【Decomposition,of,China’s,Industrial,Environmental,Total,Factor,Productivity,Sources:Analysis,Based,on,Factor,Input,and,Pollution,Treatment】 Voice of China

　　Abstract: Following slack-based inefficiency measurement method and Luenberger productivity index, this paper decomposes industrial environmental TFP index by input factors and output to estimate the sources of China’s industrial environmental TFP. Results indicate that (1) China’s industrial environmental TFP increased between 2001 and 2007 but declined in 2008 under the effect of global financial crisis; (2) input utilization productivity contributes 1/3 to industrial environmental TFP and pollution treatment productivity contributes about 2/3, which means that pollution treatment will effectively increase industrial environmental TFP; (3) capital utilization productivity contributes 1/6 to input productivity and labor utilization productivity contributes about 5/6, which means that progress of labor production technologies is an effective means to increase industrial environmental TFP; (4) COD treatment productivity contributes roughly 2/3 to pollution treatment productivity, and SO2 treatment productivity contributes about 1/3, which shows that priority on COD treatment will more effectively increase industrial environmental TFP.
　　Key words: industrial environmental TFP, slack-based efficiency loss measurement, Luenberger productivity inde
　　JEL Classifications: D24, Q52
　　1. Introduction
　　Estimation of total factor productivity (TFP) of China’s industrial environment helps gain a deeper understanding on the momentum of China’s rapid economic growth. Without considerations on environmental factors, it is difficult for conventional methods of TFP estimation to reflect the real significance of TFP. Under the backdrop of growing domestic environmental problems and increasing attention attached by the government to these issues, it is highly necessary to include environmental factors into the analytical framework1 of TFP.
　　As for the estimation of TFP, conventional studies have mainly employed such methods as cost-return method and parameter-based econometric models. For environmental problems featuring high input and output, these approaches may not apply: cost-return method has the problem of subjectivity in the selection of weight for different input factors and output; while parameter-based econometric model is subject to priori equation form and distribution assumptions. However, these problems are avoided by data envelopment analysis (DEA), which can conveniently include pollutants into analytical framework and is therefore widely applied in the analysis of resource and environmental problems2.