New Assessment of the Japanese Foreign Exchange Intervention: A Stopping Time Approach
Posted: 27 Jan 2016
Date Written: January 27, 2016
I propose a new variable to assess the effect of the Japanese foreign exchange (FX) intervention. The variable is the probability of an FX rate reaching one threshold before the other threshold is reached. Importantly, the probability depends on not only the level but also the trend and volatility of a current FX rate. When an intervention changes the probability into its intending direction, the intervention is effective. The notable feature of the probability is to consider both the level and volatility of an FX rate comprehensively, on those previous literature examine the effect of FX intervention separately.
I process the two analysis; time series and event study one. The former uses all the time series data, in which sporadic and infrequent interventions are observed. The latter uses data only from around periods of intervention.
In the time series analysis, I regresses the probability on the amounts of FX intervention. My regression result indicates that the intervention can cause a desired change in the probability. Event study processes the nearest-neighbors matching analysis. In order to implement that analysis, I compares the probability changes of the pairs. Each pair consists of two nearest-neighbors in terms of equivalent probability, and one is followed by next day’s intervention and the other is not. I expect that the former neighbor shows a desired change in the probability after the intervention, and that change is larger than that of the other neighbor. The result of the matching analysis also supports the effectiveness of the Japanese intervention although the regression and matching results are against the effectiveness in some sub-samples.
Keywords: Foreign exchange intervention; Nearest-neighbors matching analysis; Probability reaching a threshold
JEL Classification: F31; E58
Suggested Citation: Suggested Citation