Research on the impact of financial development in different regions on the decoupling of carbon emissions from economic growth

Abstract

The decoupling of carbon emissions has also become the dependence of countries’ development. Finance plays an important role in decoupling economic development from carbon emissions. This paper explores the impact of financial development in different regions on the decoupling of carbon emissions from economic growth, using the Tapio decoupling elastic model and the method of fully modified least squares (FMOLS) to study the impact of financial development on carbon emissions in six regional panels from 1995 to 2020: and Foreign direct investment (FDI), urbanization, population, and infrastructure as control variables. The results turn out that financial development will promote the decoupling of carbon emissions from economic growth in the ECA region. For EAP, SSA, AC, SA, and MENA regions, financial development will promote the growth of carbon emissions, and due to the different economic development dynamics of different countries, the positive effects of financial development on carbon emissions are heterogeneous. The impact of FDI, urbanization, and infrastructure on carbon emissions varies from region to region. The population will promote the growth of carbon emissions, regardless of the region. In addition, the ECA region is the most countries that has achieved the strong decoupling and is the first to realize the transition from weak decoupling to strong decoupling. Therefore,the ECA, EAP and AC region should accelerate the construction of a green financial system to promote the decoupling of economic growth and carbon emissions. The SSA, SA and MENA region should speed up the transformation of economic development mode and move towards weak decoupling or even strong decoupling.

Keywords

Financial development carbon emissions decoupling foreign direct investment urbanization infrastructure

Introduction

As a hot topic, global warming has always been the focus of attention of various countries. Due to the different levels of development in different countries, some countries are still consuming energy to promote local economic development, and this is the main consideration for global warming. According to the 2019 Global Carbon Emissions Report released by the International Energy Agency, the United States’ energy-related carbon emissions were 4.8 billion tons, Germany's emissions were 620 million tons, and Japan's emissions were 1.03 billion tons, which were both lower than in 2018. The energy carbon emissions of other economies increased by nearly 400 million tons, and 80% of them came from the Asian region. In 2020, due to the ragingCOVID-19 epidemic, various production and consumption activities will be stopped, making energy carbon emissions reach the lowest level of 31.5 billion tons. Although the world is calling for emission reduction actions, in the face of the economic recession that comes with emission reduction actions, some countries will choose to reduce such actions. Therefore, how to control the growth of carbon emissions while economic development has become the most concerned issue in recent years. The concept of decoupling is also used by academia to quantify the relationship between economic growth and carbon emissions.

Regarding the decoupling of carbon emissions from economic growth, academia has conducted corresponding studies, such as the effects of labour and investment,¹ trade opening,² residential buildings ³ urbanization and industrialization,⁴ tourism growth,⁵ research and development,⁶ renewable energy,⁷ technology-environmental innovation,⁸ industrial structure transformation9 and other angles have been fully studied. As a supporting industry for all walks of life, finance has played an important role in the development of the country's economy. All industries that develop the economy cannot do without the support of funds. The financial industry, which has the role of financing and fund-raising, also runs through the development of all industries and has become the general hub of social funds. Therefore, as the financial industry develops, it will inevitably generate carbon emissions. Different countries have different levels of economic development, and the development of the financial industry is also different. For example, the financial markets in the United States and the United Kingdom have been relatively complete, while Africa and parts of South Asia (SA) have not yet established a complete financial system. Therefore, the decoupling of carbon emissions from economic development is inseparable from financial development. The research on the decoupling of carbon emissions from economic growth caused by financial development has certain value.

The purpose of this paper is to study the impact of financial development in different regions on the decoupling of their economic development and carbon emissions, and to provide corresponding policy recommendations for how to reduce emissions in each region. Therefore, this study selected panel data of 173 countries from 1995 to 2020 to study this, and added Foreign direct investment (FDI), urbanization, population and infrastructure as control variables in order to more fully reflect the empirical results. This study makes contributions in the following aspects: First, this study divides 173 countries into six regional panels according to the World Bank classification standard, and uses the Tapio decoupling elasticity model to determine the decoupling status and decoupling trends of countries in different regions . The impact of financial development in different regions on the decoupling of carbon emissions is studied from the regional level. In the past, only a single region was studied, but this study integrates this and provides a reference for the decoupling of carbon emissions in various regions. Secondly, this study is based on the second-generation unit root test, which makes the research results more valuable. Compared with the first-generation unit root test, the second-generation unit root test considers the problems related to the cross-section of the same period, which can fully reflect the actual situation. Finally, this study takes infrastructure as a control variable, and quantifies and constructs infrastructure indicators according to communication, innovation and transportation, which enriches the theoretical research on carbon emission influencing factors. The rest of this article is as follows: The second part is a review of related literature, and the third part introduces the models and research methods used in this study. The fourth part shows the empirical results and discussion, and the fifth part summarizes the conclusions and puts forward some suggestions.

Literature review

Decoupling theory

The theory of decoupling carbon emissions is a problem proposed by the Organization for Economic and Cooperative Development (OECD) that describes the relationship between economic growth and carbon emissions. When the economy grows, carbon emissions decrease, or the growth rate is slower than the economic growth can be regarded as decoupling. In other words, decoupling is the ideal state of the relationship between economic growth and carbon emissions. The concept of decoupling was first proposed by Von.¹⁰ By 2002, the OECD used it as an indicator to measure the relationship between economic development and carbon emissions. There are many ways to achieve decoupling, and the Tapio decoupling model has been used more in past research. Hao et al.¹¹ studied the decoupling relationship between economic growth and carbon emissions in different provinces in China; Ning et al.¹² studied the relationship between energy-related economic growth and carbon emissions decoupling based on the Tapio decoupling method; Wang and Xu¹³ used the Tapio decoupling model to study the relationship between the decoupling of carbon emissions from economic growth in Zhejiang's transportation industry. In addition, scholars also use the Tapio model to study economic development and municipal solid waste,¹⁴ urban intensive land use and urban heat island effect¹⁵ and PM2.5 emissions decomposition analysis.¹⁶ In addition to studying China,¹⁷ also used the Tapio model to compare economic growth and carbon emissions in Pakistan, India, and China. Research on the decoupling of carbon emissions from economic growth in various countries has always been sufficient, but the analysis from a regional perspective is still blank. Therefore, this study mainly uses the Tapio model proposed by Tapio¹⁸ to describe the relationship between economic growth and carbon emissions.

Financial development and carbon emissions

Regarding the relationship between financial development and carbon emissions, scholars have different views on this. Some scholars believe that financial development promotes carbon emissions. Maji et al.¹⁹ investigated the relationship between Malaysian financial development and sectoral carbon emissions and found that financial development has increased carbon emissions in the transportation sector and the oil and gas sector, but it can control the emissions in manufacturing and construction industry. Overall, financial development has increased Malaysia's carbon emissions. Khan et al.²⁰ used panel quantitative regression method to examine the heterogeneity of global panel renewable energy consumption, carbon emissions and financial development . The results found that renewable energy consumption is negatively related to carbon emissions, but financial development has a positive effect on carbon emissions and renewable energy consumption. Shen et al.²¹ adopted the cross-sectional augmented autoregressive distribution lag (CS-ARDL) method to study the effects of natural resource rents, green investment, financial development on carbon emissions and environmental improvement in China's provincial panels . The results confirmed the positive effects of energy consumption and financial development on carbon emissions, indicating that the growth of China's financial development will cause certain harm to the environment.

Another part of scholars believe that economic development can curb carbon emissions. Lin and Sun²² studied the relationship between financial development, technological innovation and carbon emissions in 30 provinces in China from 2000 to 2017. The author decomposed financial development into financial scale and financial structure and found that financial development characterized by financial scale and financial structure is negatively correlated with carbon emissions, and financial scale controls carbon emissions, while financial structure has no significant impact on carbon emissions. Ehigiamusoe and Lean, 2019)²³ investigated the impact of energy consumption, economic growth and financial development on carbon emissions of different income groups, and found that financial development can control the carbon emissions of high-income countries, while the impact on low- and middle-income countries is the opposite. Sherazet al. 2021²⁴ investigated the impact of globalization in G20 countries on financial development, energy consumption, human capital and carbon emissions, using Fixed Effect Ordinary Least Squares (FE-OLS), Driscoll-Kraay standard error Approach (DK), and Dumitrescu and Hurlin's (2012) panel causality test method, and the results found that financial development and human capital reduced carbon emissions, and globalization also slowed down the carbon emissions of financial development and human development impact.

Even some scholars found that there is negligible effect on the relationship between financial development and carbon emissions. Neog and Yadava 2020²⁵ studied the relationship between India's carbon emissions, remittances, and financial development, and found that the impact coefficient of financial development on carbon emissions was positive, but not statistically significant.

Different scholars have different conclusions due to their different research objects and research methods. In addition, the difference in the indicators of quantitative financial development will also cause different research results. Acheampong²⁶ found that financial development measured by broad money, domestic credit to the private sector, and domestic credit to the private sector by banks increased carbon emissions, while the financial sector's FDI in the private sector, domestic financial development measured by credit does not affect carbon emissions. Gok²⁷ conducted a meta-regression analysis on the relationship between financial development and carbon emissions, which further confirmed this finding. In other words, the impact of financial development on carbon emissions depends on the country, time, research methods and selected indicators.

This study selected data from 173 countries from 1995 to 2020 and created six regional panels for this analysis. Financial development indicators are expressed in terms of domestic credit to the private sector, and choose FDI, urbanization, population and infrastructure as the control variables, which is the contribution of this article. Due to differences in geographic locations and national cultures, different regions have different development models. This study conducts research from the regional panel to provide certain guidance for countries in different regions to promote financial development.

Methodology

Data source

This study uses data from 173 countries from 1995 to 2020 to study the impact of financial development on the decoupling of economic growth and carbon emissions. In addition, according to the 2021 World Bank country classification, this study divides a sample of 173 countries into six regional panels: Europe and Central Asia (ECA), East Asia and the Pacific (EAP), Sub-Saharan Africa (SSA), and the Americas (AC), SA and the Middle East and North Africa (MENA). Among them, the ECA region is composed of 48 countries, the EAP region is composed of 25 countries, the SSA region is composed of 43 countries, the AC region is composed of 34 countries, the SA region is composed of 7 countries, and the MENA region is composed of 16 countries. See Table A1 in Appendix A for description. Table 1 shows the definitions and sources of variables in this study.

Table 1.

Definition and description of variables.

variable	Definition	Source
C	CO2 emissions (metric tons per capita)	World Bank
FD	Domestic credit to private sector (% of GDP)	World Bank
FDI	Foreign direct investment, net inflows (% of GDP)	World Bank
UZ	Urban population growth (annual %)	World Bank
POP	Population, total (Ten thousand)	World Bank
IN	1/3Individuals using the Internet (% of population) + 1/3Research and development expenditure (% of GDP) + 1/3Railways, passengers carried (million passenger-km)	World Bank

Decoupled elastic model

This study uses the Tapio decoupling model, the equation is as follows:

e = \frac{Δ C / C_{0}}{Δ G / G_{0}}

(1)

e represents the decoupling elasticity of economic development and carbon emissions,

Δ C

represents the change in carbon emissions from the end of the base period,

C_{0}

represents carbon emissions in the base period,

Δ G

and

G_{0}

are as mentioned above.

The Tapio decoupling model divides the decoupling state into 8 states according to the decoupling elasticity, as shown in Table 2:

Table 2.

Decoupling status list.

State	Definition	Description
strong decoupling	$Δ C / C_{0}$ <0, $Δ G / G_{0}$ >0,e<0	Economic growth and carbon emissions decline
weak decoupling	$Δ C / C_{0}$ >0, $Δ G / G_{0}$ >0,0 < e<0.8	The growth rate of carbon emissions is slower than the rate of economic growth
expansive coupling	$Δ C / C_{0}$ >0, $Δ G / G_{0}$ >0,0.8 < e<1.2	The growth rate of carbon emissions is basically the same as the rate of economic growth
expansive negative decoupling	$Δ C / C_{0}$ >0, $Δ G / G_{0}$ >0,e>1.2	The growth rate of carbon emissions is greater than the rate of economic growth
strong negative decoupling	$Δ C / C_{0}$ >0, $Δ G / G_{0}$ <0,e < 0	Economic recession and increased carbon emissions
weak negative decoupling	$Δ C / C_{0}$ <0, $Δ G / G_{0}$ <0,0 < e<0.8	Carbon emissions are falling at a slower rate than the economic recession
recessive coupling	$Δ C / C_{0}$ <0, $Δ G / G_{0}$ <0,0.8 < e<1.2	The rate of decline in carbon emissions is basically the same as the rate of economic recession
recessive decoupling	$Δ C / C_{0}$ <0, $Δ G / G_{0}$ <0,e>1.2	Carbon emissions are falling faster than the economic recession

Unit root test

Panel data needs to perform unit root test before regression estimation to check whether the data is stationary. If it is non-stationary, the regression estimation result may be unreliable. Since the first-generation unit root test did not consider the cross-section-related problems, its results have certain defects. Therefore, in order to overcome the defects of the first generation unit root test, this study adopts the cross-sectional enhanced dickey-fuller (CADF) and the cross-sectional enhanced Im, Pesaran and Shin (CIPS) unit root test.²⁸ The null hypothesis of four unit root tests is that the panel data has unit roots and is not stable, and the alternative hypothesis is that the data is stationary and does not have unit roots.

Its statistics are:

Δ Z_{it} = σ_{i} + δ_{i} Z_{i, t - 1} + δ_{i} {\bar{Z}}_{t - 1} + \sum_{l = 0}^{p} δ_{il} Δ {\bar{Z}}_{t - 1} + \sum_{l = 0}^{p} δ_{il} Δ Z_{i, t - 1} + ε_{it}

(2)

Where

{\bar{Z}}_{t - 1}

is averages for lagged,

Δ \bar{Z_{t - 1}}

is averages for first difference. CIPS statistics is:

\hat{C I P S} = \frac{1}{N} \sum_{i = 1}^{n} C A D F_{i}

(3)

CADF refers to cross-sections augment dickey-fuller tests.

Cointegration test

Since the previous unit root test is non-stationary at the I(0) level, it indicates that the panel data has unit roots. After the first difference of the data, the data is stationary, so before regression estimation, a cointegration test is required to check whether there is a long-term cointegration relationship between variables. This study used the cointegration test proposed by Westerlund.²⁹ Westerlund cointegration test can handle cross-section dependent issues.²⁹ The null hypothesis of these two tests is that there is no long-term cointegration relationship between variables, while the alternative hypothesis is the opposite.

Cointegration estimation

Based on the preliminary work of this study and the relationship between financial development and carbon emissions, the following estimation model was established. The model is as follows:

\ln C_{it} = α_{it} lnF D_{it} + β_{it} lnFD I_{it} + γ_{it} lnU Z_{it} + δ_{it} lnPO P_{it} + θ_{it} lnI N_{it} + ε_{it}

(4)

Where i represents a country in each panel, t represents the time period.

The regression estimation method used in this study is the Fully Modified Ordinary Least Squares (FMOLS) method of cointegration estimation. The FMOLS method is widely used in regression. Riti et al.³⁰ using this method to study the relationship between carbon emissions, economic growth and energy consumption. Rehman et al.³¹ used this method to study the impact of coal energy consumption on Pakistan's economic growth. Yorucu and Varoglu³² used FMOLS to study the relationship between energy consumption, industrial production, carbon dioxide emissions and economic growth. Compared with the ordinary least squares method, FMOLS corrects the problems caused by the ordinary least squares method to estimate the parameters, and the regression results are robust

The FMOLS estimator is as follows:

\hat{FMOLS} = \frac{1}{N} \sum_{i = 1}^{N} {{(\sum_{t = 1}^{T} X_{it} X_{it}^{^{'}})}^{- 1} \sum_{t = 1}^{T} (β_{it} X_{it} - θ_{12 i}^{^{'}})}

(5)

Empirical analysis

Decoupling state analysis

During the investigation, countries in the ECA region identified a total of eight decoupling states, as shown in Table 3. Countries in this region have strong decoupling and weak decoupling accounting for 82.7%, ranking second among the six regions. The decoupling trend is shown in Figure 1. Most countries have achieved strong decoupling between 2010 and 2020. The decoupling state of some countries such as Armenia, Azerbaijan, Czech Republic, Estonia, Georgia, Ireland, Latvia, Lithuania, Poland, Slovak Republic, United Kingdom are relatively stable. During the entire observation period (1995–2020), the decoupling state changed between strong decoupling and weak decoupling, and finally stabilized in a strong decoupling state, indicating that these countries have basically achieved the goal of decoupling carbon emissions from economic growth, which is the ultimate goal pursued by the global environment. Other countries in the region will experience multiple decoupling states during the observation period, but they basically stabilized in a weak decoupling state during 2002–2003. This may be since after 1995 the countries in the European region are mainly member countries’ own trade, and the main destination of production is also their own member countries. In ECA, countries can offset each other's deindustrialization.³³ Since the European countries began the industrial revolution in the 1860s, the economy grew rapidly. Similarly, Europe was the first to achieve the “carbon peak” around the 1980s, and gradually developed such a carbon emissions trading, carbon taxes and other policies control carbon emissions in ECA. Therefore, most European countries have also reached a state of weak decoupling earlier, and there is a trend of development towards a state of strong decoupling.

Figure 1.

Time trend of decoupling state in ECA area.

Table 3.

ECA region decoupling state frequency table.

Code	State	Frequency	Valid Percent
1	strong decoupling	563	46.9
2	weak decoupling	430	35.8
3	expansive coupling	11	.9
4	expansive negative decoupling	46	3.8
5	strong negative decoupling	69	5.8
6	weak negative decoupling	42	3.5
7	recessive coupling	10	.8
8	recessive decoupling	29	2.4

EAP countries have identified eight decoupling states during the survey period, as shown in Table 4. In this region, the weak decoupling state is the most, accounting for 45.4%. The decoupling trend is shown in Figure 2. Except for Tuvalu, the decoupling state of all countries in the region converged to a weak decoupling state in 2020, and China only experienced two decoupling states during the entire observation period and stabilized in the weak decoupling state after 2000. Tuvalu is always in a strongly decoupled state during the entire observation period. The remaining countries experienced a variety of decoupling states during the observation period. The decoupling state of EAP countries fluctuates significantly from 1995 to 2020, and the main trend is the shift from expansive negative decoupling to weak decoupling. This is consistent with the course of economic development in East Asia. The industrial revolution in East Asia started relatively late. Therefore, in the early stage, most countries developed their economies at the expense of the environment, resulting in a faster growth rate of carbon emissions than the economic growth rate, which is the expansive decoupling state. After the emergence of the topic of global warming, the world began to pay attention to the issue of carbon emissions and signed the Paris Agreement in 2016. Therefore, countries have begun to transform their economic development methods to control carbon emissions. The state of weak decoupling has been achieved today.

Figure 2.

Time trend of decoupling state in EAP area.

Table 4.

Frequency table of decoupling state in EAP area.

Code	state	Frequency	Valid Percent
1	strong decoupling	84	13.4
2	weak decoupling	284	45.4
3	expansive coupling	46	7.4
4	expansive negative decoupling	84	13.4
5	strong negative decoupling	87	13.9
6	weak negative decoupling	29	4.6
7	recessive coupling	6	1.0
8	recessive decoupling	5	.8

Countries in the SSA region have identified eight decoupling states, as shown in Table 5. In this region, the strong decoupling state and the weak decoupling state account for almost more than half. The decoupling trend is shown in Figure 3. Most countries will converge to a weak decoupling state in 2018–2020, but Benin, Cabo Verde, Gambia, The, Mali, and Sudan will eventually converge to the state of expansion coupling or expansion negative decoupling. Botswana, Mauritius, Namibia, South Africa and other high- or upper-middle-income countries have been in a stable weak decoupling state around 2015, which is consistent with the state of national economic development. These countries have better economic levels and living conditions than other countries in Africa, and they are close to the Atlantic Ocean, have superior geographical environment and rich tourism resources. They can achieve an early balance between environmental and economic growth and achieve a relatively good weak decoupling state. In addition, most of the countries in the region are low-income countries with backward economies but rich in mineral resources. Therefore, these countries mainly use resource extraction and energy consumption as a means of economic growth, which leads to these countries being in the expansive negative decoupling state or achieving of weak decoupling state later.³⁴

Figure 3.

Time trend diagram of decoupling state in SSA area.

Table 5.

SSA area decoupling state frequency table.

Code	state	Frequency	Valid Percent
1	strong decoupling	240	22.3
2	weak decoupling	400	37.2
3	expansive coupling	70	6.5
4	expansive negative decoupling	182	16.9
5	strong negative decoupling	118	11.0
6	weak negative decoupling	28	2.6
7	recessive coupling	7	.7
8	recessive decoupling	30	2.8

Countries in the AC region have identified eight decoupling states, as shown in Table 6. In the region, weak decoupling account for 60.5%. The decoupling trend is shown in Figure 4. The decoupling state of countries in the AC region is basically stable. Argentina, Belize, Brazil, Chile, Colombia, CostaRica, Cuba, Dominica, Dominican Republic, Ecuador, Honduras, Paraguay, Peru, St Lucia, Uruguay experienced no less than 4 decoupling states during the entire investigation, but basically all countries eventually converged to weak decoupling. Countries in the AC region are basically high-income and upper-middle-income countries, with relatively developed economies, and basically stabilized in a weakly decoupled state around 2010. Only Belize, Colombia, CostaRica, Paraguay, Peru have reached the recessive decoupling state. The rest of the countries have all shifted from expansive negative decoupling to weak or strong decoupling and focus on achieving this transition in 2010. However, Argentina, Brazil, Peru, and Uruguay as developing countries, have not signed the Paris Agreement on mitigation with emission reduction targets. These countries will not choose to abandon economic growth in order to reduce carbon emissions, which has led to obvious expansive negative decoupling state.³⁵ The 2019 Carbon Pricing Status and Trends Report released by the World Bank pointed out that the new carbon pricing mainly occurred in America, indicating that countries in America are also working hard to achieve emission reduction targets. Even in 2019–2020, almost all countries have achieved weak decoupling.

Figure 4.

Time trend of decoupling state in AC area.

Table 6.

Frequency table of decoupling state in AC area.

Code	state	Frequency	Valid Percent
1	strong decoupling	144	16.9
2	weak decoupling	514	60.5
3	expansive coupling	57	6.7
4	expansive negative decoupling	81	9.5
5	strong negative decoupling	30	3.5
6	weak negative decoupling	11	1.3
7	recessive coupling	3	.4
8	recessive decoupling	10	1.2

Countries in the SA region have five decoupling states, as shown in Table 7. In the region, the weak decoupling of countries in the SA region accounted for 69.1%, ranking first among the six regions. The decoupling trend is shown in Figure 5. The countries in the SA region have basically converged to a weak decoupling state during 2015–2020, and the fluctuations are relatively obvious. All countries have changed from expansive negative decoupling to weak decoupling. Countries in the SA region are basically low- and middle-income countries. Therefore, in order to develop the economy, the countries in the region will choose to sacrifice the environment as a price. This will cause the country to be in a state of expansive negative decoupling. SA's economy is dominated by agriculture, supplemented by industry, and industry is dominated by light industry. Therefore, the region will experience a state of expansive coupling, that is, the growth rate of carbon emissions is basically the same as the rate of economic growth.

Figure 5.

Time trend of decoupling state in SA area.

Table 7.

Frequency table of decoupling state in SA area.

Code	state	Frequency	Valid Percent
1	strong decoupling	1	.6
2	weak decoupling	121	69.1
3	expansive coupling	18	10.3
4	expansive negative decoupling	32	18.3
5	strong negative decoupling	3	1.7

The MENA region determines eight decoupling states, as shown in Table 8. In the MENA region, strong decoupling and weak decoupling accounted for 87.5%, ranking the highest among the six regional countries. Strong decoupling ranks second in six regional countries. The decoupling trend is shown in Figure 6. Except for Libya, all countries converged to weak decoupling or strong decoupling. The countries in the MENA region are mainly low- and middle-income countries, and the Middle East is rich in oil resources. Therefore, most countries’ economic development depends on the exploitation and trading of oil resources, and resource exploitation is one of the main sources of carbon emissions. Therefore, most countries in the region showed an expansive negative decoupling state in the early stages of development (1995–2005). As countries in the Middle East gradually recognize the risks of diversified energy sources, they begin to attach importance to the potential of renewable energy and focus on green development. What follows is the emergence of weak decoupling between economic growth and carbon emissions.

Figure 6.

Time trend of decoupling state in MENA area.

Table 8.

MENA area decoupling state frequency table.

Code	state	Frequency	Valid Percent
1	strong decoupling	106	26.5
2	weak decoupling	244	61.0
3	expansive coupling	15	3.8
4	expansive negative decoupling	13	3.3
5	strong negative decoupling	14	3.5
6	weak negative decoupling	5	1.3
7	recessive coupling	1	.3
8	recessive decoupling	2	.5

Cointegration regression analysis

Unit root test

This study took CIPS unit root tests for all variables, and the results are shown in Table B1-B6 in Appendix B. The results show that the unit root tests of all variables in the ECA,SSA and AC area panel are significant at the I(0) level, so the panel data level in this area is stable, and there is no need for the following cointegration test The unit root test of the EAP, SA, and MENA region panel remains insignificant at the I(0) order, indicating the presence of a unit root After that, first-order difference processing was performed on the data.According to the results in the table, after the first difference, all variables basically rejected the null hypothesis at the significance level (1%, 5%, 10%) and reached a stationary state. This is the next cointegration test conditions are provided.

Panel cointegration test

This study uses Westerlund panel cointegration tests, as shown in Table 9. The null hypothesis of the Westerlund cointegration test is that there is no long-term cointegration relationship between the variables, and the alternative hypothesis is that there is cointegration between the variables According to the results of the unit root test, only the EAP, SA, and MENA region panel need to be tested for cointegration. The results show that the three regional panels all pass the cointegration test, and all reject the null hypothesis. That is, there is a long-term co-integration relationship between carbon emissions, financial development, FDI, urbanization, population, and transportation infrastructure.

Table 9.

Cointegration test results.

	Statistics (Pvalue)
EAP	−1.8443 **(0.033)
SA	−1.5376 **(0.042)
MENA	2.3931 ***(0.008)

Note: ***, **, * represent significant at 1%, 5%, and 10% levels, respectively.

Panel cointegration regression

In this study, FMOLS was used to perform panel cointegration regression on the data, and the results are shown in Table 10. Model 1 studies the impact of financial development on carbon emissions without interference from other factors. The results show that for ECA regional panels, financial development has a negative effect on carbon emissions in the long run, and rejects the null hypothesis at a significance level of 10%, that is, if the level of financial development increases by 1%, carbon emissions will be reduced by 0.065%, which is similar to the empirical results of Lin and Sun²² and Sheraz et al.²⁴ While the results of the EAP region panel, SSA region panel, AC region panel, SA region panel and MENA region panel are completely opposite. In the long run, financial development in these regions is significantly positively correlated with carbon emissions, which is similar to the empirical results of Maji et al.¹⁹, Khan et al.²⁰ and Shen et al.²¹ As shown in the table, for EAP regional panels, for every 1% increase in financial development, carbon emissions will increase by 0.412%. For SSA regional panels, 1% unit of financial development will generate 0.225% of carbon dioxide, and for AC regional panels, 1% of unit financial development will result in 0.228% of carbon emissions. 1% unit financial development will lead to0.513% and 0.462% increase in carbon emissions in SA and MENA regions, respectively.

Table 10.

FMOLS regression results.

Group	Model	LnFD	LnFDI	CZ	LnPOP	lnIN
ECA	Model 1	−0.065* (0.070)
ECA	Model 2	−0.242*** (0.000)	0.004** (0.046)	−0.076*** (0.000)	4.253*** (0.001)	0.008 (0.369)
EAP	Model 1	0.412*** (0.000)
EAP	Model 2	0.062*** (0.003)	−0.004 (0.310)	0.059*** (0.000)	0.218*** (0.000)	−0.006 (0.551)
SSA	Model 1	0.225*** (0.000)
SSA	Model 2	0.239*** (0.000)	0.003 (0.680)	0.154*** (0.005)	0.159*** (0.000)	0.039*** (0.003)
AC	Model 1	0.228*** (0.000)
AC	Model 2	0.367*** (0.000)	0.014*** (0.000)	−0.103*** (0.000)	0.263*** (0.000)	0.033*** (0.000)
SA	Model 1	0.354*** (0.001)
SA	Model 2	0.093*** (0.002)	0.027*** (0.000)	0.115*** (0.000)	2.724*** (0.000)	−0.041*** (0.001)
MENA	Model 1	0.462*** (0.000)
MENA	Model 2	0.412*** (0.000)	0.030*** (0.001)	0.030** (0.039)	0.101** (0.012)	0.037*** (0.003)

Note: ***, **, * represent significant at 1%, 5%, and 10% levels, respectively.

Financial development in different regions has different effects on carbon emissions, which are closely related to the driving force of national economic development. Financial funds tend to flow more to the pillar industries of the national economy. Therefore, for developing countries, the effect of financial development on carbon emissions is generally consistent with the effect of economic growth on carbon emissions. SSA and MENA are both relatively backward regions. The positive effect of financial development in the MENA region on carbon emissions is greater than the SSA region. This may be due to the different economic development dynamics of the two regions. Due to its geographical location, the Middle East is rich in oil resources. Therefore, the economic development of the region depends more on the exploitation of natural resources. From dams to oil to natural gas development, each process promotes regional economic growth, but also causes serious pollution to produce a lot of carbon dioxide. Compared with North Africa, SSA has a slightly higher economic level and a better living environment. Due to its superior geographical location, tourism resources can be used as one of the reasons for the economic growth of the SSA region. Therefore, the economic growth brought about by financial development generates less carbon emissions than North Africa.

On the contrary, the per capita income and economic level of the EAP region is higher than that of the SA region, but the positive effect of financial development on carbon emissions is greater than the SA region. This is due to the relatively developed agriculture in the SA region, which is mainly based on agriculture and supplemented by industry to develop the economy, while the EAP region pays more attention to the development of the manufacturing industry. In 2005, the manufacturing industry showed a trend of three pillars in North America, the European Union, EAP. By 2017, the proportion of manufacturing in North America and the European Union has gradually declined, and the manufacturing industries in EAP are clearly dominant. The manufacturing industry is also a high-carbon emission industry, so the effect of financial development in the EAP region on carbon emissions is greater than that in the SA region

For the relatively developed panels in the ECA region and the AC region, the results are completely opposite. The financial development in the ECA region has a negative effect on carbon emissions, while the AC region has a positive effect. This has a certain relationship with the European region taking the lead in completing industrialization. Most countries in the ECA region are high-income countries, while high-income countries and upper-middle-income countries in the AC region account for half of them. According to Ehigiamusoe and Lean23 and Thampanya et al.³⁶ financial development in high-income countries is negatively correlated with carbon emissions, and financial development in middle-income countries is positively correlated with carbon emissions. Financial development can attract foreign investment and can accelerate economic growth through import and export trade, and developed countries will choose to establish high-polluting enterprises in countries with lower income levels.³⁷

Model 2 studies the relationship between financial development and carbon emissions under the influence of FDI, urbanization, population and infrastructure, and provides strong evidence for Model 1. In all, in Model 3, the effect of financial development on carbon emissions is consistent with Model 1. For ECA regional panels, 1% of financial development will reduce long-term carbon emissions by 0.242%, while for EAP, SSA, and AC In SA and MENA regions, an increase of 1% of unit financial development will lead to long-term carbon emissions increase of 0.062%, 0.239%, 0.367%, 0.093%, 0.412%, respectively, which supports the conclusion of Model 1 to a certain extent. In addition, it is common for the population of all regions to have a positive effect on carbon emissions, which is consistent with the results of Mujtaba et al.³⁸ All aspects of people's consumption, production and life are inseparable from carbon emissions. It is obvious that the population causes carbon emissions to increase.

Regarding FDI, urbanization, and the impact of infrastructure on carbon emissions, the panels of different regions are heterogeneous. Regarding the relationship between FDI and carbon emissions, the EAP and SSA regions are not significant, while the ECA, AC, SA, and MENA regions are significantly positively correlated with carbon emissions, which is related to Shahbaz et al.³⁹, Essandoh et al.⁴⁰, Mujtaba and Jena⁴¹ and Hamid et al.^42,⁴³ who gave consistent results. FDI will bring in technology and capital inflows, thereby promoting domestic production and consumption and energy consumption, leading to domestic economic growth, and thus generating a large amount of carbon emissions.⁴⁴ Regarding the relationship between urbanization and carbon emissions, urbanization in ECA and AC regions is significantly negatively correlated with carbon emissions, while urbanization in EAP, SSA, SA, and MENA regions is positively correlated with carbon emissions. ECA and AC regions are relative to EAP, SSA, SA and MENA regions have a higher degree of urbanization. According to Dong et al.⁴⁵ when the level of urbanization is relatively low, urbanization will promote carbon emissions. When urbanization develops to the mid-term stage, technological innovation will bring about low-carbon and green development models, which will curb carbon emissions to a certain extent. In the advanced stage of urbanization, urbanization will in turn promote carbon emissions. In other words, most countries in the ECA and AC regions are still in the mid-stage of urbanization, that is, the increase in urbanization rate will control carbon emissions. However, most countries in the EAP, SSA, SA, and MENA regions are still in a stage of low urbanization rate. Regarding the impact of infrastructure on carbon emissions, the relationship between infrastructure and carbon emissions in ECA and EAP regions is not significant. SSA, AC, MENA regional infrastructure and carbon emissions are significantly positively correlated, while SA regional infrastructure and carbon emissions are negatively correlated. This may be due to the higher proportion of communication and transportation infrastructure construction in the infrastructure of SSA, AC, and MENA, which will promote carbon emissions, while the SA area is dominated by agriculture, and its transportation and communication infrastructure are relatively backward, so it reduces carbon emissions to a certain extent.

Conclusions and policy recommendations

This study explored the impact of financial development in 172 countries around the world on the decoupling of carbon emissions from economic growth from 1995 to 2020, and combined the Tapio decoupling model with panel estimation methods, and created six panel data corresponding to 172 countries by region, and systematically studied the impact of financial development on carbon emissions in different regions. The results show that the decoupling trend of panel economic growth and carbon emissions in different regions tends to be weakly decoupled, but the performance of different regions is inconsistent. The ECA and AC region decoupling trend is relatively stable, and some countries have achieved a stable and strong decoupling state. Most countries in the EAP region, SSA region and MENA region have achieved weak decoupling. However, there are also some countries that have turned to a strong decoupling state. The weak decoupling state accounts for the largest proportion of countries in the SA region, and the decoupling state is unstable.

The panel cointegration test shows that there is a long-term cointegration relationship between carbon emissions, financial development, FDI, urbanization, population, and infrastructure in different regions. In addition, using the Fully Modified Least Squares Method (FMOLS) to estimate the cointegration of panels in different regions. The results are as follows:

(1) The effect of financial development on carbon emissions in different regions is heterogeneous . In the ECA region, financial development has a negative impact on carbon emissions, while in the EAP, SSA, AC, SA, MENA regions the growth of financial development will still promote the increase of carbon emissions, and due to the different economic levels and economic development motives in different regions, the positive effect of financial development on carbon emissions are heterogeneous. The positive effect of financial development in EAP, SA and MENA regions on carbon emissions is much greater than in SA and AC regions.

(2) In regions with a low level of economic development, the increase of various indicators will lead to an increase in carbon emissions; while in regions with a high level of economic development, different indicators have different impacts on carbon emissions. Among them, FDI in the ECA, AC, SA, and MENA regions has a positive effect on carbon emissions, while in the EAP and SSA regions, the impact of FDI on carbon emissions is minimal. The impact of urbanization on carbon emissions also presents a heterogeneous effect in the panels of different regions. In ECA and AC regions where the level of urbanization is relatively high, the impact is negative, while in EAP, SSA, SA, MENA regions, the increase in urbanization level will increase carbon emissions. The impact of infrastructure on carbon emissions in the ECA and EAP regions is not significant, while it has a positive impact on carbon emissions in the SSA, AC and MENA regions, but will reduce carbon emissions in the SA region. The population has always been positively correlated with carbon emissions and has nothing to do with the region.

In order to achieve the decoupling of carbon emissions from economic growth, this study puts forward some suggestions. First of all, for the ECA region, most countries have achieved a stable state of strong decoupling. This is the most meaningful goal of controlling global warming and a commitment to the Paris Agreement. Some countries in the ECA region and relatively developed countries in the AC region should learn from their development paths and combine their own characteristics to achieve their own emission reduction roads, and actively sign the Paris Agreement to be responsible for their emission reduction commitments. In addition, countries in the region should speed up the construction of financial development, use finance to develop the economy, and build a green financial development system, so as to realize the decoupling of carbon emissions and economic growth. For the EAP region and some countries in the AC region, although their economic level lags the ECA region, it is higher than the SSA, SA, and MENA regions. They belong to upper-middle-income countries. There is no need to sacrifice the environment to develop the economy. It is necessary to change the development model. In addition, due to the relatively high level of urbanization in this region, it can slow down the development of urbanization to a certain extent and accelerate the construction of a green financial system, transfer the funds that used to be mainly high pollution and high energy consumption to new energy and technological innovation enterprises, strengthen the construction of a green credit system, and promote the sustainable development of the national financial industry. As for the SSA, SA and MENA regions, their financial development is immature, but the potential is huge. The difference between the MENA region and the SSA and SA regions is that energy is the main driving force for development. Energy is a recognized high-carbon emission industry. As countries around the world pursue the goals of the Paris Agreement, the MENA region needs to think about the path to new energy, such as developing an energy financial system. When the world's main energy demand shifts to new energy sources, if the MENA region does not change its development mode in time, it may lag far behind countries in other regions.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Guangdong social science, (grant number 18JRZ04).

ORCID iD

Jiemin Huang

Jiemin Huang is an associate professor in Shenzhen Institute of Technology, and she is also a post-doctor in University of Electronic and Technology of China, and she research on carbon emission problems and related policy.

Liyiing Guo received the MS in Public Administration from Harbin Institute of Technology, Harbin, China, in 2022. Her main research direction is related policies and influencing factors of carbon emission governance.

Appendices

References

Wang

Jiang

. Is carbon emission growth decoupled from economic growth in emerging countries? New insights from labor and investment effects. J Cleaner Prod 2020; 248: 119188.

Wang

Zhang

. The effects of trade openness on decoupling carbon emissions from economic growth–evidence from 182 countries. J Cleaner Prod 2021; 279: 123838.

Huo

, et al. Decoupling and decomposition analysis of residential building carbon emissions from residential income: evidence from the provincial level in China. Environ Impact Assess Rev 2021; 86: 106487.

Wang

. The effects of urbanization and industrialization on decoupling economic growth from carbon emission–a case study of China. Sustainable Cities Soc 2019; 51: 101758.

Zha

Dai

, et al. How to decouple tourism growth from carbon emissions? A case study of Chengdu, China. Tourism Manage Perspect 2021; 39: 100849.

Wang

Zhang

. Does increasing investment in research and development promote economic growth decoupling from carbon emission growth? An empirical analysis of BRICS countries. J Cleaner Prod 2020; 252: 119853.

Wang

Dong

. How renewable energy reduces CO2 emissions? Decoupling and decomposition analysis for 25 countries along the belt and road. Appl Econ 2021; 53: 4597–4613.

Chen

Zhao

Liao

. The impact of technology-environmental innovation on CO2 emissions in China’s transportation sector. Environ Sci Pollut Res 2020; 27: 29485–29501.

Qiu

Yuan

Bai

, et al. Spatial-temporal heterogeneity of industrial structure transformation and carbon emission effect in Xuzhou metropolitan area. Chin Geog Sci 2017; 27: 904–917.

10.

Weizsäcker EUv. Erdpolitik: ökologische Realpolitik an der Schwelle zum Jahrhundert der Umwelt. Frankfurt: Büchergilde Gutenberg, 1990.

11.

Hao

Huang

. Do carbon emissions and economic growth decouple in China? An empirical analysis based on provincial panel data. Energies 2019; 12: 2411.

12.

Ning

Zhang

Ding

, et al. Analysis of regional decoupling relationship between energy-related CO2 emission and economic growth in China. Nat Hazards 2017; 87: 867–883.

13.

Wang

. Decoupling elasticity of carbon emissions and economic growth of transportation industry in zhejiang province and its influencing factors. J Environ Prot Ecol 2021; 22: 750–759.

14.

Wang

Zhu

Zhang

. Decoupling economic development from municipal solid waste generation in China’s cities: assessment and prediction based on tapio method and EKC models. Waste Manage 2021; 133: 37–48.

15.

Shao

. Investigation into relationship between intensive land use and urban heat island effect in shijiazhuang city based on the tapio decoupling theory. J Urban Plann Dev 2020; 146: 04020043.

16.

Lai

Zhou

. Decomposition analysis of PM2. 5 emissions based on LMDI and tapio decoupling model: study of hunan and guangdong. Environ Sci Pollut Res 2021; 28: 43443–43458.

17.

Ozturk

Majeed

Khan

. Decoupling and decomposition analysis of environmental impact from economic growth: a comparative analysis of Pakistan, India, and China. Environ Ecol Stat 2021; 28: 793–820.

18.

Tapio

. Towards a theory of decoupling: degrees of decoupling in the EU and the case of road traffic in Finland between 1970 and 2001. Transp Policy 2005; 12: 137–151.

19.

Maji

Habibullah

Saari

. Financial development and sectoral CO2 emissions in Malaysia. Environ Sci Pollut Res 2017; 24: 7160–7176.

20.

Khan

Binh

. The heterogeneity of renewable energy consumption, carbon emission and financial development in the globe: a panel quantile regression approach. Energy Rep 2020; 6: 859–867.

21.

Shen

Z-W

Malik

, et al. Does green investment, financial development and natural resources rent limit carbon emissions? A provincial panel analysis of China. Sci Total Environ 2021; 755: 142538.

22.

Lin

Sun

. Financial development, technical innovation, and carbon emissions: new evidence from China. J Environ Prot Ecol 2020; 21: 1526–1540.

23.

Ehigiamusoe

Lean

. Effects of energy consumption, economic growth, and financial development on carbon emissions: evidence from heterogeneous income groups. Environ Sci Pollut Res 2019; 26: 22611–22624.

24.

Sheraz

Deyi

Ahmed

, et al. Moderating the effect of globalization on financial development, energy consumption, human capital, and carbon emissions: evidence from G20 countries. Environ Sci Pollut Res 2021; 28: 35126–35144.

25.

Neog

Yadava

. Nexus among CO2 emissions, remittances, and financial development: a NARDL approach for India. Environ Sci Pollut Res 2020; 27: 44470–44481.

26.

Acheampong

. Modelling for insight: does financial development improve environmental quality? Energy Econ 2019; 83: 156–179.

27.

Gök

. The role of financial development on carbon emissions: a meta regression analysis. Environ Sci Pollu Res 2020; 27: 11618–11636.

28.

Pesaran

. Testing weak cross-sectional dependence in large panels. Econom Rev 2015; 34: 1089–1117.

29.

Westerlund

. New simple tests for panel cointegration. Econom Rev 2005; 24: 297–316.

30.

Riti

Song

Shu

, et al.

Decoupling CO2 emission and economic growth in China: is there consistency in estimation results in analyzing environmental kuznets curve?

J Cleaner Prod 2017; 166: 1448–1461.

31.

Rehman

Radulescu

, et al. Energy crisis in Pakistan and economic progress: decoupling the impact of coal energy consumption in power and Brick Kilns. Mathematics 2021; 9: 2083.

32.

Yorucu

Ertac Varoglu

. Empirical investigation of relationships between energy consumption, industrial production, CO2 emissions, and economic growth: the case of small island states. Environ Sci Pollut Res 2020; 27: 14228–14236.

33.

Moreau

Neves

CADO

Vuille

. Is decoupling a red herring? The role of structural effects and energy policies in Europe. Energy Policy 2019; 128: 243–252.

34.

Sarkodie

Adams

Leirvik

. Foreign direct investment and renewable energy in climate change mitigation: does governance matter? J Cleaner Prod 2020; 263: 121262.

35.

Timilsina

Shrestha

. Factors affecting transport sector CO2 emissions growth in latin American and Caribbean countries: an LMDI decomposition analysis. Int J Energy Res 2009; 33: 396–414.

36.

Thampanya

Cowton

. Carbon neutrality targets, optimal environmental management strategies & the role of financial development: new evidence incorporating non-linear effects and different income levels. J Environ Manag 2021; 297: 113352.

37.

Riti

Shu

Song

, et al. The contribution of energy use and financial development by source in climate change mitigation process: a global empirical perspective. J Cleaner Prod 2017; 148: 882–894.

38.

Mujtaba

Jena

Joshi

DPP

. Growth and determinants of CO2 emissions: evidence from selected Asian emerging economies. Environ Sci Pollut Res 2021; 28: 39357–39369.

39.

Shahbaz

Balsalobre-Lorente

Sinha

. Foreign direct investment–CO2 emissions nexus in Middle East and north African countries: importance of biomass energy consumption. J Cleaner Prod 2019; 217: 603–614.

40.

Essandoh

Islam

Kakinaka

. Linking international trade and foreign direct investment to CO2 emissions: any differences between developed and developing countries? Sci Total Environ 2020; 712: 136437.

41.

Mujtaba

Jena

. Analyzing asymmetric impact of economic growth, energy use, FDI inflows, and oil prices on CO2 emissions through NARDL approach. Environ Sci Pollut Res 2021; 28: 30873–30886.

42.

Hamid

Alam

Murshed

, et al. The roles of foreign direct investments, economic growth, and capital investments in decarbonizing the economy of Oman. Environ Sci Pollut Res 2021: 29: 22122–22138.

43.

Hamid

Jena

Mukhopdhyay

. Should China prefer more foreign direct investment inflows to environmental change? J Public 2022; 22: e2466.

44.

Xie

Wang

Cong

. How does foreign direct investment affect CO2 emissions in emerging countries? New findings from a nonlinear panel analysis. J Cleaner Prod 2020; 249: 119422.

45.

Dong

Wang

, et al. The process of peak CO2 emissions in developed economies: a perspective of industrialization and urbanization. Resour Conserv Recycl 2019; 141: 61–75.