The Integration of Commercial Real Estate Markets and Stock Markets
Ling, David C., Naranjo, Andy, Real Estate Economics
The value of U.S. commercial real estate is estimated to be in excess of $4 trillion, which represents approximately 16% of total domestic investible wealth. As of 1994, this percentage was only slightly less than the 18% share of the domestic investible universe represented by publicly traded (non-real-estate) equities (Miles, et al. 1994). Despite the importance of commercial real estate and its growing acceptance by institutional investors as an important component of a mixed-asset portfolio, little is known about the degree of integration between commercial real estate markets and other domestic asset markets. However, the notion that U.S. capital markets are perfectly integrated is central to many theoretical developments in both finance and real estate. Moreover, the degree of integration has important ramifications for private-sector decisions, including effective portfolio diversification strategies. The degree of integration also is significant for public policy issues dealing with market efficiency and regulatory impediments to capital flows among various domestic capital markets.
We employ in this study a series of commonly used asset pricing models to test whether commercial real estate markets in the US are integrated with stock markets. In particular, we use stock return data from the NYSE, AMEX and Nasdaq, return data on publicly traded real estate companies [including equity real estate investment trusts (REITs)], return data on privately-held real estate from the National Council of Real Estate Investment Fiduciaries (NCREIF), and a combination of return data from the NCREIF and the American Council of Life Insurance Companies (ACLI). We create three groups of real estate portfolios and one group of stock portfolios from these various return data.
For each real estate portfolio group, we use the stock portfolio group and data on various macroeconomic risk factors to estimate a system of nonlinear, seemingly unrelated asset pricing equations to obtain time-invariant (i.e., fixed-coefficient) estimates of risk-factor sensitivities and risk premia. If the commercial real estate market (both exchange-traded and nonexchange-traded) and the stock market are integrated, the risk premia (prices of risk) for the macroeconomic factors must be the same in both markets. That is, the price per unit of exposure to each risk factor must be the same regardless of the asset market in which the risk factor is traded (i.e., priced). Using Wald tests, we then formally investigate the integration hypothesis by analyzing the differences in the factor risk premia estimates across U.S. stock and commercial real estate markets. In addition to the time-invariant estimates, we also use a version of the two-pass regression technique of Fama and MacBeth (1973) to obtain time-varying risk-premium estimates for both real estate markets and the stock market. The two-pass procedure allows us to check the robustness of our fixed-coefficient results, as well as analyze changes in the degree of integration over time.
The results support the hypothesis that markets for exchange-traded real estate companies, including REITs, are integrated with the market for exchange-traded (non-real-estate) stocks. However, when NCREIF data (adjusted for smoothing) and ACLI data are used to construct real estate portfolio returns, the results support rejection of the integration hypothesis, although this may reflect the inability of these appraisal-based returns to accurately proxy for commercial real estate returns. Interestingly, the growth rate in real per capita consumption is consistently priced in both commercial real estate markets and stock markets, whereas previous studies have found mixed evidence on the role of consumption in explaining ex ante stock returns. This result suggests that models of commercial real estate returns and stock price returns that exclude consumption risk are misspecified.
The layout of the paper is as follows. …