Analysis of Price Behavior in Sri Lankan Vegetable Market

Abstract

Vegetables are important source of nutrient for the Sri Lankan population and both farmers and consumers are adversely affected by vegetable price volatility. The lack of price analysis and forecasting has made it difficult to establish an effective early warning system for the vegetable farming sector in Sri Lanka. This study investigates the price behaviour of selected fresh vegetables - carrot, cabbage, and tomato - and forecasts the future prices and volatilities using time series techniques. Analysis of weekly price data from 1997 to 2018 revealed that all three - price series had one structural break, but none coincided with the policy change when the government introduced fertilizer subsidies for vegetable producers in the agriculture sector. The autoregressive integrated moving average (ARIMA) model estimations show that the best model for forecasting carrot price is ARIMA (3,1,2) (0,0,2)[52]* capable of predicting retail prices at 71% accuracy while the best model for cabbage prices is ARIMA(1,1,1)(0,0,1)[52] with a prediction accuracy of 55%. All three-price series exhibit serial correlation in residuals; hence GARCH estimations were used to model and predict volatility. Of the fitted ARMA GARCH models, the best model for estimating the volatility of carrot and cabbage were GARCH (1, 2) ARMA (3, 2) and GARCH (1,1) ARMA (3 ,2), respectively. The volatility predictions for the first ten weeks for the year 2019 indicate a gradual decrease in volatility in the carrot price series whilst a gradual increase in volatility in the cabbage price series.