Probing the Effect of Governance of Tourism Development, Economic Growth, and Foreign Direct Investment on Carbon Dioxide Emissions in Africa: The African Experience

Abstract

The environmental repercussions of extensive carbon dioxide (CO₂) emissions on the environment are crucial for policymakers and scholars. The repercussions of and connection between economic growth (ECG), tourism (TOUR), and foreign direct investment (FDI) on CO₂ emission mitigation have been measured and argued from empirical and theoretical perspectives by scholars. Notwithstanding, the extant body of knowledge has failed to incorporate and investigate the function of governance in decarbonizing tourism activities and FDI from CO₂ emissions to attain a healthy and quality environment in Africa. Hence, this current research investigates governance’s role in the reduction processes of CO₂ emissions grounded in environmental Kuznets curve (EKC) conceptual assumptions for panel data spanning 2000 through 2020 for 27 African countries. This research utilized the Westerlund panel cointegration approach for the investigation of the cointegration of the selected variables. This study applied the Driscoll–Kraay regression approach for the long-term estimation. In addition, the dynamic ordinary least squares (DOLS) and the pooled mean group (PMG) were used for robustness checks. The findings of this research indicated that the governance (GOV) indicators employed have a statistically significant effect on the CO₂ emission reduction. Besides, this study found that the appreciation of the income of the nations gives credence to the formation of the EKC theory and contributes to the decline in CO₂ emissions within the selected African nations. The findings revealed that tourism, FDI, ECG, and GOV are positive and significant factors leading to increased CO₂ emissions in Africa. Furthermore, the results showed that effective governance and control of FDI inflows and tourism activities can support decarbonization. These findings suggest the merits of governance in ensuring effective decarbonization policies of the environment, and policy suggestions are accordingly put forward.

1. Introduction

The connection linking economic growth (ECG), tourism, and environmental pollution problems has been extensively argued [1]. Extant research has investigated the environmental Kuznets curve (EKC) assumptions from empirical and theoretical perspectives to measure the connection between the environment’s quality and economic growth [2,3]. Environmental issues are quantified by factors such as ECG, tourism (TOUR), governance (GOV), urbanization, energy consumption, and financial development [2,3,4]. Furthermore, most studies employed the EKC theory to reveal the significance of the connection of variables FDI, TOUR, and ECG to an economy [5,6,7]. This study is important because it is premised that African nations seeking higher economic growth are interested in higher FDI inflows, tourism development, and effective governance policies [8]. Furthermore, the enormous benefits that countries derive from FDI and TOUR engagement include employment, economic development, and increases in gross domestic product (GDP) [3]. These benefits gained from tourism and FDI have led to the greater promotion of their activities, highlighting a double-edged sword because of the accompanying CO₂ emissions in the pursuance of ECG [3]. Travel and tourism have detrimental consequences on the global climate, the ecosystem, and biodiversity [9]. According to the climate change vulnerability index within 2015, several nations in Africa are at extreme and high risk in the case of continuing life-threatening climate change. The rapid reduction in Africa’s carbon sinks is worsened by low levels of development, a large expected increase in greenhouse gases (GHGs), and significant climate change sensitivity. Deforestation and the loss of grasslands are hastened by human activity and are credited for destroying Africa’s carbon sink [10].

In recent times, African countries have rapidly grown in industrialization with a corresponding increase in GDP, approaching the level of economic development in developed nations. Considering Africa’s growing population and urban cities, with a considerable inflow of FDI and TOUR activities, this study projects a need to provide generally agreed governance structures and policies to decarbonize ECG, FDI, and tourism-related activities. Globally, tourism-associated transportation contributes to GHG emissions [11]. Additionally, the African economy continues to expand over time, resulting in a rise in energy consumption and demand [12]. Liousse et al. [12], suggested that the IPCC models understate emissions from Africa that might increase from 20% to 55% of worldwide anthropogenic gas pollution by 2030. Thus, a more credible investigation is suggested to determine the actual emission status in Africa and the role that the governments in the respective countries play in reducing emissions.

In Africa, it is realized that there is no apparent statewide emissions database that offers estimates of anthropogenic combustion [12]. Liousse et al. [12] improved the suggested IPCC-underestimated CO₂ trends in the African continent and revealed the real condition of CO₂ emission and GHG emission standards in African countries using scenarios with projections from 2005 to 2030. The reference scenarios are supposed to replicate the current condition of the globe from the standpoint of ‘technological and business transition as usual’, with fundamental economics as the only driving force. The carbon constraint scenarios are defined as the imposition of carbon fines segregated throughout the globe’s major territories and time frames. When comparing the CO₂ emission constraint scenario to the reference case, the CO₂ emissions are calculated to represent regional variances following the Kyoto objectives for 2010 to attain a worldwide carbon emission depletion of around 37 Gt of CO₂ equivalent to 10 GtC before the year 2030 [12,13,14,15]. Thus, the regional spatial dispensation of African anthropogenic gases combines black and organic carbon emissions. Organic carbon emission is a key component of atmospheric aerosol [12].

Premised on the investigation of Liousse et al. [12], it is apparent that Africa needs pragmatic policies to decarbonize ECG, TOUR, FDI, and other contributing factors of pollution to ensure a healthy environment for human survival. According to IPCC [16] findings, inadequate adaptation interventions for the emerging impact of environmental change are already degrading the foundation for sustainable economic development and sustainable tourism by extrapolation. Considering the gravity of the CO₂ emission development in Africa, this study put forward four preliminary measures to be implemented by the respective governments and other stakeholders of the selected African countries to achieve the target of decarbonizing tourism-related emissions and FDI to attain economic growth. These four key areas meriting needful concentration are as follows: (1) incorporating carbon emission programs into the education systems; (2) adopting cleaner technologies (technological improvement programs and policies); (3) creating awareness of the impact of carbon emissions on humans and the environment; (4) structured general efficient and effective governance policies to be implemented in all regions.

In addition to the above four key points, the Paris Climate Agreement (PCA), which depicts the key component of the United Nations Framework Convention on Climate Change (UNFCCC) that all 54 African countries signed, needs to be followed [17]. The UNFCCC aims to reduce GHGs and maintain global average temperatures far below 2 °C. Furthermore, the reliance of Africans on biomass for lighting and cooking indicates that about 90% of the African populace is vulnerable to the health risks posed by CO₂ emissions [18]. This emphasizes that the amount of environmental destruction is escalating in Africa and worldwide. Notwithstanding, policymakers still have an opportunity to counteract the worst effects if they act now [18].

These agreements suggest that good environmental governance is vital for safeguarding and ensuring sustainable use of resources and environmental quality [19,20]. Several indicators emphasize the dimensions of governance. Thus, institutional regulatory quality helps sustain natural resources and conserve the environment [21]. Diverse forms of governance measurement have various influences on CO₂ emissions [22,23]. In most countries, institutional governance quality is the premise for determining the quality of the environment [24,25,26]. These dimensions of governance are multifaceted: political stability (POLS), the voice of accountability (VA), regulatory quality (REQ), the rule of law (RULE), control of corruption (CC), and governance effectiveness (GOVE) [27,28,29]. The impact of these elements of governance concerning the environment varies. Notwithstanding, the governance regulatory quality performs an essential role in mitigating pollution. The regulatory quality of the government and other institutions influences CO₂ emissions mitigation directly and indirectly [30]. Democratic and social factors are connected with the quality of an environment because effective governance enhances the quality of the environment via prudent environment regulations, creating awareness for organizations and citizens greatly concerned about environmental pollution problems. Research has demonstrated the merits of good environmental governance and effective governance institutions in ensuring the environment’s quality in maintaining sustainable resource utilization [20,23]. Furthermore, studies have shown that corruption influences the environment’s quality indirectly and directly. This debilitates institutional performance, impedes ineffective environmental policy implementation and creates rent-seeking behavior [31,32]. However, controlling corruption reduces CO₂ emissions [33,34].

On the basis of the abovementioned premises, this study seeks to question whether governance matters in Africa’s environmental pollution issues. This research intends to answer whether the function of governance matters in the mitigation of CO₂, as earlier studies conducted little investigation on this crucial topic and sometimes reported inconsistent findings in the context of Africa. Although the connection between governance and CO₂ and other variables has been investigated in a few advanced countries, the essential aspect of panel analysis in African countries has been greatly ignored. This current research fills the gap created on the African continent by probing the role of governance of TOUR, FDI, and ECG in CO₂ emission mitigation by employing the EKC assumptions. A key limitation identified in the earlier research denotes that the methodologies used for the empirical evaluation do not support and capture the phenomenon of cross-sectional dependence (CSD). As a result, the regression parameters given in CSD testing were proven unreliable due to the inherent bias. The authors of the present study address this challenge by employing Driscoll–Kraay panel standard regression and pooled mean group estimation methodology. The essential characteristics of this current research are organized into three categories. Firstly, this research probes the conduct of governance in 27 African nations employing the largest accessible data spanning 2000 through 2020 for the selected countries. Secondly, the authors evaluate governance to outline a comprehensive environmental regulation. Thirdly, the authors employ a robust panel data evaluation approach for the analysis of this study.

The remainder of this paper is structured as follows: Section 2 entails a review of the literature; Section 3 presents the data, methodology, and respective empirical models; Section outlines the study results and discussion; Section 5 presents the research conclusions and core regulatory framework.

2. Review of the Literature

This section of the research elaborates on earlier literary studies concerning the connections linking carbon dioxide emission (CO₂), tourism (TOUR), economic growth (ECG), governance (GV), and foreign direct investment (FDI).

2.1. The Connection between Tourism and CO₂ Emissions

Extant research has demonstrated that tourism has expanded steadily during the last decades, accounting for 10% of world job opportunities and 10% of worldwide GDP [35,36]. The tourist industry is forecasted to create significant merits in socioeconomic growth and job avenues worldwide by 2030 [37]. Notwithstanding, these positive impacts of TOUR have their corresponding negative environmental impacts. Contemporary TOUR activities are considered the most significant source of CO₂ emissions, posing challenges to most governments. Tourism-associated economic growth is expected to be decoupled from resource consumption, leading to CO₂ emissions [37]. The preference for foreign and domestic tourists to use personal automobiles is becoming more prevalent, which significantly affects the environment [38,39,40,41]. The CO₂ footprint of the worldwide TOUR industries revealed that worldwide CO₂ emissions from the tourism industry are presently not adequately evaluated [42]. The researchers determined global tourism-related carbon flows and carbon footprints for 160 countries from both an origin and destination accounting perspective. According to the study result, the tourism industry’s worldwide CO₂ footprint soared between 3.9 and 4.5 GtCO₂ emissions between 2009 and 2013, accounting for nearly 8% of worldwide GHG emissions, four times higher than previously estimated. The findings revealed that transportation and other energy consumables are significant factors leading to CO₂ emissions. The study indicated that high-income countries might be responsible for the vast majority of this footprint [42].

Furthermore, the United Nations WTO [43] comprehensive analysis of tourism-related transport influence on environmental pollution indicated that, by 2030, transportation-related CO₂ emissions would constitute 5.3% of all artificial CO₂ discharge. Furthermore, between 1995 and 2018, Anser et al. [44] utilized extensive longitudinal research data from 132 nations to analyze CO₂ emission costs in TOUR industries. The findings revealed that inward tourism will probably fall from 19.5% to 16.8% between 2020 and 2028 due to the rise in CO₂ emissions and the stringent policies implemented by stakeholders to reduce CO₂ emissions. Tourism-related activities substantially impact climate change, which challenges Africa and the world economy. As a result, tourism and travel, land degradation, and deforestation are closely linked to GHG emissions [45].

The problem of experiencing increases in CO₂ emissions relating to the TOUR industry was portrayed much more grimly in [46]. They predicted that the tourist industry would be a significant source of GHGs. Notwithstanding, their findings revealed that many governance policies and practice reforms in travel activities might significantly reduce pollution. Thus, the adaptation of TOUR activities to reduce emissions would be prudent for the development of sustainable long-term goals [46,47]. Despite continued increases in the patronage of TOUR activities, effective governance policies may reduce CO₂ emissions. Moreover, introducing a modern lower-level emission technological system may support and sustain the decrease in CO₂ emissions. This current study posits that a positive association between tourism and carbon emissions may be established on the basis of the issues presented.

2.2. The Connection between Economic Growth and CO₂ Emissions

Studies have indicated that CO₂ emissions also increase from industrial and domestic use when pursuing economic growth [48,49]. EKC serves as a model for the interaction of economic expansion, energy consumption, and environmental pollution issues [50]. The World Bank Development Report on EKC in 1992 revealed that “increased economic activities invariably impacted the environment and premised on rigid suppositions concerning advanced technologies, preferences, and investment opportunities”. Moreover, “as income increases, the demand for enhanced environmental standards would keep rising, and so would the available resources for investment” [50,51]. With the assumption that growth in an economy is inevitably accompanied by carbon emissions, this study adopts the EKC theory in anticipation of getting varied or dynamic results from the various studies analyzed relating to the role played by governance in reducing CO₂ emissions of TOUR, ECG, and FDI activities within selected African countries. This approach can validate or refute the EKC assumption on the basis of the roles played by the governments of the specified countries.

In most sub-Saharan African countries, ECG grew from a minimal base of about 6.7 million in 1990 to 33.8 million in 2012, enhancing the economy’s growth [52]. For instance, the tourist industry accounted for 44% of aggregate GDP growth within the Seychelles and 16% of GDP in Mauritius [53]. Studies have established the link between ECG and CO₂ emissions in some sub-Saharan African nations [54]. The study revealed that, while the effects differ by nation, in the long term, high consumption of energy-related products and economic prosperity are connected with higher environmental pollution within many countries. The findings indicated further that long-term economic growth would result in minimal CO₂ emissions within Congo, Ghana, Senegal, Benin, and Nigeria [54]. Within Nigeria, Gabon, and Togo, the link between CO₂ emissions and the ECG demonstrated that the absence of environmental pollution measures might affect their economies. Furthermore, the study revealed bidirectional causation between ECG and CO₂ emissions in Nigeria in the short term. In the long-term interrelations, bidirectional causation was found to exist between ECG and CO₂ emissions in Gabon and Congo. The findings revealed that Benin, the Ivory Coast, South Africa, Togo, Nigeria, and Senegal are affected by high CO₂ emissions [54]. An investigation in six sub-Saharan African nations on the causality relationship among FDI, ECG, and CO₂ emission adopting the ARDL approach revealed that countries are cointegrated into the long-term association, giving credence to the EKC hypothesis in Zimbabwe, Kenya, and Congo [55]. This current research proposes that a positive connection between economic growth and carbon emission may be established on the basis of the critical issues presented.

2.3. The Connection between FDI and CO₂ Emissions

FDI is the acquisition of a share in a company by a corporate entity or investment firm based outside of the entity’s territory. In the broad sense, FDI alludes to a commercial decision to acquire a major stake in or purchase a foreign corporation entirely to expand its operations into a new environment [56,57]. Causal relationship analysis among CO₂ emissions, ECG, and FDI for 54 nations from 1990 to 2011 revealed that bidirectional causation exists between ECG, FDI inflows, and FDI and CO₂ in the selected nations’ panel estimations [58]. Investigations concerning the connection linking FDI and governance revealed a plethora of resources essential to implementing FDI policies [27]. Minimal focus is geared toward host nations’ governance for foreign investment policy implementation [27]. An investigation in sub-Saharan African countries demonstrated that various aspects of governance contribute to the major attraction of FDI [59,60]. Nevertheless, political instability and corruption negatively influence FDI, whereas enhancing political and institutional structures affirmatively impact FDI flows [59]. Furthermore, the capacity of governments and institutions to implement and facilitate FDI is constrained by the goals of such investments within the various economic sectors [61].

Additionally, studies have shown that CO₂ is mainly increased due to FDI activities in certain nations while having the reverse effect in others [55]. Thus, normally, a country’s pace of capital formation dictates the speed of economic expansion [62]. Economic growth is essential in generating FDI amongst emerging nations [63,64]. Thus, studies have indicated that the causation for both ECG and FDI is influenced by nation-specific characteristics [65,66]. Research has found that FDI significantly impacts Africa’s and China’s rising CO₂ emissions [54,67]. FDI and CO₂ emissions in diverse countries have a positive association. However, not all studies back up the conclusion [55,68]. Notwithstanding, it has been argued that FDI could help countries attain sustainable development goals (SDGs) (OECD, 2019). Thus, governments must explore how to combat the problems of attaining economic growth without environmental pollution [69]. Studies have portrayed that FDI helps host countries improve their energy efficiency and reduce CO₂ emissions [70,71].

Thus, FDI inflows responsive to laws elevate international shareholding of FDI to local shareholding [72,73,74]. Studies have shown that collective good governance actions may enhance FDI inflows [75,76,77]. Thus, factors that improve investors’ regulatory quality, including tax exemptions or reductions and flexible regulations, support the implementation of legal policies to safeguard the natural environment from uncontrolled exploitation of resources and pollution. Different studies have highlighted the essence of effective governance adherence to a precise selection of policies to achieve sustainable environment policies in enhancing good practices relating to FDI and governance interaction mechanisms. Similarly, institutions and stakeholders need to play an essential role in increasing efficient governance systems to support the environmental protection agencies that regulate environmental pollution issues [78]. Furthermore, empirical results have revealed that FDI inflows contribute to pollution and a corresponding increase in CO₂ emissions, corroborating the pollution haven theory. They concluded that reducing fossil-fuel usage and promoting an ecologically friendly economic growth approach in emerging countries will benefit their overall wellbeing and may further support the presence of the EKC theory [79]. This present research posits that a positive association between FDI and carbon emission may be established on the basis of the core issues presented.

2.4. The Nexus between Governance and CO₂ Emission

Governance refers to the institutional and traditional frameworks that allow a nation’s supreme powers to be enforced [28]. These institutions and traditions comprise the nation’s procedures to elect, monitor, and re-elect governments, the government’s capacity to judiciously formulate and implement prudent actions, the condition of institutional bodies that govern economic and social interconnections, and citizens’ respect for the authorities. Moreover, political stability and the fight against corruption invariably favor regionalism. In reality, conformance to groups of well-selected governments would facilitate a stable and sound regulatory framework and macroeconomic policies that motivate investors [27]. According to Ederington et al. [80], the rule of law supports environmental protection and attracts FDI and trade. Studies have demonstrated that institutional development, expediting compliance with laws, and minimizing corruption may reduce a nation’s risk of CO₂ emissions and boost the attractiveness of FDI [81,82]. Notwithstanding, limited studies have investigated the role of governance in the connection between CO₂ emissions and FDI [61].

Governance is an essential element that positively facilitates and regulates the activities of FDI and TOUR to reduce CO₂ emissions [61,83,84,85,86]. Governance ensures that a country’s resources are used efficiently by providing those activities geared toward economic productivity and can sustain environmental quality processes. Governance also strengthens access to tourism and FDI by implementing decarbonization measures [87,88,89]. As a result, this study posits that governance helps strengthen or lessen the interdependence of tourism, FDI, and CO₂ emissions. Thus, researchers have applied several indicators in measuring governance functions in FDI inflows, economic prosperity, and CO₂ emissions in diverse jurisdictions. This study employs governance (GOV) indicators comprising governance effectiveness, political stability, the rule of law, regulatory quality, the voice of accountability, and control of corruption for analysis [27,28,29]. According to the above development, it is evident that researchers have not paid great attention to the functions of governance in successfully implementing decarbonization policies that ensure zero emissions from FDI. Therefore, the authors of this research hypothesize that effective governance may play a significant part in regulating the reduction in CO₂ emissions.

3. Research Data, Study Area, and Model Construction

3.1. Research Data

The research data employed for advancing or estimating the study indicators and normalizing them were collected from accredited data sources [28,90], as shown in

Table 1. Variables chosen for the study span the period 2000 to 2020.

Variable	Symbol	Unit of Measurement and Definition	Source of Data and References
Tourism	TOUR	Tourism receipts per capita in current US dollars (USD).	World Bank [90]
Carbon dioxide emissions	CO2	CO2 emission per capita.	World Bank [90]
Economic growth	ECG	GDP per capita (constant 2015 USD).	World Bank [90]
Foreign direct investment	FDI	Inward flows (FDI per capita USD)	World Bank [90]
Voice of accountability	VA	The accomplishment and the extent of opinion of a nation’s citizens eligible in the selection and operation of government, including free media, freedom of association, and freedom of expression.	[28]
Governance effectiveness	GOVE	GOVE captures the perception and the extent of pursuing self-reliance devoid of political tensions, the civil service quality and effective administration policies, and credible implementation of structured governance policies.	[28]
Political stability	POLS	POLS assesses people’s perception and plausibility of political influence, terrorism, and political instability.	[28]
Regulatory quality	REQ	REQ involves the government’s perception and capability to furnish and administer prudent regulations that allow and motivate the development of institutions and agencies.	[28]
Control of corruption	CC	CC entails the opinion and degree to which the state exercises power for personal benefits, which could be a great or petty type of corruption.	[28]
The rule of law	RULE	RULE reveals the perception and the degree to which individuals respond to the state’s laws and the confidence people have in court systems and the police, the protection of properties, and the enforcement of contracts to eliminate the tendency of crime and violence.	[28]

. The sample was constricted to the period when yearly data were accessible from 2000 through 2020, i.e., 20 years of observations for each country. All the datasets for the period were gathered and extracted via the world development indicators (WDI) data source from the World Bank. In this research, economic growth (ECG) was measured by employing GDP per capita in (constant 2015 USD) because it encompasses the broad developmental economic health, indicating the standard of the production process and aggregated energy consumption. Furthermore, governance was proxied by the voice of accountability (VA), regulatory quality (REQ), governance effectiveness (GOVE), political stability (POLS), the rule of law (RULE), and control of corruption (CC). These proxied variables for governance were measured on the basis of the annual percentile rank of each country [86]. The authors filled the missing values in the governance data using a linear extrapolation approach [86,91]. This study’s data sources, variables, symbols, and standard units of measurement are revealed in Table 1.

3.2. Description of Study Area

This study conducts a rigorous analysis concerning the connection among ECG, TOUR, CO₂ emissions, and FDI in 27 African countries. The selection is based on the United Nations Geoscheme’s five major classifications [92] of African countries, namely, northern, western, central, eastern, and southern African countries, as shown in Figure 1. African countries differ greatly economically, culturally, socially, and geographically. These differences help determine the core functions of governance CO₂ emission control due to the zeal to attract more FDI, ECG, and TOUR expansion in the selected regions. The specific countries chosen for this study include Angola, Botswana, Mauritania, Egypt, Democratic Republic of Congo, Ethiopia, The Gambia, Gabon, Kenya, Ghana, Malawi, Mauritius, Mali, Morocco, Mozambique, Cameroon, Namibia, Nigeria, Senegal, Rwanda, Sierra Leone, South Africa, Seychelles, Tanzania, Togo, Zimbabwe, and Tunisia. The selection criteria for the African countries are crucial because most of these countries receive a high inflow of FDI and TOUR and are highly pursuing ECG activities. Notwithstanding, the enhancement of FDI, TOUR, and ECG has its accompanying negative effect from CO₂ emissions, which necessitates strong governance policies to implement effective decarbonization policies to attain environmental quality and sustainability.

3.3. Model Construction

Theoretical Underpinning

This present study is theoretically and empirically premised on the EKC theory. The EKC theory emphasizes that the heightening of economies increases environmental pollution. However, environmental pollution declines when economies reach a specific threshold. Therefore, an inverted U-shaped connection is formed between economic expansion and environmental pollution [93]. EKC is an established theory extensively employed for empirical modeling analysis by most studies to examine or verify why nations experiencing expansion in income achieve a decline or increase in CO₂ emissions. The continual adoption of EKC in many studies emphasizes its merit in making governmental environmental policies. The EKC theory depicts that the constant flow of income contributes significantly to environmental pollution due to the corresponding growth in the consumption of goods and energies. In contrast, income expansion attaining a threshold or standard contributes to decreasing CO₂ emissions as the citizens become aware of the environmental pollution issues, culminating in implementing effective environmental regulations. Although many scholars have widely tested EKC, different and ambiguous findings have mostly been recorded. For instance, some scholars affirmed the EKC theory [94,95], while others stated nonsupport in their findings [96,97]. This study posits that governance plays an essential role in environmental quality. However, behaviors such as rampant corruption in governance systems are not conducive to forming the EKC inflections [34,98].

A study on EKC prediction indicated an affirmation by taking into account the characteristics of governance such as accountability, quality of policies, political freedom, civil rights, control of corruption, and the rule of law [99]. On the basis of the above explorations, it is apparent that the role of governance in emission control is critical. However, the scanty research presents unclear findings on the policies for mitigating and implementing decarbonization measures. Therefore, this investigation aimed to conduct further in-depth analysis to obtain dynamic results in 27 African nations. Thus, this present research sought to examine the influence of governance on regulating carbon dioxide emissions by testing the EKC hypothesis through its connection with ECG, FDI, and TOUR. This research explores the role of ECG, TOUR, and FDI governance in mitigating CO₂ emissions using the panel dataset of selected African countries. Westerlund’s panel cointegration approach, panel standard error regression of Driscoll–Kraay, the dynamic ordinary least squares (DOLS), the pooled mean group (PMG), and heterogeneous panel causality testing were adopted to investigate the longitudinal and short-term elasticities of the connections in this research. Therefore, the models below were structured to assess these variables and their assumptions. The general specification model to attain the objective of this current study is demonstrated in Equation (1).

$$C{O}_{2it}={\alpha }_{0i}+{\beta }_{1i}TOU{R}_{it}+{\beta }_{2i}FD{I}_{it}+{\beta }_{3i}EC{G}_{it}+{\beta }_{5i}GO{V}_{it}+{\epsilon }_{it}.$$

(1)

The empirical models of the log-linear specification are expected to generate precise flexibility evaluations because they operate as coefficients of the research-controlled variables. The linear logarithm specification of the effects of TOUR, FDI, GOV, and ECG on CO₂ emissions and the other representations is given in Equation (2). The data were transformed into natural logarithm terms for better distribution and sharpness among the selected variables. The data transformation into a natural logarithm helps solve issues relating to heteroskedasticity and autocorrelation [100].

$$lnC{O}_{2it}={\beta }_{0it}+{\beta }_{1i}lnTOU{R}_{it}+{\beta }_{2i}lnFD{I}_{it}+{\beta }_{3i}lnEC{G}_{it}+{\beta }_{4i}lnGO{V}_{it}+{\epsilon }_{it}.$$

(2)

In Equation (2), $ln$ represents the logarithm form based on the selected variables, and $\beta$ indicates the figures or values of the coefficients, while $\epsilon$ denotes the error term deemed normally distributed.

Furthermore, Equation (3) establishes the classical reduced function of EKC with governance.

$$lnC{O}_{2it}={\beta }_{0it}+{\beta }_{1i}lnEC{G}_{it}+{\beta }_{2i}lnEC{G}^2{}_{it}+{\beta }_{3i}lnFD{I}_{it}+{\beta }_{4i}lnTOU{R}_{it}+{\beta }_{5i}lnGO{V}_{it}+{\epsilon }_{it},$$

(3)

where CO₂ denotes per capita CO₂ emissions, ECG represents economic growth measured by GDP per capita, $EC{G}^2$ represents GDP per capita squared, tourism (TOUR) is measured by international tourism receipts, FDI is measured by inward flows (FDI per capita USD), and GOV denotes the proxied governance variables. Additionally, $i$ and $t$ represent the number of nations and the respective time. The growth of GDP per capita of nations is perceived to contribute to higher CO₂ emissions and is connected with the EKC hypothesis. ECG has a nonlinear connection with CO₂ emissions [100,101]. Therefore, this study indicates that, if ${\beta }_1$ < 0 and ${\beta }_2$ < 0, there is an affirmation of the EKC assumption. Moreover, in capturing the nonlinear connection, $EC{G}^2$ was incorporated in the model estimation.

3.4. Econometric Methodologies

3.4.1. Unit Root Testing

Unit root testing is a fundamental measurement approach widely used to verify the stationarity characteristics of econometric series data. A series of unit root testing tools have been applied in countless studies by scholars [102,103,104,105]. However, some scholars have indicated the statistical limitations in applying unit root testing methods to panel data as a function of the sample size and the testing power [106,107,108]. Deciding on the level of stationarity of a series is vital in empirical estimations involving economic data, predominantly when the data comprise several economic variables [109,110]. Although extant studies have proposed different panel unit root testing approaches, determining the integration standard is vital in ascertaining effective parameters. This research selected 27 African nations with diverse economic growth and emission levels and applied panel unit root testing to determine biased and inefficacious estimations. In eliminating the plausibility of inefficiency and improving the robustness of the present research findings, the authors adopted the panel estimation approach that yields accurate robustness. We adopt the cross-sectional augmented Im–Pesaran–Shin (CIPS) and cross-sectional augmented Dickey–Fuller (CADF) [111]. The CIPS represents the augmented form of IPS which estimates by averaging the individual CADF statistical testing for every panel [112]. These two-panel unit root testing approaches effectively estimate heterogeneity and cross-sectional dependence (CSD). CIPS estimations are presented below:

$$CIPS \left(N,T\right)=t-bar=N^{-1}{{\displaystyle \sum }}_{i=1}^N{t}_i\left(N,T\right),$$

(4)

where $t_i\left(N,T\right)$ denotes the $i_{}$-th cross-sectional CADF statistical unit, and $t-bar$ denotes the average of $CAD{F}_i$ according to [111].

3.4.2. Westerlund Panel Cointegration Testing

After stationarity checking, the variables’ cointegration was estimated. Considering the heterogeneity and cross-sectional dependence obstacles, the authors applied the panel cointegration strategy that furnishes efficient and reliable estimations. The authors deemed applying the panel cointegration approach plausible and essential in short-term cross-sectional analysis. Founded on previous studies [61,75,86], this research adopts panel cointegration testing [113], which depends on cross-sectional dependence assumptions. Thus, Westerlund’s panel cointegration testing supports cross-sectional dependence and slope parameter analysis, which are critical problems with panel data [96,114]. The methodology consists of four test statistics: $G_t, G_a, P_t, and P_a$. $G_t$ and $G_a$ represent the group statistics; thus, they are not dependent on the grouped information arising from the techniques of error corrections. Notwithstanding, $P_t$ and $P_a$ represent the dependent elements in the error correction term in connection with the cross-section units. These four statistical tests possess standard characteristics and robustness in enhancing the residuals connected with the panel cointegration testing [115,116]. According to the Westerlund panel cointegration testing, the null assumption denoting the absence of a cointegration connection opposes the null theory and demonstrates the cointegration connection for the minimum unit of cross-section relating to the statistical group of CSD. It also supports testing of panel cointegration for the nations. The Westerlund panel cointegration estimation is put forward as follows:

$$\Delta z_{it}={\sigma }_i{d}_i+{\theta }_i(z_{i\left(t-1\right)}+{\pi }^{\prime }i{y}_i{}_{\left(t-1\right)}+{{\displaystyle \sum }}_{j=1}^m{\theta }_{ij}\Delta Z_{i\left(t-j\right)}+{{\displaystyle \sum }}_{j=1}^m{\phi }_{ij}\Delta y_{i\left(t-j\right)}+{\omega }_{it},$$

(5)

where, ${\theta }_i$ denotes the rate of adjustment through which the system makes corrections to revert and attain an equilibrium connection.

3.4.3. Panel Standard Error Regression of Driscoll–Kraay

This study is premised on the EKC assumption and investigates the role of governance in the mitigation of CO₂ emissions in connection with other variables contributing to environmental pollution. Adhering to the tenet of existing studies, this study applies the Driscoll–Kraay (DK) standard error panel data evaluating strategies for estimating coefficients via the fixed effects approach [117]. The Driscoll–Kraay yields consistent, robust, and effective standard error and CSD estimations [118]. The DK method is suitable for analyzing unbalanced and balanced panel data with missing values [119]. The DK standard error is also a critical approach to overriding serial independence and heteroskedasticity in the fixed effect approach [120]. Furthermore, the standard error of DK is classified as a nonparametric method that avoids restrictions and is flexible for greater time dimensional analysis. Premised on the above analysis, we estimated the pooled ordinary least squares using the linear regression approach of DK measured as

$$y_{i,t}={x^{\prime }}_{it-1}\beta +z_{it}y+{\epsilon }_{it}i=1,\dots ,N,t=1, \dots ,T.$$

(6)

Thus, $y_{i,t}$ represents the response variable (CO₂ emissions) with its corresponding scalar value, while $x_{i,t}$ denotes regressors TOUR, FDI, ECG, and GOV using a $\left(K+1\right)\times 1 vector$ representing the initial factor. Moreover, $\beta$ denotes the unknown coefficient $\left(K+1\right)\times 1 vector$, and $i$ indicates the respective units corresponding to time $t.$ To be succinct, the estimation of the standard error of DK is observed to be a ‘squared root’ $\widehat{S}T$ with regard to the asymptotic covariance and diagonal matrix estimation [117,121].

$$V\left(\widehat{\beta }\right)={(X^{\prime }X)}^{-1}{\widehat{S}}_T{(XX)}^{-1}.$$

(7)

3.4.4. Analysis of Pooled Mean Group (PMG) and the Dynamic Ordinary Least Squares (DOLS)

In conducting a robustness check in this present study, the authors adopted the renowned panel data approach denoted as pooled mean group (PMG) [122]. The PMG algorithm employs the maximum likelihood approach that handles the CSD, which depicts identical distribution characteristics across units with a zero mean and a constant variance. The maximum likelihood ratio regarding the PMG strategy is adopted to examine the error variances in a homogeneous group with short- or long-term coefficients that normally reject the equivalency of error variances at the traditional significance level [123]. With an emphasis on the homogeneous slope hypothesis, PMG is assumed to be a viable and efficient estimator for a long-term analysis. Furthermore, this study applied the DOLS panel data evaluation methodology for checking robustness. DOLS has the pliability to accommodate endogeneity problems found within autocorrelation and regressors via parametric methods, leads, and lags [124]. In addition, the DOLS approach is renowned for yielding reliable and effective estimation for small samples. In general, cross-sectional and heterogeneity issues are better countered when DOLS is applied [125].

3.4.5. Heterogeneous Panel Causality Testing

In confirming panel cointegration among the selected research variables, it is proposed that there should be the presence of bidirectional or single-sided causation between the underlined study variables. In enhancing the causality detection in this present study, the panel causation test [121,126] was followed and was effective for heterogeneous analysis. The panel causation testing approach is ideal for investigating causation in panel data because of its ability to estimate and address heterogeneous issues in panel data. An investigation by Dogan and Seker [127] emphasized that the panel causality methodology functions whether $T>N or T<N$ and is very effective in heterogenous and unbalanced panel analysis. One other merit of heterogeneous panel causation testing is its effectiveness in capturing CSD. Furthermore, the heterogenous panel causality testing yields better and a\more efficient estimation based on the assumption of homogeneity found using the standard Granger causality testing. The expression below indicates the estimation of heterogeneous panel causation testing.

$$y_{i,t}={\alpha }_i+{{\displaystyle \sum }}_{i=1}^K{\gamma }_i^{\left(k\right)}{y}_{i,t-k}+{{\displaystyle \sum }}_{i=1}^K{\gamma }_i^{\left(k\right)}{x}_{i,t-k}+{\omega }_{i,t},$$

(8)

where $K\epsilon N^{+} and K\epsilon N^{-} and {\beta }_i=\left({\beta }_i^{\left(1\right)}\dots .,{\beta }_i^{\left(k\right)}\right)$ and ${\alpha }_i$, ${\gamma }_i^{\left(k\right)}$, and ${\beta }_i^{\left(k\right)}$ demonstrate the constant terms, the coefficient slope, and lag parameters accordingly.

4. Empirical Results and Discussion

4.1. Descriptive Statistics

Table 2. Descriptive statistics for the nations and variables.

Descriptive Statistics	CO2	TOUR	ECG	FDI	GOVE	POL	REQ	RULE	CC	VA
Mean	35,585.34	1.94 × 109	2952.381	–9.53 × 108	35.93530	35.98319	35.43035	36.78418	36.02208	33.85435
Median	4490.000	4.17 × 108	1419.120	–2.87 × 108	32.82051	36.96682	35.29412	38.02817	32.38095	32.69231
Maximum	447,980.0	2.45 × 1010	15,913.95	8.75 × 109	85.85366	93.75000	84.13461	83.66337	86.05769	76.61691
Minimum	0.000000	0.000000	258.6288	–2.51 × 1010	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000
SD	83,562.42	3.93 × 109	3181.614	2.57 × 109	23.75323	23.80495	21.08252	22.26876	23.41770	20.55757
Skewness	3.251347	3.179809	1.643401	–2.919242	0.273366	0.308646	0.181659	–0.046096	0.195426	0.237545
Kurtosis	13.58561	13.99642	5.491075	25.86389	1.936119	2.299989	2.209516	1.956243	1.867471	1.985618
Jarque–Bera	3646.284	3812.274	401.8262	13155.47	33.80168	20.57891	17.88093	25.93853	33.91103	29.64187
Probability	0.000000	0.000000	0.000000	0.000000	0.000000	0.000034	0.000131	0.000002	0.000000	0.000000
Sum	20,176,890	1.10 × 1012	1674000.	–5.41 × 1011	20,375.32	20,402.47	20,089.01	20,856.63	20,424.52	19,195.42
Sum Sq. Dev.	3.95 × 1012	8.75 × 1021	5.73 × 109	3.73 × 1021	319,346.3	320,738.4	251,571.6	280,678.0	310,388.0	239,199.4
Observations	567	567	567	567	567	567	567	567	567	567

furnishes the general statistical descriptive profile information of the 27 African countries for the panel data from 2000 to 2020. The results reveal the statistical information, comprising the median, mean, minimum, maximum, skewness, standard deviation, probability, kurtosis, etc. The total number of observations in the series was 567. The mean statistical results for the series were as follows: CO₂ (35,585.34), ECG (2952.381), FDI (−9.53 × 10⁸), TOUR (1.94 × 10⁹), CC (36.02208), GOVE (35.93530), POL (35.98319), REQ (35.43035), RULE (36.78418), and VA (33.85435). Furthermore, the chosen variables investigated in this study demonstrated a relatively high standard deviation level. This research’s data composition of normality was evaluated by employing statistical results such as kurtosis, Jarque–Bera, and probability tests. Table 2 further demonstrates that all the probability values were statistically significant at the 1% level. Hence, it is prudent to reject the null hypothesis. The study results in Table 2 show that the data sample was not normally distributed, denoting variations within the data gathered. In checking for stationarity of the data, this study followed the systematic process of conducting empirical panel data estimation. Figure 2 demonstrates the magnitude of emissions in the respective countries. The research discovery of this present study demonstrates that South Africa, Egypt, Nigeria, Morocco, Tunisia, and Angola are the highest CO₂-emitting nations. The interesting outcome of this present research is in unison with the literature [54].

4.2. Cross-Sectional Dependency Test

Extant studies have indicated that failing to address cross-sectional dependency (CSD) testing issues may lead to discrepancies, lack of credibility, prejudices, and inefficiency, all of which may yield inaccuracies in analysis [128,129]. Therefore, this study employed four CSD testing approaches to test the dataset to address the inherent challenges of CSD in our study findings. The CSD summary is demonstrated in

Table 3. Summary for CSD test result.

Series	Breusch–Pagan LM	Pesaran Scaled LM	Bias-Corrected Scaled LM	Pesaran CD	p-Value
LNCO2	5362.799	189.1581	188.4831	72.23309	0.0000
LNECG	4824.993	168.8600	168.1850	52.45064	0.0000
LNFDI	1060.630	26.78327	26.10827	13.32846	0.0000
LNTOUR	2494.332	80.89487	80.21987	38.79201	0.0000
LNCC	2729.790	89.78164	89.10664	48.10628	0.0000
LNGOVE	2856.737	94.57295	93.89795	47.09715	0.0000
LNPOLS	2269.158	72.39621	71.72121	38.49664	0.0000
LNREQ	3167.025	106.2840	105.6090	51.65035	0.0000
LNRULE	3229.730	108.6507	107.9757	50.51663	0.0000
LNVA	3263.186	109.9134	109.2384	53.18886	0.0000

. Table 3 depicts that all four CSD testing approaches were significant at the 1% level, confirming the rejection of the null hypothesis regarding CSD existence among the cross-section of the variables. On the basis of these fascinating findings, the authors conclude that the selected African countries are linked through economic growth, FDI inflows, technological advancement, tourism development, and governance.

4.3. Panel Unit Root Test for CADF and CIPS

Furthermore, CADF and CIPS unit root testing were adopted because of their effectiveness in stationary checking, as posited in a series of investigations [111,112,130].

Table 4. The panel unit testing result for CADF and CIPS.

	CADF		CIPS		I(1)
Series	Level	First Difference	Level	First Difference	I(1)
LnCO2	−0.50373	−4.56455 ***	46.1395	130.940 ***	I(1)
lnECG	1.05229	−9.80334 ***	42.7889	205.104 ***	I(1)
lnFDI	−2.82903 *	−21.8558 ***	80.9592 *	458.900 ***	I(1)
lnTOUR	−2.88814 *	−9.62010 ***	80.4268 *	205.318 ***	I(1)
lnCC	−22.7518 **	−23.6729 ***	820.661 ***	499.781 ***	I(1)
lnGOVE	−27.5620 *	−24.5640 ***	1575.45 *	570.136 ***	I(1)
lnPOLS	−17.9853 *	−22.1142 ***	597.946 *	445.108 ***	I(1)
lnREQ	−21.3757 *	−28.6887 ***	679.543 *	713.316 ***	I(1)
lnRULE	−30.7023 *	−26.0936 ***	1170.88 *	570.277 ***	I(1)
lnVA	−31.7819 *	−23.5597 ***	1119.46 *	498.785 ***	I(1)

Note: *, **, *** represent 1%, 5% and 10% significance level accordingly.

demonstrates the CADF and CIPS, revealing that it is not feasible to disprove the null hypothesis; however, none of the series $\left(lnC{O}_2,lnECG,lnFDI,lnTOUR, and lnGOV\right)$ were found to be stationed at the first difference level. Given this, after the first differentiation, the remainder of the series became stationary I(1). Most studies have demonstrated that, after integration at order 1, it is relevant to assess cointegration among the variables selected [86,131,132].

4.4. Panel Cointegration Test

In achieving this crucial objective, this research employed the Westerlund Panel cointegration testing approach that counters cross-sectional dependence and heterogeneity among highly heterogeneous datasets over diverse nations, as indicated in

Table 5. Westerlund error correction model and panel cointegration test.

Statistics	Value	z-Value	p-Value
$G_t$	−1.5352 *	−0.3473	0.0033
$G_a$	−2.8663 *	1.1632	0.0093
$P_t$	−4.0633 ***	−1.4783	0.0573
$P_a$	−2.7470	−0.1733	0.3463

Note: * (1%) and *** (10%) represent the significance level.

. The two categories of cointegration approaches of Westerlund (2007) and the corresponding two-panel statistics with their respective probabilities are shown in Table 5. The results demonstrate that both categories (G_t and G_a) were statistically significant at the 1% level. Notwithstanding, the statistics of Pt revealed that a single panel value was highly significant at 10%. These exciting results demonstrate longitudinal cointegration amongst the chosen variables employed for analysis.

4.5. Panel Long-Term Elasticity Estimation

In switching to the core estimation of this study,

Table 6. Driscoll and Kraay standard error PMG and DMOL panel regression estimation results.

Dependent	Variable: Carbon Dioxide Emission
	Regression of DK Model		DOLS	Estimator	PMG	Estimator
Parameters	Coefficients	p-Value	Coefficients	p-Value	Coefficients	p-Value	VIF
lnECG	1.104483 **	0.0084	2.193806 ***	0.0353	2.193806 ***	0.0353	3.24
lnECG2	−5.33 × 10−5 ***	0.0800	−0.11260 **	0.0033	−0.1482 **	0.0012	2.42
lnFDI	4.46 × 10−7	0.5177	2.42 × 10−6 **	0.0112	2.42 × 10−6 ***	0.0112	4.54
lnTOUR	8.24 × 10−6 *	0.0000	1.53 × 10−5 *	0.0000	1.53 × 10−5 *	0.0000	1.83
lnCC	267.1152	0.3770	550.7260 ***	0.0683	550.7260	0.0683	2.88
lnGOVE	−658.2410 ***	0.0486	63.98539	0.8605	63.98539	0.8605	4.89
lnPOLS	22.29708	0.9051	−647.3316 **	0.0010	−647.3316 **	0.0010	1.79
lnREQ	405.3785	0.1612	365.8648	0.2521	365.8648	0.2521	4.44
lnRULE	−45.46990	0.9083	−1775.434 *	0.0000	−1775.434 *	0.0000	2.29
lnVA	13.30481	0.9566	1445.129 *	0.0000	1445.129 *	0.0000	4.66
Turning point	5617.28		7585.34		11,585.43

Note: * (10%), ** (5%), and *** (1%) represents rejection of the null hypothesis.

demonstrates the Driscoll–Kraay standardized panel error regression approach employed in assessing the parameters.

This study’s findings indicate that economic growth (lnECG) measured with respect to GDP per capita significantly and positively affects CO₂ emissions. Contrastingly, the GDP squared (lnECG2) for all models was statistically significant but revealed a negative impact. Hence, the current research found an inverted U-shaped connection linking income and environmental pollution, giving strong credence to the conventional environmental Kuznets Curve (EKC) assumption for the 27 selected African nations. The EKC evidence revealed that antecedent policies have opted to depend on ECG to decarbonize CO₂ emissions from tourism activities and FDI, as ecological factors may be curbed through economic development.

Studies have demonstrated the connection between pollution and income with fundamental economic variations and technological advancements. The findings of this study revealed that, as income increases, it makes the population aware of their environment; therefore, governance regulations on the environment, policies, and laws can be implemented to curb pollution within the environment [133,134]. This study revealed that the 27 selected African countries continue to experience structural transformation and speedily increase industrialization to attain economic growth. Additionally, this research assessed the turning point for the value of income after which environmental pollution begins to decline. A critical observation in the bottom section of Table 6 denotes that the turning point value falls between 5617.28 and 11,585.43 USD. On the basis of this phenomenon, it is evident that most of the 27 countries surpassed the turning point standard, and most nations are currently proceeding toward decarbonization policies based on their increase in income. In contrast, countries that are unable to reach the threshold must concentrate on the intensifying decarbonization process of CO₂ via an effective way of increasing their income. This study found that the following countries are at their peak turning point: Seychelles, Mauritius, Botswana, South Africa, Gabon, and Namibia. In contrast, some countries that have not reached the turning point are Nigeria, Egypt, Ghana, Tanzania, Mozambique, Rwanda, Sierra Leone, Congo, Kenya, Gambia, Mali, Togo, Ethiopia, and others. Analysis of Table 6 further reveals that the governance elements comprising CC (control of corruption), GOVE (governance effectiveness), POLS (political stability), REQ (regulatory quality), VA (voice of accountability), and RULE (the rule of law) may invariably yield positive and negative repercussions for decarbonizing the variables from pollution. This objective would be feasible using effective policy implementations.

This study confirmed a statistically significant and positive connection between TOUR and CO₂ emissions. Notwithstanding, an increase in TOUR activities may invariably increase the nation’s income, reducing pollution and increasing environmental quality [61,135]. This study demonstrated a statistically significant and positive connection linking FDI inflows and CO₂ emissions. Notwithstanding, this study’s findings revealed that governance policies may help regulate FDI to reduce CO₂ emissions [6,136]. Policy implementers in the selected African nations can ensure eco-friendly and advanced technology systems in captivating FDI inflows to safeguard the environment from CO₂ emissions.

Furthermore, this study’s discoveries revealed an affirmative connection for some GOV indicators, including political stability, regulatory quality, control of corruption, and voice of accountability, demonstrating that ineffective implementation of these governance systems may influence CO₂ emissions. Notwithstanding, a negative association was revealed between governance effectiveness and the rule of law. This suggests that governance effectiveness and the standardized rule of law are crucial for implementing policies that reduce CO₂ emissions [86,99,137]. Thus, the governance elements may significantly impact controlling ECG, TOUR, and FDI inflow from carbon dioxide emissions. Hence, it can be deduced that government disposition toward formulating and implementing effective regulations and policies is potent for curbing environmental pollution. Thus, instilling the rule of law and controlling corruption can significantly decrease resource wastage. Corruption is perceived to play an essential function through the link between carbon dioxide emissions and income. Notwithstanding, the connection diminishes when income reveals a statistically positive effect on carbon dioxide emissions [34,138]. A developed economic order can govern the unlawful activities that create environmental pollution. Regulatory quality may be enhanced through transparent policy standards directly accessible to individuals. Governments that consider environmental regulations as a key factor are capable of implementing environmental pollution policies. As nations experience economic growth, most governments implement proper regulations to balance any economic failure from the market that contributes to pollution growth and heeds guidelines enhancing the environmental degradation awareness of the public.

The government bodies can assist nations in implementing prudent environmental policies that reduce CO₂ emissions. Hence, effective decision making in managing and controlling resources is efficient in these nations. These fascinating results are in harmony with [139,140,141]. Moreover, the findings in this research are reliable and more robust because of the application of effective evaluation techniques that capture CSD and heterogeneity [142]. Next, this study applied the VIF (variance inflation factor) estimation technique to determine the presence of multicollinearity issues. Thus, Table 6 demonstrates that the VIF values were all below 10, indicating no multicollinearity.

In connection with the afore statistical findings, this study conducted static analysis using the Hausman fixed and random effect test to validate the results obtained for the dynamic panel approaches employed in this study.

Table 7. Fixed effect results.

Variable	Coefficient	Std. Error	t-Statistic	Prob.
C	31,954.80	11,536.73	2.769831	0.0058
LNECG	−3.703451	5.857835	−0.632222	0.5275
GDP2	−2.90 × 10−5	0.000290	−0.100100	0.9203
LNFDI	4.76 × 10−7	5.96 × 10−7	0.798624	0.4249
LNTOUR	7.53 × 10−6	8.82 × 10−7	8.536953	0.0000
LNCC	204.3822	251.6418	0.812195	0.4170
LNGOVE	−342.4472	278.3194	−1.230411	0.2191
LNPOLS	32.01174	149.4174	0.214244	0.8304
LNREQ	292.4937	241.5353	1.210977	0.2264
LNRULE	−158.2416	326.1821	−0.485133	0.6278
LNVA	6.479820	204.3414	0.031711	0.9747

and

Table 8. Random effect results.

Variable	Coefficient	Std. Error	t-Statistic	Prob.
C	−9398.541	5282.470	−1.779194	0.0758
LNECG	16.59335	2.362611	7.023311	0.0000
GDP2	−0.001188	0.000180	−6.595535	0.0000
LNFDI	−2.16 × 10−6	9.02 × 10−7	−2.396373	0.0169
LNTOUR	1.46 × 10−5	8.05 × 10−7	18.16107	0.0000
LNCC	1003.023	291.5521	3.440289	0.0006
LNGOVE	−132.4509	349.0949	−0.379412	0.7045
LNPOLS	−754.2005	187.0877	−4.031267	0.0001
LNREQ	20.36407	307.2546	0.066278	0.9472
LNRULE	−1822.108	353.1146	−5.160103	0.0000
LNVA	1477.636	187.1454	7.895659	0.0000

demonstrate the fixed and random effect results for the period, respectively, while

Table 9. Correlated random effects—Hausman test results.

Test Summary		Chi-Sq. Statistic	Chi-Sq. d.f.	Prob.
Period random		19.381319	10	0.9357
Period random effects test comparisons:
Variable	Fixed	Random	Var (Diff.)	Prob.
LNECG	17.184592	16.593355	0.203250	0.1897
GDP2	−0.001178	−0.001188	0.000000	0.6459
LNFDI	−0.000002	−0.000002	0.000000	0.0882
LNTOUR	0.000015	0.000015	0.000000	0.0671
LNCC	1199.025708	1003.023210	2462.572353	0.0001
LNGOVE	−413.620422	−132.450857	7812.034366	0.0015
LNPOLS	−872.318996	−754.200463	2245.026263	0.0127
LNREQ	16.500125	20.364073	1921.246217	0.9298
LNRULE	−1752.006568	−1822.107657	1617.448079	0.0813
LNVA	1539.169737	1477.636394	737.075598	0.0234

shows the correlated effects of the Hausman testing approach employed for comparing the fixed and random effects approaches of testing.

According to the coefficients of the fixed effect model in Table 7, it can be observed that only one variable, TOUR, was statistically significant compared with the other variables. Therefore, we employed the random effect model for further analytical checks, as shown in Table 8. The random effect test results demonstrated that most variables were statistically significant compared to the fixed effect. Thus, in considering the Hausman test assumption, the probability of the variables was used to justify the findings of dynamic panel approaches, including DOLS and PMG. This study compared the random and fixed effect results based on the correlated random effects and Hausman test results, as shown in Table 9; it can be observed that the p-value of the period was 0.9357, suggesting that the null hypothesis could not be rejected. Therefore, the assumption of the Hausman test random effect for the variables was validated, giving credence to the study results. Moreover, it can be observed from the coefficients of the random effects variables results in Table 9 that most of the variables were significant compared to fixed-effect results. On the basis of these exciting results, the authors of this study suggest that it is appropriate to conduct static panel analysis such as the Hausman fixed and random testing to validate the results of dynamic panel approaches.

4.6. Dumitrescu and Hurlin Causality Test

Furthermore, the research demonstrated that adopting the approach of long-term panel data evaluation only reflects the influence of dependent elements on the independent factors. Therefore, in outlining recommendable policies to decision makers, this study posits that it is prudent to determine the connection trajectory linking the variables of this research. In achieving the core purpose of the study, the authors employed the DH (Dumitrescu–Hurlin) pairwise testing for panel causation to identify the connection of causality within the variables selected. The exciting results of the DH panel causation are shown in

Table 10. Pairwise Dumitrescu and Hurlin panel causality testing results.

Null Hypothesis:	W-Stat.	Zbar-Stat.	p-Value	Conclusion
LNCO2 ⇎ lnECG2	59.6386 *	107.854	0.0000	LNCO2 ↔ lnECG2
LNGDP2 ⇎ LNCO2	5.44630	5.85890	5 × 10−9
LNCO2 ⇎ LNECG	40.6965 *	72.2030	0.0000	LNCO2 ↔ LNECG
LNECG ⇎ LNCO2	4.30630	3.71330	0.0002
LNCO2 ⇎ LNFDI	6.12931 **	7.14439	9 × 10−13	LNCO2 ↔ LNFDI
LNFDI ⇎ LNCO2	3.72624	2.62157	0.0088
LNCO2 ⇎ LNTOUR	37.2118 *	65.6445	0.0000	LNCO2 ↔ LNTOUR
LNTOUR ⇎ LNCO2	4.34756	3.79097	0.0002
LNCO2 ⇎ LNCC	6.90273 ***	8.60004	0.0000	LNCO2 ↔ LNCC
LNCC ⇎ LNCO2	1.16142	−2.20564	0.0274
LNCO2 ⇎ LNGOVE	20.1821 **	33.5930	0.0000	LNCO2 ↔ LNGOVE
LNGOVE ⇎ LNCO2	0.78929	−2.90603	0.0037
LNCO2 ⇎ LNPOLS	6.18530 ***	7.24975	4 × 10−13	LNCO2 ↔ LNPOLS
LNPOLS ⇎ LNCO2	1.06192	−2.39291	0.0167
LNCO2 ⇎ LNREQ	18.3444 *	30.1343	0.0000	LNCO2 → LNREQ
LNREQ ⇎ LNCO2	1.32711	−1.89380	0.4583
LNCO2 ⇎ LNRULE	11.9236 *	18.0498	0.0000	LNCO2 ↔ LNRULE
LNRULE ⇎ LNCO2	0.92299	−2.65439	0.0079
LNCO2 ⇎ LNVA	9.95596 *	14.3465	0.0000	LNCO2 ↔ LNVA

Note: ***, **, and * demonstrate the significance levels of 10%, 5%, and 1%, respectively. The symbols ⇎, ↔, and → indicate no causality, bidirectional causality, and unidirectional causality, respectively.

. The findings revealed bidirectional causation linking carbon dioxide emissions and ECG², thus highlighting the link between ECG and CO₂ emissions emerging from the growth activities. Thus, the pursuit of economic growth and income expansion invariably causes CO₂ emissions and may be reduced as income increases. Bidirectional causation links exist between ECG and CO₂, FDI and CO₂, and TOUR and CO₂, suggesting that TOUR, ECG, and FDI activities contribute to pollution. The findings also revealed that ECG has a causal link with the governance indicators. Bidirectional causation exists between the indicators of GOV (CC, GOVE, POLS, VA, RULE, and REQ) and CO₂ emissions. These fascinating results indicated that GOV policies in the selected African nations can be considered helpful in CO₂ emission control as income increases. This study’s exciting results validate the hypothesis that effective governance may play a significant role in regulating CO₂ emissions in the selected African countries.

Notwithstanding, unidirectional causality was found between LNECG² and LNCC, LNECG² and LNPOLS, LNECG² and LNREQ, LNECG² and LNRULE, LNECG² and LNTOUR, LNECG and LNCC, LNFDI and LNECG, LNECG² and LNVA, and LNTOUR and LNECG. These new discoveries in this study demonstrate that policies aimed at CO₂, ECG, FDI, TOUR, and GOV will have a cyclical trajectory impact because pf a bidirectional Granger causality effect. A noteworthy implication is that any comprehensive variations in ECG, TOUR, and FDI may increase CO₂ emissions. Bidirectional causation among TOUR, FDI, and CO₂ denotes that any critical policies implemented in the flow of FDI and regulation of tourism activities will reduce carbon dioxide emissions within the nations selected for this study. Nevertheless, due to the unidirectional causation, any strategic policy measures aimed at these factors will influence the policies for environmental protection in the selected African nations.

5. Summary and Conclusions

The current research dissected the repercussions of tourism (TOUR) activities, FDI inflows, and ECG (economic growth) on carbon dioxide emissions. This crucial investigation aimed to determine whether governance matters in controlling CO₂ emissions in Africa while creating a collaboration for FDI inflows, tourism expansion, and economic growth. In this investigation, the authors employed a panel data approach for 27 selected African nations spanning 2000 through 2020. This study employed econometrics methodologies when testing the stationarity and other relevant findings that were based on the gathered data. Panel unit testing approaches, including CADF and CIPS, were primarily adopted to examine the dynamic levels of stationarity. Furthermore, parametric estimation approaches comprising DOLS, PMG, and Driscoll and Kraay standard error estimators for panel data were used to estimate the variables’ long-term associations. This study adopted heterogeneous testing for the panel causation approach in performing causality analysis. The research findings reflect the merits of the governance indicators comprising governance effectiveness, political stability, control of corruption, the voice of accountability, the rule of law, and regulatory quality within the selected African nations, revealing a statistically significant influence on carbon dioxide emission control.

Furthermore, various government policies ensuring income expansion gave credence to the theory of EKC within the selected nations for this study. The proposed theoretical assumptions projected on the selected variables and regions in this study were validated. The robustness check conducted validated the study findings. An exciting contribution made in this study to justify the dynamic panel models used was performed by adopting the Hausman fixed and random effects. Furthermore, this study proved that governance matters in controlling and reducing CO₂ emissions within the countries selected. Eventually, the results revealed bidirectional causation between ECG and governance indicators and ECG and CO₂ emissions. The affirmative connection linking governance and carbon dioxide emissions demonstrated that effective governance performance may curb increasing patterns of corruption and unregulated rule of law, while improved governance effectiveness may propel regulations of pollution policies in the environment. It will also reduce enormous industrial environmental pollution pressure [142]. This study’s findings showed that governance plays an essential role and may adequately and efficiently address pollution problems. Thus, governance policy expansion and implementation in the selected African nations can reduce the CO₂ emissions experienced in Africa due to tourism development, FDI, and the pursuit of economic growth. This suggests that effective governance creates access to political information and freedom that may stimulate citizens’ preference for a quality environment and consciousness while promoting environmental regulations and legislation issues. Correspondingly, the citizens’ tendency for a clean environment contributes to the enforcement of better environmental regulations, thus helping to reduce emissions and declining the risk of environmental hazards to human health conditions.

5.1. Policy Recommendation

In recent decades, a faster economic growth pace has been observed in most African countries, especially within the 27 African nations selected for this analysis. Thus, the pursuance of economic expansion has emerged with corresponding environmental degradation through tourism activities, FDI, and other industrial activities that merit effective governance interventions. In addressing these mounting environmental issues, this study suggests that the selected African governments should endorse environmental regulations and laws to avert the menace by improving the governance and decarbonization systems. This study projects that a significant and persistent connection between political stability and CO₂ emission decline will permit decision-makers to structure institutional policies that would eventually contribute to decarbonizing CO₂ emissions in the pursuance of economic growth and tourism activities. Furthermore, the policymakers in nations with high CO₂ emissions must enhance the effectiveness of governance because better governance is anticipated to improve the lives of people and the environment through the stimulation of political privileges and independent flow of information. This process invariably increases citizens’ awareness and attracts their support for the effective implementation of environmental regulations. Thus, increasing citizens’ self-awareness of their environment boosts environmental quality. Policymakers must comprehensively design pathways to heighten governance and institutional quality in emission control. Hence, this study proposes that making provisions for effective governance is motivated by the demand of indigenes and procedures for efficiently sustaining industries and resources, in line with the conditions and actions outlined by UNFCCC to help achieve decarbonization goals in the African continent. This study recommends strengthening intergovernmental cooperation at the regional and subregional levels to enhance governance policies related to decarbonization at the national level. Moreover, the government, nongovernmental organizations, and other stakeholders may collaborate to implement pollution reduction and different environmental pressure reduction strategies. Governments will almost certainly need to devise novel systems and technology that support the decoupling of tourism, economic prosperity, and FDI from carbon dioxide emissions. Lastly, decarbonization can be assured when there is a reduction in the over-reliance on fossil fuels by diversifying the country’s energy sources [143,144].

5.2. Future Directions and Limitations

This current study had inherent limitations anticipated to be examined in future studies. Future studies should emphasize the nation’s nonlinear impact and suggest diverse policies regarding environmental pollution standards because of the nonlinear connection between governance and CO₂ emissions. The authors believe that following this trajectory will facilitate comprehensive policies and eliminate cumbersome policy measures for decision-makers. This research creates a vacuum for economics and environmental science investigators to empirically analyze time-series and panel data concerning governance, as well as decarbonize emissions from economic activities and institutional elements.