Modeling and Forecasting NPR-USD Exchange Rate Volatility Using GARCH Family Models

Abstract

In terms of business, exchange rates are crucial factors influencing a nation's economic prosperity, impacting investors, government bodies, policymakers, and various other components. This research aimed to examine the dynamics of the Nepalese rupee relative to the US dollar in the Nepalese foreign exchange market, recognising that exchange rates play a crucial role in competitiveness. The primary goal of this study was to assess the applicability of GARCH-type models, including GARCH, TGARCH, and EGARCH, for modelling the NPR-USD exchange rate using daily time series data provided by Nepal Rastra Bank. The analysis compared the results with ARIMA models. The data analysed spans from January 1, 2014, to March 30, 2024, with in-sample and out-of-sample datasets covering January 1, 2014, to September 30, 2023, and October 1, 2023, to March 30, 2024, respectively. The study also involved exploratory data analysis of the variables, which underwent diagnostic tests, including unit root and normality tests. A key finding is that all GARCH-type models indicate that historical exchange rate volatility has a significant impact on current volatility. Three models were developed and tested for diagnostic accuracy, with the Threshold GARCH model demonstrating suitability and stability. The findings concluded that negative shocks have a greater effect on volatility than positive shocks. Furthermore, this methodology is recommended for future studies and can be applied for predicting exchange rate volatility in Nepal.