Are we overestimating REIT idiosyncratic risk? Analysis of pricing effects and persistence

This study uses a multifactor REIT-specific model to estimate and compare REIT idiosyncratic volatility vis-à-vis the same from the Fama–French three-factor model. Estimates of conditional idiosyncratic volatility and conditional betas obtained from a multifactor REIT-returns model and a bivariate EGARCH model respectively are found to be positively and significantly related with REIT returns. Consistent with Merton's (1987) predictions, we observe that larger REITs post higher average returns when idiosyncratic risk is introduced in cross-sectional regressions. Persistence of past market-risk does not appear to be short-lived and seems to have a lasting impact on future idiosyncratic volatility. We also observe mild evidence of persistence of past idiosyncratic risk, albeit short-lived, thereby suggesting that past idiosyncratic risk has a short-term impact on future idiosyncratic risk.