Determinants of Green Innovation: The Role of Monetary Policy and Central Bank Characteristics

Abstract

The current global energy crisis has prompted a comprehensive investigation into its influencing factors. It is hypothesised that a set of monetary, macro-environmental, and institutional variables causally affect the transition to green development in a holistic model. Monetary expansion and central bank characteristics are required for economic and environmental development. The current study investigates and rigorously verifies the impact of expansionary monetary policy actions on green innovation, using a panel of 109 countries from 2010 to 2018. Overall, specific actions have a substantial positive effect on the performance of green innovation. A rise in per capita GDP, government spending, and improvement in bureaucracy all promote green economic activity. Green innovation is significantly affected by developing nations’ central bank independence and lower interest rates. Expansionary monetary policy, central bank transparency, and energy variables promote green growth in developed countries and green innovation in Latin American countries and in East Asian and Pacific countries. Finally, green innovation is more affected by expansionary monetary policy in countries with high institutional quality, industrial concentration, and energy intensity, and inflation and trade openness serve as deterrents in the monetary expansion–green development nexus.

1. Introduction

All countries, particularly developing ones, face numerous environmental difficulties while pursuing economic growth [1,2] (Beckerman 1992; Zhang and Wen 2008). Degradation of the environment and ecological disasters today incur such enormous costs that they pose important questions about sustainable development [3]. Developing a sustainable development plan that achieves a “harmonious symbiosis” between development and the environment is a global concern [4,5]. A sustainable development strategy is sought to adhere to the objectives of green economic development while actively addressing environmental concerns [6,7,8].

Transitioning to a low-carbon world will necessitate substantial investments in “green” industries [9,10,11,12]. From a macroeconomic standpoint, investment is expenditure; it consists of the purchase of investment products and services that will be employed to manufacture some consumer item. To undertake investments, businesses evaluate their risk and profitability. Companies that can quickly modify the traditional model of supplying products and services and carry out green innovation and reform will have a greater competitive advantage [13,14]. In the classic economic development model, environmental pollution, resource restrictions, and ecological degradation obstruct economic progress. To achieve a “win-win” development of economic transformation and environmental protection, investments in green innovation must be embraced [15]. Investment, like any other type of spending, necessitates the availability of sufficient financial resources. Given the up-front costs of investments, which are particularly expensive in the case of renewable energy production, businesses are often unable to finance them with their savings and, thus, require external financing. In other words, they must borrow funds from a third party before they invest.

Clean energy financing mobilisation is essential for limiting global warming to 1.5 °C and preventing catastrophic climate change [16]. A total of 1.5 trillion USD must be invested annually in green initiatives [17]. With estimates for the year 2020, the corresponding investments are at one-third of the assets required [18]. Numerous nations have established measures to encourage renewable energy investment by lowering these obstacles [19]. After the 2008–2010 financial crisis and its aftermath, the number of inventions related to environmental concerns has risen. This process is regulated by numerous factors, many of which are examined in this research study.

Monetary policy is one of the major variables affecting innovation in the green sector. It has only recently been researched in the literature, but not thoroughly enough [20]. By balancing the portfolio, an expansionary monetary policy boosts consumption and energy use by directly supplying money or lowering interest rates [21]. As a result, the demand for products will increase across all markets. Industrialists prefer to use conventional manufacturing processes if innovation-related financing is offered at a higher interest rate. An increase in pollutant emissions is caused by a rise in the consumption of less ecofriendly technologies [22]. Similarly, the supply channel contends that expansionary monetary policy encourages economic growth, motivating consumers to purchase renewable goods and pressuring producers to increase the output to satisfy the increased demand. By utilising green innovation, more renewable energy products may be produced for less money.

The signing of the Paris Agreement and the United Nations 2030 Agenda for Sustainable Development in 2015 ignited a burst of efforts by central bankers to mobilise the resources required to achieve the transition to a carbon-neutral economy. Central banks are in charge of “efficient green finance” by pricing climate risks due to financial regulatory monitoring on money and credit flows [23,24]. Central bankers have recently been addressing the risks of climate change uncertainty to monetary and financial policy [25,26]. However, it would be inaccurate to refer to these financial institutions as “green central banks” [27]. The two main characteristics of central banks are independence and transparency. They are directly related to the process and policy basis upon which the central bank makes decisions [28,29,30,31,32,33]. The present study analyses these two characteristics in the context of a mainly monetary model to determine the channels and processes through which they operate and how they relate to green innovation [34,35].

According to the research, environmental innovations are crucial to shifting to a green economy. By analysing the topic from all angles, we attempt to pinpoint the key variables affecting the quality and quantity of green innovation in this research. We also examine various other factors that appear to impact the green economy, with a particular emphasis on monetary policy and the features of central banks that are typically positively connected with it. Fiscal policy uses government spending and GDP per capita, which seem to advance green innovation [36,37]. Bureaucracy [38] and institutional quality [39] are employed in relation to institutional variables. CO₂ emissions [40], energy intensity, and losses in production, transportation, and distribution were included as energy variables. Trade openness [3,41] and economic freedom [42] were utilised as trade liberalisation factors since they appear to have an increasing impact on green innovation. The effects of population and urbanisation [43,44], industrialisation [3], inflation [45], and central government debt [46] were also tested as a final analysis of heterogeneity and moderating effects. This study specifically addresses the following research questions: Does an expansionary monetary policy impact green innovation? How do the characteristics of central banks influence the above effect? Do these relationships differ or have any similarities across different world regions? Are there economic, energy, institutional, or additional factors and moderating effects that differentiate these effects, and if so, how do they differ? These questions necessitate a more thorough and accurate examination and help in addressing the issue comprehensively.

By treating the factors that influence green innovation holistically in this research, we attempt to fill in the gaps in the literature. Our work differs from and complements the existing literature in various ways. First, by using the same effect model, we conclude that monetary factors (money supply and interest rates) and central bank features (independence and transparency) could enhance the green economy. Second, additional robustness tests are conducted. They modify the measuring indicators of green innovation to continue the empirical investigation and provide consistent results. In addition, we employ various static (FE, RE, PCSE) and dynamic (PCSE (ar1), FM–OLS) econometric techniques to further account for lag effects, cross-sectional dependence, and endogeneity concerns. Simultaneously, a model of 109 countries is examined to assess the change in the correlation of variables in the period following the financial crisis (2010–2018). Third, numerous heterogeneity analyses are performed, with geographic region disaggregation being the most prominent. We show that expansionary monetary policy boosts green innovation in Latin American countries (LAC) and East Asian and Pacific (EAP) countries and reinforces the model more than in Sub-Saharan and African (SSA) and Middle Eastern and North African (MENA) countries. At the same time, we carry out specialised analyses of factors in developing and developed countries separately. In addition, the sample is divided based on institutional, social, energy, and liberalisation policy variables (high-low) to study the model’s coherence and the differences in variable intensities. Finally, we demonstrate how certain variables, such as inflation (a result of monetary policy), population (a result of social policy), trade freedom (a result of government policy), and losses in the distribution of energy (a derivative of energy policy) act as moderating factors on the relationship between monetary policy and green innovation.

The remainder of the paper unfolds as follows: Section 2 comprises the theoretical background. The third section analyses the methodology and the data used. Section 4 elaborates on the findings and robustness checks. Finally, Section 5 concludes the study.

2. Theoretical Background

Several monetary policies, energy policies, and other main macroeconomic and institutional factors are mentioned in the existing literature that might affect green innovation. Monetary policy factors concentrate on the money supply, interest rate, inflation, and central bank characteristics, while energy policy factors encompass energy prices, energy intensity and losses, and current greenhouse gas emissions. Other leading macroeconomic and institutional factors involve economic development, government expenses, trade freedom, institutional quality, and the implementation of competitive markets.

2.1. Monetary Policy Factors

As previously stated, there is a substantial investment gap between current levels and what would be required to transition to an environmentally friendly economic structure [47]. According to Campiglio [23], a weak macroeconomic climate after the financial crisis and an unappealing risk-return profile are the primary obstacles impeding adequate financial backing for green innovation efforts. According to popular thinking, these two crucial characteristics may be closely related to monetary policies. Studies that explored whether expansionary monetary policy affects the transition to a green economy through innovation reached a positive conclusion [41,48,49]. Broad money (M2) and the real interest rate are the two most frequently cited variables in the literature for monetary policy concerning the green economy.

Lv et al. [50] showed that environmental restrictions have a substantial role in mitigating the impact of an increase in financial availability on green innovation. Therefore, firms or other economic participants subject to more stringent environmental regulations may be incentivised to reduce compliance costs, increasing the potential demand for green innovation in the economy and directing more of the additional real purchasing power generated by expansionary monetary actions into green innovation-related activities. Yin et al. [41], utilising a panel of 133 countries covering the period of 1960–2018 and employing broad money and reserve money as monetary variables, demonstrated that expansionary monetary policy has a considerable beneficial impact on the efficacy of green innovation. Zhao et al. [51] achieved similar outcomes. Wen et al. [3] demonstrated that a moderately expansionary monetary policy helps reduce the issue of corporate finance limitations and encourages businesses to engage in exploratory green innovations that are beneficial to their long-term development.

Egli et al. [52] investigated empirical data in Germany and 133 photovoltaic and wind installations over 18 years. They found that reducing the real interest rate leads to a long-term improvement in using innovative processes in renewable energy sources. In addition, Hashmi et al. [20] found that an increase in the real interest rate results in a depreciation of the currency rate. This exchange rate depreciation inhibits imports of renewable energy goods, resulting in a decline in renewable energy use in the economy. Research has also examined how monetary policy affects the environment. Qingquan et al. [53] reviewed Asian economies from 1990 to 2014 and showed that a growing money supply caused rising CO₂ emissions. A contractionary monetary policy, however, would decrease investments in technical innovation and could reduce the likelihood of producing greener technologies, which could eventually harm the environment. In addition, expansionary monetary policy may raise aggregate demand via growing consumption and investment activity. Hence, an expansionary monetary policy may contribute to environmental degradation if the energy sector does not convert to cleaner sources throughout the expansionary period [54].

Finally, the effect of inflation (as a result of monetary policy) on green growth has also been investigated. Typically, inflationary high fossil fuel energy prices encourage the search for alternate and renewable energy sources [45]. In addition, Bird et al. [55] showed that high wholesale power rates enhance the relative competitiveness of wind energy generation in the United States. According to the findings of Chang et al. [56], a greater variance in the consumer price index may be positively associated with renewables’ contribution to the energy supply.

Central banks play a vital role in implementing countries’ monetary policies. How they conduct business influences the manner in which they engage in politics. Thus, their primary characteristics, namely, independence and transparency, must be analysed in terms of their impact on the implementation of energy policy. By comparing the green policy actions of 135 central banks, Dikau and Volz [34] found that just 12% of central banks have explicit sustainability mandates, while 40% are mandated to assist the government’s policy aims, the majority of which include sustainability goals. To ensure macro-financial stability, all institutions should incorporate climate-related physical and transition risks into their policy frameworks, given that climate risks can directly impact central banks’ conventional core responsibilities. From this perspective, their independence will facilitate the adoption of greener policies. According to D’Orazio and Popoyan [57], the decision to implement green rules is dependent not only on the mandate, but also on the independence of the central bank and the nature of the interplay between monetary and prudential policy. A central bank that hosts green prudential regulation under its governance roof can prevent conflicts between monetary policy and green regulation due to linked transmission processes.

Simultaneously, central banks, whose independence is a vital instrument, should make defensive and awareness-raising efforts to guarantee resilience in the face of increasing climate-related threats and to secure the consistent and continuous conduct of monetary policy [58]. Finally, in their research, Yin et al. [41] find that expansionary monetary policy actions have a more significant positive impact on green innovation activities in economies with higher levels of central bank independence, an adequately developed stock market, and better protection of property rights. They indicate that central bank independence, stock market development, and property rights are crucial in transmitting monetary policies. In addition, they discovered that, compared to Western countries, Asian emerging economies are associated with much poorer central bank independence performance.

The preceding discussion demonstrates that expansionary monetary actions contribute to alleviating the risk premium and a depressed macroeconomic environment, the two primary factors inhibiting appropriate financial support for green innovation efforts. Hence, we draw the following central hypothesis of our study: Expansionary monetary policies improve the performance of green innovation. It also emerges from the preceding literature review that research on the effect of monetary policy on green innovation has been conducted variable by variable. We address this deficiency by employing a holistic approach and incorporating the most significant monetary policy variables into the same model to investigate their effect on green energy. By evaluating the disparities between advanced and developing countries and regions of the world, our findings are significant and contribute to the complex decision making required for monetary expansion for sustainable development.

2.2. Energy Policy Factors

This section examines the impact of energy and environmental variables on green innovation, starting with energy consumption. It is widely recognised that energy consumption positively influences green innovation through economic expansion and the pressure it imposes on environmental degradation [59].

The consumption of energy is typically associated with modernisation and economic prosperity; however, energy consumption also creates environmental issues in nations around the globe [60]. Much empirical research has been conducted on the relationship between energy use, economic development, and green innovation. In most of these studies, the subject of environmental pollution frequently arises [61,62]. Green growth is essential for lowering greenhouse gas emissions and sustaining global economic growth. Considering that 70% of emissions result from excessive energy consumption, energy efficiency is a crucial subject [59]. In recent years, governments have heavily invested in developing green technology and innovation to increase energy efficiency. Likewise, government entities have been given the authority to oversee the execution of energy programs that increase energy efficiency [60].

The overconsumption of traditional fossil fuels is thought to be the cause of climate change, which is currently the most significant environmental concern confronting the world [63]. According to several studies, environmental pollution can stimulate technological innovation. Su and Moaniba [64] used data from 70 countries to investigate whether technological innovation is sensitive to shifts in the surrounding environment. According to their research, the CO₂ emissions released by burning gas and liquid fuel promotes technological advancement. Several academics have made similar observations; for example, Costantini and Crespi [65] argued that the massive volume of CO₂ emissions have helped technological advances in renewable energy. From an energy consumption and structure standpoint, energy is the driving element for sustainable economic expansion. Coal and oil dominate global energy consumption [66]. Nonetheless, the increased pressure from carbon dioxide emissions is forcing authorities to develop new methods for producing green energy.

Energy intensity (the amount of energy consumed per unit of total output) is one variable that may influence the relationship between monetary expansion and green innovation. Energy intensity should drop in countries where renewable energy sources replace conventional ones. In a sample of 18 OECD nations, Wurlod and Noailly [67] found that green innovation contributed to the drop in energy intensity in most industries. Thus, a 1% increase in green patenting efforts in a specific sector is related to a 0.03% reduction in energy intensity. The influence is more significant in energy-intensive industries and in recent years. In addition, they show that the effect of an additional green patent on energy intensity is more powerful than that of a typical non-green patent. This study evaluates if energy intensity influences the process of monetary expansion and the impact of central bank characteristics on green innovation. By this method, we expand the existing body of literature by examining the current relationship from both sides.

Finally, one of the relationships that is explored is whether the energy losses that occur throughout production, distribution, and consumption are related to the development of green innovation. According to Oseni [68], providing new energy sources promotes economic growth and sustainability. The author provided a case study of Nigeria, where 40% of the population lacks access to electricity and relies on traditional ways. The author discovered significant energy distribution losses due to faulty equipment, as the data revealed a 4.49% decrease in the 2008 electricity distribution compared to 2007. Dokas et al. [69], based on a sample of 109 countries, determined that energy losses result in a significant decrease in energy consumption in developed economies (−3.5%). The losses may be replenished from renewable sources. Shahbaz et al. [70] came to similar conclusions. The literature has explored the correlation of energy losses with energy use indirectly impacting green innovation rather than directly. In this work, we analyse whether energy losses increase or detract from the influence of expansionary monetary policy on the development of green energy, i.e., whether they act as a moderator.

As observed from the literature review above, while the research is replete with studies on energy policy and its environmental implications, few correlate these variables with green innovation. Furthermore, these efforts appeared fragmentary, with models that employed limited variables. The present research aims to fill this gap in the literature.

2.3. Other Main Macroeconomic and Institutional Factors

In addition to the above-mentioned factors, other macroeconomic variables influence green innovation. The effect of fiscal variables, in particular, has been researched, and these variables cannot be excluded from a comprehensive framework.

In his study of 18 countries from 1994 to 2003, Sadorsky [71] discovered that real per capita income increases positively, and statistically significantly impacts per capita renewable energy usage. Long term, a 1% rise in real income per capita boosts renewable energy consumption per capita in emerging economies by around 3.5%. Chang et al. [56] found in their study of OECD countries for the period of 1997–2006 that countries with high economic growth can respond to high energy prices by increasing their use of renewable energy. In contrast, countries with low economic growth tend to be insensitive to energy price changes once they reach their level of renewable energy. Wang et al. [36], in their study of 57 developing nations during the period of 2002–2016, found, among other issues, that an increase in economic growth leads to a substantial increase in green finance and green innovation. Similar outcomes are observed by other researchers [50,72].

Green innovation may suffer from government expansion. A suitably sized government increases marginal productivity and green innovation by boosting government expenditure. Still, an excessively sized government generates a substantial administrative agency that raises the tax burden and undermines social welfare by excessive interference. So, government size may not linearly affect green innovation. Wen et al. [37] found an inverted U-shaped association between government size and green innovation in 166 countries between 1995 and 2018, suggesting that optimal government size may optimise green innovation output. They also found that countries with high organisational inertia and high R&D spending have a stronger inverted U-shaped association. Przychodzen and Przychodzen [45], in a study of 27 transition countries from 1990 to 2014, determined that economic growth, unemployment, government expenditures, and rising government debt served as stimulants for renewable energy production. Wang et al. [73] conducted research into the connection between government subsidies and green technology innovation in China’s provinces from 2003 to 2017 and revealed, among other findings, a positive correlation between government spending and green development. A similar conclusion was reached by Sun and Razzaq [74], analysing 32 OECD countries. The above variables were considered fiscal influence factors on green growth in the literature. In our research, these are employed as control factors to derive their effect in a primarily monetary model, from which we derive implications regarding the impact of monetary policy on green innovation.

Trade openness and economic freedom are additional aspects of the liberalisation of transactions that have been examined. Facilitating international transactions generally provides businesses with more opportunities, imports, and partnerships. Concurrently, increased competition from foreign enterprises drives domestic firms to seek resources (provided by monetary policy and central banks) for greater innovation [75]. Nonetheless, many research studies contend that the increased activity harms the environment. According to a scientific article by Ding et al. [76], increased trade openness benefits the ecological integrity of the G-7 countries. Ibrahim and Ajide [77] also indicated that trade openness reduced environmental problems for the G-20 nations and 48 Sub-Saharan African nations from 2005 to 2014. Yin et al. [41] and Wen et al. [3] drew similar conclusions. In contrast, Khan et al. [78] found experimental evidence that trade openness degrades environmental quality in Pakistan by increasing CO₂ emissions. Van Tran [79] showed that trade openness improved environmental quality in 66 emerging countries from 1971 to 2017, indicating similar tendencies. Khan and Ozturk [80] found that trade openness increased CO₂ emissions in 88 countries between 2000 and 2014.

Economic freedom, according to researchers, has a favourable impact on green growth [42]. Using a panel of 67 countries from 1995 to 2019, Liu and Feng [81] demonstrated that economic freedom and technological innovation favour green total factor productivity in both the global and regional panels at different income levels. In their study of 71 developed and developing countries between 1990 and 2014, Sun et al. [59] concluded, among other issues, that economic freedom enhances green innovation. Lastly, we examine the impact of institutional variables on the relationship between monetary policy and green growth. The impact of institutional quality on green development has been extensively discussed in the literature. According to Berrone et al. [82], environmental regulations benefit green technology innovation, and institutional pressure can motivate businesses to raise R&D expenditures. Across 71 nations, Sun et al. [59] discovered that institutional quality and green innovation positively influence energy efficiency. In addition, potential approaches include giving subsidies for research and development, mandating green innovation, improving communication between governments and businesses, etc., to encourage enterprises, individuals, and organisations to increase their creation of innovative technologies [83]. Finally, in a sample of 133 countries from 1960 to 2018, Yin et al. [41] demonstrated that institutional quality influences green innovation.

To sum up, efforts have been undertaken in the literature to establish the impact of monetary policy on green growth. To explore the phenomenon through the issues of endogeneity and cross-sectional dependence, however, few works have dealt with concurrently incorporating many monetary variables into a model. In the present study, we employ a holistic model that includes not just strictly monetary factors but also the impact of central bank characteristics on the relationship between expansionary monetary policy and green innovation. Simultaneously, we examine the effects using regional analysis to highlight the similarities and differences between the results in various regions. Lastly, we introduce a variety of additional variables to explore their potential impact on the investigated nexus.

3. Data and Methodology

3.1. Data and Variables

3.1.1. Data

We use panel data from 109 countries (see Appendix A) (developed and developing) for 9 years (2010–2018). The choice of the period and countries stems from our intention to study the period after the financial crisis. In addition, a sufficient and complete data set on the variables used for this period is necessary for using the highly restrictive econometric method FM–OLS.

Table 1. Descriptive statistics (981 observations, 109 countries).

Variables	Disambiguation	Source	Min.	Max.	Mean	St. Dev.
LENVPAT (per capita)	Ln(Total number of envir. Related techn. Patents p.c.)	OECD, WDI	−5.177741	6.438482	1.737711	2.103058
LPAT (p.c.)	Ln(Patent applications, all (per capita, per million))	WDI	−3.248097	8.340181	3.650523	2.230593
BRM M3% GDP	Sum of narrow money and other assets-Broad Money	WDI	10.10652	699.197	76.10276	73.81916
INTRT	Real interest rate (X 100)	WDI	−64.38082	52.43679	5.28662	8.03546
CBI	Central bank independence index (0 and 1)	Garriga [84]	0	1	0.53313	0.49916
CBT	Central bank transparency index (Eichengreen’s online database)	DEG group [85]	0.5	14.5	7.559123	3.537492
LGDP	GDP per capita fixed USD prices (2011) (PPP)	WDI	6.49193	11.45042	9.451556	1.125927
GEXP (%GDP)	General Gov. Expenses (% of GDP)	WDI	4.403315	73.57668	16.3179	6.037289
BUR	Bureaucracy index	Fraser Institute	3.272567	9.2726	6.583437	1.130936
TEMPWIN	Average temperatures over three winter months	WB, CCKP	−25.42333	28.56333	11.08565	12.18566
CO2EMMS	CO2 emissions	WDI	0.026146	22.61852	4.755074	4.571762

LENVPAT: logarithm of environmental patents, LPAT: logarithm of total patents, BRM: broad money (M3% of GDP), INTRT: real interest rate, CBI: central bank independence, CBT: central bank transparency, GEXP: government expenditure (% of GDP), LGDP: logarithm of GDP, BUR: bureaucracy index, TEMPWIN: temperature winter, CO2EMMS: CO₂ emissions, PPP: power purchasing parity. Sources: WB, OECD, Garriga [84], Eichengreen’s online database, Fraser Institute, CCKP 2022, and DEG group [85].

presents the descriptive statistics and definitions of the variables used in this research.

3.1.2. Dependent Variable

Our study’s environmental management innovation represents green advancements (proxied by Patent). According to Griliches [86], patent applications are a good indicator of innovation. Most countries track patents, making them an excellent indicator of innovation. Following Cai et al. [87] and Yin et al. [41], we use environmental patents (ENVPAT), the total number of environmental-related technology patent applications in each country (recorded in OECD statistics) per capita (1,000,000 residents and non-residents), as a proxy for green innovation. Environmental management innovation is an indicator. Its sub-indicators include air, water, waste, soil, and environmental monitoring, and represent a country’s total number of environmental inventions (primary patent families). The OECD statistics provide 1990–2019 environmental management innovation data. The dataset includes patent numbers and metrics to track environmental protection technology, help governments create environmental and innovation policies, and assess business and national innovation performance.

3.1.3. Explanatory Variables

This article assesses the impact of monetary policy and central bank characteristics on green innovation. Several researchers have utilised the real interest rate variable for monetary policy [20,21,46]; however, in recent years, M3 as a proportion of GDP (Broad Money) has also been used [3,23,41]. We use both variables as monetary policy determinants (INTRT and BRM). Central bank characteristics are determined by central bank independence (CBI) [84,88,89] and central bank transparency (CBT) [34,85]. All variables are calculated from the years 2010 to 2018.

3.1.4. Control Variables

We employ a variety of control variables following the pertinent literature on green innovation. A country’s thriving economy provides financial support for advancing environment-related technologies. Gross domestic product (GDP) is typically used to measure economic progress. Consequently, we follow Wang et al. [36] and utilise GDP per capita, evaluated in constant 2011 US dollars (LGDP), to reflect the economic development level of the nations studied (proxied by GDP–PPP). Scholars have also proven the significance of government size in innovation, green innovation, and economic growth [37,90]. Following Wen et al. [37], the size of the government is determined by the ratio of domestic consumer spending to GDP in this article (GEXP).

Many models that explore green innovation and growth include bureaucracy [3,91,92]. We anticipate that this development will improve green growth through greater transparency and more efficient resource allocation [38,39] (BUR). We also include the variable of extreme climatic events [3,93], represented by extremely cold winter temperatures [94] (TEMPWIN). According to D’Orazio and Popoyan [27], the demand for green innovation increases as environmental conditions deteriorate. The final variable used is carbon dioxide emissions [39,45], where we expect a positive association [66,95] (CO2EMMS). Table 1 shows the descriptive statistics of the variables used.

3.1.5. Data Quality

Most of the variables are drawn from the World Bank’s (WDI) database, as seen in Table 1. The variables not retrieved from this database are the bureaucracy (BUR) and central bank variables (CBI and CBT). The World Bank assessed the bureaucracy, but it took the form of corruption, which we chose not to include in the model because the institutional variable has a more significant impact. In addition, two databases (OECD and CCKP) were used to retrieve the variables ENVPAT and TEMPWIN, which use qualitative standards comparable to those of the World Bank. Because the databases mentioned above are known for their unique qualities and are frequently used in the literature, the data demonstrate completeness, consistency, and conformity. They exhibit uniqueness and are strictly balanced, both of which were verified using the STATA econometric tool. We must emphasise that only balanced data with no missing values can be successfully applied to the FM–OLS approach for panel data. Over 170 countries have variables from the WDI, 112 countries from the DEG group, 110 countries from Garriga’s CBI, and over 160 countries from the Fraser Institute’s bureaucracy. From these sources, the data were obtained for 109 countries, which is the maximum number for which there were complete data and a correct statistical distribution between regions according to the World Bank’s information (40 European and Central Asian countries (ECA), 17 Latin American and Caribbean nations (LAC), 14 East Asian and Pacific nations (EAP), 9 Middle Eastern and North African nations (MENA), and 24 Sub-Saharan and African countries (SSA)).

The percentage variables were taken as is (BRM, GEXP, TEMPWIN, INTRT, CO2EMMS, INFL, POP, TROP, ENRGLOS), and the numerical variables were taken as logarithms (LENVPAT (p.c.), LPAT (p.c.)), and as per capita for comparison reasons. The variables that represent the survey results (CBT and BUR) remained unchanged, the CBI variable was changed to a binary variable (0 or 1) because it presents temporal stability, and the variables for sample separation (INSQL, ENERINT, INDUST, URBAN, CGDEBT, and ECFRD) were obtained as high and low by dividing our samples in half so that the samples could be compared. Finally, the data are timely because they were only collected for the years following the 2008–2009 financial crisis, excluding the pandemic period, which would have changed how the results could be interpreted.

The equation of the empirical model contains broad money, real interest rate, central bank independence, and central bank transparency as essential independent variables, while GDP per capita, government expenses, bureaucracy, extreme weather conditions, and CO₂ emissions are the control variables. The equation of the dynamic OLS system is as follows:

$${\mathrm{y}}_{\mathrm{it}}={\mathsf{\beta }}_0+{\sum }_{j=1}^m{\beta }_j{x}_{it}^j+{\sum }_{r=1}^d{a}_r{z}_{it}^r+{\mathsf{\eta }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}}+{\mathrm{\upsilon }}_{\mathrm{it}}$$

(1)

where y_it is green innovation, $x_{it}^j$ is broad money (BRM), real interest rate (INTRT), central bank independence (CBI), and central bank transparency (CBT). The $z_{it}^r$ are the control variables, i.e., LGDP, GEXP, BUR, TEMPWIN, respectively, and CO2EMMS. μ_i refers to the constant individual effect of each country’s attributes, η_t signifies the time effects, and υ_it represents the time-dependent error. The study of the variables and correlations of our research necessitates more dynamic data because our results must consider both time and differences between countries. This leads us to the panel data analysis. The proposed study includes initial correlations with fixed terms and the application of the Hausmann, Woolridge, Modified Wald, Breusch–Pagan, Pesaran, and Frees tests, where the model appears to have heteroscedasticity, autocorrelation, and cross-sectoral dependence. To handle and mitigate the issue of cross-sectional dependence, the panel-corrected standard errors (PCSE) methodology is used [96]. The PCSE and PCSE (AR1) models are used for correction [97]. Finally, to address the potential effect of the variables’ past values (mainly through time lags) and the endogeneity problem, we apply a dynamic OLS methodology, which considers past and future temporal effects on the variables.

Additionally, we use the FM–OLS method to deal with endogeneity issues and the problems of omitted variables. Initially, we examine whether the variables in the model are stationary. Since the variables under investigation are cointegrated, we can estimate their relationship in the long-run equilibrium state. When applied to cointegrated panels, the least-squares method (OLS) leads to discriminatory estimators, especially when the sample size relative to periods (t) is small. Therefore, we estimate the long-run equilibrium function with the FM–OLS procedure proposed by Pedroni [98] to calculate the valid t-statistics. This approach yields more reliable results when heterogeneity in the integrated variables is diagnosed. The primary equation used in the FM–OLS model is (1). The constant β₀ includes the heterogeneity of each country, which may differ. The $x_{it}^j$ and $z_{it}^r$ variables are first-order integrated, and the error term tends asymptotically to the normal distribution.

The pooled panel FM–OLS estimators allow cross-section dependence, which also relies on the assumption of cross-sectional independence. This was proposed by Kao and Chiang [99] and Mark and Sul [100] and has provided more reliable results when heterogeneity is diagnosed in the cointegrated variables [101].

4. Results and Discussion

4.1. Preliminary Tests of Econometric Analysis (FM–OLS Methodology)

4.1.1. Unit Root Tests

We use unit root tests introduced by Levin, Lin and Chu, Breitung, Maddala and Wu, and Choi (Augmented Dickey–Fuller test). The results in

Table 2. Unit root test results.

Variables	LLC		BREITUNG		ADF–FISHER
	I (0)	I (1)	I (0)	I (1)	I (0)	I (1)
LENVPAT	−17.8638 *** (0.000)	−25.0319 *** (0.000)	−2.4942 *** (0.006)	−10.8627 *** (0.000)	−5.5783 *** (0.000)	−29.1441 *** (0.000)
LPAT	−4.2199 *** (0.000)	−23.2980 *** (0.000)	1.7826 (0.963)	−9.2141 *** (0.000)	−6.1754 *** (0.000)	−26.8931 *** (0.000)
BRM	−6.8127 *** (0.000)	−18.5995 *** (0.000)	5.2999 (1.000)	−8.0925 *** (0.000)	4.1314 (1.000)	−20.1151 *** (0.000)
INTRT	−0.1758 (0.430)	−19.2161 *** (0.000)	−2.586 *** (0.005)	−10.1021 *** (0.000)	−14.0642 *** (0.000)	−38.2400 *** (0.000)
CBT	7.6032 (1.000)	−5.795 *** (0.000)	3.6651 (0.999)	−7.5163 *** (0.000)	1.8560 (0.968)	−31.3916 *** (0.000)
LGDP	−6.8053 *** (0.000)	−23.6136 *** (0.000)	14.9109 (1.000)	−3.0600 *** (0.001)	−1.0350 (0.151)	−17.0195 *** (0.000)
GEXP	−13.4925 *** (0.000)	−23.9109 *** (0.000)	2.0076 (0.978)	−3.0600 *** (0.001)	−1.4094 * (0.080)	−22.5909 *** (0.000)
BUR	−21.4848 *** (0.000)	−22.0046 *** (0.000)	1.7007 (0.956)	−10.1051 *** (0.000)	−1.8740 ** (0.031)	−23.3768 *** (0.000)
TEMPWIN	−10.2416 *** (0.000)	−24.8454 *** (0.000)	−6.8882 *** (0.000)	−8.9330 *** (0.000)	−23.3454 *** (0.000)	−38.4535 *** (0.000)
CO2EMMS	−13.0539 *** (0.000)	−19.0209 *** (0.000)	3.2047 *** (0.999)	−8.8778 *** (0.000)	−2.8411 *** (0.002)	−14.2553 *** (0.000)

Notes: p-values in parentheses. *, **, and *** indicate 90%, 95%, and 99% statistical significance, respectively. All abbreviations and sources are the same as in Table 1 with the author’s calculations.

show that green innovation, total innovation, broad money, real interest rate, central bank independence, central bank transparency, GDP per capita, government expenses, bureaucracy, extreme weather conditions, and CO₂ emissions are stationary in the first differences. The null hypothesis of non-stationarity is rejected at a 99% significance level for most tests.

4.1.2. Cross-Sectional Dependence (CSD) and Panel Cointegration Results

The unit root tests assume that the cross-sectional units of the data panels are not strongly correlated. Next, we run the CSD tests proposed by Pesaran and Frees for the models containing 5, 7, and 10 variables.

Cross-sectional dependence is confirmed in all models, thus implying that one change in one country will likely affect others. This process occurs as similar processes between neighbouring countries across several areas such as tourism, free trade rules, legislation, governance, bureaucracy, etc. [102]. A typical example is the European Union, NAFTA, Latin American countries, and South Africa. To handle the issue of cross-sectional dependence, the PCSE methodology is used in addition to FM–OLS econometrics [103,104].

To apply FM–OLS econometrics, we adopt panel cointegration tests proposed by Kao, Pedroni, and Westerlund to determine the possible existence of a cointegrated vector in our essential variables. The results show cointegration between the model’s main variables, LENVPAT or LPAT as the dependent variable, INTRT, BRM, CBI, CBT, LGDP, and GEXP as regressors, and a long-term equilibrium relationship.

4.2. Empirical Results and Discussion

4.2.1. Key Conclusions of the Empirical Model

Table 3. Regression (PCSE variants and FM–OLS) results.

Dependent Variable	Green Innovation (LENVPAT) 981 Observations 109 Countries
Model	Model 1 FE	Model 2 RE	Model 3 PCSE	Model 4 PCSE (ar1)	Model 5 FM–OLS (5)	Model 6 FM–OLS (7)	Model 7 FM–OLS (10)
BRM	0.0007 (0.754)	0.0015 (0.306)	0.0016 *** (0.000)	0.0013 (0.130)	0.0095 *** (0.000)	0.0020 *** (0.000)	0.0023 *** (0.002)
INTRT	−0.0779 *** (0.010)	−0.0644 *** (0.010)	−0.0294 *** (0.000)	−0.0122 (0.584)	−0.0459 *** (0.000)	−0.0560 *** (0.000)	−0.1012 *** (0.000)
CBI	0.7354 (0.164)	−0.1396 (0.542)	0.0421 (0.878)	0.0313 * (0.052)	0.1101 (0.480)	0.0124 (0.865)	0.1303 (0.214)
CBT	0.0081 ** (0.025)	0.0078 ** (0.030)	0.0142 *** (0.002)	0.0047 (0.148)	0.0152 ** (0.011)	0.0328 *** (0.000)	0.0257 *** (0.000)
LGDP	0.6467 *** (0.008)	1.1353 *** (0.000)	1.2166 *** (0.000)	1.2006 *** (0.000)		1.6852 *** (0.000)	1.3932 *** (0.000)
GEXP	0.01413 *** (0.010)	0.0123 ** (0.022)	0.0145 *** (0.000)	0.0078 (0.111)		0.0473 *** (0.000)	0.0249 *** (0.002)
BUR	0.2023 *** (0.000)	0.1783 *** (0.001)	0.0913 ** (0.011)	0.0662 (0.339)			0.1502 ** (0.011)
TEMPWIN	−0.0243 (0.150)	−0.0188 * (0.054)	−0.0048 ** (0.021)	−0.0054 * (0.099)			−0.0189 *** (0.000)
CO2EMMS	0.0309 (0.401)	0.0753 *** (0.006)	0.0612 *** (0.000)	0.0780 *** (0.000)			0.01335 (0.439)
$R^2$	0.4341	0.5984	0.6288	0.4104	0.2088	0.2377	0.3109

*, **, and *** mean the significance level of 10%, 5%, and 1%, respectively; the numbers in parentheses show the p values. Variables and sources are as same as in Table 1 with the author’s calculations. In FM–OLS, lags 2.00 are used, and the linear and quadratic eqtrend effect is controlled (2).

demonstrates a positive link between broad money and green innovation in 109 countries and the econometric methodologies used at the 1% significance level. This relationship is not affected by the valuation method, even during the inclusion of the self-regulating factor in the model, and increases the explanatory power. This implies that, with a 1% increase in the broad money (% of GDP), green patents increase by 0.23%. The effect is considered significant as the above-mentioned implication would correspond with an essential contribution to environmental policy and is consistent with the literature [3,41,51]. In addition, there is a negative relation between the interest rate and green innovation. As interest rates rise, environmentally creative actions decline [20,21]. Using these two explanatory variables, we find that an expansionary monetary policy is substantially related to an increase in environmentally innovative actions.

Central bank independence is not highly connected with green innovation, although central bank transparency is considerably associated. A 1% increase in CBT correlates with a 0.39% increase in green innovation. We show that transparent methods attract more resources related to environmental benefits, which is a significant consequence. In conclusion, the characteristics of central banks influence environmentally innovative actions [34,35,105].

An increase in economic activity is related to a significant boost in environmentally friendly innovation. More specifically, a 1% rise in GDP per capita results in an approximately 1.4% improvement in green patents filed per capita. This conclusion was anticipated based on previous research [50,56,71,72]. A similar pattern can be seen with government expenditures [45,74].

In all models, an improved bureaucracy is correlated with a substantial improvement in green innovation and environmental policy [38,39]. Following the relevant general literature on natural disasters [3,27,93], the deterioration of winter temperatures is also associated with a rise in the number of environmental innovations. In conclusion, the overall model does not correlate strongly with carbon dioxide emissions or green innovation.

4.2.2. Robustness Checks

Econometric Methodologies and Endogeneity

Table 3 presents the models through various econometric methods, which generally give similar estimators to our main variables. We used fixed and random effects and error correction methods to deal with the autocorrelation and heteroskedasticity problems diagnosed in the model. Finally, we used the FM–OLS methodology after identifying integration and cross-sectoral dependence between the variables due to endogeneity and omitted variable issues. We performed static tests and found that the directions of the relations of the key variables of our model variables do not change, even though we used different static and dynamic econometric methodologies.

Models of Different Variables

In the FM–OLS Models 5, 6, and 7, we applied the gradual completion of our final model. First, we used one dependent and four independent variables (Model 5). Second, we added the two variables related to fiscal policy (Model 6), and finally, we introduced the last three control variables (Model 7). We found stability in the standard variables and a proportional increase in the completeness and the interpretive power of the models from 20.87% in Model 5 to 23.77% in Model 6 and 31.09% in the model that includes all the variables of interest.

Alternative Dependent Variable

Table 4. Distinguishing between low and moderately low-income and high and moderately high-income countries and alternative innovation proxy analysis with dependent variable LPAT.

	981 Observations in 109 Countries (Total)
Dependent	(LENVPAT) for 35 Low-Mid Low-Income Countries (Sample 1)	(LENVPAT) for 74 High-Mid High-Income Countries (Sample 2)	Robustness Analysis with Dependent Var LPAT
Models	Model 1	Model 2	Model 3
BRM	0.018493 (0.140)	0.01420 * (0.085)	0.002496 *** (0.000)
INTRT	−0.190777 * (0.054)	−0.055624 (0.119)	−0.021795 (0.061) *
CBI	0.837930 * (0.099)	−0.656027 *** (0.001)	0.177315 ** (0.037)
CBT	0.023128 (0.295)	0.066645 *** (0.000)	0.028971 *** (0.000)
LGDP	0.879847 * (0.091)	0.173666 (0.465)	1.649686 *** (0.000)
GEXP	0.034636 (0.909)	0.033528 (0.130)	0.003866 (0.559)
BUR	0.063233 (0.857)	0.350250 *** (0.002)	0.001975 (0.967)
TEMPWIN	−0.011203 (0.831)	−0.020769 ** (0.014)	−0.030779 *** (0.000)
CO2EMMS	0.496369 (0.269)	0.082267 *** (0.002)	0.014977 (0.285)
$R^2$	0.3723	0.2602	0.3654

Notes: *, **, and *** indicate a significance level of 10%, 5%, and 1%, respectively. The numbers in parentheses show the p-values. Variables and sources are as same as in Table 1 with the author’s calculations. FM–OLS controls the linear and quadratic effect of eqtrend (2).

and Model 3 show that each country’s total number of patents (LPAT) act as the dependent variable. We used the World Bank database for 109 countries. This variable includes environmental patents and is an indirect indicator of its assessment. We found that monetary policy and central bank features have a favourable impact on innovation, validating the findings of the basic model.

Low- and Medium-Low-Income Countries versus Medium-High and High-Income Countries

The sample of nations is divided into two categories based on the World Bank’s criteria, including those with low and relatively low incomes (Sample 1), and those with relatively high and high incomes (Sample 2). We aim to control the model’s power, variables, and consistency of the sample using this division and study the effect of monetary policy, central bank characteristics, and control variables on green innovation in both samples and models. Table 4 presents the estimators of these models.

The consistency and power of the model were observed. While the relations of the key variables remained in the same direction, significant differences were noted between the two models. Broad money, central bank transparency, and control variables, except for GDP per capita, do not influence green innovation in the first set of countries, as innovation incentives are limited. They are tasked with addressing more pressing concerns, such as economic development, the organisation of the state, governance challenges, and possibly a substantial informal sector. In several countries, the number of problems is so great that environmental policy represents only a small portion of the total output. Simultaneously, the authorities have little trust and weak power [36]. Thus, firms and citizens will likely maximise their utility without having significant environmental sensitivities. Keeping this in mind, we argue that the paradigm in these countries must change to adopt green technology. The countries need strong banking and funding from reliable government institutions [59]. On the contrary, interest rates, central bank independence, and economic growth are crucial to improving the green economy. These essential development elements inadvertently (and unintentionally) promote innovation and, by extension, green innovation.

In the second sample group, the impact of monetary expansion policies and central bank features on green innovation is robust and beneficial. Environmental patents increase significantly when the money supply rises directly or indirectly through interest rates. The only “paradoxical” finding is that the independence of central banks hinders green innovation. In the aftermath of the financial crisis, central banks may use their autonomy to finance more productive investments for a limited time. There are instances in the literature where the independence of central banks has little effect on green innovation [41]. In addition, all other variables, such as bureaucracy, climate, and carbon dioxide emissions, significantly impact green innovation in developed countries.

4.2.3. Heterogeneity Analysis

In this subsection, we are initially interested in dividing our sample into the geographical regions individually addressed in the literature. The objective is to examine the model’s heterogeneity and identify any significant differences to include them in policy proposals. Even though many comparative studies of geographic regions have been conducted up to this point, the vast majority focus on regions of China, and just a small number focus on regions throughout the world. This study aims to fill this gap. Based on the previous general-case analysis, which confirmed the positive causal relationship between monetary policy and central bank characteristics and green innovation, we are interested in the potential for expansionary monetary policy to improve green innovation performance in various geographic regions [59]. In recent decades, for instance, the progress of Asian developing economies has garnered growing attention due to Asia’s prominent role in the global economic recovery [41]. Their strategy for transforming the growth route from an energy-intensive to an environmentally friendly model may significantly impact the progression of global climate change.

Table 5. The heterogeneity analysis of monetary policy transmission in five regions according to the World Bank’s classification; FM–OLS estimation method.

Dep. Var.	LENVPAT, 936 Observations in 104 Countries (Total)
Model	Model 1 ECA	Model 2 LAC	Model 3 EAP	Model 4 MENA	Model 5 SSA
BRM	0.000555 (0.533)	0.014159 *** (0.000)	0.014692 *** (0.000)	−0.011983 *** (0.000)	0.004348 *** (0.000)
INTRT	−0.2965 ** (0.042)	−0.146979 *** (0.000)	−0.270896 *** (0.000)	−0.057659 *** (0.007)	−0.100880 (0.178)
CBI	−0.627043 *** (0.002)	0.718503 *** (0.000)	3.041000 *** (0.000)	−0.465513 *** (0.000)	−0.460603 (0.309)
CBT	0.046643 *** (0.000)	0.064394 *** (0.000)	0.215479 *** (0.000)	−0.013750 *** (0.001)	0.021373 (0.133)
$R^2$	0.1430	0.1071	0.1805	0.3987	0.1502
Obs.	360	153	126	81	216
CVs.	Yes	Yes	Yes	Yes	Yes

** and *** mean the significance level of 5% and 1%, respectively. The numbers in parentheses show the p values. All abbreviations and sources are the same as in Table 1 with the author’s calculations. In FM–OLS, lags 2.00 are used, and the linear and quadratic eqtrend effect is controlled (2). ECA: Europe and Central Asia, LAC: Latin America and Caribbean, EAP: East Asia and Pacific, MENA: Middle East and North America, SSA: Sub-Saharan Africa.

displays the correlation of the monetary and central bank characteristics variables with green innovation in five subsamples. The World Bank’s classification was proposed to make this division such that the criterion is the region to which the country belongs. This separation aims to reduce the estimation bias and investigate the essential factors affecting growth in each region. The first region comprises 40 European and Central Asian countries, and the second consists of 17 Latin American and Caribbean (LAC) countries. The third region includes 14 East Asian and Pacific (EAP) nations, the fourth region consists of 9 Middle Eastern and North African (MENA) countries, and the fifth region consists of 24 Sub-Saharan and African (SSA) countries. However, no analysis was conducted in North America and South Asia due to the few countries included in the model. In general, expansionary monetary policy increases green innovation and strengthens the model in the LAC and EAP countries. In the SSA countries, the only supporting factor is the increase in the supply of money, and that is relatively limited, indicating the inherent deficiencies of the individual countries. Furthermore, expansionary monetary policy and the characteristics of central banks do not foster green innovation in the MENA countries. In conclusion, the research confirms that the independence of central banks can have a negative impact on green innovation in European and some Asian countries.

As a second step, we selected a few variables that were used as controls in the literature, such as institutional quality [39], energy intensity [67], industry [3], central government debt [45,46], urbanity [43,44] and economic freedom [42]. Based on these variables, we divided our sample into low and high groups to determine whether the relationship between monetary policy, central bank characteristics, and green innovation is affected.

Table 6. The effect of monetary policy and central bank characteristics on green innovation in grouping base as follows: institutional quality, energy intensity, industry, urbanity, central government debt, and economic freedom level; FM–OLS estimation method.

Dep. Var.	LENVPAT 109 Countries (Total)
Grouping Base	Institutional Quality (INSQL)		Energy Intensity (ENERINT)		Industry (INDUST)
	Low	High	Low	High	Low	High
BRM	0.01485 *** (0.000)	0.0224 * (0.078)	0.00233 * (0.076)	0.01985 *** (0.000)	−0.00283 *** (0.000)	0.01961 *** (0.000)
INTRT	−0.059479 *** (0.003)	−0.08763 * (0.092)	−0.01774 (0.613)	−0.06447 *** (0.001)	−0.07013 *** (0.001)	−0.06312 *** (0.002)
CBI	0.80064 *** (0.000)	0.90084 *** (0.002)	0.05009 (0.783)	0.40188 *** (0.001)	−0.31368 *** (0.003)	0.80955 *** (0.000)
CBT	0.01398 *** (0.005)	0.06991 ** (0.030)	0.03023 ** (0.034)	0.01423 ** (0.011)	0.03717 *** (0.000)	0.00451 (0.532)
$R^2$	0.0809	0.1351	0.1472	0.4712	0.2877	0.2455
Obs.	512	467	490	488	468	511
Grouping Base	Urban (URBAN)		Central Gov. Debt (CGDEBT)		Economic Freedom (ECFRD)
	Low	High	Low	High	Low	High
BRM	0.02483 *** (0.000)	0.00186 ** (0.029)	0.00036 (0.638)	0.00727 *** (0.002)	0.015225 (0.112)	0.00054 (0.479)
INTRT	−0.00320 (0.918)	−0.01051 (0.680)	−0.01260 (0.638)	−0.04221 (0.137)	−0.05093 *** (0.008)	−0.08842 *** (0.010)
CBI	0.63183 *** (0.000)	0.25735 * (0.069)	0.39004 *** (0.003)	−0.17361 (0.339)	0.63599 *** (0.000)	0.95878 *** (0.000)
CBT	0.01386 ** (0.049)	0.00771 (0.458)	0.01308 ** (0.033)	0.01690 * (0.071)	0.01398 *** (0.007)	0.04210 ** (0.030)
$R^2$	0.1188	0.0877	0.1896	0.1589	0.3574	0.1836
Obs.	494	484	453	525	551	426
CVs.	Yes	Yes	Yes	Yes	Yes	Yes

*, **, and *** mean the significance level of 10%, 5%, and 1%, respectively. The numbers in parentheses show the p values. All abbreviations and sources are the same as in Table 1 with the author’s calculations. In FM–OLS, lags 1.00 are used, and the linear and quadratic eqtrend effect is controlled (2).

presents the results of these analyses. We note that the effect of the dependent variables is more significant in countries with high institutional quality and industry, most likely due to their superior transmission mechanisms. The impact of the dependent variables is considerable in countries with high energy intensity, probably because there is much space for improvement. While central banks support green innovation in nations with low debt levels, the effects are inconsistent elsewhere. Due to fewer restrictions, the dependent variables have a greater impact on green innovation in countries with a high level of economic freedom. Finally, in our samples, urbanisation did not produce distinct outcomes.

4.2.4. Moderating Effects

The analysis of regional heterogeneity and the effects of our model’s major factors confirmed the key elements that influence how monetary policy actions affect green innovation activities. Additionally, we are concerned with how fiscal, energy, and social policy factors relate to the monetary policy and central bank characteristics–green innovation nexus. Specifically, we examine the role different types of policies could have in amplifying or reducing the effect of expansionary monetary policy on environmentally friendly innovation.

To achieve this aim, we examine the correlation between monetary policy and central banks in green innovation in light of inflation (a result of monetary policy), population (a result of social policy), freedom of trade (a result of government policy), and losses in the distribution of energy (a result of energy policy).

Table 7. The moderating effects of the inflation, population, trade openness, and energy loss channels; FM–OLS estimation method.

Dep. Var.	LENVPAT, 981 Observations in 109 Countries (Total)
Model	Model 1	Model 2	Model 3	Model 4
BRM	0.003939 *** (0.004)	0.006792 *** (0.000)	0.016645 *** (0.000)	0.005982 *** (0.000)
INTRT	−0.138968 *** (0.000)	−0.064365 *** (0.000)	−0.090848 *** (0.000)	−0.117240 *** (0.000)
CBI	0.099907 (0.564)	0.126616 (0.125)	0.209241 *** (0.000)	0.085753 (0.493)
CBT	0.037457 *** (0.000)	0.006311 (0.203)	0.043332 *** (0.000)	0.039971 *** (0.000)
BRM×INFL	−0.002335 *** (0.000)
BRM×POP		−0.002924 *** (0.000)
BRM×TROP			−0.000040 *** (0.000)
BRM×ENRGLOS				0.000575 *** (0.000)
$R^2$	0.2708	0.3489	0.2285	0.2325
Control Variables	Yes	Yes	Yes	Yes

*** means the significance level of 1%. The numbers in parentheses show the p values. All abbreviations and sources are the same as in Table 1 with the author’s calculations. In FM–OLS, lags 2.00 are used, and the linear and quadratic eqtrend effect is controlled (2).

shows the effect of the above-mentioned moderating factors on the monetary policy–green innovation nexus.

The expansion of inventive activity necessitates substantial investment. Researchers claimed that inflation reflects a high level of macroeconomic uncertainty, noting a detrimental impact on investment and, subsequently, green investments [106]. It operated as a control variable in econometric models but had no statistically significant impact [19,45]. According to our analysis, inflation is detrimental to the correlation between expansionary monetary policy and green innovation, impeding the latter’s advancement.

Population growth positively impacts the quantity of innovative activity [44]. However, the outcomes are mixed when innovations are measured per capita [3,39,43]. Our research uses environmental patents per capita as a proxy for green innovation. The effect of an expansionary monetary policy on green innovation is observed to be constrained by the presence of the population variable. Possible explanations include the diminishing marginal ratios as the population increases, and the emerging state of environmental innovation.

The impact of trade openness on green innovation has been extensively studied in the academic literature. A largely positive correlation was found [3,41]. However, a negative correlation cannot be ruled out, mainly when carbon dioxide gas emissions are utilised as an intermediate variable [40,107]. According to our research, increasing trade openness limits the impact of expansionary monetary policy on green innovation. One argument for this outcome is that nations actively seek foreign investment in the years following a crisis. With the help of domestic financing and their already-registered patents, foreign companies are drawn to the country due to trade liberalisation. They partially implement the expansionary monetary policy in this way while limiting the growth of new green patents. Finally, the role of energy losses as a moderating factor for the impact of expansionary monetary policy on environmental innovation has received less attention. We observe that energy losses positively impact the effect previously mentioned. Innovative methods are needed to reduce energy losses; thus, following the current tendency, outdated methods are being replaced with more advanced, effective ones.

5. Conclusions and Policy Implications

The current study examines the effects of monetary policy factors and central bank characteristics on green innovation and transitioning to a low-carbon economy. Additionally, it examines international and regional differences in monetary policy and the green innovation nexus and analyses these differences by selected determinants of a green economy. First, the empirical data from 109 countries during the period of 2010–2018 reveals that an expansionary monetary policy action, either by expanding the money supply or by lowering the real interest rate, considerably boosts green innovation [20,21,41,51]. Moreover, we demonstrated that the characteristics of central banks, mainly their transparency, promote green growth [34,35]. In addition, institutional factors were included in our model alongside energy, fiscal, and other macroeconomic factors. We showed that a rise in per capita GDP, government spending, and improvement in bureaucracy all promote green economic activity. Increasing carbon dioxide emissions and energy losses are also essential factors in developing more sustainable technologies.

We utilised numerous static and dynamic econometric methods. However, our basic PCSE (ar1) and FM–OLS were employed to alleviate cross-sectoral dependence issues caused by endogeneity and omitted variable problems. According to the literature [97,99,101,102,103], these techniques are most effective when several countries are expected to exhibit strong similarities. Afterwards, we conducted two subgroup heterogeneity analyses, initially between developing and developed economies. First, we found that interest rates, the independence of the central bank, and economic growth significantly affect green innovation in developing nations. Particularly in developed countries, expansionary monetary policy, central bank transparency, and energy variables promote green growth. Second, we conducted a regional analysis and found that expansionary monetary policy promotes green innovation in countries in the LAC and EAP regions. However, among the countries of Sub-Saharan Africa, the only supporting element is the increase in money supply, revealing the fundamental limitations of each country. In addition, expansionary monetary policy and central bank features do not support green innovation in MENA nations.

Moreover, using differentiating factors in the sample of countries, we showed that the influence of expansionary monetary policy on green innovation is more pronounced in countries with high institutional quality, industrial concentration, and energy intensity. Finally, by employing interaction term analysis, we analysed the impact that inflation, population, trade openness, and energy losses have as moderating effects of an expansionary monetary policy on green innovation. We found the significance of the first three as deterrents and the fourth as a multiplier.

5.1. Policy Recommendations

Our research, particularly the analysis of the dynamic behaviour of the monetary policy–green innovation nexus, provides additional evidence for the non-neutrality of monetary policy actions. It provides information indicating that the change in the money supply and central bank characteristics should be viewed as something other than typical processes for resolving financial crises because it has a lasting effect on economic activities. Using the outcomes of our research’s empirical analysis, we provide several policy recommendations for economies targeting green growth. First, expansionary monetary policy acts, such as expanding the money supply or lowering the real interest rate, can reduce the risk premium and raise macroeconomic expectations. This could increase the flow of capital into businesses involved in green innovation, thereby boosting their performance in this area. This dynamic is particularly prominent in LAC and EAP countries, which stand to earn more from monetary expansion policies and better central bank features. Additionally, we have shown that inflation is a deterrent in the transmission process of monetary expansion to green innovation; therefore, this monetary development must be carried out cautiously.

Second, there should be more transparency regarding central banks, particularly in industrialised nations where it appears they must actively contribute to sustainable development. This will help promote green innovation activities that an expansionary monetary policy will have on them. Third, the fiscal environment needs to be improved, particularly in developing countries. Nonetheless, developed nations must also reduce bureaucracy issues because it helps direct the additional purchasing power that an expansionary monetary policy brings towards green innovation initiatives. Fourth, SSA nations must pay more attention to infrastructure, environmental laws, and fiscal expansion for monetary expansion to generate green growth. In the MENA region, neither monetary development nor greater independence and transparency of central banks contributed to green growth during the period under consideration.

Fifth, institutions must be optimised, and rigorous environmental regulation helps direct the additional real purchasing power brought about by an expansionary monetary policy into green innovation activities. In addition, the regulation’s legality requires robust institutions with corresponding enforcement capacity. Most countries with stricter environmental legislation can implement them, but introducing more stringent environmental regulations in developing nations does not necessarily result in their enforcement. In addition, lagging regions require additional development and industrialisation incentives to reach a higher rate of green growth due to expansionary monetary policy. Moreover, greater efforts must be made to simplify transactions. Finally, our research’s moderating analysis concludes that import and export incentives must be administered cautiously, as trade openness is a barrier to the diffusion of monetary expansions for green innovation. Inflation experiences a similar phenomenon, and expenditures are needed to update the energy products’ manufacturing, transportation, and distribution networks—a move that appears to promote green growth.

5.2. Limitations and Future Directions

The current study is undoubtedly subject to some limitations. For a more appropriate robustness test, we may use more specialised green innovation indicators, such as the percentage of environment-related technologies or environmentally related government R&D budget as a percentage of total government R&D—variables that are currently exclusively computed for OECD nations. Furthermore, what impact will energy product prices have on green innovation? Additionally, we can conduct empirical research on the factors that influence how monetary policy impacts green innovation, including public debt, bond yields, reserve funds, and others. The frameworks put forth in this paper pave the way for further investigation and the implementation of climate-related financial stability policies. Additionally, empirical research has the potential to expand the scope of policy considerations towards central bank governance models that prioritize coordination among various policy wings. It is also possible to research which financial stability governance model produces climate-related policies that are more successful. Finally, this study’s data collection period is 2010–2018, which may be extended backwards to draw long-term conclusions and make the econometric causality control techniques more reliable (with more time inertia).