作者
黃台心^* (國立政治大學金融學系)
李起銓 (國立政治大學金融學系)
郭肇軒 (國立政治大學金融學系)

Authors
Tai-Hsin Huang^* (Department of Money and Banking, National Chengchi University)
Chi-Chuan Lee (Department of Money and Banking, National Chengchi University)
Chao-Hsuan Kuo (Department of Money and Banking, National Chengchi University)

中文摘要
本文選取資訊電子業和食品業為代表，利用分量迴歸模型，估計大企業和中小企業勞動與資本的偏產量彈性與邊際產量、規模彈性、技術進步率、替代彈性等指標，再以 Blinder-Oaxaca 方法分解兩種規模廠商間生產技術的差異，做為分析與比較之基礎，盼能提供企業改進方向，增進整體產業之生產效率。實證結果發現，兩產業之勞動偏產量彈性皆大於資本偏產量彈性，且規模彈性值皆大於 1，顯示應持續擴大生產規模，藉以降低長期平均成本。另外，大企業替代彈性高於中小企業，表示台灣製造業大企業調整要素投入的能力較好。在生產技術的差異上，高科技產業的大企業，生產力優於中小企業；傳統產業差異則不明顯。

Abstract
This paper studies the productivity differentials between large and small-medium Taiwanese manufacturing firms during 1992-2005. We apply a quantile regression technique to characterize the production processes of the manufacturers under consideration at various quantiles of conditional distribution. We then employ the Blinder-Oaxaca decomposition method to measure the difference in productivity between the two types of firms. Evidence shows that the output elasticity of labor is greater than that of capital and increasing returns to scale prevail, indicating that the sample firms should extend their production scale further to lower their long-run average cost. The elasticities of substitution for large firms exceed those of small-medium firms, implying that it is easier to substitute one input for another in large firms than it would be in small-medium firms. Large firms exhibit higher productivities than small-medium firms especially in the hightech industry.

中文關鍵字
生產技術；分量迴歸法；差異分解法

Keywords
Production Techniques; Quantile Regression; Blinder-Oaxaca Decomposition