The Effect of Geographic Diversification on the Returns of Real Estate Investment Trusts

In analyzing the effect of geographic diversification on the risk and volatility of real estate returns, past studies have suffered from several flaws. They have been performed on ex post portfolios constructed from a larger sample of properties, they have used appraisal-based return series, and the geographic regions analyzed did not bear any relation to the economic characteristics which underlie real estate. The results of such research have been largely inconclusive in determining whether geographic diversification in fact represents a possibility for reducing the risk profile of real estate investment. Real estate investment trusts are a real estate ownership vehicle which is liquid, priced on a daily basis, and which represents an interest in an underlying portfolio of real estate. Since the governing legislation requires that 95% of all income be passed through to the security holders, ownership interest in a REIT represents a close proxy in income characteristics for the performance of real estate. Capital appreciation, the second major component of real estate returns, is also strongly linked to the performance of the real estate, although here stock market effects come into play as a REIT share, like any tradable security listed on an exchange, is subject to market forces. Because REITs represent an ownership interest in a defined portfolio of real estate, are priced in the capital markets on a daily basis, and can be broken down on a geographical basis to test any proposed geographic diversification scheme, they represent a chance to draw conclusions about the effect on real estate returns of such diversification, and the degree to which such diversification reduces risk. If stock market effects are accounted for and removed from the analysis, the remaining performance factors can be assumed to represent the underlying real estate, and an analysis can be made of the degree to which geographic diversification has affected the volatility of the returns. I researched and calculated a concentration index for each REIT on a quarterly basis over the period from 1980 to 1989. I then analyzed both the variance in the returns and the relationship of average returns to the variance of the returns. I also analyzed the returns themselves over one, two, and five-year periods with respect to the concentration index. My conclusions were that the Hartzell-Shulman-Wurtzebach model provides a context for genuine scientific diversification for real estate over periods of five years or longer, explaining in a statistically significant way the variance of the returns. Over shorter time periods, the data is too noisy for diversification to be significant in explaining variance. The variance of returns appears to affect the level of returns, which is consistent with expectations and which indicates that the concentration index should have a direct link with returns. Finally, the concentration index also appears to explain the level of returns over longer time periods, although this was not statistically significant. Thesis Advisor: William C. Wheaton Associate Professor of Economics The Effect of Geographic Diversification in the Returns of Real Estate Investment Trusts An Analysis ABSTRACT.....--.....---..----. ------------------------**...2 CHAPTER ONE DEFINING THE PROBLEM........... Problem definition...................... Previous research....................... Questions posed......................... .....------6 .6 . e7 .o 0 00 0 0 010 CHAPTER TWO REAL ESTATE INVESTMENT TRUSTS................ Definition of a REIT...... Legal operational Industry History.......... Legislative Operational 11 .-. . 13 Classifications of REITs.............................. Equity Mortgage Hybrid Sources of REIT Returns............... Real Estate Components Cash flows from properties Appreciation Stock Market Components CHAPTER THREE DIVERSIFICATION............ 17