Factors Affecting the Growth of Iran's Agricultural Sector: Applying the Bayesian Model Averaging Approach

The wide range of factors influencing growth in theoretical foundations and empirical studies and the weakness of conventional methods have led studies to focus on only one aspect of theoretical and empirical growth patterns. This gives rise to uncertainty about specifying or combining variables in the model and estimated coefficients. This uncertainty can lead to bias and inefficiency in estimating the coefficients resulting in inaccurate predictions and inaccurate statistical inference. Therefore, in this study, using the Bayesian averaging method, the influence of the most important factors affecting the growth of Iranian agricultural sector during 1978-2017 was investigated. Using this approach, all possible sub-models are estimated using study variables and then the coefficient of each variable is averaged across the models. The weights in this averaging are determined by the Bayesian rule or the posterior probability of each pattern. In this study, 2048 different models were estimated. The results showed that investment, financial development and oil revenues with the probability of impact of 0.81, 0.67 and 0.42, respectively, are the most important variables affecting the growth of agricultural sector and also the growth rate of agricultural imports with a probability of impact of 0.90 had the most negative effect on the growth of value added. Therefore, investing and financing producers, paying attention to domestic production and setting trade policies on imports should be a top priority in policy making and planning.