Do Actions Speak Louder than Words? A Foreign Exchange Intervention Analysis

Abstract

We revisit an old question but with a new identification strategy, namely the difference in exchange rate effects between announced (“vocal”) and secret (“dirty”) foreign exchange intervention. Using a Regression Discontinuity Design, we exploit a rule-based intervention mechanism enacted by the Central Bank of Colombia that, under observable and deterministic conditions, triggered either the issuance of FX options or the ability to exercise them. We take the former (issuance) as central bank announcements under a sharp setting, since the rule and information that triggered the issuance of options was public, and we take the latter (exercise) as secret trades under a fuzzy setting, since traders could have chosen (but were not required) to exercise their options in the following days after issuance. Our results indicate that, unconditionally, both announcements and secret trades carry similar effects. However, the effects of announcements are considerably amplified conditional on: (i) higher central bank credibility, (ii) less frequent announcements, and (iii) episodes of higher FX volatility.