Nepalese Stock Market Volatility Using ARCH and GARCH Models

Abstract

This article analyzes the volatility process of the series of returns of Nepal Stock Exchange (NEPSE) through time series econometric models. This article applies ARCH (2) and GARCH (1, 1) on NEPSE index from 2020 to 2025, which is on a daily frequency. Stationarity tests through Augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) tests make the return series stationary. ARCH tests determine the presence of autoregressive conditional heteroskedasticity effects as an attempt to validate the presence of long-run clustering volatility. ARCH (2) and GARCH (1, 1) describe volatility. In ARCH (2), it is proved that previous squared errors form an adequate description of current volatility across time in such a manner that recent shocks matter more than the previous shocks. The GARCH (1, 1) model also indicates high volatility persistence, wherein future market movement is significantly dependent on previous volatility. The results suggest that returns in NEPSE are volatile clustered, i.e., there will be continued high volatility. The findings are of significant policy implication for risk-averse investors and policy makers, wherein there is a need to apply suitable risk management measures in the Nepalese stock market.