Heterogeneous Market Efficiency in Cryptocurrency Markets: A Multi‐Frequency Memory‐Based Approach

ABSTRACT Empirical cryptocurrency researchers frequently use concepts of mean reversion, trend and cointegration to characterise price dynamics and tests of market inefficiency. This paper introduces a broader memory‐driven framework to examine how the concept of market efficiency has evolved over time, especially by characterising varied mean‐reversion strategies with slow‐paced error corrections to identify a semi‐strong market efficiency in crypto markets. We find strong evidence that market efficiency in Bitcoin (BTC) and Ethereum (ETH) is heterogeneous and time‐varying across all frequencies. Event shocks and structural breaks exert substantial influence on abrupt changes in efficiency, while results at the 240‐min sampling interval show robustness. The findings suggest that policymakers should time interventions to mitigate lag effects, release policy during low‐leverage periods and closely monitor the two distinct waves of efficiency breaks as well as cross‐frequency risk exposure. For practitioners, linking efficiency dynamics to trading strategies can enhance risk management and statistical arbitrage opportunities, while scholars can build advanced regime‐switching models informed by these empirical patterns. This paper provides new insight into price patterns and market efficiency dynamics, contributing to the understanding of the complex interplay between event shocks, structural breaks and regime switching of market efficiency in cryptocurrency markets.