台灣之林場經營效率研究-兩階段資料包絡分析之應用

Abstract

本研究旨在探討台灣林場之經營效率。本文首先透過文獻回顧探討以往在衡量經營效率的研究上所使用的方法，而後利用兩階段資料包絡分析法針對2005年台灣林業普查資料庫內林場的資料進行實證，首先在第一階段利用資料包絡分析法計算出各個林場的效率值。進一步使用差額變數分析與敏感度分析尋求林場績效的改進方向。第二階段以DEA效率值為依變數，針對經營組織、管理者特性與土地特性等各項環境變數進行截斷式迴歸分析。 研究結果顯示，台灣的林場經營效率普遍來說並不佳，經營效率平均值約0.34，代表其可節省66%的投入項以達到相同產出。總體而言，所有的林場皆需改善其林產品的銷售，對小規模的林場而言，除了需要改進林產品銷售外，林副產品的銷售也需改進；對大規模的林場而言，則還需減少林地產能閒置的情況。而在截斷式迴歸部分，結果顯示經營者的教育程度、地區以及無兼營其他產業這些環境變數對於林場的效率有顯著的正向影響。

The main purpose of this study is to analyze the operating efficiency of forest farms in Taiwan. The study first explored the methods to measure operating efficiency through literature review, then applied the two-stage data envelopment analysis (DEA) to examine the forest farms empirically based on the data of Taiwan’s forestry census in 2005. The DEA was used to calculate the efficiency scores of forest farms individually in the first stage. Moreover, slacks variable analysis and sensitive analysis were utilized to find out the way to improve their performance. In the second stage, the efficiency scores was regressed against a set of forest farms’ environment variables, such as manager and land characteristics by truncated regression. The study result shows that most forest farms in Taiwan do not operate efficiently. Their average score of operating efficiency is 0.34 which means that 66% inputs can be saved for achieving the same level of output. All forest farms need to increase their sales of forest products, and furthermore small forest farms have to increase their sales of by-products, large forest farms have to reduce their idle lands . The regression analysis shows that the educational level of managers, forest farm’s location and non-diversified business are significantly positive factors to operating efficiency.