Revisiting the environmental impact of financial development on economic growth and carbon emissions: evidence from South Asian economies.

It is a global challenge to achieve sustainable economic growth by improving the environment. The present study discussed the role of the financial development sector in achieving sustainable economic growth and environmental quality in South Asian countries from 1990 to 2020 by controlling labour force participation, globalization, industrialization, and the education sector. A feasible generalized least squares (FGLS) panel data econometric technique has been used to check the relationship among the variables. The results show that financial development has a U-shaped relationship with carbon emissions and economic growth. Furthermore, labour force participation, industrialization, globalization, and educational school enrolment significantly increase CO2 and economic growth. This study suggests that the governments of South Asian countries should take steps to increase economic growth. For this purpose, effective supervisory mechanisms of financial development through financial innovation, improving financial efficiency, maintaining financial stability, and reducing the environmental pollution.

Introduction

The steadily increasing environmental deterioration of the world has made achieving the objective of sustainable development extremely difficult. It has become a major challenge for researchers and policymakers to achieve sustainable development by controlling environmental degradation. On one side, human activities increase economic growth, while, in return, it badly pollutes our atmosphere and cause global warming Rani et al., (2022). The ice and glaciers are melting rapidly due to climate change issues. The NOAA (2020) estimated that the average global temperature has increased by 0.13 °F since 1980, but it is expected to be doubled. In 2020, WMO (2020) acknowledged that the last decade had been the warmest. Keeping in view the importance of the most recent conference on climate change, COP 26, also pays special attention to climate change issues. At this conference, thousands of envoys and more than 130 heads of different countries met and set new targets to cope with climate change issues (Friedman, 2021, November 15).

The present study explores sustainable economic growth by protecting the environment through the financial development sector in South Asian countries. This study is based on the motivation for the first and third Sustainable Development Goals (SDG) under COP 26. The South Asian region is important because it is the home of 34% world's poorest population (Islam et al., 2021). Furthermore, this region is at the edge of the catastrophic effect of climate change (Germanwatch, 2021). Out of eight, five countries are included in the list of Germanwatch (2021) of most affected countries by the climate risk. The most vulnerable countries to climate change are India, Pakistan, Nepal, and Afghanistan.

The financial development sector is indispensable for any economy because most economic activities depend on it and it serves as the ventilator to all productive units, markets, and institutions that increase economic growth (Usman & Hammar, 2021). The advanced financial system is a competitive and efficient measure to allocate savings, support trade, and helps in diversification. It also works for hedging the risk and allows for easier access to investment opportunities for economic growth. The financial development sector promotes economic growth because transportation, agriculture, and industrial sector are closely interconnected with the financial sector. These sectors are the major contributors to economic growth (Jiang et al., 2019). Several developing countries introduced major reforms after 1990 which have increased the economic growth in many countries. But oil shock of the 1990s, the 9/11 incident, the financial crisis in 2008, the Covid-19 pandemic, and the Russian-Ukrainian war have badly affected the financial sector. These shocks become hurdles in the way of economic growth of these countries which failed to handle these shocks properly (Rahman et al. 2020); on the other hand, it is pointed out that those countries that cope with these crises properly enjoyed economic growth (Ahmad et al. 2022).

To achieve sustainable growth, we cannot ignore the importance of a clean and neat environment. Due to rapid growth in the financial sector, the agriculture, industrial, and transportation sectors require more energy for work. Usually, most countries meet their energy requirements through fossil fuel consumption, thus emitting carbon dioxide (CO2). Several studies indicate that the FDsector is base of environmental degradation (Khan et al., 2022; Jiang & Ma, 2019; Tahir et al., 2021). In some developed countries, the strong FDsector causes improvements in the environment (Shobande & Ogbeifun, 2022; Abid et al., 2022).

In South Asian regions, GDP per capita increases swiftly compared to the other lower-income countries. It is noted in Fig. 1 that all selected South Asian countries' GDP per capita is rising at a consistent speed. Sri Lanka and Bhutan are at the top with higher GDP per capita.

Fig. 1

GDP Per capita in from 2000 to 2020

Furthermore, in this region, average CO2 was recorded at 0.99 metric tons in 2020, which is higher than the other lower-income countries (WDI, 2022). Figure 2 shows the CO2 in selected South Asian countries from 2000 to 2020. It is observed that CO2 is increasing consistently in all South Asian countries. India and Bhutan are with the highest carbon emissions. Due to the massive growth in carbon emissions, the average temperature rises in this region, causing many extreme events.

Fig. 2

CO2 per capita from 1990 to 2020

Previous studies have analysed the impact of financial development (FD) on a country's growth. But not many studies have explored the separate effects of FD on CO2 and economic growth in South Asian economies. This brings up the need to check the relationship between them. The current study tries to fill the gap by using the nonlinear effect of FD on economic growth and CO2 in South Asian countries. This study plays a significant role in achieving the sustainable development goals of South Asian countries. It endeavours to contribute to FD and further understanding of a sustainable environment. Therefore, the primary objective of this study is to examine the impact of nonlinear FD on economic growth and carbon emissions.

The rest of the study is organized as follows. The second section reviews both theoretical and empirical literature. The third section focuses on the methodology to be employed for analysis. The study results and analysis have been discussed in the fourth section. The final section represents the conclusion, recommendations, and suggestions.

Literature review

There is vast literature that has reported the FD relationships with different nations' economic growth and carbon emissions. In this modern world, financial development plays a vital role in the promotion of industry, agriculture, and education sectors. After World War II, this sector has gained special importance for government institutions to stabilize the exchange rate, maintain political stability, and implement fiscal and monetary policies. These factors are considered important contributors to economic growth and environmental quality.

The policymakers and researchers have failed to find the exact relationship of FD with economic growth and environmental quality. A long list of studies has shown that the FD sector improves the quality of the environment. Shobande and Ogbeifun (2022) revealed that FD reduces CO2 and improves environmental quality in OECD countries. Abid et al. (2022) found that FD negatively affects the CO2 in G8 countries by using FMOLS from 1990 to 2019. These studies argue that the developed financial sector inspires investors in environment-friendly technologies.

Apart from the above studies, various studies have explored that the FD sector is the major culprit of environmental pollution. From a global perspective, Khan et al. (2022) and Jiang and Ma (2019) explored that FD deteriorates the environmental quality by using the Generalized Method of Moments (GMM). Tahir et al. (2021) concluded that FD contributes to CO2 emissions by using Fully Modified Least square (FMOLS), Dynamic Least Square (DOLS) and Pooled mean group from 1990 to 2014. These studies found that a healthy financial sector inspired the investors to attract the cheaper resources, which pollutes the environment.

Presently, each country desires to achieve sustainable development goals without losing environmental quality. The FD sector plays a significant role in sustainable economic growth. On the other hand, there is a controversial relationship between FD and economic development. Ahmed et al. (2022) pointed out that FD increases sustainable economic development in South Asian countries from 2000 to 2018. Rahman et al. (2020) empirically estimated that FD negatively impacted economic growth in Pakistan from 1975 to 2017. Asteriou and Spanos (2019) found that FD increases economic growth in 26 European countries before the financial crisis, while after the financial crisis, it declines economic growth. Amjad et al. (2021) examined the impact of FD on economic growth and the environment from 1990 to 2020 in Pakistan. The Nonlinear ARDL econometrics technique was employed to determine the empirical results. The study concluded that positive shocks of FD improve environmental quality and economic growth. Wang et al. (2022) examined the impact of economic growth and environmental degradation in developed countries. They found that economic growth is a major cause of deteriorating environmental quality. Similarly, Rani et al., (2022) explored that FD not only increase the economic growth of SAARC countries but has badly effect on environmental quality.

Apart from financial development, there are also many determinants of economic growth and environmental quality. A skilled and high-quality labour force is critical to economic development. Countries with many skilled labour resources expand faster than those with many low-skilled workforces. Several studies have been conducted to explain the positive association between labour force participation and economic growth (Ziaei, 2022; Gouzoulis, 2022). However, a few studies concluded that a negative relationship between the labour force and economic growth and labour becomes a cause of environmental degradation (Alsamara, 2022).

Industrialization has huge impact in innovation, job creation, and optimal utilization of resources. Elfaki et al. (2022) investigated the impact of industrialization on economic growth in Indonesia. Using FMOLS and DOLS approaches, the studies concluded that industrialization positively contributes to economic growth. Ndiaya and Lv (2018) examined that industrialization increased economic growth in Senegal from 1960 to 2017. On the other hand, the study concluded that industrialization increases CO2 emissions because most of the industrial sector is based on fossil fuel consumption, which is considered as the major contributor to CO2 emissions. Zheng et al., (2021) found that industrialization increases CO2 in the 29 provinces of China.

Globalization increases the interdependence among countries by developing economic and social relationships. Santiago et al. (2020) found that globalization positively increases economic growth in 24 Caribbean and Latin American developing countries from 1990 to 2015. Similarly, Guan et al. (2022) examined the effect of globalization in G-10 countries and explored that it certainly rise the environmental degradation in these countries. Moreover, Hassan et al. (2019) examined that globalization increases economic growth in Pakistan from 1970 to 2014, and on the other hand, globalization has a strong influence on carbon emissions. Furthermore,  Khan et al. (2019) scrutinized the impact of globalization on CO2 in Pakistan from 1971 to 2016. The study concluded that globalization positively increases carbon emissions.

Education enrolment has been used as a proxy to measure the educational level. It is divided into pre-primary, primary, secondary, and tertiary education. A vast literature has discussed that education enrollment increases economic growth (Nenbee & Danielle, 2021) and education also impacts carbon emissions (Maneejuk & Yamaka, 2021). Bano et al. (2018) identified the impact of human capital through education on CO2 in Pakistan. The study found that improvement in human capital declines CO2 by accessing higher education. A low level of education declines human capital, which increases carbon emissions. Similar findings were obtained by Haini (2021) and Katircioglu et al. (2020).

Based on the above discussion, it is observed that there exists a linear relationship between financial development, economic growth and environmental quality in single-country studies. This study will contribute to the empirical literature by examining the association of FD with economic growth and CO2 emissions using the nonlinear quadratic effect in South Asian countries.

Material and methods

The present study is an attempt to examine the effect of FD on economic growth and carbon emission in six South Asian countries named Bangladesh, Bhutan, India, Nepal, Pakistan, and Sri Lanka, panel data for the period 1990–2020. The remaining two countries, Afghanistan and Maldives, are not included due to the unavailability of the data set. This study has been carried out in South Asian countries because, in these countries, the financial sector is growing rapidly and also these countries are keen to achieve sustainable economic development. Due to the massive poverty revealed in these countries, the population of these countries rely heavily on cheaper technologies which emit carbon and cause environmental degradation. This study uses the FD index as the key explanatory variable which is constructed by using the recommendations of the World Bank based on the financial stability, efficiency, depth and access to capital. For this purpose, bank credit to bank deposits, bank return on assets, domestic credit to the private sector and bank deposits to GDP annual time-series data are taken from the Global FD database. Further details of variables are listed in Table 1.

Table 1

Description of data

Symbols	Variables	Units
CO2	Carbon dioxide emission	Metric tons per capita
GDPPC	GDP per capita	Constant 2015 US$)
FD	Financial development index	Index
LFTOT	Labor force	Total
EDU	School enrollment, tertiary	% gross
INDU	Industrialization	% of GDPPC
KOF	Globalization	Index

The study uses two models for analysis purpose.

The functional forms can be written as:

The multiple regressions of both equations can be written as:

In Eqs. 3 and 4, and are coefficients of the first model whereas,, and indicate coefficients of the second model and is the residual. These models are not appropriate in the case of SAARC countries because the financial systems of these countries were not stable during 1990 to 2020. So this sector faced many ups and downs. In this equation, FD index has nonlinear behaviour. So, we use the quadratic term of FD.

The advantage of employing a square form is to detect the deviations from the constant returns to scale assumption, which explains the marginal effect (Sial et al., 2022). The square form can also be used to, country-specific effects and can compute from a single-mode by plugging country-specific incidence. After taking the partial derivative, we find the cut-off values of the FD index and GDPPC from Eqs. (5) and (6) as follows:

The quadratic specification helps in determining the optimal value of FD at which the marginal effect of the FD and gross domestic product per capita reverse signs. This study uses its benchmark value to determine whether the economy is located before or after the threshold value.

Estimation technique

Parks (1967) proposed the algorithm-based feasible generalized least square (FGLS) method popularized by Kmenta (1986). The current study uses FGLS model for panel data estimation of six South Asian countries spanning from 1990 to 2020. The advantage of this model is that the cross-sectional discrepancies can be used to reflect differences in the slope coefficients' standard errors. Furthermore with the configuration, it is ease to control time-series autocorrelation, cross-sectional autocorrelation, and cross-sectional heteroscedasticity (Sial et al., 2022).

Empirical results

Table 2 provides a summary statistics of the variables. It can be observed that all mean values of variables are higher than their standard deviation. This indicates that all the data are under dispersed (Amjad et al. 2021). Based on the normality test, LNGDPPC, CO2, LNLFTOT, LNEDU, and LNINDU are not normal at the 5% level. The kurtosis reveals the presence of outliers; it implies that variables are not homogeneous across countries, rendering the pooled OLS model obsolete.

Table 2

Descriptive statistics

	LNGDPPC	LNCO2	FD	LNINDU	LNKOF	LNLFTOT	LNEDU
Mean	7.1332	0.6966	0.3207	3.2117	3.8103	16.7377	2.1801
Median	7.0737	0.6474	0.3627	3.2769	3.8607	16.9929	2.2601
Maximum	8.3488	1.9611	2.2801	3.8089	4.1271	20.0195	3.3824
Minimum	6.2018	0.0876	− 1.3939	2.4679	3.1355	12.2770	0.0931
Std. Dev	0.5410	0.4545	0.7910	0.3051	0.2495	2.2517	0.6821
Skewness	0.4158	0.9058	− 0.1302	− 0.2480	− 0.8630	− 0.5517	− 0.5029
Kurtosis	2.4281	3.4735	2.7127	2.7023	3.1148	2.6228	2.8108
Jarque–Bera	6.6208	22.7897	0.9770	2.1757	19.4502	8.8373	6.8088
Probability	0.0365	0.0000	0.6135	0.3369	0.0001	0.0121	0.0332

Figure 3 depicts the correlation plot of both models. At the same time, there is evidence of a perfect association between paired variables, so there is no problem with multicollinearity in the models.

Fig. 3

Correlation plot

Table 3 presents the results of the Feasible Generalized Least Square (FGLS) model. The total number of years per cross-section is less than 30; assuming all variables are stationary, the results show that a one percent increase in the financial development index (FD) significantly declines LNGDPPC by 0.108 percent in model 1. Similar results were explored by Rahman et al. (2020) in Pakistan, while the quadratic coefficient of FD shows a one percent increase in the FD significantly increases economic growth by 0.080 percent. Similar results were estimated by Ahmed et al. (2022) in South Asian countries. When we trace the negative level coefficient and positive quadratic coefficient of FD by using the mean, standard deviation and constant term, we propose a U-shaped relationship as presented in Fig. 4a.

Table 3

Detail results of FGLS

Variables	Dependent variable: LNGDPPC	Dependent variable: LN CO2
FD	− 0.108* (0.021)	− 0.208* (0.045)
FD2	0.080* (0.016)	0.110* (0.034)
LNLFTOT	0.200* (0.008)	0.008 (0.017)
LNINDU	0.684* (0.046)	0.832* (0.098)
LNKOF	2.246* (0.084)	0.848* (0.179)
LNEDU	0.056*** (0.031)	0.181* (0.065)
Constant	− 0.425 (0.312)	− 5.745* (0.663)
Number of obs	156	156
Number of groups	26	26
Wald chi2(5)	2069.990	191.920
Prob > chi2	0.000	0.000

Author’s estimation

Fig. 4

Trace effect of the FD

In model 2, a 1% increase in the financial development index (FD) significantly decreases the LNCO2 by 0.208 percent. The coefficient level of FD declines CO2 (Similar results were obtained by Shobande & Ogbeifun 2022; Abid et al. 2022). At the same time, the quadratic coefficient of FD contributes positively to CO2 (Their results were in line with Khan et al., 2022; Jiang & Ma, 2019). When we trace the negative level coefficient of FD and quadratic coefficient of FD by using the mean, standard deviation and constant term, we propose a U-shaped relationship (Fig. 4b).

Table 4 shows the cut-off value of FD of model 1 and model 2, showing the minimum requirements of the FD to achieve economic growth and carbon emissions.

Table 4

Cut-off value of FD

Coefficients	Model 1	Model 2
Level	− 0.108	− 0.208
Quadratic	0.080	0.110
cut-off value	0.674	0.945

Table 5 shows the linearized affect the economic growth and CO2 by using the average value of the FD. In model 1, Bangladesh, Bhutan, India, and Pakistan lie on the negative side of Fig. 4a, while only two South Asian countries, Nepal and Sri Lanka, lie on positive side. This shows that FD increases economic growth in Nepal and Sri Lanka. Furthermore, in model 2, all South Asian countries lie on the negative side of Fig. 4b. It means that FD decreases CO2 emissions. We summarize that only Nepal and Sri Lanka are trying to achieve sustainable development. At the same time, the remaining South Asian countries failed to achieve sustainable development because these countries have lower per capita income. It is observed that most of South Asian countries have underdeveloped financial sectors and it faces many ups and downs, which failed to perform its function properly which effect the development process adversely in these countries.

Table 5

Linearized effect of economic growth and CO2 based on mean value of FD

Country	Mean value of FD	Model 1	Model2
Bangladesh	0.503	− 0.027	− 0.097
Bhutan	− 0.028	− 0.113	− 0.214
India	0.433	− 0.039	− 0.113
Nepal	0.774	0.016	− 0.038
Pakistan	− 0.535	− 0.194	− 0.326
Sri Lanka	0.777	0.016	− 0.037

After discussing the financial development, now we discuss the impact of the control variables on economic growth and CO2 emissions. Labour force participation (LNLFTOT) has a significant and positive relation with economic growth. The results show that a 1% increase in the labour force brings a 0.200% increase in the economic growth of South Asian economies. It means a lot of the labour force has much work to do; therefore, the productivity of South Asian countries is high. They earn money for the survival of their families and help reduce the burden on those countries’ economies. The results are consistent with those (Ziaei, 2022; Gouzoulis, 2022). On the other hand, the labour force (LNLFTOT) significantly and positively impacts CO2 emissions (Alsamara, 2022). In these countries, most of the workers are unskilled and prefer to work with old and cheaper machinery, which causes CO2 emissions in the atmosphere.

Similarly, industrialization (LNINDU) has also a positive impact on economic growth (Elfaki et al., 2022; Ndiaya & Lv, 2018). The industrialization process improves productivity and provides job opportunities to skilled and unskilled persons, increasing the nation's development process. Industrialization (LNINDU) significantly increases CO2 emissions (Zheng et al., 2021). All countries are under development in the South Asian region, and their industries utilized fossil fuel consumption which is the major source of CO2 emissions.

Globalization (LNKOF) has a significant and positive relation with economic growth. The results show that a 1% increase in globalization brings a 2.246% increase in the economic growth in South Asian economies (Santiago et al., 2020; Hassan et al., 2019). It means more globalization increases the recognition of diversification of the products in different countries. Furthermore, globalization (LNKOF) also significantly increases CO2 (Khan et al., 2019).

Moreover, education (LNEDU) positively impacts increasing economic growth. The results show that a 1% increase in LNEDU brings a 0.056% increase in economic growth. Information and skills have an essential role in raising awareness among the public to enhance economic growth. (The outcomes are similar to Nenbee & Danielle, 2021; Maneejuk & Yamaka, 2021). On the other hand, education positively influences the CO2 in South Asian countries. One percent increase in the LNEDU statistically and significantly increases carbon emission by an average of 0.181% (Haini, 2021).

Conclusion and policy implication

The objective of this study is to achieve sustainable economic development without environmental degradation based on financial development. For this propose, the quadratic transformation of FD is used with GDP per capita and CO2 in selected South Asian countries from 1990 to 2020. In the quadratic FD model, the level coefficient of FD has a negative sign. In contrast, the quadratic coefficient of FD has a positive sign, which proposes the U-shaped relationship between GDP per capita and carbon emissions. It means an initial increase in the FD decreases GDP per capita and CO2 because the infrastructure is not so much developed at this stage, while after the maturity stage of FD GDP per capita and CO2 increase because of the infrastructure improvement. The cut-off value of FD shows the minimum optimal value of the U-shaped curve concerning GDP per capita and carbon emissions. The linearized effect of FD shows that only two countries, Nepal and Sri Lanka, have achieved sustainable growth.

Furthermore, the remaining independent variables, labour force participation, industrialization, globalization and school education enrolment, positively increase GDP per capita and CO2 in South Asian countries. This study recommends that the financial institutions should provide funds to those industries that use environment-friendly technologies. Before providing funds, the financial institution must make sure that the investment has been made in those major projects which use renewable energy resources and restrict funds to highly polluting industries.