부트스트래핑과 절단 회귀(truncated regression)를 이용한 과수 농가의 기술효율성 분석 -Tobit 모형(censored regression)과의 비교-

This paper analyzes the effect of the land size and the family labor hours on the technical efficiencies of Korean fruit (apple, pear, grape, and peach) farms which are measured with Data Envelopment Analysis. It follows the estimation method suggested by Simar and Wilson(2007) by using truncated regression, (instead of the censored regression model which are prevalent in the empirical works) and bootstrapping, in order to adjust for the bias of the simple DEA and to estimate the confidence interval of the estimates, taking into account the serial correlation of the error terms and their correlation with the explanatory variables which are intrinsic to DEA estimates. The estimation results show that the family labor hours enhance the technical efficiency of grape growing. They also suggest that the choice between the censored regression and the truncated regression model may have important consequence on the estimation results.