Understanding infrastructure risks and returns

The recent decision by the Norwegian government not to allow unlisted infrastructure investment needs contexualising. We look at a wide range of investor stances towards the sector, and at the possible effects of insufficient and inadequate fund performance data. This can be improved, and potentially persuade more investors into the asset class, if more rigorous performance data approaches seen in other classes, such as private equity, are adopted.

Currently, infrastructure investment is seemingly in excellent health, with strong fundraising opportunities, good deal flow, and a number of new entrants into the asset class. There are more than 1,100 sovereign wealth funds, pension funds, and insurance companies who are keen investors in the asset class, with assets under management of over $35 trillion between them. On average, they currently have a 3.2 percent allocation of these funds to infrastructure, with many others targeting investment in the future, bringing the average target allocation to over 5 percent.

While the data indicates that the sector is appealing for investors, there are also signs of problems ahead, and there is a question mark over why almost 100 investors have stopped investing in infrastructure equity, despite having previously invested close to 1 percent of their AUM in this area. These investors include MetLife, which has now decided, despite having invested over $400 million in 16 infrastructure funds between 1994 and 2013(1), to focus solely on providing debt, reverting to their traditional activities in private placements.

Some high-profile investors have chosen to eschew infrastructure investment altogether, including Japan Post, Hong Kong Monetary Authority and US State Farm Insurance. Each of these has AUM in excess of $200 billion, which suggests that available funds is not the reason for their lack of investment in the sector.


The most recent investor to reject infrastructure was Norway’s Government Pension Fund (made up of two sovereign wealth funds, Government Pension Fund Global and Government Pension Fund Norway) in April this year. A white paper issued by the government discussed the pros and cons of unlisted infrastructure investment (2), marking the culmination of two years of evaluation involving the funds’ managers and outside experts.

The Norwegians’ conclusion – that infrastructure was not viable for the $900 billion portfolio – is worth examining in detail. The government identified two potential benefits from unlisted infrastructure: firstly, potentially higher risk-adjusted returns, due to differences between infrastructure’s profile compared to other asset classes; and secondly, the possibility of higher excess returns, due to factors such as longer time horizons, investment size and management expertise. The reasons given for believing that the GPFG should not attempt to secure these benefits is illuminating: “A lack of data on unlisted infrastructure makes it difficult to assess whether such investments improve risk diversification or raise expected returns for the average investor.”(3)

The government accepted infrastructure’s theoretical benefits, but it could not be sure they exist in practice, due to the absence of sufficient information on infrastructure returns and their risk profile compared to other asset classes, particularly listed equities and bonds. This is a serious indictment for an asset class that is now over 20 years old and which has raised over $375 billion through more than 600 funds. Given that the same report highlighted GPFG’s poor equity and bond performance in 2015, the rejection of an asset class that could have boosted portfolio return is all the more striking.

These thorny issues have not yet been fully addressed by either the fund industry or academic research into the infrastructure sector. In part, this knowledge gap has arisen from a basic lack of information on returns, given that unlisted fund cashflows and periodic net asset values are not widely available, nor is financial performance on the underlying investments made by funds. Likewise, there is little information on the growing number of direct investments. Some researchers have attempted to offer an analysis of the asset class by looking at the returns of listed infrastructure companies, such as US utilities and privatised European companies. However, there is little evidence to suggest that this is a good comparative class to use, in terms of their cashflow patterns, risk profiles and the mix of sub-sectors they invest in. Without these basic performance data, it becomes impossible to create effective performance benchmarks, difficult to gauge the risk of a portfolio and, ultimately, problematic to take steps to diversify that risk.


Even when some of this information is available it is not in the form needed by sophisticated chief investment officers to determine the effect of adding infrastructure to their funds’ risk-adjusted returns. In order to do this, we need different types of data, as well as more of them: in particular, IRR is a crude measure, which cannot be used to compare assets with different cashflow patterns or maturities, and which fails to adjust for risk or performance relative to the wider market. This weakness has long been recognised in the larger and more established private equity market,(4) and alternative performance measures have been developed, including tweaks to traditional IRRs and money multiples to take into account performance of the overall equity market, or specific benchmarks, such as Modified IRR (MIRR) and Public Market Equivalent (PME). MIRR and PME are now commonly used for assessing private equity, yet are not available for infrastructure.

Although the Norwegian decision came as a surprise to many, there had been warning signs. The GPFG fund’s manager Norges Bank (NBIM) advocated the expansion into unlisted infrastructure, but in 2013, NBIM itself published research that was at best equivocal stating:

“The performance history for infrastructure investments is limited and performance data are to a large extent private. The high degree of heterogeneity makes comparisons across projects, structures and jurisdictions challenging. Scholarly studies on infrastructure investments are few and the approaches taken to deal with the shortcomings in available datasets vary widely. It is therefore challenging to draw general and firm conclusions based on these studies.”(5)

Of course, data issues are not the only reasons investors choose not to invest in infrastructure: for example, the Norwegian government report also highlighted political risk. While emerging economies are often regarded as the riskiest places to invest, even mature markets can present problems for the investor, as the UK government’s wavering policy on wind energy proves. Likewise, liquidity concerns were discussed in the white paper and most infrastructure projects, with their lengthy lifespans, must be regarded as long-term holdings. Underlying sector risk remains another issue. By way of illustration, changing technology, energy policy, consumer demand, and so on, will almost certainly reflect on the returns commanded by investments in energy infrastructure, exemplified by recent negative German power prices due to the volume of renewables production.

The need for better data has been recognised by the infrastructure industry. For example through the auspices of the Long Term Infrastructure Investors Association, founded in 2014, investors are funding independent academic research to create a repository of infrastructure performance data.(6) We argue that even more is needed, with a need for greater pooling of data by investors, increased performance transparency by fund managers and the use of more sophisticated return metrics. This would allow better comparisons to be made with other asset classes. Furthermore, unlike some other investor concerns, data availability is an issue that the infrastructure industry can – and should – address.

A number of Norway’s reasons for rejecting infrastructure would seem common to other alternative investment categories and, indeed, the smaller of the two Norwegian sovereign wealth funds is no longer permitted to invest in unlisted real estate either. However, other asset classes, such as real estate and private equity, offer greater transparency and better performance tools. Perhaps it should come as no surprise that, whilst unlisted infrastructure has grown rapidly in the last decade, it continues to lag far behind the popularity of either real estate or private equity, as recently highlighted by the World Economic Forum (WEF).(7)

That said, the WEF also identifies many long-term trends favouring infrastructure, with a particular focus on the growing need for retirement investing in a lower rate environment. They also note that alternatives have increased as a proportion of major institutions’ portfolios from 5 percent in 1995 to 25 percent today, albeit with rapidly changing fund management dynamics, characterised by increasing competition and innovation.

It is our view that the infrastructure industry needs to continue to adapt to this challenge, and there is an opportunity to more clearly define what infrastructure investment actually delivers. In our ongoing research, we examine infrastructure fund performance using approaches more common in private equity, to demonstrate risk-adjusted performance compared with other asset classes. We also assess direct investment versus fund returns, and use risk-management approaches developed originally for commodities and derivatives markets, allowing the downside risk profile of infrastructure to be mapped with greater precision. These will be looked at in future editions and may help address some of the issues raised by the Norwegian white paper.

Despite Norway’s negative decision, more than 100 other institutions are still considering starting to invest in infrastructure (in the short or longer term). These investors’ aggregate AUM comfortably exceeds those of institutions that have pulled back from infrastructure, potentially boding well for the sector. However, the data concerns raised by the likes of the Norwegian government need to be addressed to ensure the continued long-term health of the sector. ?

1. Preqin Infrastructure Online
2. Norwegian Government (2016), Report No. 23: The Management of the Government Pension Fund in 2015.
3. Norwegian Government (2016), Report No. 23: The Management of the Government Pension Fund in 2015, English Translation, page 6
4. See for example, Ludovic Phalippou (2008), ‘The Hazards of Using IRR to Measure Performance: The Case of Private Equity’, Journal of Performance Management, 12(4), pp. 55-67.
5. NBIM (2013), Infrastructure Investments, Discussion Note 2-2013.
6. LTIIA (2016), ‘LTIIA endows new research chair with EDHEC’, Press Release, February 24, 2016.
7. World Economic Forum (2015), Alternative Investments 2020: The Future of Alternative Investments.