Environmental regulation, upgrading of industrial structure and economic fluctuation

This paper utilizes dynamic panel threshold technology to conduct a nonlinear test on the direct effect between environmental regulation and economic fluctuations and the effect of industrial structure upgrading by taking 35 industrial sectors in China from 2003 to 2016 . The research has found that there is an inverted "U" relationship between environmental regulation and economic fluctuations, with the enhancement of environmental regulations, the economic fluctuation increases first and then decreases. The cross-terms of environmental regulation and industrial structure rationalization or industrial structure upgrading are significantly negative, which indicates that the enhancement of environmental regulation is conducive to promoting industrial structure upgrading and reducing the economic fluctuations. While the rationalization factors of industrial structure and advanced industrial structure are significantly negative, indicating that both forms of industrial structure upgrading are conducive to reducing the economic fluctuations. Environmental regulation has technical innovation thresholds for industrial structure upgrading and economic fluctuations, but there are no human capital or FDI thresholds. In the rationalization model of industrial structure, there is a nonlinear "U" relationship between environmental regulation and economic fluctuations when the proportion of scientific research expenditure is more than 1.35%. With the enhancement of environmental regulation, the economic fluctuation reduces first and then increases, and the corresponding inflection point value is 2.398% of the environmental regulation level. At the same time, the environmental regulation can indirectly reduce economic fluctuations by pushing down the industrial structure upgrades. In the advanced model of industrial structure, there is a “U” relationship between environmental regulation and economic fluctuation when the proportion of scientific research expenditure is greater than 1.26%. With continuous enhancement of environmental regulation, the economic fluctuation reduces first and then increases. The corresponding inflection point value is 1.78% of the environmental regulation level, and environmental regulation can indirectly reduce economic fluctuations by promoting the industrial structure at the same time.

changes in the world industrial structure. The high-end modern service industry has become the driving force of economic development, and its industrial structure has been continuously upgraded, which has contributed greatly to the steady economic development. However, after the financial crisis in 2008, economic stability encountered challenges, increasing the risk of economic fluctuations. Therefore, how to upgrade the industrial structure with stable economic fluctuation and how to reduce economic fluctuation through industrial structure upgrading are important problems that every country encounters.
From the perspective of our country, the current economic society has stepped into the "new normal" of speedy shifts and structural adjustment pains. The report of the 19th National Congress of the CPC noted that construction of an ecological civilization is a millennium plan for the sustainable development of the Chinese nation, and it is necessary to resolutely strike a major offensive war on pollution prevention and control; this reflects the importance attached by the Party and the state to environmental protection. On the one hand, the environmental regulation policy increases an enterprise's cost and profit to a certain extent, impels the enterprise to adjust its production behaviour, and causes an industrial structure adjustment. On the other hand, the upgrading of the industrial structure, as the core tool to coordinate the economy and the environment, affects environmental quality and economic fluctuation by influencing the proportion of high energy consumption, high-pollution industries and investment in clean equipment, green technology research and development. Therefore, it is necessary to use environmental regulation as the intrinsic incentive of industrial structure upgrading and to analyse the mechanism of environmental regulation to reduce economic fluctuation by forcing the industrial structure to upgrade; this is an important theoretical and practical research topic.
Research on the relationship between environmental regulation and industrial structure upgrading mainly focuses on two aspects. One aspect is the influence mechanism. Environmental regulation mainly influences the upgrading of the industrial structure by three means: demand, technological innovation and international trade transmission mechanism. The overall environmental regulation intensity of our country has a positive impact on the direction and path of industrial upgrading (Xiao Xingzhi, Li Shaolin; 2013). Li Qiang et al. (2013) [2] show that the existence of environmental regulation will increase the proportion of the service sector relative to the industrial sector and thus promote the adjustment of the industrial structure. The second aspect is empirical research. Li Xiaoying (2018) [3] uses the spatial econometric model to test the impact of environmental regulation on the industrial structure upgrading and finds that environmental regulation has a forced effect on industrial structure adjustment. However, this literature basically assumes that there is a linear relationship between environmental regulation and industrial structure upgrading. In fact, there is a threshold effect on the impact of environmental regulation on industrial structure upgrading. Zhong Maochu et al. (2015) [4] show that there is a U-type relationship between environmental regulation and industrial structure upgrading. Only when the threshold of environmental regulation is crossed can the adjustment of industrial structure be promoted.
Currently, there are great differences in the influence of different changes in the industrial structure on economic fluctuation. Dry Chunhui et al. (2011) [5] hold that the effects of rationalization and upgrading of the industrial structure on economic fluctuations are mainly manifested in unpredictable cyclical fluctuations; however, their influence mechanisms are quite different. Upgrading of the industrial structure is an important source of economic fluctuation, and the rationalization of the industrial structure is helpful to restrain economic fluctuation. Li Qiang (2012) [6] studies the relationship between industrial structure change and economic fluctuation through an econometric model. It is found that the change in the industrial structure is an important reason for economic fluctuation. Rationalization and upgrading of the industrial structure are helpful to alleviate economic fluctuation; however, the higher industrial structure has more influence on economic fluctuation than the rationalization of the industrial structure. Peng Chong et al. (2013) [7] use the panel vector autoregressive model to investigate the effect of an industrial structure change on economic fluctuation and determined that the rationalization of the industrial structure has an "ironing effect" on economic fluctuation; however, the impact of a high-level impact is strongly pro-cyclical. Both short and long term have obvious positive effects.
In summary, there is no literature on the relationship between environmental regulation, industrial structure and economic fluctuation, particularly on the direct effect of environmental regulation on economic fluctuation and the effect of industrial structure change. At the same time, the related research literature mainly utilizes provincial panel data to test the lack of industry-level environmental regulation, industrial structure upgrading and economic fluctuations; in addition, the industrial industry is the main body of our economy. This industry has an important effect on the economic fluctuation; therefore, it is necessary to examine and analyse the industrial industry.
In addition, the most important item is the threshold effect of environmental regulation on industrial structure upgrading and economic fluctuation (Zhong Maochu et al., 2015 [4]). The existing static panel threshold technology ignores the dynamism and continuity of the explained variables and cannot address the endogenous problem of environmental regulation; therefore, there is much room for improvement.
Targeting the deficiency of the existing research, this paper mainly uses 35 industrial industries in China from 2003-2016 as the research object and uses dynamic panel threshold technology to directly test the direct effect between environmental regulation and economic fluctuation and the upgrading effect of the industrial structure. This paper provides the theoretical basis for establishing reasonable environmental regulation intensity to force the industrial structure to upgrade to reduce economic fluctuation and finally realizes the win-win situation between environmental protection and economic development. The contributions of this paper are as follows: First, environmental regulation, industrial structure upgrading and economic fluctuation are incorporated into the unified research framework, and the direct effect of environmental regulation on economic fluctuation and the upgrading effect of industrial structure are discussed. The second contribution is to extend the test to the industrial level. The third is to consider the dynamic nature of economic fluctuations, the continuity and the threshold effect, the use of dynamic panel threshold technology to estimate, and solving the endogenous problem of environmental regulation. The estimation conclusion is more robust.

Research and design
Among the variables, i and t represent the industry and the time, respectively; i  is the individual effect. t  is the time effect, and it  is the random disturbance term.
where I is the threshold function, thr is the threshold variables set, and  is for the corresponding threshold. This paper mainly establishes three threshold variables: technological statistics on "other mining industries", the new "other mining industry" will be merged with the "non-metal ore mining and separation industry". Some data are missing from "handicraft and other manufacturing industries", "waste resources and waste materials recovery and processing industry" and "timber and bamboo harvesting and transportation industry". "Rubber products industry" and "plastic products industry" are merged into "rubber and plastic products industry" in 5. Industry control variable. The industrial value-added rate is expressed by the ratio of the industrial added value to the industrial gross output value in the same period, and labour productivity is expressed by the ratio of industrial added value to the total number of employees.

Dynamic panel threshold estimation method
There is a fixed individual effect in the dynamic panel threshold model set in this paper, so it is necessary to eliminate the individual effect before estimating. Usually, there are two methods of intra-group transformation (within transformation) and first-order difference to deal with. The forward orthogonal deviation transformation of model (2) is conducted, in which the forward orthogonal deviation transformation model of the error term is as follows: The forward orthogonal deviation transformation of other variables is consistent with the

Direct effect Test
According to the established basic model, this paper first uses the dynamic panel estimation method to test the direct effect of environmental regulation on economic fluctuation and the upgrading effect of industrial structure. Considering the strong endogeneity of environmental regulation, the first order lag of environmental regulation is used as the tool variable to solve the endogeneity problem, and the robust estimation result is obtained using the system GMM two-step robust standard error estimation.   fluctuation, the coefficient of environmental regulation, ER, is significantly positive, and the square term is significantly negative in both the industrial structure rationalization and the industrial structure advanced model. The result shows that there is an inverted "U"-type relationship between environmental regulation and economic fluctuation. With the enhancement of environmental regulation, economic fluctuation first increases and then decreases. When the environmental regulation is lower than the specific threshold, the increase of the non-productive environmental cost in the short term will urge enterprises to reduce their investment, but the main body of the investment decline consists of enterprises with poor pollution control equipment and serious pollution. When the environmental regulation is further raised beyond the threshold, polluting enterprises will be gradually "cleaned", and the green innovation ability of the industry will be continuously improved. Once the innovation investment income exceeds the cost, the whole industry will choose to increase the innovation investment, thus improving the overall level of investment. Therefore, there is a direct "U" relationship between environmental regulation and investment. At the same time, the impact on economic growth is also obvious; when investment decreases, economic volatility increases. When investment increases, particularly when the level of innovation investment increases, it will promote steady economic growth and reduce economic volatility. Therefore, the relationship between environmental regulation and economic fluctuation also shows a direct "U"-type relationship. According to the results of the threshold estimation and the significance test, if TI is used as the threshold variable, there is only one threshold value in both the industrial structure rationalization model and the industrial structure advanced model. In the industrial structure rationalization model, the threshold is the technological innovation level of 1.35, while in the industrial structure advanced model, the threshold is the technological innovation level of 1.26. By calculating the average value of the technological innovation water for 35 industrial industries between 2003 and 2015, it was found that the industries that crossed the 1.35% threshold were mainly medicine (1.76%), manufacturing of specialized equipment (1.64%), manufacturing of transportation equipment (1.60%), manufacturing of communications equipment, computers and other electronic equipment (1.59%), manufacturing of instruments and cultural office machinery (1.55%), and electrical machinery and equipment manufacturing industry (1.46%), a total of 6 industries. The other 29 industries did not cross the threshold level of technological innovation. In addition to the six sectors above the 1.26 percent threshold, there is the general equipment manufacturing industry (1.26 percent) for a total of seven industries. In addition, according to figs. 1 and 2, a single threshold exists, which indicates that environmental regulation has a technological innovation threshold for industrial structure upgrading and economic fluctuation. If human capital HU is used as the threshold variable, in both the industrial structure rationalization model and the industrial structure advanced model, the significance test results show that there is no threshold effect. The result shows that the enhancement of industry environmental regulation will not restrict the level of human capital or the influence of human capital on the upgrading of industrial structure and the fluctuation of the economy. Alternatively, the finding shows the restriction of human capital on this kind of effect is relatively small, which is not sufficient to constitute a threshold. At the same time, if FDI is used as the threshold variable, the significance test results of the threshold effect in the two models also show that there is no threshold effect; this indicates that the enhancement of industry environmental regulation does not restrict FDI on the upgrading of the industrial structure and the impact of economic fluctuation. Alternatively, FDI has less restriction on this effect and is not sufficient to form a threshold. In fact, the key to the impact of environmental regulation on the upgrading of the industrial structure and on the economic fluctuation is whether the innovation investment income caused by the enhancement of environmental regulation is greater than the cost, thus affecting the innovation investment choice and the investment level improvement in the whole industry. All of these impacts depend on the level of technological innovation in the industry, but when the level of technological innovation is low, it is difficult to promote the innovation investment income to exceed the cost through environmental regulation. However, only the industry that surpasses the threshold has better technical innovation strength. When the environmental regulation strengthens, these enterprises will raise income through innovation investment; once this exceeds the cost, it will raise the investment level of the whole industry. Therefore, it is not difficult to understand why the impact of environmental regulation on the upgrading of the industrial structure is only the threshold of technological innovation and not the threshold of human capital and FDI.  In this paper, upgrading of the industrial structure is divided into rationalization and upgrading, and the corresponding threshold TI is 1.35% (rationalization) and 1.26% (advanced). This paper uses the lag of environmental regulation more than one order as the tool variable to solve the endogenous problem and uses the dynamic panel GMM two-step robust standard to test the threshold effect of environmental regulation and industrial structure upgrading on economic fluctuation. In the estimates shown in tables 5 and 6, the Sargan test results show that the tool variables are set reasonably, and the AR (2) test shows that there is no second order sequence correlation. In addition, the lag first order coefficient of economic fluctuations is above 0.90 and significant; this shows that the economic fluctuation is dynamic.

Disclosure statement
The authors declare no conflict of interest