創新、人力資本投入與企業經營效率－－以我國IC設計業為例

The Effects of Innovation and Human Capital Investments on the Operating Efficiencies of Taiwan's IC Design Industry

本研究之目的係利用資料包絡分析法探討2002－2004年我國上市上櫃IC設計廠商之人力與創新資本投入的經營效率，並提供無效率廠商在資源管理上之具體建議。再利用隨機效果Tobit長期追蹤模型探究影響IC設計廠商經營效率的重要因素，藉此效率評估結果，供管理者改善缺失與未來營運方針之重要參考依據。主要實證結果有以下三項。第一，全體IC設計樣本廠商的總平均技術效率為0.487。表示在這三年期間，就技術效率而言，IC設計樣本無效率相對於有效率廠商平均有51.3%之創新與人力資本投入要素形成浪費。經營無效率之原因分別來自於純技術無效率與規模無效率，且多數之廠商處於規模報酬遞增階段，顯示各IC設計廠商營運規模過小，有擴大規模或進行整併之必要性。第二，差額變數分析結果顯示IC設計樣本中的無效率廠商有將近一半的創新資本投入是被浪費的，而在員工人數、學歷比及員工薪資等人力資本投入約有四成是被浪費的。其次，在績效產出方面，市場資本仍有六成的成長空間，而營業收入淨額及市場占有率則約有兩成的成長空間。第三，隨機效果Tobit長期追蹤模型迴歸結果發現每位研發人員擁有的資源、員工生產力、專利權核准數及負債比率等皆為影響IC設計廠商經營效率之重要因素。

This study evaluated the effects of innovation and human capital investments on operating efficiencies of the IC design industry in Taiwan from 2002 to 2004. The data envelopment analysis (DEA) was utilized to measure the efficiencies of IC design companies and the slack variable analysis was used to provide some suggestions about resource distributions for the inefficient companies. The study also used the random-effects Tobit panel regression model to examine the determinants of the operating efficiencies of the IC design industry. The results showed that: (1) the average global technical efficiency for IC design industry was 48.7%, indicating that there were about 51.3% inefficiencies. Most of inefficiencies in operation came from pure technical and scale inefficiencies, and most companies were to lie in increasing returns to scale. The results demonstrated that the operational scales of most IC design companies were too small and the expansions of scales were required; (2) the slack variable analysis appeared that inefficient firms tended to spend too much on innovation capital investment such as research and development (R&D) and human capital investment such as labor costs and employ too many employees and R&D employees for their operations; (3) The random-effects Tobit panel regression results indicated that the important determinants of operating efficiencies for IC design companies included resources each R&D employee owned, employees productivity, number of patents owned, and debit ratio.