Are the Green TBTs a Stimulus or a Trap for Enterprises’ Green Technology Development?

2.1 The Impact of Green Trade Barriers on Green Technology of Enterprises

Compared with the total control and market incentive environmental regulations, green trade barriers are technical standard environmental regulations, which are fast to implement and have strong rigidity, and if enterprises fail to meet the standards in the short term, they will lose their market access rights and even face the penalty of shutdown (Peng et al., 2021). Wan et al. (2021) found that in the face of technical standard-based environmental regulations, most enterprises will choose the path of technological transformation to achieve green transformation. They believe that this is due to the long cycle of technological innovation, high uncertainty, and the need for a certain amount of technology accumulation, so when enterprises expect to quickly meet the technical standards set by environmental regulations, it is more realistic and feasible to carry out green technology transformation than green technology innovation. Based on the above analysis, this paper proposes:

Hypothesis 1: Green trade barriers will encourage Chinese enterprises to increase investment in green technology transformation.

In addition to technological transformation, some studies believe that technical barriers to trade can have a forcing effect on the technological innovation of enterprises in exporting countries. On the one hand, the willingness of enterprises in exporting countries to innovate will be strengthened after encountering technical barriers. On the other hand, because the technical regulations and standards given by technical barriers to trade can provide enterprises with R&D ideas and reduce R&D risks, thus playing a role in guiding enterprise technological innovation (Swann, 2010). Wu and Liu (2007) demonstrated through the theoretical model that the technical trade barriers of importing countries can promote the enterprises to carry out technical innovation and passive industrial upgrading of exporting countries when the expected innovation investment and technical barriers are not too high. Li et al. (2014) empirically examined the impact of different types of trade barriers on technological innovation, and found that tariff barriers had a negative impact on technological innovation, while technical barriers had a positive impact on technological innovation.

Another support for the positive effect of green trade barriers on green technology innovation comes from the “Porter hypothesis”. According to this hypothesis, well-designed environmental regulations will bring about an “innovation compensation” effect, which will motivate enterprises to actively innovate green technology and improve production efficiency and market competitiveness (Porter and Linde, 1995). Rubashkina et al. (2015) empirically supported the Porter hypothesis using panel data from the manufacturing sector in 17 European countries. Qi et al. (2018) tested Porter’s hypothesis with a sample of listed companies in China and confirmed that emission trading policies can stimulate green technology innovation in polluting industries in pilot areas. Accordingly, this paper proposes:

Hypothesis 2a: Green trade barriers will have forcing effect on Chinese enterprises to innovate in green technology.

However, some scholars hold the opposite view, believing that technology-standard environmental regulations inhibit the green technology innovation of enterprises. Palmer et al. (1995) argue that the need to internalize the compliance costs of environmental regulations has forced enterprises to divert funds from their budgets that could be used to invest in innovation projects, thus hindering enterprises from improving their productivity. Gao and Yuan (2020) conducted research on the quasi-natural experiment of the implementation of cleaner production standards, showing that the implementation of cleaner production standards not only reduces the probability of R&D and innovation of enterprises, but also reduces the patent output of green technology innovation of enterprises. Accordingly, this paper proposes:

Hypothesis 2b: Green trade barriers inhibit green technology innovation of Chinese entreprises.

Finally, when encountering green trade barriers set up by overseas countries, in addition to actively adapting to maintain the right to access the country’s market, enterprises can also transfer exports to countries where there are no green trade barriers, or withdraw from export markets, so as to avoid green trade barriers. Fontagne et al. (2015) examined the impact of SPS on export behavior of French enterprises, and found that SPS significantly reduces the export probability and export scale of enterprises.Further, Fontagne and Orefice (2018) found that TBTs can lead enterprises to export transfer and look for new markets that are not subject to TBTs. Bao and Zhu (2015) examined the trade restriction effect of exporting countries after encountering technical barriers, and pointed out that the lower the per capita income level of a country, the stronger the restrictive effect of technical barriers on its exports. In other words, enterprises in developing countries are more likely than developed countries to respond through changing their export decisions when they encounter technical barriers (Chen et al., 2006). At this time, enterprises do not need to transform or innovate their own technology, so this paper proposes:

Hypothesis 3: Green trade barriers will not affect the green technology transformation and green technology innovation of Chinese enterprises.

2.2 The Heterogeneity of the Impact of Green Trade Barriers on Enterprises’ Green Technologies

Enterprises with different characteristics may adopt different strategies when encountering green trade barriers. Strong R&D capabilities mean that enterprises may be able to achieve a leap over technical trade barriers with relatively low innovation investment. This increases the probability of enterprises in exporting countries choosing independent innovation (Xu and Liang, 2010). Accordingly, this paper proposes:

Hypothesis 4a: If green trade barriers have a promoting effect on enterprises’ investment in green technology transformation, the effect is more significant in enterprises with weak R&D capabilities.

Hypothesis 4b: If green trade barriers have a promoting effect on green technology innovation, the effect is more significant in enterprises with strong R&D capabilities.

Differences in the ownership attributes of enterprises may also lead them to make different choices in the face of green trade barriers. Liu (2000) pointed out that due to the low efficiency of production management and the lack of sensitivity to profit fluctuations, state-owned enterprises have a serious time lag in information collection and product adjustment after being hit by the market environment. At the same time, the innovation strategies of SOEs are usually guided by domestic policies, which is a long-term decision-making and has a strong path-dependent effect, which is difficult to change significantly due to a single external environment regulation (Liu and Xiao, 2022). The combination of these two factors will make SOEs more reactive in dealing with technical barriers to trade, more inclined to reduce exports, and try to avoid drastically adjusting their green technologies. Hu et al. (2019) used the introduction of the EU Child-Resistant Open Packaging Regulation as a quasi-natural experiment to verify this tendency of state-owned enterprises in the face of technical barriers to trade.

In contrast, non-SOEs are more likely to respond positively to green trade barriers because they are more efficient in decision-making and have a stronger willingness to adjust themselves flexibly to maintain market competitiveness and market share. At the same time, non-state-owned enterprises are generally subject to a higher degree of financing constraints, and it is more difficult to ensure the stability of R&D investment when encountering negative capital shocks (Aghion et al., 2012), and the possibility of passively abandoning green technology innovation due to insufficient funds is higher. Accordingly, this paper proposes:

Hypothesis 5: If green trade barriers have an impact on enterprises’ investment in green technology transformation and green technology innovation, these effects are more significant in non-state-owned enterprises.

3.1 Data Sources and Processing

This paper takes China’s 2009–2021 A-share heavily polluting industry listed companies as a sample to examine the impact of green trade barriers on green technology transformation and green technology innovation. Among them, the determination of heavily polluting industries refers to the provisions of the Notice on Printing and Distributing the Catalogue of Classified Management Industries for Environmental Protection Verification of Listed Companies issued by the Ministry of Environmental Protection in 2008. The data on green trade barrier measures are from the WTO environmental database.

When matching the affected industries of environment-related TBT notification with the listed industries, this paper refers to the practice of Zhao and Liu (2022), first matching the 6-digit HS codes involved in TBT published in the WTO environmental database with the standard industrial classification codes, and then matching the standard industrial classification codes with the 2012 industry classification codes of the China Securities Regulatory Commission, so as to obtain the industries of listed enterprises affected by TBT.

The data on the scale of investment in green technological transformation of enterprises comes from the annual reports of listed companies in heavily polluting industries. Referring to the practice of Zhang et al. (2019), this paper manually sorts out and counts the project expenditures directly related to green technological transformation in the details of construction projects in the annual report, including desulfurization projects, denitrification projects, energy-saving technological transformation, solid waste and sewage treatment, pollution control upgrading, exhaust gas technical transformation, environmental protection facility transformation, dust removal and waste gas treatment, etc. By summing up the above project data, the scale of investment in green technology transformation of the enterprise in the current year is obtained, and the total assets of the enterprise at the end of the year are used to standardize it to control the difference caused by the scale of the enterprise. Similar to Zhang et al. (2019), in order to enhance the readability of the regression coefficient in the following text, the investment in green technology transformation after standardization is multiplied by 100.

The patent data used to measure the green technology innovation of enterprises comes from the State Intellectual Property Office of the People’s Republic of China. Based on the “International Patent Classification Green List” launched by the World Intellectual Property Organization (WIPO) in 2010, this paper screens the green patent data of listed companies in combination with the International Patent Classification symbols, and obtains the annual number of green patent applications of the sample enterprises, including the number of green invention patents and green utility model patents.

Other company-level and industry-level feature data are obtained from the CSMAR database. The data on urban characteristics and environmental indicators are obtained from the China Urban Statistical Yearbook. In order to eliminate the influence of outliers, this paper winsorize all continuous variables at the 1% level, and excludes ST enterprises, enterprises with asset-liability ratios greater than 1, and samples with missing indicators. Finally, a total of 3019 observations are made from 2009 to 2021.

3.2 Benchmark Regression Model Setting and Variable Description

The benchmark regression model constructed in this paper is as follows:

Among them, the subscript i represents the enterprise, j represents the industry to which the enterprise belongs, r represents the province where the enterprise is located, and t represents the time. μ_j and λ_j represent the fixed effect of enterprises and industries, which is used to control the characteristics of enterprises and industries that do not change with time, γ_rt is the fixed effect of region-year, which is used to control the regional trend effect of province level over time, ε_ijrt is the stochastic disturbance term. In order to eliminate the possible heteroskedasticity and autocorrelation, this paper clusters the standard errors at the industry level during regression.

Envst_it is the scale of investment in green technology transformation of enterprise i in period t, and Enpa_it is the number of green patents applied for by enterprise i in period t, and they are logarithmic processed. TBT_jt-1 is the number of green trade barrier measures encountered by industry j in t-1 year, measured by the logarithm of the number of environment-related TBT notification in industry j in that year. Controls is a series of control variables, among which the control variables of enterprise characteristics include: enterprise size (Size), return on assets (ROA), asset-liability ratio (Lev), government subsidies (Subsidy), and Tobin Q (Tobin). The control variables of urban characteristics include: local gross domestic product (GDP), fiscal expenditure (Fiscal), and industrial sulfur dioxide emissions (SO2). The control variable of industry characteristics is the Herfindahl Index (HHI), which reflects the degree of competition in the industry market. Tables 1 and 2 provide definitions and descriptive statistics for each of the main variables, respectively.

4.1 Benchmark Regression Results

Table 3 reports the full-sample regression results for (1) and (2) of the benchmark model and the subsample regression results for enterprises whose main business description includes exports. Columns (1) and (2) show the impact of green trade barriers on enterprises’ investment of green technology transformation. It can be seen that the estimation coefficient of the TBT variable is significantly positive at the level of 1%, both for the whole sample and for the sub-sample of export enterprises, indicating that the green trade barrier measures will lead to increased investment in green technology transformation, supporting the research hypothesis 1 and rejecting the research hypothesis 3.

Columns (3) and (4) report on the impact of green trade barriers on green technology innovation of enterprises. The results show that the estimation coefficient of the TBT variable is significantly negative at the level of 5% in both the whole sample and the sub-sample of export enterprises, indicating that green trade barriers will have a significant negative impact on the number of green patent applications of enterprises, which supports research hypothesis 2b and also rejects research hypothesis 3.

4.2 Robustness Tests

In order to test the robustness of the regression results, the explanatory variables are first replaced. Referring to Zhang et al. (2019), this paper adds the greening fee and sewage fee in the “management expenses” section of the income statement of heavily polluting enterprises to the investment of green technology transformation of enterprises, and uses the operating income to standardize them. At the same time, only the number of green invention patent applications with higher innovation content is used as a measure of enterprise green technology innovation.

Secondly, in order to take into account the intensity of the industry’s encounter with green barriers, this paper refer to the practice of Li et al. (2014) and multiplies the number of environment-related TBT notification encountered by the industry’s export dependence as a proxy variable for green trade barriers.

In addition, it is considered that the green technological transformation and green technology innovation of enterprises are not only affected by the green trade barriers faced by their own industries, but also by the green trade barriers encountered by their downstream industries. This paper also uses the number of environment-related TBT notification in the entreprises’ industry and its important downstream industries as a surrogate measure of the enterprises’ encounter with green trade barriers.

Finally, if there is a linkage between the environmental regulatory policies of domestic industries and the green trade barrier measures of overseas countries, it may lead to endogeneity due to the omission of variables. In order to solve this problem, this paper uses self-sampling (Bootstrap) to repeatedly extract 3000 subsamples from the whole sample for regression estimation, and then investigates the Bootsrap standard error and confidence interval of the estimation coefficient of TBT variables.

The logic of this test is that the domestic industry environmental regulation policies are less likely to overlap with the green trade barrier measures within each sub-sample interval. Therefore, if the estimation coefficient of the TBT variable obtained by Boostrap is still significant after randomly selecting enough subsamples for regression, the possibility of confusion in the domestic industry environmental regulation can be ruled out. The results of the above four types of robustness tests support the conclusions of this paper.

5.1 Heterogeneity of R&D Capabilities

In order to examine the differences in the impact of green trade barriers on enterprises with different R&D capabilities, for each year this paper ranks enterprises in the top one-third of the total number of past green patent applications since 2009 as enterprises with high R&D capabilities, while those in the bottom one-third of the total number of applications as enterprises with low R&D capabilities. This rolling grouping method can better capture the dynamic changes of the enterprsie’s R&D capabilities. Table 4 reports the regression results grouped by enterprises’ R&D capabilities. It can be seen from columns (1) and (2) that green trade barriers have a significant positive impact on the green technology transformation investment of enterprises with low R&D capabilities, but have no significant impact on the green technology transformation investment of enterprises with high R&D capabilities, supporting the research hypothesis 4a. In terms of green technology innovation, columns (3) and (4) show that green trade barriers have a trap effect on enterprises with low R&D capabilities, but have no significant impact on green technology innovation of enterprises with high R&D capabilities, which cannot provide support for hypothesis 4b. There are two possibilities for this result: one is that the technology of enterprises with higher R&D capabilities has a higher level of green technology, and the probability of green trade barriers constituting substantial constraints on them is low, so the impact on its green technology transformation and innovation is not significant. Second, the role of technical barriers in promoting enterprise technological innovation may take longer to reflect.

5.2 Ownership Heterogeneity

In order to investigate the heterogeneity effect brought by the difference of enterprise ownership. This paper adds the interaction term of the enterprise ownership variable Owner and the green trade barrier variable TBT to the benchmark model, in which when the enterprise is a non-state-owned enterprise, the Owner variable is taken as 1, otherwise 0. Table 5 reports the regression results of the heterogeneity test. According to column (1) of Table 5, when the explained variable is green technology investment, the estimation coefficient of the interaction term is significantly positive, indicating that compared with state-owned enterprises, non-state-owned enterprises will increase more green technology investment after encountering green trade barriers. The regression results in column (2) show that compared with state-owned enterprises, green trade barriers significantly inhibit the green technology innovation of non-state-owned enterprises. In fact, according to the estimation coefficient of the TBT variable, the inhibitory effect of green trade barriers on green technology innovation of state-owned enterprises is not significant. On the one hand, this supports Liu and Xiao (2022) that the innovation strategy of SOEs is long-term and less affected by external shocks, and on the other hand, it also suggests that financing constraints may be a mechanism leading to the trap effect of green trade barriers.

6.1 Medium- and Long-Term Impact of Green Trade Barriers

In order to analyze the technological innovation effect of green trade barriers over a longer period of time, this paper replaces the explained variables in the original model (2) with the number of green patent applications filed by enterprises in the second, third and fourth years after the notification of green trade barrier measures.

There are two main reasons why the focus is not paid to the longer time period: first, green trade barriers are usually based on the production technology used by enterprises in countries with environmental technology advantages (Lu and Yang, 2003), so the technical uncertainty is relatively limited, and the innovation cycle is not too long. Second, the innovation measure indicator used in this paper is the number of green patent applications, which has the advantages of timeliness and authenticity compared with other innovation measures (Li and Zheng, 2016), and the probability of a long time lag between this indicator and innovation behavior is low. Therefore, this paper argues that the four-year period after the notification of environment-related TBT should basically cover the impact period of green trade barriers on green technology innovation of enterprises.

According to the regression results reported in Table 6, the impact of green trade barriers on the number of green technology patent applications in the next 2 to 4 years is not significant. One possible explanation is that green trade barriers are a way for countries with advantages in environmental technology to implement green rules and standards, enhance their “soft power” and economic discourse (Yang, 2020). Therefore, its positive effect on innovation is more reflected in the direction of guidance, guiding the green technology innovation direction of exporting countries to converge with that of countries with advantages in environmental protection technology, and the impact on the number of innovations is relatively weak.

Another possible explanation has to do with the falsity of green trade barriers. In other words, there may be a large number of green trade barriers encountered by Chinese enterprises that increase their cost burden and restrict market access in the name of environmental protection, which will not help them improve their green technology innovation.

6.2 Alleviating Financing Constraints

In order to investigate whether the alleviation of financing constraints can alleviate the inhibition of green trade barriers on green technology innovation, this paper divides enterprises into high financing constraint group and low financing constraint group according to the degree of financing constraint faced by enterprises, and the measurement method of enterprise financing constraint degree comes from the KZ index constructed by Lamont et al. (2001). When the enterprise KZ index is below the median of all enterprise KZ indexes in the current year, the enterprise belongs to the low financing constraint group in that year, and the financing constraint group variable NFC is taken as 1; On the contrary, it is considered that the enterprise belongs to the high financing constraint group in that year, and the financing constraint group variable NFC is taken as 0.

Table 7 reports the regression results after adding the financing constraint grouping variable NFC and the interaction terms of green trade barrier TBT to models (1) and models (2). According to column (2), when the explained variable is green technology innovation, the estimation coefficient of the interaction term is not significant, indicating that the reduction of financing constraints fails to reduce the inhibiting effect of green trade barriers on green technology innovation. In fact, according to the estimation coefficient of the interaction term in column (1) of Table 7, the reduction of the degree of financing constraints significantly strengthens the promotion effect of green trade barriers on enterprises’ investment in green technology transformation.

This result supports the conclusion of Wan et al. (2021) that enterprises are more inclined to achieve green transformation through technological transformation. The relaxation of financing conditions will only encourage enterprises to further expand investment in green technology transformation in order to maintain export market share more quickly. In other words, in the face of green trade barriers, the decision of enterprise to transform or innovate green technology depends not only on funding, but also on managerial myopia, which is more likely to focus on the outcomes and benefits that can be achieved at the moment (Stein, 1989). Hu et al. (2021) found that shortsightedness of managers inhibits enterprise R&D spending.

In order to verify that managerial short-sightedness is the reason why it is difficult to reduce the inhibition of green trade barriers on green technology innovation by the relaxation of financing constraints, this paper draws on the practice of Hu et al. (2021) to carry out text analysis based on the discussion and analysis of managers of listed companies (MD&A), and calculates its proportion in the total number of words in MD&A by counting the word frequency of words involving “short-term vision” in MD&A content, so as to construct a measure indicatior of the degree of myopia of managers. Each year, we rank the sample enterprises according to this indicator, and the top one-third of companies with the highest Myopia value are regarded as those with high manager shortsightedness, while the bottom one-third of companies with the smallest Myopia value are regarded as those with low manager short-sightedness. Subsequently, we examine the effects of alleviating financing constraints in each of the two groups.

As can be seen from Table 8, the coefficient of the interaction term of TBT and NFC is significantly positive in the group of enterprises with low short-sightedness, indicating that the alleviation of financing constraints can weaken the inhibitory effect of green trade barriers on green technology innovation. However, in the group of enterprises with a high degree of short-sightedness, the interaction term coefficient is not significant, and the reduction of financing constraints does not reduce the inhibition of green trade barriers on green technology innovation. Therefore, the trap effect of green trade barriers is the result of the combined effect of financing constraints and short-sightedness of managers, and reducing the degree of financing constraints faced by enterprises is not enough to eliminate the negative impact of green trade barriers on green technology innovation.