Ginger price dynamics in the Eastern Himalayan Region: A case study of Meghalaya

Ginger contributes substantially to the agricultural sector in Meghalaya, yet comprehensive forecasting analyses that integrate both traditional and volatility-sensitive time series approaches to study its price movements are still scarce. In this study, we explore common time series yet very powerful forecast models, namely the autoregressive moving average (ARMA), autoregressive integrated moving average (ARIMA) and ARIMA with exogenous inputs (ARIMAX), along with autoregressive conditional heteroscedasticity (ARCH) and generalized autoregressive conditional heteroscedasticity (GARCH) models, to forecast the monthly price of ginger in Meghalaya. During the estimation process, attention was given to the intrinsic forecasting strengths and limitations of these models, as well as to essential time series diagnostics, including stationarity, parsimony and overfitting. The discussion in this study is based on the forecast results obtained from a real-time monthly price dataset spanning 10 years. While fitting the models to the dataset, special care was taken to select the most parsimonious model. To evaluate forecast accuracy and compare the performance of the different models applied to the time series of monthly ginger prices, we used seven forecast performance measures: ME (mean error), RMSE (root mean square error), MAE (mean absolute error), MAPE (mean absolute percentage error), MASE (mean absolute scaled error), AIC (Akaike information criterion) and BIC (Bayesian information criterion). The GARCH (1,1) model outperformed the others, yielding the lowest MAE (1,212.28), MASE (0.2817) and a high persistence in volatility (β₁ = 0.99772). The average price during the study period was ₹4400.47. The forecasts indicated a decline in prices from ₹11910 in March 2024 to ₹10938 in July 2024.