This paper estimates the effects of different forms of inflation targeting (IT) in the exchange rate pass-through (ERPT). To this end, we first estimate the ERPT for a large sample of countries using state-space models. We then consider the adoption of an inflation targeting framework by a country as a treatment to find suitable counterfactuals to the actual targeters. By controlling for self-selection bias and endogeneity of the monetary policy regime, we confirm that the ERPT tends to be lower for countries adopting explicit IT. However, we uncover that older regimes, adopting a range or point with tolerance band and keeping inflation close to the target, outperform other IT regimes. We also show that IT is effective even with a relatively high inflation target or low central bank independence.