The world reportedly needs hundreds of billions of long-term investment in new infrastructure projects, but should long-term investors like insurers or pension funds care? Simply put, infrastructure investment, whether equity or debt, can be relevant to investors only if it can improve asset allocation decisions, which remain institutional investors’ first-order problem.
Institutional investment in infrastructure assets is related to a broader trend to improve portfolio diversification through alternative investments. This trend involves investing increasingly outside of capital markets, finding sufficiently long-dated instruments with a more attractive performance than government bonds, and investing in inflation-linked securities other than Treasury Inflation Protected Securities (TIPS). One of the most salient features of these emerging investment choices is the decision to buy assets that are infrequently traded and to hold them until maturity.
Still, investment demand (for new infrastructure) and supply (from institutional investors) have not easily come together and promoting that match requires the development of new knowledge about the expected behaviour of these assets. Without this, large allocations to infrastructure investment by institutional players will remain out of reach.
To answer the question of infrastructure investment’s relevance to investors, we propose a step-by-step “roadmap” requiring a multi-stakeholder effort to reveal the characteristics of infrastructure assets at the underlying and portfolio levels, and reduce existing information asymmetries between investors and managers.
1. Definition: The first step on the road to relevant investment solutions in infrastructure for institutional investors is an unambiguous definition of the underlying assets.
2. Valuation and risk measurement methodology: With a clear and well-accepted definition of underlying instruments, adequate valuation and risk measurement methodologies can be developed that take into account infrequent trading. By “adequate” we mean that such methodologies should rely on the rigorous use of asset pricing theory and statistical techniques to derive the necessary input data, while aiming for parsimony and realism in terms of data collection. The proposed methodologies should lead to the definition of the minimum data requirement (MDR), necessary to derive robust return and risk estimates.
3. Data collection requirements: While ensuring theoretical robustness is paramount to the reliability of performance measurement, a trade-off exists with the requirement to collect real-world data from market participants. In particular, proposed methodologies should aim to minimise the number of inputs in order to limit the number of parameter estimation errors. Adequate models should also focus on using data points that are known to exist and are already collected/monitored or could reasonably easily be collected. In all cases, data requirements should be derived from the theoretical framework, not the other way around. In turn, whether the necessary data already exists or not, this process will also inform the standardisation of investment data collection and reporting.
4. Reporting standard: The standardisation of infrastructure investment data collection should allow the emergence of an industry-wide reporting standard, which can be recognised by investors and regulators alike. Such a reporting standard would increase transparency between investors and managers, who would now be mandated to invest in a well-defined type of instrument and commit to report enough relevant data for investors to benefit from their specialised monitoring.
5. Investment benchmarks: Once the investment profile of the underlying asset has been documented as well as existing data allows, spanning expected returns, risk and market correlations, investment benchmarks can be designed to reflect the performance of a given strategy (e.g. Maximum Sharpe ratio) for a given horizon. If such benchmarks are found to be a potential improvement of the investable set then investment solutions can be designed.
6. Investment solutions: These investment benchmarks can serve as the basis for the development of various standard or tailored investment solutions by the industry, including different types of funds with explicit horizons and risk profiles.
7. Regulation: The robust performance benchmarking of unlisted infrastructure equity portfolios also has direct regulatory implications for risk-based prudential frameworks like Solvency 2. It should allow the calibration of, for example, a dedicated unlisted infrastructure sub-module in the context of the Standard Formula, or usefully inform investors’ internal risk models.
8. Public procurement: Finally, documenting the financial performance of unlisted infrastructure is relevant for the design of public infrastructure tenders and contracts. It is the opportunity for the public sector to involve investors early in the design of public infrastructure contracts on the basis of an academically-validated and industry recognised measure of investment performance.
In recent publications, we have begun to suggest some of the elements to this roadmap, in particular the definition of underlying assets and the methodologies required to measure value and risk adequately, with parsimonious data inputs.
DEFINITION: PROJECT FINANCE AS THE INFRASTRUCTURE INVESTMENT BENCHMARK
It is often said that there is no universally accepted definition of infrastructure. In our view, the absence of definition of what constitutes physical infrastructure is unproblematic. As we argued in a recent paper, the sectoral/industrial terminology is close to meaningless from an investment point of view, since it does not refer to a specific type of financial instrument or investable asset.
For instance, two European road projects can have radically different risk profiles (e.g. a Hungarian real toll road vs. a Finish availability payment road) while a French social infrastructure project and an Australian coal terminal (both receiving pre-committed cash flows, controlling their operating costs through fixed price contracts, and with comparable initial balance sheet leverage and debt amortisation schedules) can have very similar equity cash flow volatility and duration.
Defining “infrastructure” investment from a strategic asset allocation perspective aims to determine expected returns, risk and market correlations, or for short, the existence of a distinctive or remarkable beta. In turn, the identification of attractive investment properties requires the definition of underlying financial instruments that are used to finance infrastructure projects.
Financial instruments, as opposed to industrial sectors, have a clear definition and regulatory treatment, and their expected behaviour can be modelled and documented empirically. Of course, there is a trade-off between clarity and scope. A definition of infrastructure investment as a finite set of financial instruments will almost certainly restrict our definition to a subset of what is commonly understood to be “infrastructure”. Again, from a pure investment point of view, this is unproblematic. What matters is to frame the question in terms of asset allocation.
Our proposed solution in a forthcoming paper (Blanc-Brude and Ismail 2014) is to define the underlying asset in infrastructure investment using the Basel 2 definition of Project Finance (BIS 2005) i.e. a special purpose entity (SPE) dedicated to the construction and operation of a new infrastructure project over a given period, typically 25 to 30 years. In effect, most infrastructure investment and the immense majority of new or ‘greenfield’ projects are financed using such structures.
Of course, other definitions of “infrastructure” are possible as long as they a) correspond to well defined financial instruments, b) are expected to behave differently than existing investments from the point of view of a large well-diversified investor, c) correspond to pools of assets that are sufficiently large to be relevant at the strategic asset allocation level.
METHODOLOGY: MEASURING VALUE AND RISK IN INCOMPLETE MARKETS
Standard asset pricing theory relies on a demanding framework of market completeness and efficiency, which is at odds with the valuation of infrequently traded assets like infrastructure equity. Indeed, while the equity investor in an infrastructure project can always be described as writing a call option on the project’s free cash flow (with the debt service as the strike price), there is no traded instrument representing the free cash flow process that could be combined with a risk-free bond to replicate the pay-off of the call option. If investors cannot continuously hedge the call option, their valuations will not converge towards a unique pricing measure.
When markets are incomplete, individual investors arrive at different valuations of the same asset. Instead, individual investor risk preferences determine their cost of capital for a given expected payoff. Hence, the proportion of infrastructure equity returns that cannot be explained by objective systematic factors must lie within a range of discount rates, determined by individual investors’ attitudes towards risk, liquidity, inflation, duration etc.
For example, consider a basket of infrastructure equity in infrastructure projects with a well-documented duration and inflation hedging property. Imagine two potential investors: an insurance firm with a duration mismatch but no inflation-linked liabilities and a minimum liquidity constraint, and a sovereign wealth fund with a mandate to preserve the purchasing power of national savings and a very-long investment horizon.
With no duration or liquidity constraint, the SWF is only interested in investing in the inflation-hedging property of these assets. Conversely, the insurer values the opportunity to invest in assets with a duration but must also price its liquidity constraint and, in this case, has no particular preference for inflation hedging. Both buyers are unlikely to offer the same price and there is no reason why their valuation should converge, expect incidentally.
In a forthcoming paper, Blanc-Brude and Ismail (2014) propose a set of methodologies aiming to measure risk and valuations when the law of one price does not apply i.e. when a range of valuations is to be observed in the market, and how this range may be modelled and observed empirically.
NEXT STEP: DATA COLLECTION
Relevant methodologies to value infrastructure equity stakes will continue to be developed, but their contribution to the “roadmap” will depend on their data collection requirements. Blanc-Brude and Ismail (2014) also define the minimum amount of data that needs to be collected to implement our proposed methodologies. Widespread data collection using a standardised format, leading up to an industry-recognised reporting standard for infrastructure investment managers, is the next important step on the road to building investment solutions in infrastructure that are relevant to institutional investors.
A major effort remains to be made by investors, managers, regulators and academics to contribute to the development of relevant investment solutions in infrastructure assets, including state-of-the-art methodologies and the definition of the minimum data required to implement them and answer benchmarking and asset allocation questions seriously. Market players must also embrace the standardised and transparent reporting of underlying cash flow data for these benchmarks to become a reality.