Stochastic Frontier Model with Distributional Assumptions for Rice Production Technical Efficiency

Abstract

Efficiency in agricultural production is indicative of the efficiency level of farm households in their farming activities. Farmers in developing countries do not make use of all the potential technological resources, thus making inefficient decisions in their agricultural activities. Herein, technical efficiency in relation with the production of three types of rice crop (Boro, Aus and Aman) was evaluated, with some determinants of technical efficiency identified, in Bangladesh. It was attempted, throughout this study, to access the status of technical efficiency in rice production in Bangladesh for panel data while using the Stochastic Frontier Production Model with either of truncated normal or half-normal distributional assumptions. Both time-variant and time-invariant inefficiency effects models were estimated, one at a time. Collected data from agricultural sector pertaining to three main rice crops in Bangladesh for the period of 1980 to 2008 were made used of throughout the study. The results revealed that technical efficiency gradually increased over the reference period with the half normal distribution being found preferable to the truncated normal distribution as regards the technical inefficiency effects. The value of technical efficiency was found high for Boro rice while low for Aus in comparison with Aman rice in Bangladesh for both distributions in either of time-variant or invariant ones. It was observed that the most efficient rice production system has occurred for the case of Boro with a technical efficiency of 0.98. Yearwise mean technical efficiency increased during the reference time periods. Keywords: Bangladesh rice production, Cobb-Douglas stochastic frontier model, Panel data, Time variant and Time invariant efficiencies.