Most pension funds use external managers to select their assets, but this appears to be a costly choice, write Piet Eichholtz and Thomas Heijdendael 

Real estate has long been the third-largest asset class in pension portfolios. Pension funds have traditionally allocated between 5% and 10% of their wealth to the asset class, which makes it the biggest of the alternative asset classes. At the same time, it is an asset class that can sometimes be off the radar screen of the institutional investment industry, where debate focuses mostly on equities and bonds. This article provides an overview of the motivation for pension funds to invest in real estate, of real estate investment styles and approaches, and of their performance consequences. 

Pension funds have long-dated liabilities, often sensitive to inflation risk, and they tend to have a – legally enforced – fiduciary responsibility that requires them to diversify their risk. Their choices for certain assets reflect that, and real estate plays an important role there. When asked about their motivation for investing in real estate, pension funds tend to mention the risk-return profile of the asset class, inflation hedging and the diversification benefits.

To see whether that argument makes sense we provide some numbers concerning these issues. Table 1 shows real estate risks and returns for a selected group of six countries, and compares these with the performance of stocks and bonds.

1. Risk and return 2001-2013

The numbers are total annual returns and standard deviations covering 2001-13. This has, of course, been a tumultuous period for investors, during which there the global financial crisis. That is clearly visible in average stock returns, which have been quite poor, with very high risk. 

The MSCI Global Stock index had a total return of just 1.3% over the period, with a standard deviation of 20%. Bond returns have been much better, mostly because falling yields implied increasing bond prices. But since yields are not likely to fall much further, this high bond return is unlikely to continue. Real estate has done quite well, especially considering that the global crisis was a real estate crisis: IPD’s global index returned 7%, with a standard deviation of 6.6%, so having a much lower risk than stocks.

Figure 1 looks at whether these returns have offered protection against inflation risk. The table presents the correlations of real estate returns with local inflation for the same group of countries – high correlations imply stronger inflation protection, with a maximum of 1. Figure 2 tells a clear story: for stocks and bonds, correlations have hardly been positive over the period, implying that returns have fallen even while the nominal value of indexed pension liabilities went up. Real estate returns tend to have much stronger correlations with inflation than stocks and bonds. This holds for all countries, except the UK. However, the inflation correlation never gets very close to 1, so a perfect hedge against inflation risk is not on offer. Of the main three investment choices a pension fund has, real estate is not a very effective inflation hedge but it is the best on offer.

2. Correlations with local inflation 2001-13

3. Cross-asset correlations 2001-2013

Figure 3 concerns whether real estate offers diversification benefits to an investor that already has a portfolio of stocks and bonds, for the same selected group of countries. The table contains the cross-correlations between real estate, stocks, and bonds, with a low correlation signalling good diversification. The correlation’s minimum is -1, but a value of 0 already implies solid risk reduction potential.

The figure shows that real estate is indeed a good diversifier: the correlations between real estate and the other two asset classes (in columns 1 and 2) tend to be lower than those between stocks and bonds for most countries. Especially for pension funds with large bond portfolios, the risk-reduction effect of property investments has been remarkable, as the negative correlations in column 2 show – the average property-bond correlation was -0.17. For property and stocks, it was 0.16, and for stocks and bonds it was 0.3. So real estate can substantially reduce the risk of portfolios consisting of stocks and bonds. This illustrates that an investor should not regard the risk-return profile of real estate in isolation, but should rather take the portfolio view. 

The result of these investment characteristics is that pension funds tend to allocate a small but meaningful part of their wealth to real estate. Figure 4 illustrates this. It depicts the allocation to real estate for a worldwide sample of almost 1,000 pension funds from 1990 to 2011. The graph shows that the long-term average allocation was 5.5%, and that the average allocation was just under 7% in 2011. It also makes clear that listed property companies have come to play an increasingly important role in this. The allocation to listed real estate was almost non-existent until 1997, and since then has increased to about 1% of the overall asset portfolio, or 13-22% of the real estate portfolio on average.

4. Global pension fund allocation to real estate

The returns and return characteristics described in the tables are ultimately driven by the rents paid by the users of buildings. But these rents hardly ever flow into the coffers of pension funds directly. To implement an investment strategy into real estate, pension funds often use external service providers and vehicles, as figure 5 illustrates.

5. Investment approaches to real estate

Essentially, a pension fund first needs to decide to invest in real estate directly, ie, by getting ownership of the properties – either in whole or in part – or to do that indirectly, by getting ownership of shares in property companies. In both cases, the ultimate selection of buildings or property companies can be done in-house or by an external manager. In some cases, even the selection of the external manager is delegated to so-called fund of funds.

Andonov, Eichholtz and Kok look at these issues for a global sample of almost 1,000 pension funds and find that the larger funds tend to do more in-house, and are more likely to avoid external fund managers and fund-of-funds managers. These large funds also tend to combine direct real estate investments with investments in listed and unlisted property companies. Andonov, Eichholtz and Kok also find that a large majority of pension funds use external fund managers in real estate, and that fund-of-funds usage is trending upwards. That trend is especially visible among smaller pension funds, and it is strongest in the US. 

For the investment industry as a whole, this increased use of investment intermediaries adds costs but not value: the rental cash flows of the buildings that ultimately generate the returns are unlikely to be affected by the use of intermediaries. But maybe there is added value for pension funds that retain these service providers, compared with those that do not. In other words, it might be possible that external managers and fund-of-funds managers provide a competitive edge in accessing the best buildings and tenants, and consequently generate the best performance. 

The choices for specific investment styles have clear consequences for costs and performance. Not surprisingly, adding investment layers adds costs. Andonov, Eichholtz and Kok look at costs, excluding performance fees, and report that internal investment is cheapest, especially in listed property companies. The mean annual investment costs of that approach are just 12bps. Internal investment in directly held real estate is not much more expensive at 32bps. But using an external fund manager in direct real estate boosts these costs to 88bps, and using the services of a fund-of-funds manager implies 197bps in annual base fees. So, if costs ultimately decide the choice of an investment approach, internal investment would be the preferred option.

But, of course, the choice for an investment approach should be based on performance, and it is possible that the higher costs of outsourcing are justified by better performance. Unfortunately, that does not turn out to be the case, on average. Andonov, Eichholtz and Kok study the extent to which pension funds are able to outperform their benchmarks, with and without the use of an external fund manager. They report that pension funds investing internally tend to outperform their benchmarks with an average of 70bps per year.

The use of external fund managers – who are supposed to select the best assets – leads to an underperformance of 98bps relative to the benchmark, and the addition of fund-of-funds managers to help select the best managers further decreases the performance to 538bps below benchmark. In other words, adding layers adds costs but not performance.

Real estate is an important alternative investment category for pension funds, having good risk-return characteristics, inflation-hedging potential that is stronger than for stocks and bonds, as well as good diversification benefits relative to an existing stock and bond portfolio. So including real estate in the strategic asset allocation seems to make sense for pension funds.

To translate that strategy into viable execution, only a small minority of pension funds take the internal route. Most pension funds use external fund managers to select their assets, and even use fund-of-funds managers to select these managers. That appears to be a costly choice, and these costs do not tend to be justified by better performance. On the contrary, only pension funds that invest internally outperform their benchmarks, on average. Funds retaining external managers underperform, and those retaining fund-of-funds managers underperform terribly.

In favour of global diversity