Modelling and Forecasting of Asymmetric Price Volatility of Onion for Lucknow Market, UP, India

Abstract

Onion prices in the Lucknow market exhibit very high instability, or in other words, volatility of UP. When price volatility affects differently with positive and negative shocks of the equivalent size, it is said to be asymmetric. Since GARCH is a symmetric model, it will be unable to account for asymmetric volatility in prices. EGARCH, GJR-GARCH and APARCH models are popularly used to capture asymmetric price volatility. The present study aimed to model and forecast the price volatility of the monthly modal prices of onion for the Lucknow market of UP. The study is based on the secondary time series data on the monthly price of onion from January 2007 to December 2021. Augmented Dicky-Fuller (ADF), Philips Peron (PP), and Kwiatkowski-Phillips-Schmidt-Shin (KPSS) tests have been used for testing the stationarity of the series. The best ARMA model has been selected for the individual series after confirming the stationarity of the series. Residuals have been examined for the presence of autocorrelation, heteroscedasticity, and nonlinear dependence in them, and it has been found that residuals have all those properties verification has been done using Ljung-Box test, ARCH-LM test and BDS test, respectively. Upon analysis of data, the ARMA (1,0)-APARCH (1,1) model outperformed the other forecasting model, and it is deemed to be the best fit model for the data under consideration. The R software version 4.2.3 has been used for data analysis.