Usually, literature relates financial development and innovation to growth and supports the idea that they are positive as well as strongly related to one another. It has been argued by King and Levine, (1993) that more efficient and effective financial systems can significantly enhance innovation and, as a result, accelerate economic growth. Benfratello et al. (2008) examined a range of Italian firms from the 1990s and found that financial development was associated with process innovation.

In recent years, some scholars have demonstrated that finance and innovation may not have a monotonic relationship. Zhu et al. (2020) found that higher levels of financial development are correlated with lower levels of innovation based on a dynamic panel threshold model. Trinugroho et al. (2021) also revealed an inverted U-shaped non-linear relationship between finance and innovation by utilizing panel data from 68 developed and developing countries during 1995–2018. Further studies have shed light on what role equity markets and credit markets play in the development of innovation. Regarding the equity market, there are two opposing viewpoints regarding its development, both positive and negative. On the one hand, since firms’ equity financing is not subject to strict collateral requirements, it provides a set of risk management options for investments in riskier, yet innovative projects (Levine, 2005). This makes equity financing particularly suitable for young, high-tech companies (Carpenter and Petersen, 2002). Le et al. (2019) demonstrate that stock market-based financial development is positively correlated with patent applications, which is considered a proxy for innovation. On the other hand, some literature suggests that the development of equity markets may hinder innovation. In advanced countries, fully functioning markets could lead to a market for corporate control, which might encourage firms to acquire other companies instead of expanding organically, which could be detrimental to the real economy (Singh, 1991).

Concerning the credit market, two implications relating to its role in innovation may also be considered. Given the pledgeable nature of credit financing, credit markets tend to favor conservative investments, which discourage firms from undertaking innovative and valuable projects (Morck and Nakamura, 1999). Xiao and Zhao (2012) have shown that bank credit promotes innovation in countries with low levels of government ownership of banks; however, in countries with a higher level of government ownership of banks, it can sometimes be detrimental. Hsu et al. (2014) examine the effect of the equity market versus the credit market in their study, demonstrating that the development of the credit market seems to inhibit innovation as a result of the lack of price signals.

Nonetheless, existing studies neglect the importance of the interactive effect of financial development and innovation which are required to be seen in conjunction with the environment.

An effective and functional financial sector contributes to economic growth (Schumpeter, 1911; King and Levine, 1993). According to Levine (2005), the development of financial systems (banks, capital markets, laws and regulations about banks, and capital markets) is geared towards reducing barriers to financing, monitoring, and gaining access to financial information at a lower cost, thereby supporting environmental sector activities through this channel. However, its precise impact on environmental quality remains unclear. One strand of the literature suggests that well-developed financial sectors may lead to environmental degradation. For example, Park et al. (2018) reported that environmental quality is adversely affected by financial development in selected European Union (EU) countries because domestic credit to the private sector is invested in non-environmentally friendly projects. In addition, private credit growth has been associated with an increase in CO₂ emissions as a result of higher energy consumption. Further, based on a comparison of developed and developing countries, Shoaib et al. (2020) investigated the impact of financial development on CO₂ emission over the period 1999 to 2013. The study shows that financial development has a more significant and positive impact on carbon emissions in developing countries. Sunday Adebayo et al. (2022) demonstrate that the effects of financial development vary between countries owing to differences in political, social, and economic shocks. Moreover, studies by Al-Silefanee et al. (2022) on Islamic countries, Haug and Ucal (2019) on Turkey, Nasir et al. (2019) on ASEAN-5, Fang et al. (2020) on China, and Yang et al. (2021) on BRICS countries, Abid et al. (2022) on G8 countries, documented negative effects of financial development on the environment.

Another strand of literature reported that in fact financial development positively affects the environment by decreasing carbon emissions. Specifically, Paramati et al. (2017) reported stock market growth reduces CO₂ emissions in developed economies since it provides investors with access to additional funding sources and can therefore result in higher levels of investment in clean energy projects instead of conventional energy. Likewise, Anees et al. (2019) hold that higher financial development decreases environmental degradation since greater financial sector development contributes to the development of energy-efficient technologies. Kirikkaleli and Adebayo (2021) document that within a global context, financial development contributes to environmental sustainability over the long run. Apart from that, a surplus of empirical studies has also reported a negative relationship between financial development and CO₂ emissions. (see, e.g., Abid, 2017; Shahbaz et al., 2018; Umar et al., 2020; Kirikkaleli et al., 2021; He et al., 2021; Brown et al., 2022).

Indeed, as evidenced by the literature, studies on the linkage are mixed and not yet conclusive. Despite this, it also indicates some limitations in the existing body of knowledge. This study intends to get a more comprehensive picture of financial development not only by incorporating the credit creation as well as equity market but also demonstrate its specific mechanism through which act on innovation via different financial markets, thereby influencing the environmental sector.

Accordingly, from the foregoing, studies of these links are mixed and far from being conclusive. To reiterate, existing studies neglect the importance of the interactive effect of financial development and innovation which are required to be seen in conjunction with the environment. Currently, no existing research efforts have been conducted on the specific mechanism through which acts on innovation via different financial markets, thereby influencing the environmental sector. Thus, the present study expands on previous research by systematically analyzing the relationship between innovation and environmental quality under different levels of financial development, both in the credit market and equity market.

As the first objective of this research is to investigate the effect of innovation on environmental quality, the baseline model is as follows. Next, we categorized the sample into two categories according to the level of financial development. Initially, the countries are sorted by level of financial development. Following this, we designate the 26 top countries as those with high financial development, while the 26 bottom countries are considered to have low financial development. By following this strategy, we can gain an understanding of how financial development may moderate the relationship between innovation and CO₂ emissions. Within each group, the following equation is considered: E N Q U c t = α 0 E N Q U c t − 1 + α 1 I N N O c t + α 2 C O N T R O L c t + ε c t (1) Where c , t denote country and year, respectively. E N Q U c t and E N Q U c t − 1 are the current and lagged indicator of environmental quality proxied by CO₂ emissions; I N N O c t denote innovation; C O N T R O L c t denotes the control set including GDP, Trade, Energy, Human capital, Population, and Urbanization; ε c t captures the stochastic error term. Considering possible endogeneity issues in this study, we implement a two-step dynamic system generalized method of moments (GMM) using lagged dependent variables and regressors as instruments.

Thus, from Equation 1, innovation is supposed to increase pollution if the estimated coefficient ( α 1 ) is positive and significant. Moreover, the study also explored whether the financial development of a country moderated the impact of innovation on environmental quality. Therefore, we introduce an interactive term of I N N O c t and F D c t thus producing the equation below: E N Q U c t = β 0 E N Q U c t − 1 + + β 1 I N N O c t + β 2 F D c t * I N N O c t + β 3 C O N T R O L c t + ε c t (2) Where c , t denote country and year, respectively. E N Q U c t and E N Q U c t − 1 are the current and lagged indicator of environmental quality which is proxied by CO₂ emissions; I N N O c t denotes innovation; F D c t denotes the financial development indicators, including credit market development and equity market development. C O N T R O L c t denotes the control set including GDP, Trade, Energy, Human capital, Population, and Urbanization; ε c t captures the stochastic error term.

The coefficient for the interactive term-- β 2 is expected to provide insight as to how financial development impacts innovation–pollution nexus. The marginal effect of innovation on environmental quality is β 1 ^ + β 2 ^ F D c t . Theoretically, in countries with higher levels of financial development, innovation has a smaller impact on CO₂ emissions. Consequently, if there were any ''diminishing effect'' on innovation-pollution caused by financial development, it is expected that β 1 will be greater than 0 and β 2 will be negative.

To give a precise estimated threshold value where the effects of innovation on the environment under the impact of financial development begin to change, this study adopts a novel GMM method proposed by Hansen (1999), created by Seo and Shin (2016), and Seo, Kim, and Kim (2019), where the transitional variable and other covariates are allowed to be endogenous and were extended from prior panel model (Caner and Hansen, 2004; Kremer et al., 2013).

Financial development is considered a threshold indicator in the study, whereas innovation is a regime-dependent variable that varies based on the estimated threshold of financial development. Our notation here is similar to that used in Eq. 1. Accordingly, the model given in Eq. 2 follows: E N Q U c t = γ 0 + γ 1 E N Q U c t − 1 + γ 2 x c t ′ + k ( F D c t − γ ) I { F D c t > γ } + ε c t (3) Where c , t denote country and year, respectively. E N Q U c t and E N Q U c t − 1 are the current and lagged indicator of environmental quality proxied by CO₂ emissions; F D c t denotes the financial development indicators, including credit market development and equity market development. x c t ′ represents all the explanatory variables, including INNO, GDP, Trade, Energy, Human capital, Population, and Urbanization. The variable I (·) operates as an indicator, giving a value of one provided that the condition in parentheses holds true and zero otherwise. Finally, the bootstrap test of linearity developed by Seo et al. (2019) is adopted to test if there is a threshold or not. If there is a threshold, H 0 : δ 0 = 0 should be rejected for any γ ∈ Γ , where Γ denotes the parameter set, W n ( γ ) is the standard Wald statistic for each fixed γ , and δ ^ ( γ ) is the GMM estimator of δ for a given γ , as shown in the equation below. s u p W = s u p γ ∈ Γ W n ( γ ) = s u p γ ∈ Γ n δ ^ ( γ ) ′ Σ ^ δ ( γ ) − 1 δ ^ ( γ )

Next, we examine how innovation impacts CO₂ emissions under different samples. We split the full sample into the low and high levels of credit market and equity market development countries. As shown in Table 4 and Table 5, the effect of innovation on CO₂ emissions for low financial development countries is higher than that of high financial development countries. Throughout subsamples, we find that the coefficient estimate of innovation (INNO) is of greater magnitude and statistically more significant among countries with a low level of financial development (LFD). Additionally, we observe that for credit market development groups there is a greater dampening effect of innovation than for equity market development groups, given the coefficients of innovation. Below the Tables, we record the p-values for the AR (2) and Hansen J-tests. It is apparent from the AR (2) test that there is no significant correlation between the error term and the lagged dependent variable, indicating the use of two lags as a valid instrument. The Hansen test indicates that the specifications are not over-identified. This is in accordance with Figure 1.

The interaction analysis is summarized in Table 6. For both credit market development and equity market development, there is consistent evidence that the interaction coefficient is negative, and that the coefficient of innovation (INNO) is positive and significant. As shown in Eq. 2, the average marginal effect of innovation on CO₂ emission is β 1 ^ + β 2 ^ F D i t , since β 1 ^ > 0 and β 2 ^ < 0, the impact of innovation on environmental quality is decreasing in proportion to financial development. The results indicate that the marginal effect follows a downward trend and is mostly positive. These findings are in accordance with Tables 4 and 5.

We can observe from Table 7 that a threshold is identified where γ (financial development) equals 4.181 and the bootstrap linearity test confirms the existence of the threshold effect. In simple terms, it is evident that innovation has a significantly different impact on CO₂ emissions in the two regimes. In more detail, as shown in Figure 1 and Figure 2, innovation has a statistically significant and stronger positive effect on CO₂ emissions in the lower financial development regime but is less correlated with CO₂ emissions in the higher financial development regime. Therefore, it can be concluded that, with the growth of countries’ level of financial development, the CO₂ emissions will increase at first and then reduce more slowly once the financial development exceeds the threshold value. The estimated threshold value of financial development is 65 and 16% as a share of GDP for the credit market and equity market respectively. The threshold value doesn’t show much difference when we use patent applications and grants as two indicators.

As discussed above, equity markets and credit banks appear to exert different influences on the innovation-pollution nexus. Inspired by the work of Yao and Tang (2021), we introduce the financial structure to further investigate whether more equity-funded countries have more stringent pollution controls on innovative activities. We define Financial Structure (FS) as the proportion of stock market financing in total financing through credit and stock markets (De Haas and Popov, 2019): F i n a n c i a l   S t r u c t u r e c , t = M a r k e t   c a p i t a l i z a t i o n   c , t P r i v a t e   s e c t o r   c r e d i t c , t + M a r k e t   c a p i t a l i z a t i o n   c , t

The countries are divided into two groups by value of Financial Structure on an ascending scale. In our sample of countries, the top half is defined as more equity-based, while the bottom half is defined as more bank-based. We examine the structure of financial systems because the potential effect of innovation on pollution may be shaped by the financing mechanism of innovative activities in a country.

Table 8 provides the effect of financial structure on the innovation-pollution nexus. The results show that INNO has the expected signs. Specifically, innovation has a significant and positive effect on CO₂ emissions in countries that are relatively more bank-based while in equity-based countries, innovation is found to have a significant and negative impact on CO₂ emissions. This suggests that stock markets are better at funding innovation in an environmentally friendly way than banks.