Does green innovation crowd out other innovation of firms? Based on the extended CDM model and unconditional quantile regressions

Yi Jiang,Richard S.J. Tol
2024-02-05
Abstract:In the era of sustainability, firms grapple with the decision of how much to invest in green innovation and how it influences their economic trajectory. This study employs the Crepon, Duguet, and Mairesse (CDM) framework to examine the conversion of R&D funds into patents and their impact on productivity, effectively addressing endogeneity by utilizing predicted dependent variables at each stage to exclude unobservable factors. Extending the classical CDM model, this study contrasts green and non-green innovations' economic effects. The results show non-green patents predominantly drive productivity gains, while green patents have a limited impact in non-heavy polluting firms. However, in high-pollution and manufacturing sectors, both innovation types equally enhance productivity. Using unconditional quantile regression, I found green innovation's productivity impact follows an inverse U-shape, unlike the U-shaped pattern of non-green innovation. Significantly, in the 50th to 80th productivity percentiles of manufacturing and high-pollution firms, green innovation not only contributes to environmental sustainability but also outperforms non-green innovation economically.
General Economics
What problem does this paper attempt to address?
The core problem that this paper attempts to solve is whether green innovation will have a crowding - out effect on other innovation activities of enterprises, and how this effect is manifested in enterprises in different industries and with different levels of pollution. Specifically, through the extended Crepon, Duguet, and Mairesse (CDM) model and the unconditional quantile regression method, the research explored the impact of green innovation on productivity and made a comparative analysis with the effect of non - green innovation. ### Research Background and Problems With the advent of the era of sustainable development, enterprises are faced with the decision - making challenge of how to invest in green innovation and its impact on the economic trajectory. This paper uses the CDM framework to study the process of converting R & D funds into patents and its impact on productivity, with special attention to the differences in economic effects between green innovation and non - green innovation. The research aims to answer the following key questions: 1. **Does green innovation have a positive impact on enterprise productivity?** 2. **Will green innovation crowd out other types of innovation?** 3. **Are there differences in this crowding - out effect among enterprises in different industries and with different levels of pollution?** ### Main Findings - **Overall Trend**: Non - green patents mainly drive productivity growth, while the impact of green patents in non - heavily polluting enterprises is limited. - **Specific Industries**: In highly polluting and manufacturing sectors, green innovation and non - green innovation have an equivalent effect on productivity improvement. - **Productivity Distribution**: The productivity impact of green innovation is an inverted U - shape, while that of non - green innovation is a U - shape. At the 50th to 80th percentile productivity levels in manufacturing and highly polluting enterprises, green innovation not only contributes to environmental protection but is also economically superior to non - green innovation. - **Heavily Polluting Enterprises**: Only at the median and above of the productivity distribution can heavily polluting enterprises obtain productivity gains from green innovation, indicating the existence of a certain operational efficiency threshold. - **Non - Polluting Enterprises**: For enterprises that do not face pollution problems, green innovation is negatively correlated with productivity at the highest 20% productivity level, showing a shift from benefit to burden. ### Methodology The research adopts an extended CDM model and analyzes the relationship between R & D investment, patent output, and productivity through three - stage equations. The specific steps are as follows: 1. **R & D Equation**: Use the Heckman selection model to correct potential self - selection bias and predict R & D intensity. 2. **Patent Equation**: Use the count data model to estimate patent intensity and overcome simultaneity and omitted variable bias. 3. **Productivity Equation**: Combine the predicted patent intensity and enterprise productivity to provide an accurate estimate of the impact of innovation output on productivity. In addition, the research also uses the unconditional quantile regression (UQR) and regression - based inference function (RIF) methods to evaluate the non - linear impact of patent intensity on different productivity levels. ### Conclusion The research results show that in some cases, green innovation can promote productivity growth as much as non - green innovation, especially in highly polluting and manufacturing enterprises. However, in non - polluting enterprises and enterprises with high productivity levels, green innovation may face high opportunity costs, thus resulting in a crowding - out effect. These findings provide an important reference for enterprises when formulating innovation strategies, especially when considering the balance between environmental protection and economic benefits.