經環境因素調整後之台灣各縣市能源效率

Abstract

本研究是應用Fried at el. (1999) 所提出的四階段資料包絡分析法，研究對象為台灣二十三縣市，研究期間由1999年至2005年，主要探討台灣總要素能源效率的狀況。主要的投入變數為各縣市政府年度歲出、就業人口、垃圾處理量、家庭用電量、其他用電量、汽油銷售量及柴油銷售量，而產出變數為各縣市年度所得，名目變數皆以1999年為基期的GDP平減指數轉換為實質變數。經過第一階段的分析後，再由所選入環境變數，包括道路長度、機車數量、營利事業登記數、平均非營利組織數目、高教育水準人口比率（大學以上畢業）與製造業及服務業工作之人口比例來探討影響差額值的環境因素。經過本階段調整後的投入項，再重新計算一次技術效率，此時的效率值是排除相異客觀環境，能更準確反應使用狀況。主要的研究結果為：第一階段分析時，台北市與台中市有較佳的技術效率。而台東縣、雲林縣、嘉義縣及彰化縣有較差的技術效率。經過Tobit迴歸後所得到的結果，發現道路長度、機車數目、製造業與服務業人口比例及NPO數量顯著提高能源差額變數，進而惡化能源效率。而營利事業登記數及高教育水準比例則顯著降低能源投入差額進而改善能源效率。經過環境因素調整後，台北縣市擁有最佳的技術效率。而台東縣、澎湖縣、嘉義市及花蓮縣有較差的技術效率。擁有最佳效率的縣市數量不變，但經投入調整後之平均效率值減少。

This research applies proposed by Fried at el. (1999) the four-stage DEA to study the energy efficiency of twenty-three administrative regions in Taiwan. The sample period is from 1999 to 2005. The seven inputs are real local government expenditure, employment, solid waste, household electricity consumption, non-household electricity consumption, gasoline sales and diesel sales. The single output is the total real income. All nominal variables are adjusted into real variables by GDP deflators in the base year of 1999. After first-stage analyzing, we use seven environmental variables to explain input slacks, including road length, registered motorcycles, population ratio of graduates and college, the number of NPOs, the number of profit organizations, and the rates of employment in service industries or in manufacturing industries. This efficiency concept avoids environment influence and hence is more accurate. The major findings are as follows: In first-stage, Taichung City and Taipei City have the best energy efficiency. Taitung County, Yunlin County, Chiayi County and Chunghua County have worse energy efficiency. The second-stage uses Tobit regression to explore the effects of environmental variables and input slacks. Road length, registered motorcycles, the employment rate on manufacturing and service industries and the NGO number significantly increase energy input slacks, hence worsening energy efficiency. The Number of profit organizations and population of high educational ratio have significantly negative effects on energy input slacks, hence improving energy efficiency. Taipei County and Taipei City have the best environment-adjusted technical efficiency. Taitung County, Chiayi City, Penhu County Hualien County have the worst environment-adjusted technical efficiency.