

However, holding a large number of assets in an infrastructure portfolio is not feasible. While infrastructure investments are very heterogeneous and different from one another, it is important to recognise the existence of common risk factors in a portfolio of such investments. We show that achieving a well-diversified portfolio of infrastructure investments is nonetheless possible with a limited number of investments – as long as the key risk factors found in these investments are used to build the portfolio accordingly.

2) What role do infrastructure investments have in a multi-asset class portfolio?

We construct a portfolio that includes nine asset classes. We then add infrastructure to the portfolio and examine the implications using different optimisation techniques such as return targeting, risk targeting and equal risk contribution. We show that infrastructure investments can have a weight ranging from 4.4% to 13.1% depending on investors’ risk profiles.

Building infrastructure portfolios using the ‘Smart Infra’ approach:

To reduce idiosyncratic risk within the portfolios, we use the logic of factor investing in the context of unlisted infrastructure investments. This strategy involves assessing individual infrastructure investments for their exposure to key systematic risk factors and tilting our portfolios toward assets that have high exposure to these systematic risk factors. These five key risk factors are proxied by firm-level financials and a country risk factor.

We refer to this approach as ‘Smart Infra’, as it follows a multi-step approach that not only tilts towards a given factor or group of factors, but also achieves diversification within the factors tilt through the combination of alternative weighting schemes.

Our approach filters the number of assets in the universe to the assets that have a strong exposure to individual risk factors and a strong multi-factor intensity overall. This makes the portfolios less sensitive to the underperformance of one specific factor and enables them to benefit from higher potential for outperformance over the long run. We apply a diversified weighting scheme to diversify idiosyncratic risks and achieve the highest possible risk-adjusted return.

The outcome of this approach is a ‘High Factor Intensity (HFI)’ portfolio that achieves high risk-adjusted return with small number of assets. A 25-asset HFI portfolio achieves higher risk-adjusted returns, measured by Sharpe ratio, than both brute force portfolios built with 100 randomly chosen assets across 12 different sectors and 100 randomly chosen assets across 12 different geographies.

This confirms that, by applying our Smart Infra approach, fund managers can achieve diversification through concentration. Using our ‘Smart Infra’ approach, investors can construct portfolios that are not only theoretically viable but also practically achievable, leveraging the systematic risk factor exposure for each individual asset.

Conclusion

The Smart Infra approach described in our research makes diversification of unlisted infrastructure investments feasible within portfolios. Hundreds of bets can be necessary to build a portfolio with fully diversified idiosyncratic risks unless these assets are selected based on their risk-factor exposures. 

Trying to add decorrelation to the portfolio by adding more assets in different sectors and countries ignores the fact that investments are not only linked by sectors and countries but also by their risk profile as a business, i.e. the risk factors described in this paper. These risk factors are universally available in all assets because they represent the systematic risk that the market prices in these assets. This universal availability enables investors to access exposure to these factors much more easily than sector and country bets.

Defining a diversification strategy as ‘We need to add 20 new transport investments in 10 different countries to the portfolio’ is a non-starter for any deal team. Instead, ‘We need to add 10% of exposure to the size factor to the portfolio’ is relatively easily implemented.

The principles of risk factor diversification illustrated in our research demonstrate that in private markets, where investors are restricted to holding a limited number of assets, they can still achieve diversification through concentration. Our approach is not limited to unlisted infrastructure equity alone but can be extended to other illiquid asset classes such as infrastructure debt, real estate, and private equity. By applying similar strategies, investors can leverage the benefits of systematic risk factor exposure to achieve more efficient diversification, thereby enhancing portfolio performance and mitigating idiosyncratic risks across a broader range of investments.