The Role of Property in a Portfolio, Given Recent Experience

Investors with portfolio allocations to commercial property (real estate) today confront three unpleasant facts: (1) During the 2008 downturn, real estate investment trusts (REITs) and other forms of securitized commercial property (e.g., real estate operating companies) underperformed equities, contrary to what theory and experience led us to expect; (2) Global correlations across commercial property markets have generally been quite high; and (3) Based on reported results, directly owned commercial property seems to have outperformed REITs.

These facts raise three questions for investors and their advisers: (a) What is the future role of commercial property in a portfolio? (b) For the purpose of asset allocation, does it make sense to distinguish between regional markets, or should commercial property be treated as a single global asset class? And (c) what is the best way to invest in commercial property: directly, via REITs, or a combination of both?

Before answering these questions, we first need to develop a better understanding of what has happened over the past two years. The first issue is why global REITs have underperformed global equities. Logically, the roots of REITs' underperformance lie in some combination of causes related to fundamental valuation and investor behavior. To assess the former, let us start with our basic valuation model, in which the value of an asset is equal to the following: (Current Free Cash Flow) x (1+ Forecast FCF Growth) divided by (Current Yield on Real Return Bonds + Risk Premium - Forecast FCF Growth). In this case, let us further assume that FCF refers to the cash flow that is actually distributed to investors - dividends and stock buybacks. Across multiple countries, one of the defining characteristics of the REIT structure is the requirement to pay out to investors a very high percentage of free cash flow. Almost always, this means that the current cash flow (as a percentage of market value) from companies in a REIT index is higher than the current cash flow yield from the companies that comprise the public equity market. So this cannot be the source of REITs underperformance.

What about the rates at which these cash flows are expected to grow in the future? As we note in our Asset Class Valuation Section, over the long-term, the real rate of growth in REIT market FCF is quite low. This is logical, as the barriers to expanding the supply of commercial property are quite low - rising rents tend to trigger new construction. In contrast, the long-term real growth rate for equity market FCF is higher, because it is ultimately based on total factor productivity growth, which, because of its complex root causes, is harder to duplicate than a building, and hence less susceptible to being competed away by new entrants (however, as we note in this month's Economic Update, public policy decisions can significantly affect a country's long term total factor productivity growth rate). Moving from the long-term to the short-term view, the relevant question is whether there is a logical basis for assuming that the rate of FCF growth for REITs in aggregate over the next five years would be significantly less than the rate of FCF growth for equities. On the one hand, there seem to be some strong arguments favoring REITs - while corporations under growing financial stress could be expected to cut their dividends and buybacks, REITs remain legally required to payout a substantial portion of their cash flow. On the other hand, one might argue that this payout advantage would be offset (or perhaps more than offset) by forces that would cause commercial property FCF to suffer greater declines than FCF for the equity market as a whole. One concern is the higher average level of operating and financial leverage in commercial property relative to the equity market as a whole. In good times, many real estate managers were no doubt tempted to raise leverage, in order to provide higher dividends to income-oriented investors during a period of falling yields on many fixed income assets. Yet during the downturn, high leverage magnifies the FCF impact of a fall in rents. Beyond this, additional FCF concerns may have been raised by specific structural characteristics of the major global commercial property and REIT indices. For example, just five countries account for about 75% of the market capitalization of both the Dow Jones Global Real Estate Index and the FTSE EPRA NAREIT Global REIT Index - the United States (at, respectively, 33% and 40%), Japan (14% in both), Hong Kong (14% and 6%), Australia (8% and 9%) and the UK (5% and 7%). As you can see, the Anglo Saxon countries, where the 2008 credit and economic crisis hit hardest, have (except for the US) different weights in these commercial property indices than they do in global equity indices. Moreover, from a sector point of view, both global real estate indices give relatively heavy weight to retail property (32% and 29%), which will be particularly hard hit by a prolonged decline in consumer spending. In sum, an expectation that commercial property in general, and in certain countries and sectors in particular, would suffer worse reductions in free cash flow than the global equity market could have accounted for the underperformance of global property indices compared to global equity indices.

The final question is whether changes in discount rates could also help to explain commercial property's poor performance. We think of the discount rate as being composed of three parts. The first is the yield on real return government bonds. This is common to all asset classes. The second is a premium for risk - payment for accepting variability of future returns that an investor believes he or she understands - i.e., a range of possible future outcomes whose probabilities can be roughly estimated. The third part of the discount rate is a premium for bearing uncertainty - returns variability whose range of possible outcomes and associated probabilities cannot be estimated. We believe that risk aversion is a more rational construct that is relatively constant over time. In contrast, as we noted in last month's issue, uncertainty triggers the brain's fear circuits (i.e., the amygdala) and is an emotional and social phenomenon that varies greatly over time. In light of this view of discount rates, the key question to ask is whether there is any basis for believing that uncertainty with respect to commercial property (and hence the discount rate) has been higher over the past year than uncertainty about equities. We believe such a basis exists, due to heightened fears about REITs' ability to maintain debt payments and/or refinance maturing loans in the face of declining rents, falling credit availability, rising lending standards, and, if worries about future inflation prove accurate, rising nominal interest rates. In sum, there seem to be logical reasons - involving both the relative rate of future free cash flow growth and the relative rate at which it is discounted to present value - that can explain global property's underformance compared to global equities over the past year.

The second issue we must examine is the high correlation between commercial property returns across multiple regions in 2008. In the past, some researchers have noted the common exposure of all commercial property markets to global GDP growth. However, the data used in these analyses did not include a period characterized by a decline in global GDP - hence, the asymmetric impact of this factor (much stronger on the downside than the upside) was not picked up. Similarly, to our knowledge, no previous analysis fully captured the exposure of commercial property to a global liquidity risk factor, which only came into play - violently - on the downside. Finally, recent years have seen a number of developments that collectively have made the commercial property market much more globally integrated, including the spread of the REIT structure, increasing cross-border investment flows, greater institutional and retail interest in diversifying across a wider range of asset classes, and improved information availability due to the internet. Collectively, these go a long way toward explaining the significant rise in correlations across global property markets over the past year. That said, there have also been exceptions in the other direction, with Switzerland being the most notable example, as Swiss property appears to have receive a significant amount of "safe haven" inflows.

The third issue is why direct property investments (e.g., via a limited partnership structure) have apparently outperformed securitized commercial property vehicles like REITs. This is an issue about which reasonable people can and do disagree. Some of the difference is undoubtedly due to the different ways that direct property return indexes (e.g., from NCREIF or IPD) and securitized property indexes (e.g., from Dow Jones or EPRA/NAREIT) are calculated. Direct property indexes typically reflect a mix of actual cash flows (for net operating income or funds from operations) and appraised capital values. Multiple studies have shown that appraised values tend to be much more correlated over time than values based on purchases and sales in a continuously operating market (which is the way REIT values are calculated). As a result, the appraisal approach tends to understate both volatility and correlations. REIT indexes include the effect of leverage, while direct property indexes are typically reported on an unleveraged basis (note too that directly owned real estate is generally believed to employ higher leverage than REITs). In addition, neither direct nor REIT indexes explicitly take into account the fact that securitized real estate is much more liquid than directly held property. Finally, REIT and direct property indexes are also based on different sector weights (e.g., retail, industrial, office, multifamily residential, etc.), and typically have different management fees (higher in the case of direct property). In short, simple comparisons (which usually claim direct property ownership is superior) are usually badly flawed.

A number of statistical techniques can be used to "unsmooth" appraisal based returns. Invariably, they result in an increase in estimate volatility for directly held property and higher return correlations with other asset classes. Perhaps the best adjustment to appraisal based data has been undertaken by the Center for Real Estate at MIT. They have constructed a transactions based index from the NCREIF directly owned commercial property database. As expected, this also resulted in an increase in the volatility of commercial property returns, and their correlations with returns on other asset classes (see "A Quarterly Transactions-Based Index of Institutional Real Estate Investment Performance and Movements in Supply and Demand" by Fisher, Geltner, and Pollakowski). Two other studies have started with this MIT data, and taken further steps to make it comparable with the NAREIT index (e.g., by adjusting sector weights and leverage). Both of these studies ("Privately Versus Publicly Held Asset Investment Performance" by Riddiough, Moriarty, and Yeatman, and "A Successive Effort on Performance Comparison Between Public and Private Real Estate Equity Investment" by MIT's Jengbin Tsai) have found that the returns on directly owned and securitized property are very similar. Riddiough et al found a 3% average return advantage in favor of REITs, which they conjectured could reflect differences in liquidity or geographic differences in the underlying properties whose impact was not taken into account. Tsai found (using a different time period) that the two series were essentially the same - between 1995 and 2005, the geometric average return on REITs was 13.12% and the standard deviation (of annual returns) was 8.99%, compared to 13.03% and 9.18% on the adjusted MIT Transactions Based Index for directly owned properties. In sum, differences in reported results for directly owned versus securitized property seem to be due to a variety of factors that mask the essential economic similarity of these two different approaches to investing in commercial property. However, this is not to say that in some circumstances these differences are unimportant. For example, the delayed impact of changes in appraisal based indexes can smooth out reported fluctuations in a portfolio's value. If one believes that public markets are more susceptible to overreactions driven by investor emotions and herding, this may, in fact, be beneficial from a decision making perspective.

Having examined the past, let us now turn to our three questions about the future role of commercial property in investment portfolios. First, given what we have learned over the past 18 months, does commercial property still have a role in a diversified investment portfolio? In a 2005 paper, Dhar and Goetzmann report the results of a survey study of "Institutional Perspectives on Real Estate Investing: the Role of Risk and Uncertainty." They find that "the main reasons for investing in commercial property given by institutional investors are diversification and inflation hedging, while liquidity, lack of reliable valuation data, and poor management were cited as the main risks...The expected return and risk of real estate is perceived as mid-way between U.S. stock and bonds." After the events of the last 18 months, these assumptions are worth reexamining.

We'll start with the basics: are the underlying economic drivers of commercial property returns different from those on other asset classes? We think the answer is quite clearly yes - what other asset class involves site selection, architectural excellence, construction management, and the careful marketing and management of the non-storable services that a building provides? On the other hand, commercial property shares some common activities with equity (e.g., making good decisions about debt levels and structures) and with fixed income (e.g., making good decisions about rental contract duration). At least in the past, studies using sophisticated quantitative techniques like cointegration analysis have reached similar conclusions.

Now let's look at how the commercial property asset class has actually performed. In last month's issue, we reviewed a recent IMF study of the inflation hedging properties of different asset classes over short and medium term time horizons. At best, commercial property (because of rising replacement costs and rents) provides a partial hedge in the medium term, but not the short term. Yet other inflation hedges are available, including real return bonds, commodities, timber and gold. On the other hand, apart from the first of these, none provide the current income stream that is available from property. Moreover, as generations of European investors have confirmed over long periods of historical ups and downs, there is a sense of psychological security that comes from owning a property that, with the exception of gold, no other asset class can match.

We also need to consider how property has performed in two other regimes. In periods of normal growth, equities should deliver the highest return of any asset class (note that we are not including uncorrelated alpha strategies here, because they are active management approaches, not asset classes). However, in the years before the 2007 - 2009 crash we saw equities outperformed by commercial property and commodities. Arguably, the former reflected both fundamental factors (debt financed growth in the economy in general and consumption in particular), lenders' underpricing of risk, and poor decisions by too many property managers to raise leverage ratios higher. Similarly, high commodity returns were due to a difficult to disentangle combination of (a) supply/demand imbalances in the physicals market - e.g., the rising marginal cost of finding and producing a barrel of oil; (b) supply/demand imbalances in the futures markets - e.g., the growth of "long-only" commodity index funds relative to the supply of contracts; and (c) the availability of leverage - e.g., via low margin requirements on futures and options contracts. And what about a regime with high uncertainty, deleveraging, deflation, and severe economic contraction? Commercial property has delivered worse performance than real return and nominal government bonds, gold, timber and investable volatility products.

These observations lead us to conclude that, at best, the case for including commercial property in a portfolio is weaker than it was before the events of the past 18 months. Some studies have concluded that it is not persuasive at all (see, for example, "Real Estate in an ALM Framework" by Brounen, Porras Prado, and Verbeek). Others have concluded that use of a shortfall risk instead of a maximum volatility constraint also weakens the case for property in a portfolio (see "The Maximum Drawdown as a Risk Measure: The Role of Real Estate in the Optimal Portfolio Revisited" by Hamelink and Hoesli). As we head into the 2009 reassessment of our model portfolios (which will be based on a three regime model), we have decided that recent events merit an increase in the required risk premium on commercial property as an asset class, from 2.5% to 3.0% above the yield on real return government bonds. Along with our use of a shortfall risk constraint, we suspect that this will result in lower allocations to commercial property in our revised portfolios.

The second issue, assuming one invests in the asset class, is it worth distinguishing between domestic and foreign commercial property? Researchers were divided on this issue before the most recent crisis (for contrasting views of international real estate markets’ cointegration, see "Global Property Market Diversification" by Gallo and Zhang, and "Random Walks and Market Efficiency: Evidence from International Real Estate Markets" by Kleiman, Payne and Sahu). However, apart from Switzerland, the evidence from the past 18 months suggests a higher than expected degree of integration on the downside. This argues for treating developed country commercial property markets as a single asset class - as might also be the case for developed equity markets.

Finally, there is the question of whether to invest in commercial property directly or via securitized vehicles like real estate investment trusts. We are persuaded by the studies that find, as common sense suggests, that in economic terms the two are very similar, apart from the superior liquidity of securitized vehicles. However, we also recognize that the flip side to illiquidity is a lower chance of investor herding causing significant over and undervaluations, and that the time lag caused by appraisal-based valuations can impart stability to a portfolio and help to minimize client overreactions and, quite frankly, fear.

Over the past 18 months, perhaps no asset class has proven more disappointing - and frustrating - than commercial property. In a period of unprecedented uncertainty, other asset classes have delivered better diversification benefits, as property returns around the world were crushed by declining GDP and relatively high leverage. Moreover, studies have shown that if we return to a high inflation regime, other asset classes may better preserve the real value of investors' capital. And under normal circumstances, equities should deliver higher returns. In sum, the case for investing in commercial property has grown much weaker.

| June 2009 Issue: Key Points | This Month's Letters to the Editor: Fidelity vs Schwab vs TD Ameritrade - Opinion?; Index Investors' Allocation Models vs MVO Models - Superior? Yes; Asset Class Valuation Updates: Possible, Likely and Probable - What They Mean; and Commodity Valuations Long/Short (LSC) or Long-only Fund? | Global Asset Class Returns | Uncorrelated Alpha Strategies Detail | Asset Class Valuation Update | The Role of Property in a Portfolio, Given Recent Experience | June 2009 Economic Update | Product and Strategy Notes: Powerful Impact of Regret; More Research - Why Successful Actively Managed Funds are Rare; Did the Media Do a Good Job Predicting the 2008 Crisis?; and How Rigorous is you Investing Logic? | |