Real Estate Securities and a Filter-Based, Short-Term Trading Strategy

Article excerpt

Michael Cooper*

David H. Downs**

Gary A. Patterson***

Abstract. Anecdotal evidence provides overwhelming support to the belief that sophisticated real estate investors profit by timing long-run real estate cycles. This article examines the investment performance benefits that sophisticated investors may derive from short-run cycles in real estate, specifically, through the publicly traded real estate markets. Using a simple strategy that filters out noise in real estate investment trust (REIT) price reversals, this study shows that a contrarian strategy is many times more profitable than the associated execution costs. Furthermore, the study demonstrates that the REIT market has been sufficiently liquid to execute this trading strategy. This last point is directly related to the filter strategy since only REITs with large price movements satisfy the hypothetical investor's selection criteria.

Introduction

Academic research has identified a predictable pattern of overreaction behavior in real estate investment trust (REIT) returns, though questions remain whether these price reversals can be economically exploited (Mei and Gao, 1995). Yet, if a predictable component exists within future returns, sophisticated real estate investors should refine their trading strategies to exploit these short-term reversals. This article examines such a refinement and employs a simple trading rule that filters out marginal price movements to more accurately reflect the trading behavior described in DeBondt and Thaler's (1985) overreaction hypothesis.

This article highlights two essential features of DeBondt and Thaler's overreaction hypothesis. The first characteristic emphasizes the direction of price movement and proposes that extreme movements in REIT prices will be followed by extreme movements in the opposite direction (i.e., price reversals).1 The second feature emphasizes the magnitude of price changes and states that the more extreme the initial movement in a REIT's price, the greater will be the subsequent price reversal. These features become the hypotheses as this study tests for economically exploitable profits from a contrarian trading strategy that takes long (short) positions in REIT that experience extreme price declines (increases) during the previous trading period. Additionally, the current concept of exploitation is potentially more robust than the prior literature's criteria for economic significance, since this study investigates whether REIT liquidity is sufficient for an investor to execute a contrarian trading strategy that filters out marginal price-movements.

A contrarian strategy may be intuitively appealing for a real estate investor since it is based on specific characterizations of the general investment community. Substantial documentation exists that real estate markets overreact, and DeBondt (1995) presents an interesting overview of real estate cycles from the perspective of the investor psychology literature. However, transaction costs in the direct property markets generally preclude short-term flipping strategies. It is also impractical to construct a short position in the property markets to take advantage of situations where prices deviate too far above market fundamentals. For these reasons, sophisticated real estate investors are more likely to implement a short-term trading strategy in the publicly traded real estate markets.

To test for the economic viability of REIT market overreaction, portfolios are formed based on a ranking of the REITs' returns from the previous week. The ranking is determined by a filter-rule designed to boost the signal-to-noise ratio in the real estate security selection process. The portfolio construction process involves both features of the DeBondt and Thaler overreaction hypothesis. The hypothesis' first characteristic is addressed by taking either a long- or short-position in a security based on the direction of its price movement. …