A Kalman Filtering Approach to Event Study Analysis when Performance Variables are Nonstationary
A Kalman filtering regression model is proposed to resolve nonstationarity problems commonly found in certain performance variables, e.g., trading volume, of event study analyses. Resolution is possible when the expected performance variables are allowed to move according to random walk processes. The model can be used for cases in which performance variables have deterministic or stochastic trends. The model is applied to examine the trading turnover behavior in the Thai stock and bond markets in the time around the military coups of 2006 and 2014. The model is successful; it passes validity tests, namely, the nonstationarity and parameter constancy tests. The findings suggest that the results reported by previous studies that failed to treat the stationarity problems are misleading.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.