The Bennet Decomposition and Predictability of the U.S. REITs’ Profitability

This paper examines empirically the predictability of operating profitability and whether any observed predictability stems from the asset or debt management policies of a portfolio of REITs. Return on assets (ROA), return on equity (ROE), Change in ROA and Change in ROE are the profitability measures of the sample portfolio, which covers, on average, about 84% of U.S. REITs included in the FTSE NAREIT All Equity Index between 1989 and 2015. While the asset management policies of sample REITs engenders ROA and Change in ROA, their asset and debt management policies jointly engender the ROE and Change in ROE. Our empirical work focuses on the coefficient estimates of (i) the own lags of each of these four profitability measures, and (ii) the lags of the “between,” “within,” “entry,” and “exit” effects, obtained from the first-ever application of the Bennet (1920) dynamic decomposition to the temporal changes - between (t) and (t-1) - in the ROA and ROE of the sample portfolio. A comparison of the estimates -- between the ROA and ROE as well as between the Change in ROA and Change in ROE estimations -- in (i) and (ii) provides evidence about the root of the predictability. Our work repeats all the estimations above under the funds from operations (FFO) and net income (NI) metrics, which are used in computing the ROA and ROE measures and also their temporal changes. A comparison of the FFO- and NI-based results at the portfolio level is important since there is a growing literature and debate on whether the information content of FFO differs incrementally from that of NI. We find that (i) the predictability of profitability of the sample portfolio of REITs is highly visible and statistically strong; (ii) the estimates of the first own lags of the dependent variables or the first and second lags of some of the Bennet (1920) dynamic decomposition effects - especially the “within” effect - provide strong evidence of predictability; and (iii) the use of FFO unearths evidence that the sample REITs’ asset management policies, as embodied in ROA and Change in ROA, have more to do with predictability than a combination of their asset and debt management policies, as embodied in ROE and Change in ROE, does. These findings should be useful to the investors and REIT managers and the REIT literature.