Portfolio Construction Framework for Infrastructure

By Michael Landman, Investment Director, Infrastructure - 10 June 2012



Institutional investors seeking exposure to infrastructure typically seek a balanced portfolio with a bias towards core infrastructure.

This objective raises the question: How can fund managers and investors quantify what constitutes a balanced core infrastructure portfolio? IFM recognised that there was scope to step up the level of sophistication in this area, leading us to develop an integrated suite of tools for portfolio analysis and construction, known as InFRAME™.

InFRAME draws on IFM’s internal research and experience of over 17 years of investing in and managing infrastructure. It is based on taking a comprehensive ‘bottom-up’ perspective on risk through identifying and analysing the underlying revenue streams from which infrastructure assets derive their income. This provides a framework for portfolio construction that goes beyond traditional sub-sector classifications and increases the depth of analysis into the risk profile of infrastructure investment.

Benefits of InFRAME

The benefits of IFM’s approach to portfolio construction include:

  • a deeper understanding of the risk profile of infrastructure within an investment portfolio through the development of risk profiling tools
  • quantitative insights into the way infrastructure assets, and the sub-sectors to which they belong, respond to macro-economic drivers and scenarios
  • the development of a strategic allocation within infrastructure based upon revenue characteristics of the underlying assets.

IFM has successfully applied InFRAME to our Australian and global infrastructure funds. Our investment analysis and decision making routinely benefit from the insights provided by having a deeper understanding of portfolio investment risk.

Objectives of the portfolio construction initiative

In early 2011, IFM set out to develop a new approach to portfolio construction for the unlisted infrastructure asset class.

At the time, we analysed portfolio diversification though sub-sector analysis and the use of a limited set of revenue characteristic classifications. We had also experimented with applying the modern portfolio theory approach (‘mean-variance’ analysis) to infrastructure sub-sectors. Mean variance analysis is the basis of most conventional analysis of listed equities and fixed interest portfolios, and essentially relies on historical performance and correlation data, usually with the benefit of decades of historical data. However, the lack of a long-term dataset and the distinct nature of infrastructure sub-sectors limits the robustness of this approach.

By developing InFRAME, our objective was to enable a greater bottom-up perspective on portfolio risk exposures and characteristics. This recognises that the risk profile of an infrastructure asset is largely related to its revenue characteristics, by virtue of the normally relatively high profit margins which characterise infrastructure.

InFRAME can be split into three phases which are interlinked and supported by a range of analytical tools – Risk Profiling, Scenario Analysis and Portfolio Optimisation.

Risk Profiling

Risk Profiling provides a bottom-up assessment of risk at the asset and portfolio levels.

The approach recognises that targeting allocations by sub-sector does not fully reflect the range of asset types that can be found within a particular sub-sector, or the complexity of multiple sources of income that can co-exist within a single asset.

For each asset, we identify constituent revenue streams and classify these based on a defined set of representative revenue types. The seven revenue types, illustrated in Figure 1, are further grouped into four major categories which are based upon relative price versus volume risk.


Figure 1. InFRAME revenue types

For example, an airport can typically derive its income from up to half a dozen distinct revenue streams ranging across landing charges, property leases and ground transport, each with its own response to risks and macro-economic factors.


Figure 2. Characterisation of typical airport revenue streams

The risk profiling concept entails an assessment of the impact of underlying risk drivers affecting each distinct revenue stream (or more accurately, cash flow stream) of each existing and prospective portfolio asset. IFM utilises 20 key portfolio-wide risks and value drivers for our risk profiling analysis.

The analysis of revenue types and aggregation of risks across our portfolios allows us to identify common risk factors and risk concentration, with the ability to do ‘what-if’ scenarios for potential portfolio acquisitions/divestments.

Defining core infrastructure

One of the benefits of portfolio risk profiling is that it allows us to quantify what a ‘core’ strategy looks like. IFM targets investments in core infrastructure assets which typically possess the following five key qualities:

  • monopolistic characteristics
  • long life
  • stable and predictable revenues
  • inflation protection
  • exposure to economic growth.

Based on an aggregation of the risk profiles of portfolio asset revenue streams, we calculate a Core Infrastructure Rating according to the five key qualities which contribute to making a core infrastructure portfolio, as per Figure 3.

Figure 3. Core Infrastructure Rating for the current global portfolio and after a proposed portfolio addition.

We target maximum exposure to the first two of the five factors, but seek a balance between cash flow stability and exposure to inflation and economic growth, recognising there is a natural trade off between these last three factors.

Scenario Analysis and Optimisation

A key objective of the portfolio construction initiative was to develop an analytical tool which would guide strategic asset allocation within an infrastructure portfolio. Following significant research, we decided to use a holistic scenario-based approach to risk rather than deal with single point sensitivities. The underlying principle was that the portfolio should be sufficiently robust to provide an acceptable return under a range of economic scenarios.

This approach involves:

  • Scenario development: Macro-economic scenarios relevant to infrastructure portfolios are developed, applying a five-year time frame. The five-year time frame is a medium-term horizon over which a reasonable range of economic outcomes can be conceived.
  • Infrastructure response: We look at how portfolio assets (or representative prototypes as a proxy) respond to each of the scenarios, in terms of five year equity return.
  • Portfolio optimisation: We determine the possible range of portfolio weightings that ensure reasonable minimum portfolio return thresholds are met over the range of scenarios considered.

We have developed a number of macro-economic scenarios relevant to infrastructure portfolios. For each scenario, economic variables are varied in each of the asset models – directly for variables such as inflation and interest rates, and indirectly for variables such as GDP, where “rule of thumb” relationships are used to derive passenger growth at an airport for example.

This enables us to analyse how representative assets respond to macro-economic scenarios, both upsides and downsides to the expected base case. By having scenarios which affect sub-sectors and geographies in different ways, the approach implicitly captures correlations between asset responses.

At the last stage of the process, our main focus is to identify what asset allocations based upon revenue characteristics produce optimal and robust equity returns. The portfolio optimisation model uses Monte Carlo simulation to generate a large number of randomly weighted portfolios. For each set of weights and for each scenario, we generate a portfolio return. If the portfolio return is above a defined threshold, the portfolio is considered to be admissible; otherwise it is rejected.

Figure 4 illustrates the target portfolio allocation for the IFM Global Infrastructure Fund based upon the four key revenue categories. Again, “what-if” analysis can be carried out for revenue type analysis. Various self-consistency checks have been made between the revenue framework and the core infrastructure rating target, providing comfort that the framework is robust.

Figure 4. Target portfolio allocations based upon revenue characteristics

Concluding remarks

The InFRAME analysis provides IFM with a number of proprietary tools which assist us in our strategic approach and decision making with respect to building out balanced infrastructure portfolios. The approach has become an integral part of our investment process and the key elements of the InFRAME analysis are included in every investment committee submission.

The analysis is of course reliant on some simplifying assumptions. For example: the sensitivity analysis is based on a limited number of scenarios and does not account for volatility of asset valuations caused by factors such as market sentiment, which can affect unlisted infrastructure as it does listed stocks (though based on historical experience, to a lesser extent).

‘Overweight/underweight’ type conclusions need to be considered in the context of markets that are illiquid, subject to market cycles, and are somewhat opportunistic. As a result, relatively long timelines may be required to achieve desired portfolio outcomes. The conclusions from the InFRAME analysis assist in guiding our deal origination strategy, which focuses on strategic relationships across asset owners, operators and service providers, to stimulate deal flow.

We note however, that while portfolio construction tools provide an analytical framework to inform decision makers, the methodology is only one part of the process and does not replace the decision making process itself.

Finally, although the concepts and tools developed as part of this work programme were tailored to IFM’s unlisted infrastructure portfolios, elements of this work may be applicable to end-investor’s own portfolios both in infrastructure and other asset classes. We are happy to engage with investors to assist them with their own portfolio construction considerations.