Fitting of appropriate statistical model for study of growth and instability in cereal production of Odisha

Abstract

Cereals are the most important kharif season crops of odisha. Agriculture of Odisha is affected by the economic reforms in the 1991-92, following which our study period (1970-71 to 2013-14) has been divided into pre-reform (1970-71 to 1991-92) and post reform (1992-93 to 2013-14) period. Attempts have been taken to make a comparative study of the growth rate and instability of area, production and yield of rice in the two periods. For studying the growth rate and instability, appropriate model which could best possibly describe the behavior of the phenomenon are fitted by applying spline regression technique with kink or knot placed at the year of transition from pre–reform period to post- reform period which is considered to be 1991-92. Then by using the best fit model, identified with the help of residual diagnostics and model fit statistics, the average growth rates of area, production and yield are found. Coefficient of variation is used as a measure of instability. The difference in growth rates from the prereform to post–reform period is found to be negative but non-significant in case of area, negative and highly significant in case of production and significantly positive in case of yield. The difference in coefficient of variation between the two periods is highly significant and positive only for yield whereas, for both area and production, it is negative, but significant only in case of area.