Analysis of the relationship between Cotton Production and Prices by using Koyck Approach in Turkey

Abstract

In this study, the relationship between the amount of cotton production and real cotton prices was investigated for Turkey in the framework of Koyck approach, one of the distributed lag models. For this purpose, data for the 19 cotton grown provinces of Turkey and panel data methods were used. Cotton real prices were calculated using the nominal cotton prices of each province and the consumer price index of Turkey in general. Panel data crosssectional dependency analyses show that there are positive correlations between both the amount of cotton production and real price series of different provinces. This correlation was found to be 21.6% for cotton production and 90.4% for real cotton prices. Findings of panel unit root tests supported that cotton production and cotton real price series were stationary. This result supports that both series return to their long run levels after external shocks. Taking into account the problems related to the basic assumptions of the fixed effects model, the cotton production function was estimated in the context of an approach, which is based on robust standard errors and provides efficient results in the presence of autocorrelation, heteroscedasticity and cross-sectional dependence. The findings of the panel data fixed effect model estimated within the framework of the Koyck approach showed that a 1% change in the real cotton price affected cotton production in the same direction by 0.31% in the same year. Furthermore, this change in real cotton price affects cotton production by 0.26% and 0.22% after 1 and 2 years, respectively. Finally, the average time required to reflect the real cotton price changes on the production was calculated as 4.99 years.