Hide/Show Apps

Time varying beta estimation for Turkish real estate investment trusts(REITs): an analysis of alternative modeling techniques

Altınsoy, Gözde
This study investigates the time varying behavior of the betas (systematic risk) for the Turkish REIT sector in an attempt to identify whether the betas for the Turkish REITs are stable and if not whether the declining trend valid for the REIT betas of many developed and developing countries is also observed for the Turkish REITs. Three different techniques; namely, Diagonal BEKK (DBEKK) GARCH model, the Schwert and Seguin model and the Kalman Filter algorithm, are employed in order to estimate and analyze the time varying betas of the Turkish REIT sector over the period 2002-2009. The empirical results suggest that, similar to many other countries, betas are not stable in the Turkish REIT sector. The general view of a declining beta trend for the REITs appears to prevail for Turkish REITs as well, reinforcing the defensive characteristics of these publicly traded real estate companies. Comparing the relative forecast accuracy of the three techniques employed, Schwert and Seguin model performs the worst both for weekly and daily data; whereas the Kalman Filter and the DBEKK Garch models provide the lowest forecast errors for the weekly and the daily data, respectively. This study also shows that the use of the data sets with different frequency could lead to different empirical findings.