|Abstract: ||本研究探討2004年至2007年台灣地區上市之塑膠工業廠商的經營績效，以資料包絡分析法（Data Envelopment Analysis, DEA）做為績效評估模式；此分析方法能處理多項投入與產出，進而演算出相對效率模式。本研究以各廠商之營業費用、資產總額、固定資產、其他資產做為投入變數，以營業收入淨額、營業利益、營業外收入及每股盈餘做為產出變數，透過投入與產出的模式，推論出廠商的生產資源與績效關係。藉由效率分析、差額變數分析、敏感度分析及麥氏生產力指數分析，用以瞭解各廠商對生產資源使用之情形，以及效率改善的方向，並衡量跨期總要素生產力的變動，提供給管理者改善缺失與未來營運方針及對資源做適當的分配之參考依據。
This study uses some plastics companies of its industry in Taiwan to assess a performance model by Data Envelopment Analysis (DEA) among years as from 2004 to 2007. We propose a input-output model to conduct its resource-performance relationship, including four input variables as operating expenses, total assets, fixed assets, and the other assets; for output variables as net operating income, extra operating income, profit income, and earnings per share. Through efficiency analysis, slack variable analysis, sensitivity analysis, and Malmquist productivity index analysis, these help enhance to understand how the various manufacturers can be used to the case of productive resources and the direction of efficiency improvement to measure the inter-period changes in total factor productivity. These analysis also support managers to improve the operating principles, future loss, and appropriate, resource allocation.
The results showed as the followings: (1) The value of 0.9468 in the average pure technical efficiency of the industry; the value of 0.9435 in the average scale efficiency. The former one is greater than the latter one, which refers to improve its scale efficiency can be considerable from size adjustment; (2) The four plastics companies (Asia Polymer, Taiwan Styrene Monomer, UPC Technology, and Yem Chio) showed the unnecessary adjustment to its of input-output model because of the relative efficiency value of 1.00 to reach the necessary efficiency; (3) Through the sensitivity analysis, the most influential inputs was the total operating income; (4) By the Malmquist productivity index analysis, the overall plastics industry manufacturers, productivity, and technical efficiency showed the growth phenomena. However, technology growth still improves necessarily by their appropriate input-output adjustments.