This study empirically verifies the evolving and time-varying efficiency of Indian foreign exchange market using the framework of adaptive market hypothesis (AMH). Whether market efficiency is time varying or static, and if time varying, identification of possible events causing such time-varying efficiency are the two major agenda of this study. We employ a set of recent methods which are robust and possess stronger power properties. Moreover, we follow a fixed-length rolling window approach to explore time-varying nature of market efficiency and to avoid data-snooping bias. Our overall findings suggest that market efficiency is not an all-or-nothing condition; it varies over time. We also find that episodes of efficiency coincide with emergence of major events and market microstructure issues. Particularly, changes in exchange rate regime, financial turbulence, major central bank interventions and trade volume are the prominent causes for time-varying efficiency in INR–USD exchange rate. The evidence of swing between efficiency and inefficiency can prompt currency traders to exploit arbitrage opportunities that emerge with different market conditions.