Insights

    Infrastructure portfolio allocation: What could optimal look like?

    16 April 2025
    By Aizhan Meldebek 

    Executive summary

    • Adding private infrastructure equity to a portfolio can help extend the efficient frontier and unlock new optimal portfolios for investors.
    • Applying minimum-volatility and mean-variance optimisations to a standard portfolio with a 30% constraint on private markets suggests a theoretical optimal allocation to infrastructure in the range of 7.9% to 9.5%,1 above the current average institutional allocation of 4.3%.
    • For long-term investors who are not constrained by a 30% limit on private markets and have lower liquidity needs, the evidence suggests that private infrastructure could have an even higher allocation.
    • Despite the recent weakness in fundraising, we expect it to pick up during 2025 and remain strong over the coming years on the back of rising institutional allocations to infrastructure, democratisation of the asset class, and ongoing increases in wealth levels.

    Infrastructure and efficient frontier: Unlocking new optimal portfolios

    The efficient frontier represents a set of optimal investment portfolios offering the highest expected return for a defined level of risk. In this paper, we show how adding private infrastructure equity2 to an investment portfolio can help expand the efficient frontier due to infrastructure’s attractive risk-return characteristics.

    It can often be challenging to find the right allocation to private asset classes as they typically exhibit low volatility due to serial correlation that can be present in fund-level indices. To tackle this issue, we unsmooth the private infrastructure index.3 Unsmoothing helps improve the measurement of risk exposure and risk-adjusted performance. Figure 1 shows that private infrastructure on an unsmoothed basis lies outside a traditional bond-stock efficient frontier, meaning that an allocation to private infrastructure can help achieve a set of optimal portfolios that were previously inaccessible to investors.

    Figure 1:
    Private infrastructure has potential to expand the bond-stock efficient frontier

    Source: Cambridge Associates, Bloomberg (March 2025). US equities: S&P 500 Index; Unlisted infrastructure: Cambridge Associates Infrastructure Index (unsmoothed); Global equities: MSCI World Index; Global corporates IG: Bloomberg Global Aggregate Corporate Index; Global corporates HY: Bloomberg Global High Yield Corporate Index; Global Treasuries: Bloomberg Global Aggregate Treasuries Index. Analysis conducted from 1Q 2005 to 2Q 2024. Efficient frontier assumes a set of optimal portfolios under mean-variance optimisation model with full investment at any time and no short selling allowed. Past performance is not indicative of future returns.

    Infrastructure’s role in portfolio under different investment objectives

    There is no one-size-fits-all optimal allocation. Determining the optimal allocation to infrastructure in an investment portfolio depends on various factors, including the investor's risk tolerance, investment goals, and time horizon, and the overall composition of their existing portfolio. However, we would like to show two theoretical examples of how private infrastructure can complement portfolios with different investment objectives.

    First, we start with a standard optimised traditional portfolio consisting of 60% listed equities and 40% listed bonds. We then incorporate private infrastructure along with other alternatives, including private equity, private credit and private real estate (core), under different objectives using historical data and subject to a maximum 30% allocation to illiquid asset classes (Figure 2):

    • Portfolio A objective: Minimise volatility. This optimisation technique aims to achieve the lowest possible portfolio volatility over the analysed period. This results in volatility of Portfolio A being 186 basis points lower than that of a 60/40 portfolio. The optimisation suggests that the optimal allocation to private infrastructure is 7.9% of the total portfolio.4
    • Portfolio B objective: Maximise risk-adjusted returns. This is a portfolio optimisation using a Markowitz (1952) approach that aims to maximise the Sharpe ratio. The optimisation results in a Sharpe ratio of 0.70 for Portfolio B, significantly higher than the 0.57 of a traditional 60/40 portfolio. The optimisation suggests that the optimal allocation to private infrastructure is 9.5% of the total portfolio.5

    This analysis shows an optimal allocation to private infrastructure could be in the range of 7.9% and 9.5%, which is higher than both the current average allocation (4.3%) and the current average target allocation (5.5%) of institutional investors (more detail on this in the sections that follow). For long-term investors that are not constrained by a maximum 30% allocation to illiquid asset classes, private infrastructure could take a higher allocation of a total portfolio.

    Figure 2:
    Illustrative optimisation of portfolios with private infrastructure under different investment objectives based on historical data

    Sources: Cambridge Associates, Cliffwater, INREV, Bloomberg (March 2025). US equities: S&P 500 Index; Private infrastructure: Cambridge Associates Infrastructure Index (unsmoothed), returns are net of fees, expenses and carried interest; Global equities: MSCI World Index; Global corporates IG: Bloomberg Global Aggregate Corporate Index; Global corporates HY: Bloomberg Global High Yield Corporate Index; Private credit: Cliffwater Direct Lending Index (unsmoothed), adjusted to account for management and incentive fees; Global Treasuries: Bloomberg Global Aggregate Treasuries Index; Private equity: Cambridge Associates US Private Equity Index (unsmoothed), returns are net of fees, expenses and carried interest; Private core real estate: INREV Global Real Estate Fund Index (GREFI) refers to core property performance gained via fund structure with low levels of leverage, and excludes land, developments, and alternative property sectors. Analysis conducted from 1Q 2005 to 1Q 2024. Subject to full investment, no short-selling, liquidity (max 30% in illiquid asset classes) and concentration constraints. Does not constitute investment advice or recommendation. Based on historical performance and therefore are not suitable for forward-looking interpretation. Past performance is not indicative of future returns. For illustrative purposes only.

    *Compared with a traditional 60/40 portfolio.

    Why private infrastructure meets various investment objectives

    Since 2005, unlisted infrastructure has delivered a robust 9.3% annualised return, performing better than global equities (8.0%), global corporate bonds (6.2% for high yield and 3.0% for investment grade), private credit (6.3%), private core real estate (4.9%) and global Treasuries (1.3%) over the same period (Figure 3).6 While infrastructure returns have been below US private equity (12.9%) and US listed equity (9.8%), infrastructure has been far less volatile on an unsmoothed basis (10.5% versus 14.9% for US private equity and 16.2% for US listed equity, respectively).7 Figure 4 shows that historical volatility of infrastructure returns8 is more comparable to corporate bonds than to listed equities.

    In other words, private global infrastructure historically has delivered returns comparable to equity markets at a level of volatility comparable to bond markets courtesy of its investment characteristics:

    • Bond-like characteristics: Infrastructure assets typically generate a stable yield as a component of total return, provide downside protection during economic downturns due to the essential nature of services, and exhibit lower volatility compared with general listed equities.
    • Equity-like characteristics: At the same time, private infrastructure has a relatively high exposure to secular trends such as digitalisation, decarbonisation and demographics, which results in significant growth opportunities and provides capital appreciation potential.

    As a result, a well-structured infrastructure portfolio can offer both defensive characteristics through stable income returns as well as growth potential through capital appreciation.

    Figure 3:
    Long-run returns by asset class

    Figure 4:
    Volatility by asset class

    Sources: Cambridge Associates, Cliffwater, INREV, Bloomberg (March 2025). US private equity: Cambridge Associates US Private Equity Index (unsmoothed), returns are net of fees, expenses and carried interest; US listed equity: S&P 500 Index; Private infrastructure: Cambridge Associates Infrastructure Index (unsmoothed), returns are net of fees, expenses and carried interest; Global equities: MSCI World Index; Global corporates IG: Bloomberg Global Aggregate Corporate Index; Global corporates HY: Bloomberg Global High Yield Corporate Index; Global Treasuries: Bloomberg Global Aggregate Treasuries Index; Private credit: Cliffwater Direct Lending Index (unsmoothed), adjusted to account for fees, expenses and carried interest; Private core real estate: INREV Global Real Estate Fund Index (GREFI) refers to core property performance gained via fund structure with low levels of leverage, and excludes land, developments, and alternative property sectors. Analysis conducted from 1Q 2005 to 2Q 2024, except for private credit, which is to 1Q 2024.

    Institutional investors and current allocations to infrastructure

    In 2024, institutional investors have reported an average allocation to private infrastructure of 4.3% of a total portfolio.9 However, about 60% of institutional investors continue to remain underallocated to infrastructure, with average allocations being 123 basis points below the target of 5.5% of a total portfolio (Figure 5).

    From a regional perspective, Canadian and Australian institutions report the highest allocations to infrastructure at 10.1% and 6.7% respectively, while US investors have the lowest allocation to infrastructure at 3.0% (Figure 6).10 Within Europe, Italian investors allocate only 2.9% to private infrastructure equity on average versus their reported target of 3.3%.11 In Germany, institutional investors tend to allocate 4.3% of their portfolio compared with their target of 6.0%.12

    Figure 5:
    Allocations to infrastructure by institution type

    Figure 6:
    Allocations to infrastructure by investor location

    Sources: Hodes Weill & Associates, Cornell Brooks Public Policy, “Institutional Infrastructure Allocations Monitor 2024”, Preqin database (January 2025). Regional allocations are based on median allocations reported by Preqin.

    Australia’s super funds: Longest-standing investors in infrastructure

    Over the past decade, Australia has recorded the highest increase in its ratio of pension assets to gross domestic product (GDP) globally.13 While the growth in pension assets in Australia has been driven by favourable demographic trends, according to a study of long-term superannuation funds, asset allocation has been one of the key drivers of risk and performance.14

    Australian pension funds are among the longest-standing investors in infrastructure, as they have been pioneers in the space since the early 1990s. As of 2024, Australian pension funds allocate 6.8% to unlisted infrastructure (Figure 7), an increase from 2.8% in 2013 (Figure 8).15 Additionally, several pension funds have indicated an intention to increase allocations further.16 By comparison, the UK Defined Contribution (DC) schemes are estimated to have a 3% allocation to unlisted infrastructure.17

    Figure 7:
    Australian super funds reported a 6.8% allocation to private infrastructure equity

    Figure 8:
    Australian super funds have been increasing private infrastructure allocations since 2013

    Source: Australian Prudential Regulation Authority (APRA) (September 2024).

    Canadian pension funds favour private markets and infrastructure

    In Canada, the average allocation to private markets across the largest pension funds exceeds 50% (Figure 9), while infrastructure investments comprise about 11% of an average portfolio (Figure 10).18 This represents one of the largest allocations globally. One of the reasons behind large allocations to infrastructure is the capabilities to manage assets internally. The ‘Canadian pension fund model’, also known as Maple-8, is well known for using direct investments and internal asset management.

    Infrastructure is an asset class that requires deep specialist expertise and active asset management in order to deliver alpha and required returns. Going forward we believe this will only intensify. In the new world of higher interest rates, delivering on operational performance will become an even more important return driver, and the energy transition will require expertise and skills to execute on the emerging opportunities.19 Therefore, large allocations to infrastructure will necessitate either strong internal asset management teams or working with specialist infrastructure managers.

    Figure 9:
    Canadian pension funds tend to favour private markets

    Figure 10:
    Allocations to infrastructure by largest Canadian pension funds

    Source: Hymans Robertson, Policy Briefing Note: “The Canadian model” (2024). Data as of 2022.

    Implications for infrastructure fundraising

    Despite the recent weakness in fundraising, we expect it to pick up during 2025 and remain strong over the coming years on the back of rising institutional allocations to infrastructure.

    As mentioned earlier, current institutional allocations stand at 4.3% of total portfolio with an underallocation of 123 basis points. Infrastructure assets under management (AuM) stood at $US1.4 trillion as of June 2024. This implies that if investors reach their reported target allocation of 5.5% over the coming quarters, then the infrastructure AuM should grow by another $US387 billion. Under the assumption that an average investor increases their allocation to a theoretical optimal range between 7.9% and 9.5% in the future, then the infrastructure AuM could reach somewhere between $US2.6 trillion and $US3.1 trillion (Figure 11).

    Figure 11:
    Private infrastructure AuM under current 4.3%, target 5.5% and potential at 7.9% and 9.5% institutional allocation assumptions

    Sources: Macquarie Asset Management, Hodes Weill & Associates, Preqin (March 2025). Calculations based on AuM data from Preqin as of March 2024 and institutional allocations in 2024 as reported by Hodes Weill & Associates. AuM excludes infrastructure debt. Based on the following: (1) assuming that 2024 infrastructure AuM corresponds to 4.3% allocation to infrastructure, and (2) conservatively assuming zero growth of total global AuM in the future.

    Authors

    Aizhan Meldebek Global Infrastructure Strategist – Client Solutions Group