studying the efficiency of industrial dairy farms of saqqez and divandarreh cities: using super-efficiency approach

introduction: in the developed world, particularly in developing countries, livestock is the most important agricultural sub-sector.livestock of primary and secondary industries has an especial place in the national economybecause of their greatvalue of products, creating job opportunities, providing health products for consumers, increasing export income of the economy throughaccessing global markets of livestock products and finally their undeniable role in acquiring food security.the demand for milk in iran increased due to an increase in population and the amount of milk production was also increased. the great share of increased produced milk goes to the industrial dairy farms. one of the major methods to increase the amount of milk production continually is to make its production efficient and improve economic conditions. the current study attempts to determine the efficiency and ranking of industrial dairy farms in saqqez and divandarreh cities using super-efficiency model. materials and methods: the statistical populations of the study are all active industrial dairy farms of saqqez and divandarreh cities which are about 19 farms. the required data for calculating the efficiency were gathered by surveying and completing questionnaires for the year 2013. in this study first, for each farm data envelopment analysis (dea) method and gams software package were used to estimate super efficiency. super efficiency is a form of modified dea model in which each farm can get an efficiency greater than one. then in order to make sure about being unbiased the obtained super-efficiency scores, the modified model of banker and gifford, was re-estimated and the conventional efficiency scores of farms were compared by normalizing and removing some of the scores of outlier farm based on pre-selected screens. the model has suggested conditions for which some of the estimates for dairy farms might have been contaminated with error.as a t result, it has been ranked as an efficient farm. results and discussion: the statistical description of the farms studied showed that, the highest and lowest amount of produced milk from one dairy were in the range of 845 and 250 kg per month. the super efficiency estimation showed that the mean of farms'super efficiency based on the assumption of variable return to scale for input-oriented is 1.01. about 58 per cent of the studied dairy cattle farms were inefficient and the super-efficiency of about 42 per cent of total farms got scores under the average. the amount of λ_k^* for all farms except for farms 3 and 5 has become 0. furthermore, these two farms in primitive super efficiency model have become infeasible. therefore, they have been considered as reference farms. the results of super-efficiency method and efficiency conventional dea method were compared together and inefficient farms got the similar efficiency and super-efficiency scores, and efficient farms whose super-efficiency score had been equaled or more than 1, stand over frontier production function in the conventional model. to determine the sensitivity of results to removing the outliers, three different levels (1<, 1.5≤, 1≤), were considered as base to identify outliers. then, in the second stage the conventional bcc model was estimated for the remaining farms. the results obtained showed that, before (super efficiency) and after (conventional efficiency) removing outlier farms, the calculated efficiency mean for industrial dairy farm at1< and 1.5≤levels did not show a significant difference to each other due to a high correlation coefficient. the correlation coefficient was low only at 1≤ which indicates outlier effects on efficiency scores. also about 42 per cent of the farms have been identified at 1≤ outlier and after removing these farms the mean of super-efficiency scores has become about 1.22 so, they have increased by 28 per cent withrespect to the initial super-efficiency score. in this state, 6 farms have obtainedsuper-efficiency scores more than the average. conclusion: the suggested model in this study which followed the banker and gifford super-efficiency method may help to find and remove those farms that have been contaminated with error and then earn more reasonable scores than the conventional dea method. so, these farms in determiningthe level of milk production can project on to frontier production function and become efficient by decreasing using inputs. on the other hand, recognizing and removing outlier farms decreasesor increasesthe amount of biased results and confidence coefficient of calculated super-efficiency scores, respectively. this result can give the right information to the managers who wish to follow the correct way of managing their farms.