Is there a ‘green’ risk factor in infrastructure investment?

• In recent research, we examined the impact on realised performance of the permanent shift in investor preferences for low carbon energy investments, and how it relates to the expected returns of green power investments.
• While green infrastructure has outperformed the ‘Core’ infrastructure market over the past decade, this is largely the result of excess demand for such assets that has pushed asset prices up and discount rates down.
• After controlling for a number of risk factors present in the returns of unlisted infrastructure equity investment, there is no persistent ‘green’ risk factor, but instead a ‘green price premium’ that investors have been willing to pay to increase their holdings of such assets.
• We showed the impact of excess demand for green power investments on yield compression by building a measure of the liquidity of the market for green power investment. When too few green infrastructure investments are available in the market, asset prices increase and yields compress. Controlling for this effect, any outperformance of the green power sector over the considered period disappears.
• This phenomenon peaked in 2019 and the expected returns of green power investments are now much lower than they used to be. As a result, realised returns should not be used directly as a proxy of the future performance of green power investments.

It is often argued that more sustainable investments should coincide with better financial performance. This raises two distinct questions:

Firstly, is there any empirical evidence of superior performance by more sustainable or greener investments? And, if so, what might explain such outperformance, and can it be expected to persist in the future?
Alternatively, is any superior performance the result of an identifiable transition in investor preferences resulting in a positive shift in asset prices (higher realised returns) but not in higher expected returns?

In recent research (see Amenc and Blanc-Brude, 2022), we show that there is indeed empirical evidence of historical outperformance of green infrastructure investments (defined narrowly as wind and solar power projects). We then consider whether this finding implies continued future outperformance. In line with the literature, we argue that more sustainable infrastructure investments should in fine have lower expected returns than less sustainable ones, but that the recent shift in investor preferences in favour of greener power investments temporarily created excess demand, explaining realised performance over the past decade.

The existence of a systematic difference in pricing and expected returns between sustainable and less sustainable investments is examined in recent academic research (see Pastor, Stambaugh and Taylor, 2022; Alessi, Ossola and Panzica, 2021). Pastor, Stambaugh and Taylor summarise the reason why greener investments should have low expected returns: either investors bid up asset prices because they have increasing preferences for them, or the customers of greener businesses shift their demand towards their services, increasing their revenues and profits, and consequently their market value. As asset prices rise in response to greater demand, their cost of capital falls. In other words, the premise that greener companies and services – and the positive externalities they create – are increasingly valuable to investors and desirable to consumers (and the reverse for less green companies) implies that the market price of their equity must be higher, their cost of capital lower and their expected return (which, in equilibrium, must equal their cost of capital) also lower. As long as we accept the hypothesis of weakly efficient financial markets, in equilibrium risk must be adequately priced, which leaves little hope for the continued strong performance of green infrastructure investments in the near-to-long term.

Of course, in this context, it is still possible for greener investment to outperform during a period of persistent changes in investor preferences; for example, excess demand can drive up asset prices because investors expect preferences for green assets to have durably shifted from their previous level. As market prices increase and capital gains accrue to investors, these investments outperform but also exhibit increasingly lower expected returns.

As Pastor, Stambaugh and Taylor (2021) and others point out, the inverse relationship between price and expected return or yield is at its simplest in the case of bonds. For a buy-and-hold investor, the yield of a bond is the best estimate of its expected return, as bond prices change, its yields and expected returns change inversely. This is because bonds have no exposure to the upside i.e., the growth of the borrowers’ business. The same mechanism applies to the price and yield one the most clear-cut types of sustainable investments: green power infrastructure.

Green infrastructure can take several forms but, at its greenest, it can be narrowly defined as wind and solar power projects: new investments producing electricity (largely) without emitting greenhouse gases and potentially displacing existing power sources that do. In other words, with constant energy needs, wind and solar power projects are carbon-negative investments. This category of investments thus provides a convincing case of what the greenest types of green infrastructure investments might look like.

The way such projects are created and financed is what makes them resemble a bond. Solar and wind farms are typically incorporated as a standalone special-purpose company with a finite life based on the economic life of the physical asset and on its business model, typically a long-term power purchase agreement (PPA) or a regulated electricity market. Such projects raise asset-backed finance once, sink capital into a finite physical asset, and its investors are repaid over a period of 25 to 30 years. Like bonds, such a company has very limited upside or growth options. Wind farms can be repowered and PPAs extended, but infrastructure assets are capacity-constrained by design. Infrastructure companies thus have a maximum potential revenue defined mostly by ex-ante choices of size and technology. Hence, like many other project-based infrastructure investments, wind and solar project equity investments are akin to a bond with risky coupons.

It follows that if increasing demand for green infrastructure leads to better performance through capital gains, it must be because their yield or costs of capital is falling. Once excess demand has been absorbed by the market, the long-term performance of greener infrastructure should be lower than that of less green infrastructure investments.

In our research, we consider the question of what drives the past and future financial performance of green infrastructure in several steps. We first review the historical performance of investments in unlisted wind and solar project equity using the infraGreen index[1]. We show that green infrastructure investments have indeed outperformed the market, including Core infrastructure, which is a natural benchmark for such projects. Until 2019, they also outperformed Core+ infrastructure, a riskier subset of unlisted infrastructure investments. In effect, over the past 10 years, green infrastructure has exhibited a very attractive risk-adjusted return profile, with higher annualised returns than Core infrastructure and lower volatility than Core+ infrastructure.

We then follow the literature and examine the difference of performance between two portfolios created using asset-level data available in the EDHECinfra database: a green power portfolio of unlisted equity investments in wind and solar projects only, and a brown power portfolio of unlisted equity investments in coal and gas power projects only. As argued above, we consider all the investments in the first portfolio to be equally (and highly) green. Likewise, coal and gas power projects are unequivocally brown[2]: coal and gas power projects are net contributors to greenhouse gas emissions. Conventional power generation emitted 13.5GtCO2-eq in 2020, i.e., it is the leading contributor to total energy-related emissions (31GtCO2-eq, IEA (2021)) ahead of the transportation and industry sectors. Even though the greenhouse gas emissions of coal and gas power projects vary and can, to some extent, be reduced or captured, even with constant energy demand, these investments are always carbon positive. In other words, our green power portfolio is always greener than our brown power portfolio.

Over a period extending from 2011 to 2021, the brown power portfolio outperformed green power by a cumulative 138bp. However, during that period, green power outperformed or matched the performance of brown power between 2012 and 2015 and also between 2018 and 2020. We show that these are also the two periods during which the cost of capital spread between green and brown power widened significantly as the market value of green power assets increased.

Next, we examine the differential performance of green and brown power investments through a “green minus brown” (GMB) portfolio of their returns over the past decade. Controlling for the effect of well-documented risk factors like size, leverage and profits, this portfolio produces a statistically significant negative “alpha”. The realised green or brown power excess returns are also better explained by adding a GMB ‘effect’ to the usual set of risk factors. Prima facie, this result could be interpreted as the presence of a ‘green’ risk factor in the returns of green and brown power infrastructure investments.

To determine the potential persistence of this effect, we examine the expected returns of green and brown power using data from infraMetrics and show that there is a significant and increasing spread between the weighted average cost of capital of the two portfolios. The weighted average cost of capital (or WACC) spread or green price premium between the green and brown power portfolios is consistently negative and growing: in 2021, it had widened to almost -350bp from about -100bp a decade earlier.

High realised performance has been accompanied by a significant decrease in the cost of capital of green power infrastructure. In effect, all infrastructure investments have become more popular amongst investors in the past decade and have seen a reduction in their cost of capital, including brown power. However, the green power has seen a much larger decrease. Between December 2011 and December 2021, the infrastructure market saw a global reduction in WACC of 177bp (from 7.23% to 5.45%), while green power saw a greater reduction of 263bp, but the WACC of brown power is only 11bp lower in 2021 than it was in 2011.

We show that the evolution of the cost of capital spread of the two legs of the GMB portfolio explains away its negative alpha. In other words, taking yield compression into account, standard pricing factors suffice to explain the realised performance of the GMB portfolio.

We argue that the yield compression observed since 2011 is at least in part due to excess demand in the market for green power infrastructure i.e., demand that cannot be met immediately by a supply of green power investments. To show this effect, we construct a measure of excess demand for green power investments using the share of secondary transactions in all investments made by infrastructure investors in green energy. We argue that periods during which secondary transactions represent a smaller fraction of the overall market transaction volume are periods of lower liquidity – during which excess demand for green power assets is likely to have been higher. We show that this measure of the green power market liquidity is strongly related to the performance and WACC spread of the GMB portfolio, as well as the realised performance of the green power portfolio. In other words, when the market for renewable power projects is less liquid and excess demand is more likely to build up, we tend to see an increase in the performance of the GMB portfolio and in the WACC spread between green and brown assets.

We conclude that, while green power assets have experienced a period of strong performance (realised returns), they are likely to deliver lower returns going forward since this performance was largely driven by the compression of their cost of capital, itself largely related to the build-up of excess demand in the market for green assets. Moreover, while the green price premium has increased in line with excess demand, the supply of green power investments has also increased considerably and the GMB WACC spread has been flat since 2019. As green infrastructure plays an increasingly important and ubiquitous role in investors’ portfolios, a consensus on the price and expected returns of green power is increasingly likely and new shifts in demand for such assets less so. In effect, green power may be one the few asset classes in which green pricing has already peaked (around mid-2019).

These results are important in understanding the role that renewables and conventional energy are likely to play in investors’ portfolios going forward, since increasing allocations to green energy should not be based on returns assumptions derived from historical returns. Indeed, as the supply of renewable investments has increased and, in some markets, become one of the dominant sources of energy, investor preferences for such assets should stabilise and excess demand disappear. A recent peer-group survey of asset allocations within the infrastructure asset class found that renewable energy already represents one quarter to one third of most investors’ infrastructure portfolios (Blanc-Brude et al., 2022). While investment in green infrastructure is likely to keep increasing on aggregate, its weight in infrastructure portfolios is unlikely to keep increasing monotonically.

Durably lower expected returns and cost of capital for green power is of course a good thing, since it reduces the overall cost of the energy transition. However, investors should not expect to receive high returns while contributing to the energy transition (have a positive impact) as long as they are only exposed to a pure, unleveraged basket of green power investments.

Conclusions

The premise that green investments may have different returns than brown ones partly springs from the notion of climate ‘transition risk’: the expectation of higher future costs or lower future revenues for firms that emit greenhouse gases due to new regulations and shifts in consumer behaviour. However, the manner, timing and magnitude with which transition risks may materialise have been and remain largely unknown to investors. Today, it can seem unlikely that asset prices already fully reflect these risks when they remain very hard to assess and quantify.

When it comes to renewable energy projects and their fossil-fuel (coal and gas) equivalents, however, the writing is already on the wall: wind and solar projects will be impervious to carbon taxes and coal and gas will not. In effect, coal projects are already being divested and phased out by large utilities, implying that their future value is considered to trend towards zero. This knowledge has already impacted asset prices in the case of green and brown power investments. The gradual realisation by investors that they have an increasing preference for green power investment and want to hold less conventional power investment has taken place over the past decade. In our 2022 survey of c.350 large investor portfolios of infrastructure assets, EDHECinfra found not only that renewable energy corresponds to between one quarter and one third of investors’ infrastructure holdings by value at the end of 2021, but also that conventional gas and coal power projects represent as little as 1 to 3% of their portfolio, with the notable exception of North American investors, who hold 10% of their infrastructure investments in brown power assets. In other words, brown power investments have largely been divested by mainstream investors already and green ones have already been integrated into portfolios on a significant scale. The shift in demand for green and brown power assets has already occurred.

One might add that higher demand for green power is not the only possible reason for the yield compression observed. For instance, infrastructure investment has been characterised by a significant evolution in the nature of investors valuing such assets, with the principal market increasing in size and scope and new cohorts of buyers and sellers showing increasing comfort with long-term, illiquid investments i.e., different risk preferences to previous generations of investors in infrastructure equity, who faced higher hurdle rates e.g., construction firms.

In 2011, green power projects had expected returns of ~8% and brown power projects ~9%. Their 10-year annualised total returns in 2021 were 16% and 17% respectively. These two figures may seem related but correspond in fact to very different economic fundamentals. The high historical performance of green power is explained by a significant compression in yields (expected returns) especially between 2012 and 2015 and the corresponding capital gains. Conversely, the performance of brown power was more driven by cash returns and less by yield compression. In effect, unlike other infrastructure investments, brown power investments have seen a slight increase in their expected returns since 2018.

Hence, we find that the impact on performance of such shifts in the demand for green and brown investments cannot be equated with the appearance of a new ‘green’ asset pricing risk factor. Instead, as predicted by theory (see Pastor, Stambaugh and Taylor, 2021), demand shocks have led to relatively high realised performance in the green power market but also lower expected returns.

For this situation to persist, there needs to be continued disagreement in the market about the future value of greener investments. Once all investors agree about the future value of greener or less green investments, investors are left holding the market portfolio, which includes current and future preferences for greener assets.

Going forward, as excess demand for green power investments is gradually met with additional supply of green power assets and effective allocations to green power become significant, our findings suggest that both the realised and expected returns of green power investments can be expected to converge.

Such a convergence, which reflects a long-term pricing equilibrium, leads us to conclude that there is no reason for superior performance by green infrastructure investments to continue. The so-called “green premium” observed in the past does not correspond to the reward for a superior risk factor but instead to a temporary phenomenon of excess demand, which the supply side of the market eventually satisfied.

Footnotes

[1] The infraGreen index is available on EDHECinfra’s infraMetrics platform..

[2] Irrespective of the debate on the inclusion of natural gas generation in the EU taxonomy (see Blanc-Brude et al. 2021).

References

Alessi, L., E. Ossolaw and R. Panzica. (2021). What Greenium Matters in the Stock Market? The Role of Greenhouse Gas Emissions and Environmental Disclosures. Journal of Financial Stability 54.
Amenc, N., and F. Blanc-Brude. (2022). The Pricing of Green Infrastructure, EDHEC Infrastructure Institute Publication. See: https://edhec.infrastructure.institute/paper/the-pricing-of-green-infras...
Blanc-Brude F., A. Gupta, L. Luna and A. Wee. (2021). The Volatility of Unlisted Infrastructure Investments, EDHEC Infrastructure Institute Publication.
Blanc-Brude F., W. Schmundt, T. Bumberger, R. Friedrich, B. Georgii, A. Gupta, L. Lum and M. Wilms. (2022). Infrastructure Strategy 2022. Boston Consulting Group & EDHECinfra.
Pastor, L., R. Stambaugh and L. Taylor. (2021). Sustainable Investing in Equilibrium. Journal of Financial Economics 142(1): 550-571.
Pastor, L., R. Stambaugh and L. Taylor. (2022). Dissecting Green Returns, NBER Working Paper 28940.