利用隨機邊界法估計台灣各縣市之總要素能源效率

Abstract

自Hu and Wang (2006) 首次提出總要素能源效率（TFEE）指標以來，後續應用TFEE指標的實證研究一直利用資料包絡分析法（DEA）來進行能源投入目標值（或總能源投入差額）的計算。在引述Hu and Wang (2006) 下，Zhou, Ang and Zhou (2010) 利用將單一方程式、隨機邊界生產函數移項的方式，推導出總要素能源效率的估計式。Zhou, Ang and Zhou (2010) 的版本目前仍是橫斷面（cross section），並只利用2001年21個OECD國家的資料作分析。本計畫進一步將之延伸改寫為縱橫面資料（panel）型式，並將用以估計台灣23縣市在1999-2008年期間的總要素能源效率。本計畫可利用 Battese and Coelli (1992, 1995) 的縱橫面資料、隨機邊界法模型進行估計，並可同時納入影響總要素能源無效率之環境變數進行聯立估計。

Since Hu and Wang (2006) first proposed the total-factor energy efficiency (TFEE) index, the follow-up empirical papers have been applying the data envelopment (DEA) approach to compute the energy input target (or the total energy input slack). By referring to Hu and Wang (2006), Zhou, Ang and Zhou (2010) derive an econometric equation for estimating TFEE via re-arranging a single-equation stochastic production function. The current version of Zhou, Ang and Zhou (2010) is still for only cross section data and analyzes 21 OECD countries in the year 2001. This project further extends the cross-section SFA model of Zhou, Ang and Zhou to a panel data, stochastic frontier analysis (SFA) model. The total-factor energy efficiency scores of 23 administrative regions in Taiwan from 1999 to 2008 will be estimated by this extended panel data SFA model. The panel data SFA approach of Battese and Coelli (1992, 1995) can be used to estimate by this extended panel data SFA model. Moreover, both of the total-factor energy efficiency frontier and the inefficiency equation containing environmental variables can be simultaneously estimated.