Common risk factors and risk premia in direct and securitized real estate markets

Abstract

This study empirically examined the effects of systematic market and common risk factors in explaining the variations in excess returns of securitized and direct real estate using multifactor asset pricing models (MAP). Two estimation methodologies were used to test the market integration hypothesis. The constant risk premia model that imposes time-invariant restrictions on the coefficients of the macroeconomic risk factors across different real estate portfolios in each market was estimated using the seemingly unrelated regression (SUR) technique. Credit risk, unexpected inflation and spread between government and commercial bonds were significantly priced in the securitized real estate market, whereas real T-bill yields and unexpected inflation were the two risk factors affecting the excess returns of direct real estate. It was found that risk factors were priced differently in the two real estate markets. The second model that relaxes the time-invariant constraints was estimated using the standard Fama-MacBeth two-pass regression technique. In the model, credit risk factor remained significant in the pricing of excess securitized real estate returns, whereas term structure risk and unexpected inflation were the two factors significantly priced in direct real estate returns. The significance of the homogeneity of various risk premia across the two markets was also tested. The tests rejected the null hypothesis of integration of the two real estate markets in both fixed and time-varying MAP frameworks. © 2004 Taylor & Francis Ltd.