產業能源效率之跨國比較

Abstract

自Hu and Wang (2006) 及 Hu and Kao (2007) 建構總要素能源效率指標(Total-factor energy efficiency; TFEE) 以來，受到後續文獻的引用與改進（Zhou et al., 2008）。既存的總要素能源效率指標的應用 (包含後續引述國際期刊文獻) 幾乎都利用地區別資料，以進行跨區域或跨經濟體的能源效率比較。總要素能源效率指標在跨國的產業別能源效率上的應用，仍有待進一步的推廣。此外，之前總要素能源效率的計算，幾乎都採用資料包絡分析法中最原始的CCR 模型 (Charnes et al., 1978) 及BCC模型(Banker et al., 1984)。對效率前緣的計算或估計，缺乏更先進且周延的資料包絡分析法或隨機邊界法等方法上的應用。本兩年期計畫擬就效率前緣之計算或估計，應用更進階的新方法。本計畫第一年擬應用Zhou, Ahn and Zhou (2012) 所提出的隨機邊界法模型，來進行總要素能源效率的跨國能源效率比較。本計畫第二年擬採用Liang, Wu, Cook and Zhu (2008) 所提出的「資料包絡分析法-賽局交叉效率」(DEA Game Cross Efficiency) 來求解權重，並藉由這些客觀權重計算出每年個別產業效率前緣的位置及各國各產業的整體效率值及節能比例。目前初步蒐集資料來源主要為EU-KLEMS資料庫 (網址：http://www.euklems.net)，裡面包含歐盟72個產業的跨國資料。產出項有附加價值，投入項包含資本、勞動與能源。每次執行SFA或DEA估計或計算時，只有屬於同一個產業的觀察值會被納入。

Since Hu and Wang (2006) and Hu and Kao (2007) construct the total-factor energy efficiency (TFEE) index, it has been widely applied and improved by the following literature (Zhou et al., 2008). However, most of the existing literature applies the TFEE index to analyze regional or economy-wide energy efficiency. The application of TFEE to industry-wide energy efficiency still remains to apply and promote. Moreover, most of the existing papers applying TFEE use only the traditional CCR model (Charnes et al., 1978) and BCC model (Banker et al., 1984). More advanced DEA approaches or the use of stochastic frontier to compute or estimate TFEE can still be further tried, as this two-year project will do. In the first year we will apply the SFA model proposed by Zhou, Ahn and Zhou (2012) to compare the industry energy efficiency across countries in different years. In the second year we will apply the DEA game cross efficiency model proposed by Liang, Wu, Cook and Zhu (2008) to solve for the weights and then construct the efficiency frontier, hence finding the overall efficiency score and efficient energy-saving ratios of a country’s industry in a specific year. The main data source is the EU-KLEMS Database (Website: http://www.euklems.net). This database contains 72 industries of EU in various years. The output is the value added and the inputs include capital, labor, and energy. In either SFA or DEA, only the observations in the same industry will be included in a model.