全球IC設計產業生產力與效率分析

Abstract

台灣的IC設計產業總產值全球僅次於美國，自從九零年代中期成為我國極受矚目的產業。因此本研究所探討的主題為全球IC設計產業的生產效率分析，以2003年全球營收前卅大IC設計廠商為樣本。透過資料包絡法(Data Envelopment Analysis，DEA)，針對不同地區與個別廠商進行2000年至2002年的效率分析，最後進行Tobit迴歸分析，求得影響廠商生產績效的因子，提供本國廠商改善效率的參考，實證結果為：\r\n\r\n1.DEA跨國比較的結果顯示，在2000年至2002年整體技術效率值最高的地區為美國，在2000年與2001年為台灣整體技術效率值僅次於美國，但在2002年卻落居第三。觀察個別廠商績效，Qualcomm、MediaTek、Marvell、ICS、Lattice、Pmc-Sierra與DSP Group等廠商績效最好，其整體技術效率值為1。\r\n\r\n2.Malmquist生產力指數(MPI)跨國比較的結果顯示，臺、美、加三國總要素生產力均呈現退步的情形，2001至2002年，臺、加總要素生產力呈現進步，美國呈現退步。就個別廠商而言，在2000年至2001年間有四家廠商總要素生產力進步，在2001至2002年年則有十五家廠商總要素生產力呈現進步的狀況。\r\n\r\n3.Tobit迴歸分析的結論為：存貨週轉率inventory turnover對整體技術效率有顯著的正向影響；負債比率與平均收帳期間均對整體技術效率有顯著的負面影響。研發費用率、經營年限與整體技術效率呈現正向關係，至於每人配備率則呈現負相關。

The total output value of Taiwanese IC design industry is the globally second following United States. It had become the domestic gazed industry since mid 90s. Therefore, this research studies the production efficiency of global IC design industry and chooses the firms which sales globally ranked top 30 as samples. Through DEA, this thesis analyzes the efficiency focused on different regions and individual firms from 2000 to 2002. And finally, the Tobit regression model is proposed to find out the factors that influenced performances of firms and it could be reference for the domestic firms to improve their production efficiency and productivity. The results display:\r\n\r\n1. By comparison with regions, the DEA results display that U.S.A. United States has the highest overall technical efficiency(TE) value during 2000 to 2002. Taiwan ranked behind U.S. both in 2000 and 2001 but dropped to the 3rd position in 2003. Observing performances of individual firms, Qualcomm, MediaTek, Marvell, ICS, Lattice, Pmc-Sierra and DSP Group performed as the best and their efficiency value is 1.\r\n\r\n2. By comparison with regions, Malmquist Productivity Index(MPI) results display that the U.S.., Taiwan and Canada showed their Total Factors Productivity (TFP) to degenerate from 2000 to 2001. However, Taiwan and Canada showed their progressive TFP, progress in TFP but U.S.A. had a regressive TFP from 2001 to 2002. For one individual firm, there were only 4 firms’ TFP being aggressive from 2000 to 2001 and there are 15 firms’ TFP aggressive from 2001 to 2002.\r\n\r\n3. The result from running Tobit regression models display that Inventory Turnover has significant positive effect to TE; Debate Ratio and Average Collection Period have significant negative effect to TE; R&D Ratio and Incorporated Period have direct relation with TFE and Equipment Per Employee has negative relation with TE.