台灣的再生能源效率與綠色金融政策效果之比較研究：基於數據包絡線與迴歸模型之分析

Abstract

面對氣候變遷與淨零轉型壓力，台灣自2017年起推動三代綠色金融行動方案，以投資誘因與制度設計引導資金流向再生能源與永續產業。然而，相關政策對碳減量與產業升級的實質效益，仍缺乏系統性量化評估。 本研究整合普通最小平方法（Ordinary Least Squares, OLS）、數據包絡線分析（Data Envelopment Analysis, DEA）與Tobit迴歸，建構涵蓋2010至2022年之跨方法分析架構，實證檢視綠色金融工具（如太陽能躉購費率、裝置容量、政府投資、僑外投資）對環境績效與產業結構之影響。OLS 分析結果指出，太陽能躉購費率對減碳具高度穩定與顯著成效，裝置容量與政府投資亦具負向效果，惟對產業升級之影響有限，甚至呈現輕微負相關，顯示再生能源政策短期內未能驅動第二、三級產業成長。僑外投資的正向估計值則反映外資進入初期可能伴隨建設能耗與產業轉型方向不一。 DEA 效率分析顯示，整體效率值平均達 1.06，反映台灣能源資源配置趨近最佳化，惟2010與2011年效率異常偏高，顯示初期補助與政策推力效應強烈。Tobit 分析進一步揭示，政府環境治理支出對綠色金融效率具顯著正向影響，而人均 GDP 亦有助提升效率。反觀教育變數則呈顯著負相關，推測與台灣高等教育人力過度集中於高排放產業有關。 本研究貢獻在於首度以跨方法架構系統性量化綠色金融政策之環境與產業雙重成效，填補現有文獻多聚焦單一面向之不足。結果指出，台灣再生能源技術已趨成熟，惟綠色金融商品與制度仍有發展空間。建議未來應聚焦於高效率、低邊際減碳成本之政策工具，並強化綠能投資與產業升級間的連結，以兼顧環境與經濟之永續發展。

In response to climate change and the imperative of net-zero transition, Taiwan has launched three phases of Green Finance Action Plans since 2017, aiming to channel capital into renewable energy and sustainable industries through investment incentives and institutional design. However, the actual effectiveness of these policies on carbon reduction and industrial upgrading has yet to be systematically quantified This study establishes an integrated empirical framework combining Ordinary Least Squares (OLS), Data Envelopment Analysis (DEA), and Tobit regression, covering the period from 2010 to 2022. It evaluates the impacts of green financial instruments—including solar feed-in tariffs (FIT), installed capacity (CAP), government investment (GI), and foreign direct investment (FDI)—on environmental performance and industrial structure. OLS results show that FIT has a consistently significant effect in reducing carbon emissions, while CAP and GI also exhibit negative effects on emissions. However, their contributions to industrial upgrading remain limited or even slightly negative, indicating that renewable energy policies have not yet effectively promoted secondary and tertiary sector growth in the short term. The positive coefficient of FDI suggests potential construction-related energy consumption and misalignment with domestic transition objectives during early-stage capital inflow. DEA results indicate an average efficiency score of 1.06, reflecting near-optimal allocation of energy-related resources in most years, while 2010 and 2011 show abnormally high efficiency likely due to early-stage policy surges. Tobit regression further reveals that environmental governance expenditure significantly enhances green finance efficiency, and per capita GDP also plays a supportive role. Conversely, education level shows a significant negative association, possibly due to the concentration of highly educated labor in emission-intensive industries. This study makes a novel methodological contribution by integrating multiple empirical tools to evaluate both environmental and industrial effects of green finance policies—addressing a critical gap in existing literature, which often isolates single-policy or single-outcome analyses. We recommend that future policy efforts prioritize instruments with high implementation efficiency and low marginal abatement costs, while reinforcing the alignment between green energy investment and industrial upgrading—thereby fostering an integrated pathway toward environmental and economic sustainability.