Non-linear properties of nepalese capital market: A multifractal detrended fluctuation analysis approach

A convergence of behavioral dynamics (herd behaviour) and institutional constraints (thin trading) is hypothesized to generate the multifractal properties of the Nepalese stock market (NEPSE). To test this hypothesis, this paper employed Multifractal Detrended Fluctuation Analysis (MFDFA) using the daily NEPSE index from July 20, 1997, to March 30, 2024. First, shuffling and surrogate data procedures were employed to empirically distinguish the source of multifractality. Both long-range correlations (memory) and flat-tailed distributions (shocks) were found to contribute to the market's inefficiency. Second, the Bai-Perron test was employed to confirm the market's nonstationary nature due to the massive surge in retail investors after 2020. Finally, a dynamic sliding-window analysis reveals that market inefficiency is highly time-varying. The generalised Hurst exponent h(q) shows a clear dependence on q, confirming that the NEPSE return series is multifractal. The Hurst exponent ranges from 0.903 for small fluctuations to 0.460 for large ones, indicating the presence of both long-term memory and short-term dependencies. The width of the multifractal spectrum is 1.146, and the difference in Hurst values (Δh = 0.443) suggests strong multifractality. The capacity dimension D (0), information dimension D (1), and correlation dimension D (2) further support the complex structure of the market. The findings suggest that the NEPSE is inefficient, behaves differently across time scales and is influenced by both minor and extreme events, making it less predictable. The structural break analysis and Sliding Window MFDFA provide evidence that the timing of the surge of retail investors coincides with altered market dynamics and increased inefficiency. These findings offer a quantifiable relationship between behavioral shifts and market stability.