A Test of Weak-Form Efficiency in Residential Real Estate Returns

This paper tests weak-form efficiency of residential real estate returns for the city of Memphis, Tennessee. The database for the study is comprised of the population of all sales of single-unit residential property over a fifteen-year period, 1970–1984. The city was divided into ten submarkets based on Memphis City Planning Commission planning districts. An analysis of variance procedure was utilized to stabilize the variance both within and across submarkets and nonmarket financing was partially controlled by eliminating transactions with loan-to-value ratios greater than 95%. The remaining transactions were then used to generate a mean return series. The advantage of the mean over the single transaction series used by Gau is that it represents the most likely outcome for the investor trying to duplicate investment performance since "property unique" features would be expected to cancel. Seven of the ten submarkets exhibited time patterns; however, after adjusting for transaction costs, all ten submarkets were determined weak-form efficient for the period 1970–1984. This was not true for the short horizon holding period, 1970–1975. In four sub-markets an asymmetric version of Alexander's filter rule was able to outperform a buy-and-hold, even for round- trip transaction costs as high as 10%.