Author(s): Gadhe Kiran and TD Pansara

Abstract: The agricultural sector faces persistent challenges due to price volatility in commodity markets affecting a wide range of commodities both domestically and internationally. Market intelligence has become crucial for farmers in decision-making processes related to agricultural production and marketing. This study focuses on cumin, a significant spice in Gujarat, analyzing price trends over sixteen years (January 2006 to December 2021) in markets of Unjha, Jamnagar and Gondal. Using data primarily from Agmarknet various forecasting models including ARIMA, SARIMA and ARCH/GARCH were employed with model selection based on AIC and MAPE criteria. The study found that different models performed best in different markets. For Unjha, the ARIMA(0,1,0) model was optimal while Jamnagar showed better results with the ARIMA(2,1,0) model and Gondal performed well with the ARIMA(0,1,0) model. Additionally, SARIMA (1,0,0) (1,1,2) was effective in Unjha, SARIMA (1,0,0) (0,1,1) in Jamnagar and SARIMA (2,0,1) (3,1,0) in Gondal. The ARCH/GARCH model revealed superior performance with ARMA (1,0)+GARCH(2,1) in Unjha and ARMA (1,0)+GARCH(2,2) in both Jamnagar and Gondal markets. Among the models, ARIMA demonstrated the best forecasting accuracy with MAPE values of 22.92%, 38.52% and 16% for Unjha, Jamnagar and Gondal, respectively. Following the selection of the most suitable ARIMA models for each market, these models were applied to forecast future price of cumin for selected markets. These findings provide valuable insights for farmers and stakeholders offering a robust policy instrument to navigate cumin price fluctuations. This research contributes to effective agricultural policies, assisting farmers in optimizing their produce sales and market strategies amid price volatility.

Pages: 18-26 | Views: 769 | Downloads: 9

Download Full Article: Click Here