Econometric modelling for harvested area, production and productivity of natural rubber (Hevea brasiliensis)

Abstract

The objective of present investigation was to study the trends in harvested area, production and productivity, and growth rates of natural rubber (Hevea brasiliensis) in Tamil Nadu from April 1991 to March 2016 by employing parametric models. The annual time-series data for the period from April 1991 to March 2016 was used for the study. The analysis was carried out during the year 2016-17. Different linear, non-linear and time-series models of parametric models were used. The statistically best fitted parametric models were chosen based on adjusted R2, significant regression co-efficients and co-efficient of determination (R₂). After analysing the data for stationarity, the suitable time-series models were chosen based on various goodness of fit criteria, viz. Bayesian information criterion, Akaike’s Information Criterion, Root Mean Square Error, Mean Absolute Error and assumptions of normality of residuals. Trends in harvested area, production and productivity were analysed based on time series models using ARIMA. Among them, the best model was extracted and tabulated. Natural rubber production increased at the rate 1.31% per annum during the period, which was a result of the combined effect of rise in harvested area and productivity at a rate of 1.09% and 0.91% per annum, respectively.