The effectiveness of currency intervention: Evidence from Mongolia

Abstract

There is a long-running debate on the effectiveness of currency interventions with the issue of how to overcome the two empirical manifestations of the problem of endogeneity (simultaneity and self-selection bias), a constant challenge in this literature. To address this problem of endogeneity, we employ a recent extension of the potential outcomes approach of micro-econometrics to a time-series setting. Another crucial contribution of our study is that we made every effort on a high-frequency basis in controlling for the effects of monetary policy on the exchange rate, a crucial aspect often neglected in this literature. We conduct our analysis using unique daily data on currency interventions in Mongolia to examine the impact of these interventions on the changes in the MNT/USD exchange rate and its volatility. Our results confirm some of the previous findings from the literature, but also offer some new key findings. Both purchases and sales of US dollars are effective in moving changes in the MNT/USD in the desired direction when implemented frequently, and the effects are quite persistent. Mainly sales of USD when also implemented frequently can help reduce the volatility in the MNT/USD exchange rate. Although, our full sample estimates using quantile local projections show that both sales and purchases of USD can reduce volatility at low and moderate levels, while frequent and smaller sales of USD can help reduce high volatility in the exchange rate.