Academic journal article Journal of Real Estate Portfolio Management

Real Estate in the Real World: Dealing with Non-Normality and Risk in an Asset Allocation Model

Academic journal article Journal of Real Estate Portfolio Management

Real Estate in the Real World: Dealing with Non-Normality and Risk in an Asset Allocation Model

Article excerpt

Executive Summary. Quantitative models of asset allocation are increasingly used by institutional commercial real estate investors as a guide for investment strategy. Real estate as an asset class, however, does not conform well to many of the assumptions underlying standard mean-variance optimization. This paper outlines a model of allocation that addresses two important "real world" violations of these assumptions. First, the assumption that returns are normally distributed is relaxed; instead, returns are modeled using a distribution that allows for both the "fat-tailed" behavior and skewness seen in asset returns. Second, an alternative to the traditional MPT optimizer is employed-the so called "downside deviation" model-that better reflects the observed behavior of investors.

Introduction

Since their introduction by Markowitz (1959) over forty years ago, the use of quantitative meanvariance models of asset allocation has become routine in many, if not all, of the liquid, publicly traded investment sectors. Commercial real estate investors have also recognized the value of incorporating these models at some level in their ongoing investment decision-making practices (Worzala and Bajtelsmit, 1997). In addition, a growing body of research has demonstrated the statistical diversification value of including different types of commercial real estate in both mixed-asset and single-asset portfolios (e.g., Mueller and Mueller, 2003).

Institutional investor interest in the use of quantitative asset allocation models in commercial real estate portfolios has risen sharply since the property markets in the United States experienced a significant downturn in the mid-1980s. The crash of the real estate markets in the 1980s, born out of distortionary tax laws that led to massive over-building, was unprecedented. Since then, diversification and asset allocation have evolved as important tools to mitigate risk in real estate portfolios.

The quantitative approach most widely adopted is based on Markowitz' Modern Portfolio Theory (MPT). MPT contends that asset allocation and diversification of a portfolio across different asset classes with different performance characteristics will minimize risk. Indeed, the MPT approach seemed to offer the "optimal" solution of maximizing total portfolio return for a given level of risk. The quantitatively derived asset allocation parameters seemed to provide investors an algorithm for constructing portfolios that included assets with varying degrees of risk. Real estate professionals have applied MPT to build real estate portfolios diversified across both property sectors and geographies. Since the bitter experience of the 1980s, diversification has evolved as a dominant issue among real estate professionals including plan sponsors, consultants, managers and researchers (e.g., Pagliari, Webb and Casino, 1995; and Worzala and Bajtelsmit, 1997).

MPT has developed a large following among real estate professionals who believe this technique provided an objective and quantitative approach in determining which property sectors to over or under weight in a real estate portfolio. This is the "within real estate" application of MPT. The "within real estate" approach has largely focused on diversification across property types, since it is generally perceived that a large share of the variability of a real estate portfolio is a result of allocation across property types rather than allocation across broadly defined geographic regions (e.g., Viezer, 2000).

Despite the appeal of using quantitative asset allocation methods in commercial real estate, there are significant hurdles to overcome. Often, the Capital Asset Pricing Model (CAPM), the more general Arbitrage Pricing Theory (APT) model, or a related multi-factor pricing model, is used as the behavioral foundation to forecast the returns used in mean-variance optimization. These asset pricing models make some strong assumptions about market structure, statistical pricing dynamics, the use and dispersion of pricing information, and investor behavior; collectively these assumptions imply that mean-variance optimization will, on average, result in risk-minimizing portfolios for rational investors. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.