基於GM（1,1）及DEA模型預測臺灣營建業上市公司經營績效

DEA模型雖可評估決策單元相對效率，所衡量效率值為客觀環境下對受評單位最有利結果，惟所評估資料係針對決策單元目前之事實數據作分析，卻無法預測決策單元未來可能發生之相對效率值。本研究主要貢獻在於結合GM（1,1）預測模型有效預測各投入與產出指標項目之未來值，將所預測之各指標項目未來值提供DEA模型作未來經營績效評估，以突破DEA模型無法預測未來相對效率值之限制。最後依據實證結果比較未來預測績效與現行績效差異，並提出有利決策單元相較僅就現行績效更具參考價值之經營政策建議。本文應用GM（1,1）及DEA模型預測臺灣營建業上市公司之經營績效，依據2019年預測值評估與分析後發現，全坤建、宏盛、皇翔、華固、總太及達麗等6家營建業之總效率是有效，而寶徠、華建等2家營建業雖非總效率有效，惟技術效率仍是有效。除總效率有效的營建業其規模收益是保持不變，冠德、興富發等2家營建業的規模收益為遞減外，其他營建業的規模收益則都是遞增。而規模收益由原本2018Q2之維持現狀至2019年時須改變為遞增，如潤隆、國揚及根基等3家營建業，或規模收益已反向轉變，即由遞減（或遞增）須改變成遞增（或遞減）狀態，如達工與冠德等2家營建業，顯示該等營建業因應環境及競爭對手未來之經營績效改變，原本有效營運規模也須再作適度調整。

關鍵字

GM（1,1）； DEA模型；績效預測；資料包絡分析；規模收益； CCR效率分析； BCC效率分析

並列摘要

Although DEA model can evaluate the relative efficiency of decision making unit and the measured efficiency value is the most favorable result for the unit under evaluation in an objective environment. However, the assessed data is analyzed for the current factual data of the decision-making unit. It is impossible to predict the relative efficiency value of the decision unit in the future. The contribution of this paper is using GM(1,1) model to effectively predict future values of each input and output for construction listed companies, and proposed future values to DEA model for future performance evaluation of decision making unit. The combination method has break through the future predictive limit of DEA model. This paper compared current and future performance and found differences between them according to evidence-based results. Finally we proposed recommendations of more reference value for management policies rather than current performance only. This paper applied GM(1,1) and DEA model to investigate the performance of listed companies in Taiwan’s construction industry. The results were found that the overall efficiency of six construction corp. containing Chainqui, Hongsheng, Huang Hsiang, Huaku, Zongtai and Dali was effective. Although the overall efficiency of BLG(Better Life Group) and Huajian construction corp. was not effective, the technical efficiency of that two companies was still effective. Additionally, the return to scale of construction corp. which overall efficiency was effective is constant. The return to scale of Kindom and Highwealth construction companies is decreasing. Besides, the return to scale of other construction listed companies is increasing. The original scale return status has changed from maintaining the status to incremental condition during 2018Q2-2019 such as Runlong, Kuoyang and Kedge. Or the scale return status has reverse changed, meaning from increasing(or decreasing) into decreasing (or increasing), such as Dagong and Kindom. This phenomenon shows that the original effective operating scale must make certain appropriate readjustments when these construction listed companies have to cope with environmental changes and more effective performance of competitors.

並列關鍵字

GM(1,1) ； DEA Model ； Performance Forecast ； Data Envelopment Analysis ； Return to Scale ； CCR Efficiency Analysis ； BCC Efficiency Analysis

參考文獻

尤韻涵(2009)。台股指數開盤價格之預測─應用類神經網路及灰預測模型。輔仁大學經濟學研究所=Department of Economics, Fu Jen Catholic University。

Google Scholar

吳姿儀(2005)。從效率觀點評估建設公司財務績效之研究。台灣科技大學營建工程研究所=Graduate Institute of Civil and Construction Engineering, National Taiwan University of Science and Technology。