Foreign Commercial Property

Updated February, 2006
PDF February 2006 Journal

Like commodities, commercial property is another asset class that has seen a sharp increase in prices over the past few years. One part of its attraction is clear. More and more countries have legalized the use of vehicles known (in the U.S. and Canada) as real estate investment trusts (REITs). These vehicles own commercial property, trade publicly, and are usually exempt from corporate taxes provided they pay out a high percentage of their earnings (e.g., 90%) as dividends to their shareholders. In an era where current income returns on many other asset classes are quite low, the relatively high dividend yields offered by REITs have attracted the interest of many investors.

Once again, however, the question must be asked: have the high returns on commercial property securities in recent years represented too much of a good thing? Are they now overvalued?

We will begin our analysis with a brief overview of the classic commercial property cycle, which is summarized in the following table.

Let us start from the third column, when an economy begins to come out of recession. Vacancy rates begin to fall, and rents begin to rise, while interest rates are low. Since the market value of a commercial property is equal to the capitalized value of its expected rental income stream, the rise in rents leads to in an increase in property values. As the economy nears the peak of demand growth, rising property values (driven by further increases in rental income) have triggered an increase in new construction activity. Some of this comes onto the market after the economy has passed its peak, which accelerates the fall in rents and (along with rising interest rates), causes a decline in property values which continues through the bottom of the economic demand cycle.

To be sure, it can be argued that different types of commercial property pass through this cycle at different speeds. For example, the valuation of retail properties (where rents are driven by consumer spending) seems to track the economic cycle more closely than the valuation of office properties (that tends to track employment growth, which lags demand growth). That being said, one way to answer the valuation question is to ask which stage of the economic cycle we are in today. The recent experience of strong demand growth and rising interest rates suggest column four, which implies that commercial property values are peaking.

Another way to approach this question is via The Index Investor equity valuation model (since REITs are traded on the public equity market). As you recall, this valuation model has two parts: the returns companies are expected to supply, and the returns investors logically demand. In a market in equilibrium, these two will be the equal; however, as we have noted, since the financial markets are a complex adaptive system, they are usually not in equilibrium (though they are strongly drawn towards it).

Let's start with the returns that companies are expected to supply. They are estimated as the sum of the current dividend yield on a stock (or market) and the rate at which these dividends are expected to grow in the future. Unfortunately, while the current dividend yield on commercial property securities is easy to obtain, the rate at which dividends are expected to grow in the future can only be assumed. This task is made much harder by the relative scarcity of historical data for commercial property securities, which are relatively new in many countries.

The returns that investors demand are also composed of two parts. The first is the current yield on real return government bonds, which is the basic building block for all financial asset returns. The second is a premium that reflects the relative riskiness of the asset class in question. In this case, like many others we have judged the riskiness of liquid commercial property securities to lie in between investment grade bonds and equities; the specific risk premium we use in our asset pricing model is 2.5%.

Since the future rate of dividend growth is so hard to estimate, one way to approach the valuation question is to assume the market is in equilibrium, and that the returns the market is expected to supply equal those rationally demanded by investors. This allows you to derive the rate of growth by subtracting the current dividend yield from the sum of the current real bond yield plus the assumed commercial property risk premium. This calculation is shown in the following table for five markets with significant trading volume in commercial property securities.

As you can see, this approach yields negative expected dividend growth rates. On the one hand, this is consistent with the view that we are approaching the top of a commercial property cycle. On the other hand, it is probably inconsistent with the expectations of a lot of people who have been investing in commercial property securities on the assumption that they are not currently overvalued.

To put this issue in perspective, the following table shows the implied real growth rates that result from using different assumptions about investors' required risk premium for holding commercial property.

As you can see, it is not until the assumed risk premium reaches 4% to 5% that the implied growth rates all get into a range that many commercial property investors might consider a reasonable assumption. This strikes us as unreasonable for two reasons. The first is that an excellent recent study estimated the risk premium of four percent for directly owned commercial property, which is significantly less liquid, and therefore riskier than commercial property securities (see "The Performance of Real Estate Portfolios: A Simulation Approach" by Fisher and Goetzmann of Yale University).

Second, assuming a 4% to 5% risk premium for liquid commercial property securities also implies that investors simultaneously believe that although we have not reached the peak of the commercial property cycle, equity securities (which should require an even higher risk premium than commercial property securities) are already extremely overvalued.

Based on the rule, "choose the simplest hypothesis", we conclude that it is most likely that we are approaching, or at, the top of a commercial property cycle. This conclusion is corroborated by a new report from HSBC Bank, "A Froth Detecting Mission: Detecting U.S. Housing Bubbles." It finds that "about half of the US housing market is frothy and that this 'bubble zone' may be overvalued by as much as 35-40%, after taking into account low interest rates and tax advantages. Current valuations imply [either] a large permanent reduction in the risk premium and/or a sizable step up in future capital gains, not all of which, we think, is justified. The 'bubble zone' accounts for 50% of US GDP, or over US $ 6 trillion, nearly the size of the German, French, and UK economies put together. In other words, it's big. Therefore, when these housing bubbles begin to deflate, it is likely to have substantial macroeconomic consequences." Moreover, as The Economist global house price index has repeatedly demonstrated, this is not a phenomenon limited to the United States. Arguably, housing bubbles are already deflating in Australia, the United Kingdom, South Africa and Spain. If residential housing markets are at (or beyond) peak valuations, why should we not expect the same to be true of commercial property valuations?

Again, we note the difficulty of trying to time markets. If an investor has already made his or her allocation to commercial property, and if that allocation is currently above its target portfolio weight, this would be a good time to rebalance, perhaps to a level somewhat below the target weight. On the other hand, if an investor has not yet made his or her allocation to commercial property, we believe that the prudent course of action would be to defer any reallocation until the valuation of commercial property securities have declined from their current levels.

Updated, January 2006

It has long been known that over half the world's commercial property, measured by value, lies outside the United States. Until recently, however, it has been hard for retail investors to access this market, with the same ease that they could invest in real estate investment trusts (REITS) in the United States. With the notable exceptions of Australia and Canada, few countries had approved the use of this type of securitized vehicle for investing in commercial property. That is now changing, at an accelerating pace. Hence more and more retail investors around the world will be confronted with two important questions: Is foreign commercial property a separate asset class, and, if it is, should I invest in it?

Arriving at definitive answers to these questions is probably not possible, due to the limitations of the data we have to work with. There are four principal problems. First, the regional property market indexes that are available include different mixes of property types. For example, one might have a greater proportion of warehouses, while another might have relatively more industrial, retail, office, or hotel properties. Second, until the advent of securitized property investment vehicles, most of these indexes were based on directly owned property, which is only valued at irregular intervals by appraisers. While they no doubt try their best, analysis has shown that many appraisers' valuations are anchored on the most recent appraisal result. This causes successive valuations to be more closely related to each other (the technical term is “serially correlated”) than is the case with securities traded in a continuous market. It also causes the actual riskiness (volatility) and correlation of returns of commercial property with other asset classes to be underestimated. While securitized vehicles correct this problem, they raise another one: the length of time that they have been available is relatively short, making it hard to draw strong conclusions from data on their returns. To some extent, this is compounded by a fourth problem, that has close parallels in other asset classes. The issue is this: due to the differing percentages of total commercial property value that has been securitized in different countries, “global” securitized property indexes may present a distorted picture of the “true” risks and returns on this asset class. For example, the global securitized property indexes published by EPRA/NAREIT and Standard and Poor's tend to give relatively more weight to Australia and less to continental Europe in comparison to proportion of the total value of the world's commercial property found in these regions. Of course, this is also true of global bond indexes (as we have described in our previous article) and equity indexes as well (where public companies account for differing percentages of the total value of corporate equity in different countries).

With those appropriate warnings, let's move on to take a look at the data. The following table shows the nominal local currency returns on securitized commercial property indexes over the 1989 to 2004 period:

The next table shows the correlations between these local currency returns:

The average correlation in this table is quite low, at .32, which suggests that, assuming correlations don't change dramatically over time, substantial potential exists for achieving diversification benefits by investing in foreign as well as domestic commercial property.

Despite these apparently attractive statistics, the results of a number of analyses sound a cautionary note about the size of the potential diversification benefits from investing in foreign commercial property. All of them reach a similar conclusion: that real estate returns in different countries are driven by a mix of a global factor (generally taken to be global GDP growth) and more idiosyncratic local factors. Regarding the global factor, Edward Kane's paper asks about the late 1980s, “Has U.S. Overbuilding Affected Construction Activity Globally?” He concludes that the answer was yes. And in “The Global Real Estate Crash”, Goetzmann and Wachter analyze the resulting downturn that arrived in the early 1990s. They find clear evidence that the global commercial property crash of 1992 was preceded by declining property values from the end of the 1980s, that were closely related to declines in world GDP that occurred at the same time. In a subsequent paper (“Global Real Estate Markets, Cycles and Fundamentals”), Case, Goetzmann and Rouwenhorst analyzed commercial property returns data from 22 countries between 1987 and 1997 and confirm they are affected by a common global GDP factor.

However, this is not to say that local factors are also important determinants of commercial property returns. In “A Fundamental Comparison of International Real Estate Returns”, Pagliari, Webb, Canter and Lieblich analyze the office, retail and warehouse sectors in Australia, Canada, the UK and the US between 1985 and 1995. They conclude that the “space market” (as reflected in rental income) is more local in nature, while the rate at which they are capitalized to generate market values is subject to more global influences. In “Evidence of Segmentation in Domestic and International Property Markets”, Wilson and Okuner use a sophisticated statistical technique (cointegration analysis) and also find that potential international diversification benefits exist. Similar conclusions are reached by Wilson and Zurbruegg in their paper “Does It Pay to Diversify Real Estate Assets?” and by Conover, Friday, and Sirmans in “Diversification Benefits From Foreign Real Estate Investments.”

In light of these findings, as well as the increasing range of products that enable investors to access diversified international commercial property portfolios, we have included international commercial property as an asset class our model portfolios.