22 , 2018-03 , Faculty of Economics and Business, Hokkaido University
This study analyzes the effect of a central bank’s intervention in stock markets, while allowing for nonlinearities and state dependencies, using a semi-parametric approach. A causal inference on such intervention is difficult because of the selfselective behavior of central banks. To address these problems, we apply the propensity score method in a time series context, exploiting stock price information of a single day. We find that first, there are demand pressure effects in stock markets if an intervention is large enough. Second, the effects are state-dependent and stronger during market downturns. Finally, a central bank’s interventions have a considerable impact on stock prices only when we take permanent demand pressure effects into consideration.