بررسی بهره‌وری کل عوامل تولید برنج در استان‌‌‌های مختلف: کاربرد تابع تجزیه‌پذیر مرزی تصادفی

Introduction: In this study, total factor productivity growth components in rough rice production are assisting by econometric approach and stochastic frontier function during 2005-2013 for several provinces (Mazandaran province, Guilan province, Golestan province, Fars province, Khoozestan province). According to Food and Agriculture Organization statics, Iran is the 3rd importer and 20th exporter of rice in the world. But, during the study years (2005-2013), Iran has been one of the 6 largest importers of these products that on average about 33 percent of domestic needs are provided by imports. Annual per capita consumption of rice during 1990-2012 had been changed from 38.6 to 43.9 Kilogram. So, it can be concluded that rice has a special place in the Iranian consumption bundle. But in the production sector, cultivation area has been decreasing 15 percent during 2005-2009 and was fixed during 2009-2013. These matters indicate that domesticate production cannot provide domesticate consumption. One of the suitable ways of increasing production is increasing in total factor productivity. This strategy is needed to identify components of TFP growth sources. So, the main goal of this study is the decomposition of rice TFP in Iran.
Materials and Methods: TFP decommission growth can be analytically by four approaches included econometric estimation of production and the cost function, TFP indices of Divisia and Turnqvist, Data envelopment approaches (DEA) such as Malmqvist and stochastic frontier analysis. This study uses a stochastic frontier analysis to decompose total factor productivity (TFP) growth into four components: technical change (TC), technical efficiency (TE) change, scale efficiency (SE) change, and allocative efficiency (AE) change. For this new approach, at first, Translog production function is estimated by gathering data. So, by estimation of Translog production function, total factor productivity growths are decomposed to TC and changes in TE, SE, and AE. For the total factor productivity decomposition, we employ the time-varying model for technical inefficiency. Firm inefficiency is assumed to be distributed as a generalized truncated–normal random variable which is distributed independently of the normally distributed random errors.
Results and Discussion: Results indicate technical efficiency have been 0.86, 0.79, 0.69, 0.73 and 0.86 for Mazandaran, Guilan, Golestan Khoozstan and Fars provinces, respectively for the year of 2012. That is, most technical efficiencies were for Mazandaran and Fars provinces. Also, technical efficiency has been 0.73, 0.75, 0.77, 0.81 and 0.81 for a farm with size less than 0.5 Ha, between 0.5 and 1 Ha, between 1 and 2 Ha, between 2 and 3 Ha and more than 3 Ha, respectively for years of 2012. That is, most technical efficiencies were from a farm with the size of more than 2 ha. The annual growth rate of technical efficiency changes during 2005-2013 have been 2.3, 1.6, 0.3, 0.9 and 1.6 percent for Mazandaran, Guilan, Golestan Khoozstan and Fars provinces, respectively. For Iran, also has been 1.5 percent. The annual growth rate of scale efficiency change during 2005-2013 have been 1.5, 1, 1, 1.2 and 2.3 percent for Mazandaran, Guilan, Golestan Khoozstan and Fars provinces, respectively. Also, for Iran it has been 1.9 percent. Annual growth rate of Allocative efficiency change during 2005-2013 have been 0.01, 0.6, 0.3, 0.5 and 0.5 percent for Mazandaran, Guilan, Golestan Khoozstan and Fars provinces, respectively. Also, for Iran it has been 0.8 percent. Finally, annual growth rate of TFP change during 2005-2013 have been 4.8, 3.8, 2.05, 3.1 and 4.7 percent for Mazandaran, Guilan, Golestan Khoozstan and Fars provinces, respectively. For Iran, also has been 4.3 percent. The most and least growth were for Mazandaran and Golestan Provinces. Differences in rough rice total factor productivity growth rates in the provinces were found to be explained primarily by differences in scale efficiency and technical efficiency. Scale elasticities for a year between 2005 and 2013 were between 1.13 and 1.12 for Mazandaran, between 1.12 and 1.13 (with fluctuation) for Guilan, between 1.14 and 1.13 for Golestan, between 1.18 and 1.19 for Khoozestan and between 1.14 and 1.18 for Fars. So, scale elasticities average between the sum of farms (between 1.12 and 1.18) shows that economics of scale exists in rough rice production technology. Scale elasticities for a year between 2005 and 2013 was between 1.24 and 1.27 for farm with size of less than 0.5 ha, between 1.17 and 1.22 for farm with size of between 0.5 and 1 ha, between 1.0.8 and 1.18 for farm with size of between 1 and 2 ha, between 1.05 and 1.13 for farm with size of between 2 and 3 ha and between 1.01 and 1.09 for farm with a size of more than 3 ha.
Conclusions: With the assumptions that rough rice production technology is similar in all provinces, approximately, differences between provinces in scale elasticities are about the size of the farm. That is, smaller farms in comparison with larger farms have more economics of scale. Finally, it can be noted that by increasing in size of farms, we can increase technical efficiency and TFP of rice production.