中文摘要：

结合方向性距离函数和跨期数据包络分析法，测度绿色全要素生产率增长时尽可能地避免了出现技术倒退结论。中国省份工业的经验分析表明：忽略非期望产出的全要素生产率增长被高估，并混淆了技术进步和技术效率变化的相对贡献；技术无效率普遍存在且省际差异和年际波动性大，但这与经济发展水平无关；绿色全要素生产率变化也存在省际差异和年际波动特点，其中技术进步是其增长的主要源泉，而技术效率恶化是其倒退的根本原因；绿色全要素生产率增长存在明显的倒U型趋势和趋同性，而且外资利用印证了“污染天堂”假说。

英文摘要：

Combining Inter-Temporal Data Envelopment Analysis with Directional Distance Function, the method of measuring Green Total Factor Productivity（GTFP） growth can avoid technology regression as much as possible. The empirical analysis results of China provincial industries are as follows： if we don＇t consider undesirable output, we would over-estimate TFP change and confound those contributions of technology change and technical efficiency change;there was technology inefficiency widely which had large inter-provincial and annual difference but wasn＇t correlative with economic development level;the GTFP growth also experienced large inter-provincial and annual difference, and the most important source of GTFP growth was technology change but its regression was leaded by technical efficiency deterioration; there is an inverted U-shaped and converging tendency of GTFP growth, and the Pollution Heaven Hypothesis in utilizing FDI was confirmed