本文的目的,是透過經濟史料的考察以及實證分析,對長久以來一直被普遍認為在分益式(sharecropping)租佃契約下的農場經營效率會較自耕農場經營效率來得差的觀點進行討論。首先,從經濟史料的考察中發現,臺灣在1940年以前,普遍存在的租佃契約型態是定額租(fixed rent)制而不是分益租制。但是,在日本殖民政府並未干預租佃市場運作,使得佃租額會隨著土地生產力等因素的變化而調整的情況下,定額租制與分益租制基本上並沒有明顯的區別。接著本文採用臺灣總督府殖產局於1926年對稻作農場的經營調查資料,應用資料包絡法(DEA)對稻作農場的經營效率進行實證分析,並且利用jackknife抽樣方法檢測界外值避免其對實證分析造成影響。經由對自耕稻作農場與佃耕稻作農場經營效率的比較發現,其結果並不支持在分益式租佃契約下的農場經營效率較差的論點。
Based upon examination of historical economic data and empirical analysis, this paper discusses the issues whereby the performance of the sharecropping tenant system is worse than the owner farming system. From investigation of historical economic data, we found that there are not any differences between fixed rent and sharecropping tenancy. The Data Envelopment Analysis (DEA) is adopted to measure the technical efficiency of paddy farms by using official data. In order to reduce the impact of outliers, we employ jackknife-sampling techniques. The result of the empirical study cannot sustain the viewpoint that the sharecropping tenant system efficient.