Files
Abstract
Rainfall and temperature are the two important weather factors that affect crop yields due to their direct
and indirect influences on agricultural practices. This study has negated the method of direct use of
meteorological factors (either monthly or seasonal), in multiple regression analysis to measure weather
impact on crop yield where rainfall and temperature are incorporated in the model as increasing monotonic
functions of yield. With evidences from Odisha, where agriculture is rainfed and weather-dependent, the
study has advocated the incorporation of ‘aridity index’ variable in the regression model. The use of
composite aridity index variable in econometric model has made the analysis more easy and logical.
More importantly, the use of aridity index saves the ‘degrees of freedom’ which is very crucial in
econometric analysis. In addition, the ambiguity of using the linear trend to proxy for technological
progress is taken care of adequately by using cubic function of time. The testing of hypothesis of changing
rainfall dependency has established the fact that the dependence of agriculture on rainfall in Odisha has
declined slightly possibly because of the developments in irrigation and other facilities.