THE KALMAN FILTER IN THE EVENT-STUDY METHODOLOGY

Abstract

The purpose of this paper is to extend the event-study methodology, into a richer dynamic environment, by including time-varying parameters. We use the Kalman filter to model parameters depending on time in a state-space representation of the statistical market model of the event-study analysis. We also apply Bayesian inference to updating relevant information, and we use information theory to choose the initial distribution of parameters. The proposed extension leads to a more robust set-up in appraising the impact of economic, and financial events on the market value of firms.