Abstract: Given the lack of empirical evidence from China regarding whether carbon emission trading as a typical market-driven environmental regulation mechanism can trigger the Porter effect and improve the total factor productivity of enterprises, this paper uses the method of difference in difference in differences (DDD) to design a quasi-natural experiment of the pilot policy of carbon emission trading, and applies the spatial econometrics to explore, from the micro level, the influence of carbon emission trading on the total factor productivity (TFP) of enterprises and the mechanism underlying the influence by constructing a vertical and horizontal spillover weight matrix between industries. The results show that carbon emission trading can improve the total factor productivity of enterprises and induce inter-industry policy spillover effect, and the direct impact of carbon emission trading on the total factor productivity of enterprises presents an inverted U-shaped dynamic characteristic. The impact of carbon emissions trading on the total factor productivity of enterprises has typical heterogeneity characteristics. The total factor productivity of state-owned enterprises and small-scale enterprises is affected more significantly by policies. The mechanism analysis shows that carbon emission trading can affect the total factor productivity of enterprises through R&D effect, transformation effect and cost effect. Meanwhile, the impact of carbon emission trading on the total factor productivity of enterprises is moderated by enterprises’ internal characteristics (financial mismatch, human capital) and external characteristics (administrative intervention).

Key words: carbon emission trading, total factor productivity, industry spillover, DDD