Evaluating foreign exchange market intervention: Self-selection, counterfactuals and average treatment effects

Studies of central bank intervention are complicated by the fact that we typically observe intervention only during periods of turbulent exchange markets. Furthermore, entering the market during these particular periods is a conscious self-selection" choice made by the intervening central bank. We estimate the counterfactual" exchange rate movements that allow us to determine what would have occurred in the absence of intervention and we introduce the method of propensity score matching to the intervention literature in order to estimate the average treatment effect" (ATE) of intervention. Specifically, we estimate the ATE for daily Bank of Japan intervention over the January 1999 to March 2004 period. This sample encompasses a remarkable variation in intervention frequencies as well as unprecedented frequent intervention towards the latter part of the period. We find that the effects of intervention vary dramatically and inversely with the frequency of intervention: Intervention is effective over the 1999 to 2002 period, ineffective during 2003 and counterproductive during the first quarter of 2004.