12. Infrastructure Equity as an Asset Class

1With many large publicly listed companies, index funds, and ETFs, infrastructure has recently established itself as a new investable asset class for institutional and retail equity investors. Likewise, private equity funds have raised many billions of dollars from institutions and high-net-worth individuals to deploy in infrastructure financings.

2One driver of the surge in interest has been a “great rotation” out of standard fixed-income products into investments that are less sensitive to rising interest rates yet deliver “bond-like” cash flows.

3But are infrastructure investments more like bonds or more like stocks? Here we look at the risk and return of listed infrastructure assets. We consider three listed infrastructure indexes:

1. MSCI World Core Infrastructure Index (WCII)

2. MSCI World Infrastructure Index (WII)

3. MSCI Emerging Markets Infrastructure Index (EMII)1

4We next detail their composition. Complicating the analysis, the sample periods are not only short but also characterized by the most serious financial crisis in several generations during 2007–2009. The longer sample additionally includes the technology boom and bust of 1999 through 2001 as well as the aftermath of late-1990s Asian Financial Crisis.

5We compare infrastructure equity’s performance against that of (1) the MSCI World Core Real Estate Index (CREI), which invests in listed companies that own and operate commercial real estate worldwide; (2) global stocks (the world market portfolio from Kenneth French’s data library);2 and (3) global bonds (the Barclays Aggregate Bond Index).3

6Table 4 shows the summary statistics. For the full sample period from 1999 through 2015, the WII performed quite poorly, with an average annual return of 4.0%, annual volatility of 14.4%, and an annual Sharpe ratio of only 0.14.4 For comparison, global stocks as a whole had average returns of 6.9% with 15.7% volatility, implying a Sharpe ratio more than twice as high as for infrastructure.

7A good deal of this poor performance occurred between 1999 and 2002, when stocks performed quite poorly as well. In sharp contrast, the EMII had much higher returns of 10.4% per year over the full sample, albeit with higher volatility of 20.7%. This performance is almost as good as real estate, which logs the highest return and the highest Sharpe ratio among all asset classes during the 16.5 years covered by the data.5

8In the shorter 2003–2015 subsample (Panel B), the WII recorded much higher average returns of 9.0% per year. The annual volatility was 12.7%, and the annual Sharpe ratio was an impressive 0.61. Both the WCII and the EMII performed even better, however, with annual returns of 12.1% and 12.7% and volatilities of 13.4% and 18.4%. The WCII has an impressive Sharpe ratio of 0.80. For comparison, global stocks recorded 9.1% average annual returns with 15.5% volatility, and global real estate stocks had average returns of 11.0% but with a much higher volatility of 20.5%.

9The reason the WCII outperformed the WII comes from the comparatively strong returns in the transportation sub-sector (which the WCII overweights) and weaker returns in telecommunications infrastructure (which the WCII underweights). The low volatility of infrastructure is noteworthy, especially in light of the financial crisis, which weighs heavily on this short sample period.

Table 4. Comparative Infrastructure Returns — Summary Statistics (Source: own calculations)

10The next question is whether listed infrastructure stocks and bonds perform similarly. To address this question, we estimate a regression of the excess return on the infrastructure indexes against the excess return on global stocks, the excess return on global bonds, and a constant. Table 5 presents the results.

11For the full sample, we find that about 70% of variation in both the WII and EMII returns is accounted for by global stock and bond returns. The WII has a stock beta of 0.73 and a bond beta of 0.24. The EMII has a much higher stock beta of 1.08 and a lower bond beta of 0.07.

12For the shorter sample, stocks and bonds explain 70% to 75% of return variation in the WII and EMII, and even 85% of variation in the WCII. The WII stock beta falls to 0.61 and its bond beta rises to 0.47. The WCII stock beta is 0.69 and its bond beta is 0.52. The EMII has the highest stock beta (0.91) and the lowest bond beta (0.45).

13We conclude that exposure to two risk factors, global stock and bond market risk, goes a long way toward explaining the observed returns on the various infrastructure indexes.

14Infrastructure investment is much more stock-like than bond-like, particularly for emerging markets infrastructure. This important finding may help define the boundaries of the investor community as well as pools of capital that may be available for infrastructure finance.

Table 5. Risk Characteristics of Infrastructure (Source: own calculations)

15We also note that the EMII’s and WCII’s performance are not only stellar in absolute return terms but also strong in risk-adjusted terms. We address this issue in terms of alpha, which measures the monthly abnormal return of an asset or portfolio of assets after accounting for global stock and bond risk.

16Based on our dataset, EMII had a 0.24% per month or 2.9% per year return in excess of what would be justified based on its risk (see the last row of Table 5). So, rather than generating a 7.5% return, it generated 2.9% more for a total return of 10.4%. For the shorter sample, the WCII even outperformed the index by 33 bps per month.6 Core Infrastructure delivered 4.0% more return than the required rate of return of 8.1%.

17With this cost of capital in hand, we can make sense of current valuation ratios on listed infrastructure equities.

18Figure 4 plots the price-dividend ratios on the three indexes. Infrastructure assets were very expensive during the 1999–2000 period, probably because of a combination of the euphoria in stock markets and the novelty of the infrastructure asset class. Valuation ratios fell by half in the ensuing stock market crash. Valuations rebounded in the mid-2000s, only to crash again during the 2007–2008 financial crisis. Since 2009, they have rebounded strongly. At the end of our sample in May 2015, Core Infrastructure traded at a price-dividend ratio of 31.25, Emerging Market Infrastructure at 27.4, and the WII at 26.3.

19Given its good fit, we take the expected return given by the global two-factor model as a good measure of the required rate of return to discount future cash flows. The question then becomes, what dividend growth rates do the current valuations imply? We focus on the WCII for the 2004–2015 sample period.

20Dividend growth on the WCII averaged 7.25% per year. If that growth rate were to continue this pace, the WCII should trade for a price-dividend ratio double the level we observed in May 2015. Under a more conservative scenario that dividend growth will remain high in 2015 (10%) and then fall by 1% every year until it hits a long-term mean growth rate of 4.15% in 2021 and beyond, the price is right. Given the strong need for global infrastructure, this set of growth forecasts seems eminently reasonable.

21In sum, after a rocky start in the late 1990s and early 2000s, infrastructure performed remarkably well as an asset class during the past decade. Despite the strong rise in prices of listed infrastructure assets since the depth of the financial crisis, infrastructure assets remained attractively valued, especially in light of the very high cash flow growth rates in the recent past. Such high cash-flow growth is predicted to continue in the short term but eventually to mean-revert to lower levels.