Time-Varying Efficiency and Volatility Regimes in Nepal Stock Exchange (NEPSE): Evidence from Daily Data (1995-2025) under the Adaptive Market Hypothesis

This study evaluates the Adaptive Market Hypothesis (AMH) in the Nepalese stock market using daily index returns from July 1995 to February 2025, representing nearly three decades and the longest high-frequency analysis of the market to date. Comprehensive efficiency diagnostics reveal clear time-varying dynamics: persistent inefficiency from 1999 to 2019 is followed by emerging signs of efficiency from 2021 to 2025, coinciding with financial digitization and regulatory reforms. Employing GARCH and Markov-switching models, the results indicate highly persistent volatility (α + β ≈ 0.90) and the presence of two distinct volatility regimes. Low-volatility states persist for approximately 15 trading days, whereas high-volatility episodes last around 9 days, and are associated with major macroeconomic shocks, including the 2015 earthquake and the COVID-19 pandemic. Unlike developed markets, NEPSE does not exhibit a statistically significant leverage effect (γ₠= 0.0279, p = 0.120), suggesting symmetric volatility responses to positive and negative shocks. These findings contradict the assumption of static market efficiency and provide empirical evidence in support of the AMH's prediction of evolving efficiency in emerging markets. The results carry important implications for regime- dependent investment strategies and countercyclical regulatory policies during high-volatility states.