動態網絡資料包絡分析法探討台灣半導體產業效率

Dynamic Network Data Envelopment Analysis Method to Study the Efficiency of Taiwan Semiconductor Industry

廖峰毅(Fong-Yi Liao)

指導教授 ： 邱永和 ； 李命志

淡江大學/商管學院/財務金融學系博士班/博士(2016年)

https://doi.org/10.6846/TKU.2016.00416

摘要

本研究是以動態網絡資料包絡分析法(Dynamic Network DEA)探討台灣半導體廠商之效率評估，並藉由總要素生產力(Malmquist Productivity Index; MPI)模型討論台灣半導體產業2008年至2014年生產力變動，並提出建議希望可提供台灣半導體廠商經營管理階層做決策時之參考。 首先我們將整體半導體產業區分為生產階段與市場階段分別設定模型準備實證生產力變化並檢定兩階段的差異。其後我們再將實證模型設定第一階段投入項為員工人數、固定資產、研發費用，產出項為營業收入、其他收入、無形資產，第二階段產出項為EPS與股價;跨期間(Carry over)為稅後盈餘。再將半導體產業分成IC設計群族與製造封裝群族探討兩群族的效率與生產力變化，將整體半導體產業分成第一階段生產階段與第二階段市場階段、IC設計業與製造封裝產業兩群族生產力變化給予整合並檢定，用以探兩群族在不同階段與不同年度間生產力的變化。 最後以Wilcoxon檢定DN-DEA效率值與Malmquist生產力指數變化結果，以了解兩群族在兩階段之效率表現及生產力變動是否有所差異。 經由實證結果發現2008年至2014年在效率來論IC設計優於製造封裝業與市場階段優於生產階段，但是在生產力效率成長方面生產階段優於市場階段代表生產階段效率有所提升，而且檢定結果顯著無法推翻兩階段無差異之虛無假說。就兩群族比較生產力來說，製造封裝產業在生產階段與市場階段生產力皆優於IC設計業，代表在2008年至2014年間製造封裝業生產力提升是小幅優於IC設計業。

關鍵字

動態網路資料包絡分析法 ； 總要素生產力指數 ； 半導體製造產業 ； IC設計業 ； 效率

並列摘要

This study is based on a dynamic network data with envelopment analysis (Dynamic Network DEA) to assess the efficiency of Taiwan's semiconductor manufacturers and productivity (Malmquist Productivity Index; MPI) model to discuss Taiwan's semiconductor industry from 2008 to 2014 changes in productivity and recommendations Taiwan semiconductor manufacturers hope to provide the management decisions as a reference.
First, we divided into the overall semiconductor industry into two groups of the production stage and market stage for set model ready to empirical verification productivity changes and differences between the two stages.
Then we set the empirical model input items of the number of employees, fixed assets, research and development costs, output items as operating income, other income, intangible assets, the second stage output of EPS and stock price; Carry over is the net profit. Finally, we divided the semiconductor industry into the two groups of manufacture & package group and IC design group to discussion productivity of the two groups, and then after we first stage of production in overall semiconductor industry in the production stage and the second stage of the market stage, IC design industry and manufacturing packaging industry two groups of integration and verification productivity change given to explore the two groups at different stages and different changes in 2008 to 2014 years.
Finally,using Wilcoxon test the DN-DEA and Malmquist Index efficiency value change results in order to understand the performance of the two groups and the Productivity Change in efficiency of the two-stage are verified. The empirical results from 2008 to 2014 on the efficiency of the market in efficiency mean better than production stage, but in Malmquist Index growth in empirical results of the production stage is better than market stage, and Verification results no significant unable to overthrow the two-stage hypothesis nothingness differences. Compare the two groups in productivity of Malmquist Index, the manufacture & package industry in the production stage and market stage better than IC design industry, manufacture & package industry representatives manufacturing productivity improvement is slightly better than the IC design industry from 2008 to 2014.

並列關鍵字

Dynamic Network DEA ； Malmquist Index ； The semiconductor manufacturing industry ； IC design industry ； efficiency

參考文獻

[11]吳立文(2014)，「以空間語法結合DEA-DA為基礎分析建築工程施工作業要徑空間與衝突解析之研究」，博士論文，台灣大學，土木工程學系。

[38]蕭瑛傑(2015) ，「兩岸晶圓代工製造業之生產力分析」。碩士論文，政治大學，經濟學系。

[3]白忠哲(2010)，「整合型績效評估模式研究-以薄膜電晶體液晶顯示器產業為例」。博士論文，國立交通大學，工業工程與管理學系。

[7]李揚與鐘棠祺(2015) ，「半導體產業之績效評估－跨國分析與比較」。東吳經濟學報89期 頁61-88。

[10]林聖齡(2011)，「探討台灣智慧型手機產業之競爭策略」。碩士論文，國立中央大學，產業經濟學系。