Time series analysis and forecasting of the prices of Indian natural rubber

Abstract

The study entitled “Time series analysis and forecasting of the prices of Indian Natural Rubber” is primarily intended to forecast the prices for Indian Natural Rubber (NR). For forecasting the prices, firstly, domestic NR price was decomposed it into time series components. Evaluation of growth, instability and relationship between the domestic and international prices in the pre WTO and post WTO periods were carried out in this study. For decomposition of domestic NR prices into time series components, additive decomposition was tried. The data were decomposed into trend, seasonal and cyclic components. The trend values proved that there was quadratic trend over the years. Seasonality indices revealed that the highest price was in June and lowest price in December. Cyclic components showed three cycles over a period of time under investigation. For evaluation of growth and instability, volatility and instability analyses were carried out for pre-WTO, post-WTO and overall periods in terms of rupees as well as dollars. Two types of volatility i.e., intra-annual volatility (within year dispersion) and inter annual volatility (between year dispersion) were calculated. Intra-annual and inter annual volatility were higher in post- WTO for international and domestic NR price series and the crude oil price showed higher volatility in pre-WTO period in terms of rupees as also in dollars. GARCH (1,1) model gave an additional evidence for persistence of volatility. It proved that the volatility persisted in the overall period in terms of rupees and dollars for domestic and international NR price. Instability analysis showed that the price instability in post-WTO period was almost double than that of pre- WTO period and it tripled in the overall period in terms of rupees. In terms of dollars, the instability in post-WTO and overall period was almost triple than pre-WTO period for domestic and international NR prices and crude oil prices showed almost double instability than pre-WTO period. iii The relationship between domestic and international NR prices were analysed through cointergration analysis and Vector error correction model (VECM). The direction of relation was drawn by Granger Causality test. Cointegration and Granger Causality test proved that there was at least unidirectional relationship among the variables. VECM analysis proved that there was long run relationship between domestic NR price, international NR price and crude oil price. It revealed that, a speed rate of adjustment 14.3 per cent was required for domestic NR price series to correct its previous period. There were many general factors affecting the prices of domestic NR like synthetic rubber production, crude oil prices, international rubber demand and supply, international transactions, exchange rates, natural factors and development of automobile industries. Stepwise regression analysis was used to sort out the factors affected in pre-WTO and post-WTO periods. In pre-WTO, domestic NR price was affected by international NR prices and in post-WTO by international NR prices and SR consumption. Domestic NR prices were forecasted with three different models like Stepwise regression method, ARIMA and SARIMA models. Stepwise regression method could be predicted when the variables like international NR prices and import value of NR were available. Among ARIMA and SARIMA models, ARIMA (4,1,4) and (4,1,4) (1,0,1) 12 was found to be best judged as per different statistical criteria for assessing the model fit and model adequacy.