On Stochastic Volatility and More Powerful Parametric Tests of Event Effects on Unsystematic Returns

Econometricians often emphasize that the use of a larger information set results in better parameter estimates and stronger hypotheses tests. The use of information on the stochastic behavior of the volatility of asset returns results in the formulation of more powerful parametric tests of the impact of a certain event (stock split, corporate restructuring, change in regulation, etc.) on assets' unsystematic returns. The key assumption in this study is that return volatility follows a mean-reverting diffusion whose discrete-time filter is a GARCH model. Using test statistics derived under this more general system of stochastic prices and volatility results in up to 18% higher rates of rejection of the false null hypothesis than the rejection rates of the pre-existing parametric tests. At the same time, the true null is rejected at the correct levels. The methodology is also applied to corporate spin-offs, resulting in some findings at variance with those obtained using the traditional test