本文係以1998年至2003年台灣電子業工廠與員工的合併資料作為實證基礎, 觀察工廠與勞工特性對於工廠生產力與平均薪資之影響。本文所討論的生產力指標包括勞動生產力與總要素生產力。另外,為考量生產要素投入與生產力兩者間關聯性所造成的估計偏誤,本文進一步利Levinsohn and Petrin (2003)的生產力估計方法來解決生產要素投入的內生性問題。實證結果顯示, 整體來說, 電子業工廠規模較大、工廠營運年數較多、員工的男性比例與人力資本指標較高的工廠, 其生產力越高, 我們發現員工的教育程度組成為大型工廠生產力的重要決定因素,但員工的年齡結構與性別組成對其生產力則無明顯影響。至於工廠平均薪資方面, 各項特性對於薪資的影響方向大略與生產力相同,僅員工年齡組成對於生產力與薪資有相反的效果。

This paper investigates the effect of workforce composition on plant productivity and wages using a unique employer-employee matched data set with data from Taiwan electronics plants. Specifically, we adopt the methodology proposed by Levinsohn and Petrin (2003) to address the endogeneity caused by the correlation between inputs and unobservable productivity shocks when estimating production functions. Our results suggest that plant productivity increases with plant size, plant age, and the proportion of male, younger and more-educated workers within a plant. Moreover, the impact of the education composition of the workforce on plant productivity is more significant for large-scale plants. In addition, while younger workers are more productive than older ones, there is an age-wage premium with older workers being paid higher wages. Finally, employing foreign workers is beneficial to plant productivity, yet such benefit decreases as the ratio of foreign workers increases.