Shadow Economy and Environmental Sustainability in Global Developing Countries: Do Governance Indicators Play a Role?

Abstract

Environmental sustainability has been a challenging issue all over the globe, with air pollution posing a significant threat. One main factor contributing to air pollution is the growth of the shadow economies. This study investigates the effect of the shadow economy on air pollution and explores how these effects depend on the levels of governance indicators. We utilize key air pollution indicators: carbon dioxide (CO₂) and nitrous oxide (N₂O) emissions. Furthermore, we examine the role of key governance indicators: corruption control, the rule of law, and regulatory quality. The study utilizes an annual panel dataset of 107 selected developing countries worldwide, spanning from 2002 to 2020, and employs the System GMM technique, which effectively tackles the omitted variable bias, potential endogeneity, and simultaneity issues in the model. The estimation results indicate that a sizeable shadow economy significantly increases the levels of CO₂ and N₂O emissions. Moreover, the results reveal that robust governance frameworks, evidenced by enhanced corruption control, a stronger rule of law, and superior regularity quality, mitigate the adverse effects of the shadow economy on CO₂ and N₂O emissions. This highlights a significant substitutability between the shadow economy and governance indicators, indicating that improvements in governance formworks will not only reduce the size of the shadow economy but also weaken its harmful impact on air pollution. Policy initiatives should thus focus on strengthening governance mechanisms, particularly enhancing control of corruption and the rule of law to effectively reduce the environmental impact of the shadow economies in developing countries. Additionally, governments should prioritize reforms in regulations and legal frameworks to limit the expansion of the shadow economy, thereby decreasing their contribution to air pollution.

1. Introduction

Environmental sustainability is important for guaranteeing the long-run health and well-being of human societies and ecosystems by promoting the responsible use of natural resources. It highlights sustaining ecological balance, decreasing environmental degradation, and protecting the planet for future generations [1]. One main factor of environmental sustainability is the execution of air quality, as air pollution (AIRP) indicates a significant threat to sustainable environmental goals. AIRP is caused by fossil fuel consumption, industrial emissions, and deforestation, which harms human health, contributes to climate change, and reduces biodiversity [2]. The deteriorated air quality disrupts the balance of nature by facilitating matters such as ozone layer depletion, acid rain, and global warming that directly appear to undermine the efforts at environmental sustainability [3]. Thus, addressing air quality is essential for fostering environmental resilience and protecting the planet’s natural systems. According to the United Nations Environment Programme report, AIRP is seen to affect the health of the entire world population, accounting for more than 8 million premature deaths annually. AIRP and climate change are closely interrelated because all major pollutants have an impact on the climate, while most share common sources with greenhouse gases (https://www.unep.org/interactives/air-pollution-note/ (accessed on 3 May 2024)).

Environmental pollution in the developing world is the consequence of poor environmental measures, which lead to the transfer of dirty industries there, and these countries usually underrate the environmental aspects in order to attract multinational corporations, which causes a much larger pollutant atmosphere [4]. Many countries are taking important environmental protection steps that have been effective for developed countries, but the presence of a polluting, unregulated shadow economy (SECO) is a major issue for developing countries that regulate environmental principles. Therefore, it can be claimed that the main reason behind air pollution in the developing world is the large number of shadow economic activities [5]. Many developing countries face significant challenges related to greenhouse gas (GHG) emissions. The primary contributors to these emissions are CO₂ and N₂O, which are typically generated from agro-food processing and agro-industrial activities that largely operate within the shadow economic sector [6,7]. The SECO frequently encompasses small-scale manufacturing or industrial activities such as small workshops, backyard factories, and artisanal production units; these operations may use outdated technologies, inefficient equipment, or low-quality fuels, all of which can generate higher levels of air pollutants [8].

In the developing world, a major part of the shadow sector is engaged in resource extraction, manufacturing, bleaching and dyeing of clothes, servicing, and retailing activities such as craft mining, car repair, leather tanning, brick making, and metalworking. Most of these practices have significant adverse effects on the AIRP [9]. Furthermore, brick kilns in a traditional way are mostly done in the SECO, which is a main cause of AIRP in developing countries. These brick kilns are burned with a number of cheap and highly environmentally harmful fuels, such as used motor oil, tires, wood scrap, manure, feces, and plastic waste [10]. Moreover, the informal sector units usually manufacture intermediary products for formal sector firms on a subcontracting basis [11] for example, the leather-tanning and dyeing process is mainly done in the informal sector for the garment industry. During this process, waste from bleaching, dyeing, and burning produces hazardous chemicals that pollute groundwater and air [5,9]. In short, the shadow economy covers all the production stages that cause pollution.

Figure 1 presents the trends of CO₂ emissions, N₂O emissions, and SECO over time. The data indicates that the SECO grows alongside CO₂ and N₂O emissions in selected 107 developing countries. This positive correlation suggests that shadow economic activities, which often operate without stringent regulatory oversight, contribute significantly to air pollution.

The shadow economy’s size is closely tied to the quality of governance indicators, with poor governance fostering its growth. Corruption within bureaucratic and administrative systems, weak enforcement of legal frameworks, and overall governance inefficiency create an environment where informal activities can thrive. Particularly, it has been evidenced that businesses tend to be informal as a form of escaping complex and costly systems related either to corruption or the weak rule of law (RLAW) [5,12]. On the other hand, such forms of good governance, wherein the latter has also been characterized as having strong RLAW, a governance framework with effective practice combined with stringent corruption control (CORC), may be an obstacle to the SECO activities. Under such conditions, businesses are more likely to adhere to formal regulations and operate within the legal economy because they face fewer barriers to compliance and greater protection and incentives [13,14].

Good governance strengthens the confidence of the businessmen, guarantees justice, and endorses equitable distribution of recourses, which further decreases the reasons for contributing to the SECO [15,16]. With good governance, the official economy endorses reducing tax evasion and enhances transparency, encouraging enterprises to operate legally. Formalization provides more economic stability and can reduce air pollution since firms in the official economy are typically under strict environmental standards that ensure lower levels of pollution [17,18]. Poor governance worsens the situation as it allows firms to operate outside of regulations that compromise environmental laws. If there is poor enforcement of environmental regulations, then firms that work in the shadow sector usually avoid pollution controls that increase the level of emissions and degradation in the environment [9]. Strict environmental policies, such as high taxes on pollution, cannot bind firms, as this may be strategically used to shift them into the shadow sector in order to avoid compliance costs [19]. In such cases, weak governance not only undermines economic stability but also compromises environmental sustainability since informal enterprises contribute disproportionately to pollution while operating beyond the reach of regulatory oversight.

In short, SECO will be harmful to environmental sustainability if governments do not take the necessary actions. Therefore, any empirical analysis on environmental sustainability, as well as a policy recommendation ignoring the existence of the SECO and governance indicators, would be incomplete, if not misleading. Therefore, the present study has developed two main objectives. Firstly, we examine the impact of the SECO on AIRP. Secondly, we explore the moderating role of governance indicators, including CORC, RLAW, and regulatory quality (RQ) in the SECO-AIRP nexus. Only a few studies [20,21] empirically investigated the role of governance indicators in the SECO-AIRP nexus. However, the scope of the present study is vast. We focus on developing countries as a case study. The developing countries are a focal point of discussion because they provide a valuable context for examining the current study, given their challenging characteristics. These characteristics include the large size of the SECO, higher level of AIRP, and underdeveloped governance framework. However, we include developing countries worldwide, encompassing 107 countries across different income levels. This broader scope is important because the SECO is a concern for all types of developing economies [22]. Furthermore, we utilize the most recent dataset available, covering the period from 2002 to 2020. Additionally, we employ two different measures of the AIRP, including CO₂ and N₂O emissions. Furthermore, we use three robust measures of governance indicators, including CORC, RLAW, and RQ, to achieve the study’s objective. This approach adds to the robustness of the results, as the use of different datasets provides a more thorough analysis.

The paper is organized as follows: Section 2 offers an in-depth review of the related literature. Section 3 highlights theoretical underpinnings from previous research. Section 4 describes the data sources and outlines the methodological approach employed in the study. Section 5 presents the results. Section 6 discusses the findings in detail. Section 7 summarizes conclusions and offers insights into potential policy implications.

2. Literature Review

2.1. Shadow Economy and Air Pollution

In the past literature, many studies have explored the link between SECO and AIRP. In this context, Blackman [10] inspected the informal brick industry of Mexico and found that the brick industry utilized propane gas, which is a source of pollution. From his point of view, the developing countries’ SECOs usually comprise unlicensed and low-tech small businesses, which lead to AIRP and pose a challenge to environmental authorities. To a further extent, the research [23] studied the SECO sector in Mexico and observed that shadow sector activities are a major source of pollution because, in these activities, firms use a wide range of low-technology and shadow economic activities operate outside of formal regulatory frameworks. This indicates that there are often no environmental regulations or standards imposed on these activities. As a result, informal enterprises may not have to comply with pollution control measures, leading to increased emissions and pollution. By constructing a theoretical model, [9] claimed that the intensive environmental rules induce the official economy to shift its production activities into the informal economy; hence, this act produces pollution emissions.

By developing a theoretical model, Biswas et al. [17] combine AIRP, corruption, and the SECO into an integrated framework to expose the dynamics of how the SECO expands AIRP under a certain level of corruption. They indicated that to reduce the environmental impact of the SECO, the government should control the spread of corruption. In the same manner, [6] investigated the linkages between SECO and CO₂ emissions in developing countries, indicating a positive impact. Furthermore, Chen et al. [24] observed this relationship in 30 Chinese provinces, which broadened the analysis and contributed to more evidence of the presence of a positive synergistic relationship between SECO and CO₂ emissions. To extend the scope, [25] performed a detailed analysis of 106 economies worldwide and compared the effect of the SECO on various pollutants, N₂O, CH4, and CO₂ emissions. According to the empirical findings, SECO leads to enhanced emissions of N₂O and CH4. Surprisingly, SECO has no significant impact on CO₂ emissions. In a similar vein, [26] has conducted research across 160 countries globally and observed that the SECO put relentless pressure to enlarge the level of CO₂ emissions. Furthermore, Huynh and Ho [12] show that the SECO increases CO₂ emission in 43 developing countries.

The prior literature has also supported some strong arguments highlighting the non-linear association between the SECO and AIRP. In this context [27], panel data from 152 countries showed that there is an inverted U-shaped affiliation between the two variables. In the same vein, [7] evidenced the inverse N-shape association between CO₂ emissions for income levels when even controlling for SECO effects. Furthermore, Wang et al. [28] investigated the moderating role of corruption in the SECO and AIRP nexus. It is noted that AIRP is significantly increased in the presence of SECO and corruption in the case of China. Likewise, Huynh [18] explored the interaction between the SECO, CO_2, and fiscal policy in developing Asian countries. The study focused on the impact of taxation and government expenditure on CO₂ emission. By analyzing the findings of the study, it was found that the SECO affects CO₂ positively. Contrary to this, some evidence was found that expansionary fiscal policies, through the reduction in the size of the SECO, reduced CO₂ emissions. Similarly, Pang et al. [29] observed that the SECO and AIRP had U-shaped non-linear relationships. Most recently, Ahmad and Hussain [5] concluded that the SECO indeed raises the level of CO₂ emissions in 127 developing countries.

2.2. Governance Indicators and Air Pollution

Environmental sustainability can be enhanced through good governance. Many studies have examined the relationship between governance indicators (GOVI) and air pollution (AIRP) in the literature. For example, Usman et al. [30] analyzed the impact of corruption control (CORC) and income on CO₂ emissions in Africa. They found that CORC and higher income levels lead to an increase in CO₂ emissions, but their combined effect has the opposite effect. Similarly, Zhang et al. [31] examined the effects of GOVI on CO₂ emission in the BRICS economies. They also noted that the enhancement of CORC, RLAW, and government stability has a negative correlation with carbon emissions in the long run. These findings hold implications for the need to strengthen the governance framework in order to effectively address carbon emissions in the case of BRICS countries. Moreover, Huynh and Ho [12] found that better institutional quality significantly reduces the level of air pollution in 43 developing economies.

Furthermore, Khan et al. [32] have also revealed that better governance through factors like political stability, RLAW, and RQ are vital in reducing air pollution in developing nations. Similarly, Karim et al. [33] explored the role played by six governance factors on CO₂ emissions in African countries. According to their findings, CO₂ emissions are substantially reduced when CORC, RQ, and RLAW are enhanced. In the same way, Sultana et al. [34] exposed that corruption control plays a role in the reduction of CO₂ emissions. Similarly, Lapatinas et al. [35] stated that in countries where corruption prevails, politicians might decide to dedicate a large share of public money to environmental programs to extract resources instead of enhancing environmental quality. This detrimental practice is also found to have a direct bad impact on the quality of the environment.

Moreover, Chen et al. [24] have pointed out that the presence of a growing number of corrupt officers would lead to the weakening of environmental laws and the constant increase in the emissions of pollutants caused by illegal production. Sinha et al. [36] revealed that corruption hinders or dilutes the enforcement of environmental policies, resulting in higher emissions of pollutants. Biswas et al. [17] built an integrated model that included AIRP, corruption, and the SECO. They showed that corruption enables the growth of AIRP within certain levels and that controlling corruption would reduce the impact of the SECO on AIRP. Similarly, Ivanova [37] conducted a study on European countries and realized that polluting enterprises pay bribes to government officials not to report their pollutant release; they avoid paying pollution taxes and are also contributing to pollution emissions. Furthermore, Liu et al. [38] found that corruption and a poor RLAW lead to more AIRP.

2.3. Governance Indicators and Shadow Economy

Several empirical and theoretical works in the literature examine the link between governance indicators (GOVI) and the shadow economy (SECO). In this regard, Khan and Rehman [39] revealed that a weak governance framework increases the size of the SECO, on the other hand, a strong governance environment enables firms to move from the SECO to the formal financial sector. Furthermore, Butt et al. [40] proved that environmental degradation is highly sensitive to governance quality. Based on their findings, they concluded that governance quality can positively influence ecological change for better environmental quality. Hence, the strength of the governance framework is necessary in the formulation and implementation of laws that can address environmental challenges that are prevalent across the world today. Likewise, Khattak et al. [41] also looked at the effect of governance quality and banking competition on the SECO for 127 countries. According to their results, more banking competition and stronger governance, in general, reduces the SECO. Furthermore, the effect of competition on the SECO is stronger in countries with relatively weak governance.

Moreover, Dang et al. [42] provided a quantitative analysis of Asian countries, proving that good governance reduces the size of the SECO. In the same way, Lacobuta et al. [43] investigated the determinants of the SECO in the EU and found that the RLAW, RQ, and labor freedom negatively affect the SECO. Their primary finding is that the efficiency of governance indicators has to be increased in order to combat the SECO. Similarly, Dada et al. [44] showed that in the context of poor governance, the SECO is likely to increase. In addition, Barra and Papaccio [45] revealed that countries with good governance tend to exhibit lower levels of SECO, which contributes to the enhancement of the legal tendencies in the labor market in Italy. Further, Mveng and Henri [46] indicated that states with higher levels of governance development that feature strict anti-corruption measures usually record lower levels of the SECO. In this line, Dreher et al. [47] undertook a study across 145 economies. Based on their analysis, they concluded that corruption and the SECO are complements. Moreover, their study pointed out that a better governance framework is playing a key role in the reduction of the SECO size. Johnson et al. [48] used data from OECD, Latin American, and transition economies, where findings showed that corruption and a weak RLAW play a role in causing SECO growth. In a recent study, Huynh and Ho [12] reveal that a higher level of governance reduces the size of the SECO and also reduces the impact of the SECO on CO₂ emissions in 43 developing countries.

3. Theoretical Framework

The shadow economy, which operates outside the legal framework and lacks accountability and legal requirements, has been cited as having a link with environmental degradation. In the shadow economy (SECO), the cost of implementing environmental laws is not incurred, and hence high emissions of CO₂ and N₂O are observed [6,25].

Improving governance quality, including corruption control (CORC), the rule of law (RLAW), and regulatory quality (RQ), is instrumental in reducing the size of the shadow economy. Public Choice Theory emphasizes that government officials, driven by self-interest, may exploit their roles to foster informal economic activities unless strong anti-corruption measures are in place. By reducing opportunities for rent-seeking, effective corruption control disincentivizes participation in the SECO [49,50]. Institutional Theory similarly posits that well-defined and consistently enforced rules of the game support formal economic engagement, as these structures offer stability and predictability [51]. Deterrence Theory complements this view, proposing that rigorous enforcement of the rule of law raises the costs associated with informality, thereby discouraging shadow activities [52]. Collectively, these theories illustrate how robust governance mitigates SECO growth by fostering an environment where formal business is more efficient and trustworthy. Reducing pollution also requires a strong governance framework and regulatory systems that align market incentives to encourage firms to operate within the formal sector [17,53].

The relationship between the SECO and the different governance indicators is not straightforward. The level of governance quality mediates the level of air pollution caused by shadow economic activities. In countries with good governance structures, the SECO contributes little to pollution as good governance structures ensure that environmental laws are followed [12]. Other sources that also affect the AIRP include renewable energy (RNEC), growth of GDP (GGDP), population density (POPD), and industrialization (IND). Chen et al. [54] confirm that adoption of RNEC reduces the level of AIRP in China, as evidenced by the reduction of CO₂ emissions. On the other hand, non-renewable energy sources are known to cause pollution levels to rise. Common measures of economic development include GGDP, which usually results in higher pollution as a consequence of increased industrialization and energy use [55]. Furthermore, the studies [56,57] have also identified that the IND and high POPD lead to higher emission levels, mainly due to the demands for energy and higher production rates.

This theoretical construction demonstrates the complex relationships among the SECO, GOVI, and AIRP. It highlights how the SECO contributes to AIRP levels, especially in developing countries with poor governance. On the other hand, good governance can mitigate the adverse effects of the SECO by enforcing environmental regulations effectively. Additionally, the inclusion of control variables such as RNEC, GGDP, POPD, and IND further refines the analysis by acknowledging their direct roles in influencing AIRP in developing countries. The model describing the interconnectedness among these factors is specified in the equation below:

$${\mathrm{A}\mathrm{I}\mathrm{R}\mathrm{P}}_{\mathrm{i}\mathrm{t}}=\mathrm{f}({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}},{\mathrm{G}\mathrm{O}\mathrm{V}\mathrm{I}}_{\mathrm{i}\mathrm{t}},\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}\times \mathrm{G}\mathrm{O}\mathrm{V}{\mathrm{I}}_{\mathrm{i}\mathrm{t}},\mathrm{R}\mathrm{N}\mathrm{E}{\mathrm{C}}_{\mathrm{i}\mathrm{t}},{\mathrm{G}\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{i}\mathrm{t}},{\mathrm{P}\mathrm{O}\mathrm{P}\mathrm{D}}_{\mathrm{i}\mathrm{t}},{\mathrm{I}\mathrm{N}\mathrm{D}}_{\mathrm{i}\mathrm{t}})$$

(1)

where, AIRP stands for air pollution which is measured in terms of CO₂ and N₂O emissions, SECO represents the shadow economy, GOVI belongs to governance indicators including corruption control (CORC), the rule of law (RLAW), and regulatory quality (RQ), SECO × GOVI reflects the interaction between the shadow economy and governance indicators, RNEC refers to renewable energy, GGDP indicates economic growth rate, POPD represents to population density, and IND identifies industrialization. Equation (1) is expanded and rewritten as follows:

$${\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}}=\mathrm{f}({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}},{\mathrm{C}\mathrm{O}\mathrm{R}\mathrm{C}}_{\mathrm{i}\mathrm{t}},\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}\times {\mathrm{C}\mathrm{O}\mathrm{R}\mathrm{C}}_{\mathrm{i}\mathrm{t}},\mathrm{R}\mathrm{N}\mathrm{E}{\mathrm{C}}_{\mathrm{i}\mathrm{t}},{\mathrm{G}\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{i}\mathrm{t}},{\mathrm{P}\mathrm{O}\mathrm{P}\mathrm{D}}_{\mathrm{i}\mathrm{t}},{\mathrm{I}\mathrm{N}\mathrm{D}}_{\mathrm{i}\mathrm{t}})$$

(2)

$${\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}}=\mathrm{f}({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}},{\mathrm{R}\mathrm{L}\mathrm{A}\mathrm{W}}_{\mathrm{i}\mathrm{t}},\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}\times {\mathrm{R}\mathrm{L}\mathrm{A}\mathrm{W}}_{\mathrm{i}\mathrm{t}},\mathrm{R}\mathrm{N}\mathrm{E}{\mathrm{C}}_{\mathrm{i}\mathrm{t}},{\mathrm{G}\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{i}\mathrm{t}},{\mathrm{P}\mathrm{O}\mathrm{P}\mathrm{D}}_{\mathrm{i}\mathrm{t}},{\mathrm{I}\mathrm{N}\mathrm{D}}_{\mathrm{i}\mathrm{t}})$$

(3)

$${\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}}=\mathrm{f}({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}},{\mathrm{R}\mathrm{Q}}_{\mathrm{i}\mathrm{t}},\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}\times {\mathrm{R}\mathrm{Q}}_{\mathrm{i}\mathrm{t}},\mathrm{R}\mathrm{N}\mathrm{E}{\mathrm{C}}_{\mathrm{i}\mathrm{t}},{\mathrm{G}\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{i}\mathrm{t}},{\mathrm{P}\mathrm{O}\mathrm{P}\mathrm{D}}_{\mathrm{i}\mathrm{t}},{\mathrm{I}\mathrm{N}\mathrm{D}}_{\mathrm{i}\mathrm{t}})$$

(4)

$${{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}}=\mathrm{f}({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}},{\mathrm{C}\mathrm{O}\mathrm{R}\mathrm{C}}_{\mathrm{i}\mathrm{t}},\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}\times {\mathrm{C}\mathrm{O}\mathrm{R}\mathrm{C}}_{\mathrm{i}\mathrm{t}},\mathrm{R}\mathrm{N}\mathrm{E}{\mathrm{C}}_{\mathrm{i}\mathrm{t}},{\mathrm{G}\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{i}\mathrm{t}},{\mathrm{P}\mathrm{O}\mathrm{P}\mathrm{D}}_{\mathrm{i}\mathrm{t}},{\mathrm{I}\mathrm{N}\mathrm{D}}_{\mathrm{i}\mathrm{t}})$$

(5)

$${{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}}=\mathrm{f}({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}},{\mathrm{R}\mathrm{L}\mathrm{A}\mathrm{W}}_{\mathrm{i}\mathrm{t}},\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}\times {\mathrm{R}\mathrm{L}\mathrm{A}\mathrm{W}}_{\mathrm{i}\mathrm{t}},\mathrm{R}\mathrm{N}\mathrm{E}{\mathrm{C}}_{\mathrm{i}\mathrm{t}},{\mathrm{G}\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{i}\mathrm{t}},{\mathrm{P}\mathrm{O}\mathrm{P}\mathrm{D}}_{\mathrm{i}\mathrm{t}},{\mathrm{I}\mathrm{N}\mathrm{D}}_{\mathrm{i}\mathrm{t}})$$

(6)

$${{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}}=\mathrm{f}({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}},{\mathrm{R}\mathrm{Q}}_{\mathrm{i}\mathrm{t}},\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}\times {\mathrm{R}\mathrm{Q}}_{\mathrm{i}\mathrm{t}},\mathrm{R}\mathrm{N}\mathrm{E}{\mathrm{C}}_{\mathrm{i}\mathrm{t}},{\mathrm{G}\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{i}\mathrm{t}},{\mathrm{P}\mathrm{O}\mathrm{P}\mathrm{D}}_{\mathrm{i}\mathrm{t}},{\mathrm{I}\mathrm{N}\mathrm{D}}_{\mathrm{i}\mathrm{t}})$$

(7)

4. Materials and Methods

4.1. Data and Variables

In order to empirically verify the above-mentioned models, we used a balanced panel data of 107 developing countries out of 193 countries in the world from 2002 to 2020. The choice of time and country is driven by data availability. The list of selected developing countries is presented in

Table A1. List of selected developing countries.

Albania	Chad	Guinea-Bissau	Mauritius	Solomon Islands
Algeria	China	Haiti	Mexico	South Africa
Angola	Colombia	Honduras	Moldova	Sri Lanka
Argentina	Comoros	India	Mongolia	St. Lucia
Armenia	Congo, Dem. Rep.	Indonesia	Morocco	St. Vincent and the Grenadines
Azerbaijan	Congo, Rep.	Iran, Islamic Rep.	Mozambique	Sudan
Bangladesh	Costa Rica	Jamaica	Myanmar	Suriname
Belarus	Côte d’Ivoire	Jordan	Namibia	Syrian Arab Republic
Belize	Dominican Republic	Kazakhstan	Nepal	Tajikistan
Benin	Ecuador	Kenya	Nicaragua	Tanzania
Bhutan	Egypt, Arab Rep.	Kyrgyz Republic	Niger	Thailand
Bolivia	El Salvador	Lao PDR	Nigeria	Togo
Bosnia and Herzegovina	Equatorial Guinea	Lebanon	North Macedonia	Tunisia
Botswana	Eswatini	Lesotho	Pakistan	Türkiye
Brazil	Ethiopia	Liberia	Papua New Guinea	Uganda
Bulgaria	Fiji	Libya	Paraguay	Ukraine
Burkina Faso	Gabon	Madagascar	Peru	Vietnam
Burundi	Gambia, The	Malawi	Philippines	Yemen, Rep.
Cabo Verde	Georgia	Malaysia	Russian Federation	Zambia
Cambodia	Ghana	Maldives	Rwanda
Cameroon	Guatemala	Mali	Senegal
Central African Republic	Guinea	Mauritania	Sierra Leone

.

Air Pollution (AIRP): In the models, AIRP is the dependent variable and is estimated by carbon dioxide (CO₂) and nitrous oxide (N₂O) emissions. The factors considered under CO₂ emissions are mainly activities carried out by people that release CO₂ into the environment. These factors include the use of coal, oil, and natural gas for transport, electricity generation, and other uses. Process emissions, mainly carbon dioxide, are emitted as a result of the manufacturing process of cement. Different industrial processes like chemical manufacturing, metallurgy, and the utilization of solvents are known to emit CO₂ into the atmosphere. In addition, emissions of CO₂ result from the burning of fossil fuels for heating, cooking, and operating electrical appliances in households and companies [5,58]. Furthermore, the N₂O emissions are also from agricultural biomass burning, industrial processes, and livestock management.

Shadow Economy (SECO): The models use the SECO as a core independent variable. SECO is defined here as the hidden production of goods and services that occurs outside the official regulation by public authorities for reasons of money, regulation, or institutional. Monetary motives include tax evasion and social security exclusion; regulatory motives include avoiding bureaucracy or onerous regulations; institutional motives include bribery, which is closely related to weak political institutions and a poor rule of law [59].

Governance Indicators (GOVI): Governance indicators are the moderator variables in the models and are controlled by corruption control (CORC), the rule of law (RLAW), and regulatory quality (RQ). CORC relates to people’s attitudes towards the level of misuse of power by those in leadership positions. This encompasses all sorts of corruption that include petty corruption, grand corruption, corrupt intentions of privileged persons, and private entities over governmental transactions [5]. Furthermore, the RLAW also includes people’s beliefs and compliance with the rules of society. They show the importance of compliance with contracts, property rights, efficient policing, and the need for accurate and fair judicial systems. It also considers the level of crime and violence in a society [40]. In addition, RQ reflects the views on the government’s capacity to formulate and implement proper policies and regulations that allow and encourage private business activities [45,53].

Control Variables: Taking inspiration from prior relevant literature, in addition to our primary independent variables, our models incorporate a set of control variables. We employ renewable energy consumption, growth of GDP, population density, and industrialization as control variables based on studies [60,61]. The details of variables are also presented in

Table 1. Description of variables.

Type of Variables	Descriptions	Signs	Sources
Dependent Variables
Carbon Dioxide Emissions (CO2)	Log of CO2 emissions (metric tons per capita)		WDI, WB
Nitrous Oxide Emissions (N2O)	Log of N2O emissions (metric tons per capita of CO2 equivalent)		WDI, WB
Core Variables
Shadow Economy (SECO)	Shadow economy (% of GDP)	+	IE, WB
Moderating Variables
Corruption Control (CORC)	Control of corruption index; −2.5 (lowest corruption control) to +2.5 (highest corruption control)	−	WGI, WB
Rule of Law (RLAW)	Rule of law index; −2.5 to 2.5 (higher score indicates a stronger rule of law)	−	WGI, WB
Regulator Quality (RQ)	Regulator quality index; −2.5 to 2.5 (higher score indicates a stronger regulatory quality)	−	WGI, WB
Control Variables
Renewable Energy (RNEC)	Renewable energy consumption (as % of the total final energy consumption)	−	WDI, WB
Growth of GDP (GGDP)	Growth of GDP (annual %)	+	WDI, WB
Population Density (POPD)	Log of number of people per square km of land area	+	WDI, WB
Industrialization (IND)	Value added in industry (% of GDP)	+	WDI, WB

.

Table 2. Summary statistics.

Variables	CO2	N2O	SECO	CORC	RLAW	RQ	RNEC	GGDP	POPD	IND
Observation	2033	2033	2033	2033	2033	2033	2033	2033	2033	2033
Mean	2.24	0.11	38.33	−0.62	−0.60	−0.57	43.80	3.73	117.66	31.23
St. Dev	2.55	0.31	11.23	0.57	0.56	0.54	32.04	5.64	189.45	14.22
Minimum	0.03	0.02	11.40	−1.72	−1.83	−1.65	0.03	−50.33	1.47	5.34
Maximum	15.24	2.56	69.35	1.75	1.31	1.81	98.34	88.97	1714.7	83.66

presents the summary statistics of the variables. The average value of CO₂ emissions is 2.24 with standard deviations (St. Dev) of 2.55. Similarly, the average of N₂O emissions is 0.11 with St. Dev of 0.31. Both CO₂ and N₂O emissions statistics indicate substantial variability. The SECO has an average value of 38.33, ranging from 11.40 to 69.35, reflecting significant differences across observations. The CORC, RLAW, and RQ have mean values of −0.62, −0.60, and −0.57, respectively, both with moderate variability as indicated by St. Dev of 0.57, 0.56, and 0.54. These statistics provide a foundational understanding of the data’s characteristics, highlighting the variability and range of key variables.

4.2. Econometrics Model Specifications

To achieve the objective of the study, we propose a model based on the above theoretical framework section and follow the studies [5,21,25,40]. We first explore the direct effects of the shadow economy on air pollution. The model is formulated as follows:

$$\mathrm{A}\mathrm{I}\mathrm{R}{\mathrm{P}}_{\mathrm{i}\mathrm{t}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{A}\mathrm{I}\mathrm{R}\mathrm{P}}_{\mathrm{i}\mathrm{t}-1}+{\mathsf{\beta }}_2{\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }\prime \mathrm{Z}}_{\mathrm{i}\mathrm{t}}+{\mathrm{v}}_{\mathrm{i}}+{\mathsf{\omega }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$$

(8)

where AIRP identifies the air pollution. In this study, we use two indicators of air pollution: CO₂ and N₂O emissions. SECO denotes the shadow economy. Z is the vector of control variables, such as renewable energy, growth of GDP, population density, and industrialization. The subscripts $\mathrm{i}$ and t denote the country and time, respectively. ${\mathrm{v}}_{\mathrm{i}}$ and ${\mathsf{\omega }}_{\mathrm{t}}$ are the country and time-specific effects respectively, while ${\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$ is the usual error term. The Equation (8) is rewritten in the extended form as follows:

$${\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}-1}+{\mathsf{\beta }}_2{\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }\prime \mathrm{Z}}_{\mathrm{i}\mathrm{t}}+{\mathrm{v}}_{\mathrm{i}}+{\mathsf{\omega }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$$

(9)

$${\mathrm{N}}_2{\mathrm{O}}_{\mathrm{i}\mathrm{t}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}-1}+{\mathsf{\beta }}_2{\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\beta }\prime \mathrm{Z}}_{\mathrm{i}\mathrm{t}}+{\mathrm{v}}_{\mathrm{i}}+{\mathsf{\omega }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$$

(10)

To explore the role of governance indicators in the SECO-AIRP nexus, the Equations (9) and (10) are rewritten as follow:

$${\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}-1}+{\mathsf{\alpha }}_2{\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_3{\mathrm{G}\mathrm{O}\mathrm{V}\mathrm{I}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_4({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}\times {\mathrm{G}\mathrm{O}\mathrm{V}\mathrm{I}}_{\mathrm{i}\mathrm{t}})+{\mathsf{\alpha }\prime \mathrm{Z}}_{\mathrm{i}\mathrm{t}}+{\mathrm{v}}_{\mathrm{i}}+{\mathsf{\omega }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$$

(11)

$${{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}-1}+{\mathsf{\alpha }}_2{\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_3{\mathrm{G}\mathrm{O}\mathrm{V}\mathrm{I}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_4({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}\times {\mathrm{G}\mathrm{O}\mathrm{V}\mathrm{I}}_{\mathrm{i}\mathrm{t}})+{\mathsf{\alpha }\prime \mathrm{Z}}_{\mathrm{i}\mathrm{t}}+{\mathrm{v}}_{\mathrm{i}}+{\mathsf{\omega }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$$

(12)

where, GOVI denotes the governance indicators including corruption control (CORC), the rule of law (RLAW) and regulatory quality (RQ). SECO × GOVI represents the interaction between shadow economy and governance indicators. We further extend Equation (11) to individually check the role of governance indicators in the SECO-CO₂ emissions nexus as follows:

$${\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}-1}+{\mathsf{\alpha }}_2{\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_3{\mathrm{C}\mathrm{O}\mathrm{R}\mathrm{C}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_4({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}\times {\mathrm{C}\mathrm{O}\mathrm{R}\mathrm{C}}_{\mathrm{i}\mathrm{t}})+{\mathsf{\alpha }\prime \mathrm{Z}}_{\mathrm{i}\mathrm{t}}+{\mathrm{v}}_{\mathrm{i}}+{\mathsf{\omega }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$$

(13)

$${\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}-1}+{\mathsf{\alpha }}_2{\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_3{\mathrm{R}\mathrm{L}\mathrm{A}\mathrm{W}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_4({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}\times {\mathrm{R}\mathrm{L}\mathrm{A}\mathrm{W}}_{\mathrm{i}\mathrm{t}})+{\mathsf{\alpha }\prime \mathrm{Z}}_{\mathrm{i}\mathrm{t}}+{\mathrm{v}}_{\mathrm{i}}+{\mathsf{\omega }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$$

(14)

$${\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}-1}+{\mathsf{\alpha }}_2{\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_3{\mathrm{R}\mathrm{Q}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_4({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}\times {\mathrm{R}\mathrm{Q}}_{\mathrm{i}\mathrm{t}})+{\mathsf{\alpha }\prime \mathrm{Z}}_{\mathrm{i}\mathrm{t}}+{\mathrm{v}}_{\mathrm{i}}+{\mathsf{\omega }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$$

(15)

To explore the marginal effect of the SECO on CO₂ emissions at different levels of CORC, RLAW, and RQ, we take the partial derivatives of Equations (13)–(15) with respect to the SECO. The derivations are performed as follows:

$${\displaystyle \frac{\partial \left({\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}}\right)}{\partial \left({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}\right)}}={\mathsf{\alpha }}_2+{\mathsf{\alpha }}_4{\mathrm{C}\mathrm{O}\mathrm{R}\mathrm{C}}_{\mathrm{i}\mathrm{t}}$$

(16)

$${\displaystyle \frac{\partial \left({\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}}\right)}{\partial \left({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}\right)}}={\mathsf{\alpha }}_2+{\mathsf{\alpha }}_4{\mathrm{R}\mathrm{L}\mathrm{A}\mathrm{W}}_{\mathrm{i}\mathrm{t}}$$

(17)

$${\displaystyle \frac{\partial \left({\mathrm{C}{\mathrm{O}}_2}_{\mathrm{i}\mathrm{t}}\right)}{\partial \left({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}\right)}}={\mathsf{\alpha }}_2+{\mathsf{\alpha }}_4{\mathrm{R}\mathrm{Q}}_{\mathrm{i}\mathrm{t}}$$

(18)

Now, we extend the Equation (12) for individually check the role of governance indicators in the SECO-N₂O emissions nexus as follow:

$${{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}-1}+{\mathsf{\alpha }}_2{\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_3{\mathrm{C}\mathrm{O}\mathrm{R}\mathrm{C}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_4({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}\times {\mathrm{C}\mathrm{O}\mathrm{R}\mathrm{C}}_{\mathrm{i}\mathrm{t}})+{\mathsf{\alpha }\prime \mathrm{Z}}_{\mathrm{i}\mathrm{t}}+{\mathrm{v}}_{\mathrm{i}}+{\mathsf{\omega }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$$

(19)

$${{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}-1}+{\mathsf{\alpha }}_2{\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_3{\mathrm{R}\mathrm{L}\mathrm{A}\mathrm{W}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_4({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}\times {\mathrm{R}\mathrm{L}\mathrm{A}\mathrm{W}}_{\mathrm{i}\mathrm{t}})+{\mathsf{\alpha }\prime \mathrm{Z}}_{\mathrm{i}\mathrm{t}}+{\mathrm{v}}_{\mathrm{i}}+{\mathsf{\omega }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$$

(20)

$${{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1{{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}-1}+{\mathsf{\alpha }}_2{\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_3{\mathrm{R}\mathrm{Q}}_{\mathrm{i}\mathrm{t}}+{\mathsf{\alpha }}_4({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}\times {\mathrm{R}\mathrm{Q}}_{\mathrm{i}\mathrm{t}})+{\mathsf{\alpha }\prime \mathrm{Z}}_{\mathrm{i}\mathrm{t}}+{\mathrm{v}}_{\mathrm{i}}+{\mathsf{\omega }}_{\mathrm{t}}+{\mathsf{\mu }}_{\mathrm{i}\mathrm{t}}$$

(21)

To explore the marginal effect of SECO on N₂O emissions at different levels of CORC, RLAW and RQ, we take the partial derivatives of Equations (19)–(21) with respect to the SECO. The derivations are performed as follows:

$${\displaystyle \frac{\partial \left({{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}}\right)}{\partial \left({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}\right)}}={\mathsf{\alpha }}_2+{\mathsf{\alpha }}_4{\mathrm{C}\mathrm{O}\mathrm{R}\mathrm{C}}_{\mathrm{i}\mathrm{t}}$$

(22)

$${\displaystyle \frac{\partial \left({{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}}\right)}{\partial \left({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}\right)}}={\mathsf{\alpha }}_2+{\mathsf{\alpha }}_4{\mathrm{R}\mathrm{L}\mathrm{A}\mathrm{W}}_{\mathrm{i}\mathrm{t}}$$

(23)

$${\displaystyle \frac{\partial \left({{\mathrm{N}}_2\mathrm{O}}_{\mathrm{i}\mathrm{t}}\right)}{\partial \left({\mathrm{S}\mathrm{E}\mathrm{C}\mathrm{O}}_{\mathrm{i}\mathrm{t}}\right)}}={\mathsf{\alpha }}_2+{\mathsf{\alpha }}_4{\mathrm{R}\mathrm{Q}}_{\mathrm{i}\mathrm{t}}$$

(24)

where, the signs of ${\mathsf{\alpha }}_2$ and ${\mathsf{\alpha }}_4$ in Equations (16)–(18) and in Equations (22)–(24) reflect whether there are complementarity or substitutability effects between shadow economy and governance indicators (CORC, RLAW, RQ). If both coefficients have the same signs, there will be a complementarity effect; otherwise, there will be a substitutability effect.

4.3. Econometrics Method

It is important to note that when dealing with models related to CO₂ emissions and N₂O emissions, the issue of endogeneity often arises [5,25]. Endogeneity arises from multiple factors, including omitted variables, simultaneity, measurement errors, or reverse causality [62]. Regarding our current study, the main explanatory variables we are examining appear to be endogenous, which introduces the possibility of encountering the endogeneity problem in our models. For example, the shadow economy and CO₂ emissions cause one another [5,21,63].

Many shadow economic activities cause a large extent of CO₂ emissions. For example, in the developing world, a major part of the shadow economy is engaged in resource extraction, manufacturing, bleaching and dyeing of clothes, servicing, and retailing activities such as craft mining, car repair, leather tanning, brick making, and metalworking. Most of these practices have significant adverse consequences for the environment and create CO₂ emissions [9]. While governments imposing strict regulations or implementing high taxes on CO₂ emissions can result in increased expenses for energy production and consumption within the formal sector, this could prompt both businesses and individuals to explore more cost-effective options, such as utilizing informal channels with lower energy costs, thereby leads to increasing the shadow economy [63]. It is also found in the literature that lagged levels of CO₂ emissions and N₂O emissions may affect their current values [5,25,44].

As mentioned in the above paragraph, there is a risk of potential endogeneity problems in our models. In this case, both Ordinary Least Square (OLS) and fixed effect estimations are prone to bias. Neglecting the endogeneity bias can lead to a significant issue, as the resulting estimates may be misleading or invalid [64]. Therefore, in the current study, instead of relying on static models, we utilize the Generalized Method of Moments (GMM) estimator to perform our estimation. The GMM estimator, developed by [65,66], is commonly used for panel data analysis. It yields reliable outcomes even in the presence of different sources of endogeneity, such as simultaneity, unobserved heterogeneity, and dynamic endogeneity [67]. In simpler terms, this method allows for controlling the effects specific to each country and time period while addressing the issue of endogeneity in variables, specifically when dependent variables are used on the independent side by taking one or more lags of the dependent variable.

5. Results

5.1. Correlation Matrix

The correlation matrix in

Table 3. Correlation matrix.

Variable	CO2	N2O	SECO	CORC	RLAW	RQ	RNEC	GGDP	POPD	IND
CO2	1.000
N2O	0.29 **	1.000
	(0.031)
SECO	0.61 ***	0.54 ***	1.000
	(0.001)	(0.004)
CORC	−0.39 ***	−0.28 ***	−0.44 ***	1.000
	(0.000)	(0.002)	(0.001)
RLAW	−0.46 **	−0.42 **	−0.56 ***	0.67 ***	1.000
	(0.025)	(0.042)	(0.000)	(0.001)
RQ	−0.29 **	−0.32 ***	−0.34 **	0.11 **	0.092	1.000
	(0.022)	(0.000)	(0.013)	(0.013)	(0.235)
RNEC	−0.24 ***	−0.35 ***	−0.091 *	0.060	0.23 ***	0.24 ***	1.000
	(0.004)	(0.000)	(0.073)	(0.131)	(0.001)	(0.004)
GGDP	0.23 ***	0.21 **	−0.86 ***	0.34 ***	0.34 ***	0.12 **	0.23 ***	1.000
	(0.002)	(0.032)	(0.001)	(0.000)	(0.000)	(0.024)	(0.000)
POPD	0.82 ***	0.72 ***	0.34 ***	0.045	0.12 ***	0.41 ***	0.56 ***	−0.82 *	1.000
	(0.003)	(0.000)	(0.001)	(0.154)	(0.003)	(0.001)	(0.001)	(0.048)
IND	0.87 ***	0.64 ***	0.36 ***	0.32 ***	0.23 **	0.23 ***	0.74 ***	−0.09	−0.13	1.000
	(0.000)	(0.000)	(0.001)	(0.000)	(0.014)	(0.000)	(0.000)	(0.143)	(0.245)

Notes: p-values in parentheses. ***, **, * show 1%, 5%, 10% level of significance respectively.

reveals the relationships between key variables, showing strong associations between the SECO, CORC, RLAW, RQ, and AIRP, including CO₂ and N₂O emissions. Both CO₂ and N₂O emissions are positively correlated with the SECO (0.61 and 0.54, respectively), highlighting that the SECO activities significantly contribute to AIRP. Conversely, CORC, RLAW, and RQ negatively correlate with CO₂ (−0.39, −0.46, and −0.29) and N₂O (−0.28, −0.42, and −0.32), suggesting that better governance leads to lower air pollution. Notably, RNEC has a minimal influence on SECO (0.09), emphasizing a weak relationship. Additionally, IND shows a strong positive correlation with CO₂ (0.87) and a moderate positive correlation with N₂O (0.64), indicating that increased IND enhances AIRP.

5.2. Variance Inflation Factors

It is worth noting that a strong correlation exists among the independent variables themselves. In Table 3, a substantial 86% correlation is observed between GGDP and SECO. Likewise, the correlation coefficient between POPD and GGDP stands at approximately 82%. This correlation scenario raises concerns about significant cross-sectional dependence or multicollinearity within the models, potentially compromising the reliability of our estimated results. To mitigate this concern, we utilize the Variance Inflation Factor (VIF) as a diagnostic tool to evaluate multicollinearity within our models. The VIF values are presented in

Table 4. Variance inflation factors.

Variables	VIF	1/VIF
SECO	2.754	0.363
CORC	4.346	0.230
RLAW	3.536	0.282
RQ	2.876	0.350
RNEC	2.745	0.364
GGDP	1.543	0.648
POPD	2.867	0.348
IND	4.536	0.220

. Notably, all variables show VIF values below 5, indicating a lack of significant multicollinearity and confirming their suitability for the analysis.

However, the correlation analysis has limited power in depicting the true nature of relationships among the variables. Therefore, to obtain more accurate results, a more systematic empirical analysis is necessary.

5.3. Regression Analysis

In order to perform a comprehensive and rigorous empirical analysis, this study uses the GMM estimator system to estimate the models. The system GMM approach is especially useful in dynamic panel data analysis since it can help eliminate possible biases related to unobserved heterogeneity, simultaneity, and dynamic endogeneity. In order to minimize the potential for heteroscedasticity and autocorrelation in the standard errors, we employ robust standard errors to increase the validity of the results. This is done by using the ‘robust’ option in our Stata commands, which makes the standard errors corrected for the specific features of the data, thus enhancing the reliability of the results. To estimate the system GMM test, we employ the ‘xtabond2’ command in Stata 17, which is famous for handling panels with dynamic data. This procedure enables us to address the concern of endogeneity of the independent variables and unobserved panel-level effects, thus enhancing the efficiency of the coefficients’ estimates.

To confirm the reliability of the instruments and the appropriateness of the model, we performed two important diagnostic checks. First, the Sargan and Hansen J-tests are used to check the over-identifying restrictions in order to confirm whether the instruments used are indeed valid and meaningful. These tests are useful in establishing the reliability of the instrument set used in the study. Second, we conducted the Arellano-Bond test for first-order (AR1) and second-order (AR2) autocorrelation of the residuals to check for autocorrelation. While AR1 is expected due to the dynamic nature of the panel data model, the absence of AR2 is essential for the consistency of the system GMM estimator. Together, these diagnostic tests offer a robust assessment of the empirical model, ensuring that the assumptions of the GMM technique are upheld.

5.3.1. Impact of Shadow Economy on CO₂ and N₂O Emissions

In this section, we examine the effects of the shadow economy (SECO) on air pollution (AIRP) measured by CO₂ and N₂O emissions. The results of the estimation, along with the corresponding test statistics, are presented in

Table 5. Shadow economy effects on CO₂ and N₂O emissions.

System GMM
Variables	CO2 Emissions		N2O Emissions
	(1)	(2)	(3)	(4)
CO2 (−1)	0.773 ***	0.744 ***
	(0.089)	(0.240)
N2O (−1)			0.993 ***	0.735 ***
			(0.045)	(0.148)
SECO	0.380 ***	0.172 ***	0.440 ***	0.309 ***
	(0.132)	(0.006)	(0.074)	(0.097)
RNEC		−0.025 **		−0.255 ***
		(0.012		(0.077)
GGDP		0.048 *		0.312 ***
		(0.023)		(0.106)
POPD		0.066 ***		0.461
		(0.019)		(0.287)
IND		1.075 *		0.632 *
		(0.569)		(0.373)
Constant	1.419 ***	3.660 ***	2.553 *	4.203 **
	(0.232)	(0.988)	(1.405)	(2.066)
Model Diagnostics
Obs.	1926	1926	1926	1926
Countries	107	107	107	107
Instruments	33	35	45	49
F-test	25.23	16.54	8.65	13.74
AR(1) p-value	0.024	0.050	0.002	0.000
AR(2) p-value	0.756	0.169	0.894	0.874
Sargan p-value	0.191	0.354	0.464	0.265
Hansen-J p-value	0.254	0.198	0.289	0.524
Time Dummies	Yes	Yes	Yes	Yes

Notes: Robust standard errors in parentheses. ***, **, * show 1%, 5%, 10% level of significance respectively.

. A total of four different specifications are estimated. To retain the core variable and the dynamic nature of the models, all four specifications included the SECO and lag of dependent variables. The first and third specification serves as a simple baseline model, comprising only the lag of dependent variables and the SECO. In the second and fourth specifications, control variables such as RNEC, GGDP, POPD, and IND are added to the models. This step allows us to observe the effects of these control variables separately. It is noticeable that all four specifications fulfill the requirements of AR1, AR2, Sargan, and the Hansen J-test, ensuring the validity of the estimations and the reliability of the results.

The empirical results indicate that our core variable SECO appeared to be statistically significant and have a positive relationship with CO₂ and N₂O emissions in all the models. This result is expected because many SECO activities cause a large extent of CO₂ and N₂O emissions. In the context of a growing SECO, the environmental impact is worsened by the informal and unregulated nature of these activities. Businesses operating outside the formal sector may not adhere to environmental standards as they seek to cut costs and maximize profits. Additionally, the lack of transparency and accountability in the SECO makes it challenging for authorities to monitor and regulate environmental practices effectively [5,25].

5.3.2. Moderating Role of Governance Indicators in Shadow Economy-CO₂ Emissions Nexus

As highlighted in the literature, both SECO and governance indicators (GOVI) have a significant impact on AIRP. However, what is the impact of their interaction effect on AIRP? To address this gap, we incorporate an interaction term between the SECO and GOVI (SECO × GOVI). We analyze six specifications, presented in

Table 6. Moderating role of governance indicators in shadow economy-CO₂ emissions nexus.

System GMM
Variables	Corruption Control		Rule of Law		Regularity Quality
	(1)	(2)	(3)	(4)	(5)	(6)
CO2 (−1)	0.786 ***	0.799 ***	0.814 ***	0.816 ***	0.813 ***	0.949 ***
	(0.038)	(0.032)	(0.034)	(0.0123)	(0.034)	(0.036)
SECO	0.022 **	0.065 *	0.649 ***	0.399 ***	0.178 ***	0.828 ***
	(0.011)	(0.039)	(0.199)	(0.138)	(0.038)	(0.042)
CORC	−0.684 **	−0.684 **
	(0.289)	(0.289)
RLAW			−0.027 ***	−0.314 ***
			(0.005)	(0.123)
RQ					−0.067 **	−0.949 ***
					(0.029)	(0.031)
SECO × COR	−0.161 ***	−0.091 **
	(0.043)	(0.036)
SECO × RLAW			−0.009 **	−0.276 *
			(0.005)	(0.143)
SECO × RQ					−0.049 *	−0.098 **
					(0.029)	(0.048)
RNEC		−0.012 ***		−0.251 *		−0.098 ***
		(0.003)		(0.131)		(0.034)
GGDP		0.091 **		0.276 *		0.094 ***
		(0.036)		(0.143)		(0.033)
POPD		0.039		0.312 ***		1.851
		(0.031)		(0.123)		(1.218)
IND		1.323 ***		−0.032		0.178 ***
		(0.254)		(0.023)		(0.038)
Constant	−1.324 ***	−3.553 ***	3.019 ***	3.052 ***	2.061 ***	12.158 ***
	(0.422)	(0.971)	(0.916)	(0.919)	(0.752)	(3.385)
Model Diagnostics
Obs.	1926	1926	1926	1926	1926	1926
Countries	107	107	107	107	107	107
Instruments	27	28	34	36	32	41
F-test	7.93	26.34	17.34	3.45	34.23	15.87
AR(1) p-value	0.012	0.011	0.012	0.010	0.000	0.003
AR(2) p-value	0.656	0.676	0.699	0.679	0.583	0.895
Sargan p-value	0.651	0.423	0.243	0.345	0.294	0.234
Hansen-J p-value	0.123	0.343	0.124	0.212	0.323	0.190
Time Dummies	Yes	Yes	Yes	Yes	Yes	Yes

Notes: Robust standard errors in parentheses. ***, **, * show 1%, 5%, 10% level of significance respectively.

. The first specification includes the lag of CO₂ emissions, the SECO, CORC, and the interaction term between the SECO and CORC (SECO × CORC). The second specification builds on the first by incorporating control variables such as RNEC, GGDP, POPD, and IND. These control variables are selected based on their significance in the baseline model (see Table 5). The third specification includes the lag of CO₂ emissions, the SECO, RLAW, and the interaction term between the SECO and RLAW (SECO × RLAW). The fourth specification extends the third by adding the same set of control variables used in the second specification. Similarly, the fifth specification includes the lag of CO₂ emissions, the SECO, RQ, and the interaction term between the SECO and RQ (SECO × RQ). The sixth specification extends the fifth by adding the same set of control variables used in the second and fourth specifications.

The estimation results, presented in Table 6, indicate that all six specifications satisfy the conditions of the AR1 and AR2 tests for autocorrelation, as well as the Sargan and Hansen J-tests for instrument validity, ensuring the robustness of the obtained results. The SECO shows a significant positive impact on CO₂ emissions in all the specifications. Furthermore, the governance indicators (GOVI) such as CORC, RLAW, and RQ significantly reduce CO₂ emissions. Notably, the interaction term between the SECO and CORC (SECO × CORC) is negative and significant, suggesting that better corruption control mitigates the adverse environmental impact of the SECO. Similarly, the interaction term between the SECO and RLAW (SECO × RLAW) is also negative and significant, indicating that a stronger rule of law reduces the negative environmental impact of the SECO. Moreover, the interaction term between the SECO and RQ (SECO × RQ) is also negative and significant, indicating that a better regularity quality reduces the negative environmental impact of the Shadow economy.

The results indicate the existence of significant substitutability between SECO and GOVI. These findings also underscore the importance of governance indicators in moderating the environmental impact of the shadow economy.

For a detailed understanding, we estimate the marginal effect of the shadow economy (SECO) on CO₂ emissions at three different levels of CORC, RLAW, and RQ: the 25th, 50th, and 75th percentiles. The results presented in

Table 7. Marginal effects of the shadow economy on CO₂ emissions at different levels of governance indicators.

System GMM
Percentile Levels	Corruption Control		Rule of Law		Regularity Quality
	(1)	(2)	(3)	(4)	(5)	(6)
P25	0.172 ***	0.204 ***	0.299 ***	0.318 **	0.153 ***	0.788 ***
	(0.006)	(0.037)	(0.084)	(0.128)	(0.042)	(0.038)
P50	0.089 ***	0.182 ***	0.198 ***	0.154 ***	0.121 ***	0.318 ***
	(0.019)	(0.032)	(0.056)	(0.042)	(0.035)	(0.099)
P75	0.018 ***	0.178 ***	0.128	0.089 ***	0.067 *	0.145
	(0.003)	(0.038)	(0.196)	(0.036)	(0.036)	(0.106)

Notes: P25, P50, and P75 are the 25th, 50th and 75th percentiles respectively. ***, **, * show 1%, 5%, 10% level of significance respectively.

indicate that the marginal impact of the SECO on CO₂ emissions decreases as the levels of CORC, RLAW, and RQ increase. At the 25th percentile of CORC, RLAW, and RQ, the marginal effect of the SECO on CO₂ emissions is higher compared to the 50th and 75th percentiles. This finding demonstrates that countries with weaker governance frameworks experience more severe environmental damage from shadow economic activities. Conversely, at the 75th percentile of governance indicators, the marginal impact of the SECO on CO₂ emissions is substantially lower, indicating that improvements in governance significantly reduce the shadow economy’s impact on CO₂ emissions. This suggests that an improved governance framework, reflected through better CORC, stronger enforcement of the RLAW, and better RQ, not only diminishes the size of the SECO but also mitigates its harmful impact on CO₂ emissions [5,20,21].

5.3.3. Moderating Role of Governance Indicators in Shadow Economy-N₂O Emissions Nexus

To check the robustness of our estimates, we further explore the moderating role of governance indicators (GOVI) such as CORC, RLAW, and RQ in the relationship between SECO and N₂O emissions. The findings in

Table 8. Moderating role of governance indicators in the shadow economy-N₂O emissions nexus.

System GMM
Variables	Corruption Control		Rule of Law		Regularity Quality
	(1)	(2)	(3)	(4)	(5)	(6)
N2O (−1)	0.992 ***	0.799 ***	0.823 ***	0.919 ***	0.916 ***	0.884 ***
	(0.004)	(0.013)	(0.054)	(0.013)	(0.023)	(0.067)
SECO	0.191 ***	0.163 ***	0.112 *	0.168 ***	0.319 ***	0.207 ***
	(0.069)	(0.062)	(0.062)	(0.062)	(0.065)	(0.069)
CORC	−0.107 *	−0.098 ***
	(0.062)	(0.036)
RLAW			−0.068 *	−0.215 ***
			(0.037)	(0.058)
RQ					−0.245 *	−0.103 **
					(0.141)	(0.044)
SECO×COR	−0.018 *	−0.086 **
	(0.009)	(0.039)
SECO×RLAW			−0.058 *	−0.025 ***
			(0.036)	(0.003)
SECO×RQ					−0.006 *	−0.005 *
					(0.004)	(0.003)
RNEC		−0.025 ***		−0.166 ***		−0.038 **
		(0.003)		(0.059)		(0.016)
GGDP		0.119 **		0.278 **		0.106 **
		(0.053)		(0.128)		(0.043)
POPD		0.051		0.738 *		0.065 **
		(0.062)		(0.429)		(0.032)
IND		0.284 ***		0.188 ***		0.126 ***
		(0.064)		(0.057)		(0.044)
Constant	3.912 ***	3.250 ***	3.567 ***	3.908 ***	2.023 **	3.812 ***
	(0.413)	(0.222)	(0.423)	(0.414)	(0.669)	(0.413)
Model Diagnostics
Obs.	1926	1926	1926	1926	1926	1926
Countries	107	107	107	107	107	107
Instruments	34	45	37	54	28	36
F-test	26.63	56.36	6.75	34.36	25.99	16.35
AR(1) p-value	0.013	0.013	0.000	0.014	0.013	0.000
AR(2) p-value	0.712	0.706	0.674	0.702	0.677	0.265
Sargan p-value	0.145	0.899	0.671	0.284	0.483	0.773
Hansen-J p-value	0.135	0.156	0.108	0.224	0.384	0.423
Time Dummies	Yes	Yes	Yes	Yes	Yes	Yes

Notes: Robust standard errors in parentheses. ***, **, * show 1%, 5%, 10% level of significance respectively.

indicate that in all six specifications, SECO significantly increases the N₂O emissions. Moreover, better CORC, RLAW, and RQ significantly reduce the N₂O emissions. Furthermore, the coefficient value of the interaction between the SECO and CORC (SECO × CORC) is negative, which highlights the substitutability between the SECO and CORC. In addition, the coefficient value of the interaction between SECO and RLAW (SECO × RLAW) is also negative and significant, which also highlights the substitutability between the SECO and RLAW. In addition, the interaction between SECO and RQ (SECO × RQ) appeared negative, which also indicates the substitutability between the SECO and RQ. These results are consistent with the estimate given in Table 6.

Moreover, we examine the marginal effects of the SECO on N₂O emissions at the 25th, 50th, and 75th percentiles of CORC, RLAW, and RQ, respectively. The results presented in

Table 9. Marginal effects of the shadow economy on N₂O emissions at different levels of governance indicators.

System GMM
Percentile Levels	Corruption Control		Rule of Law		Regularity Quality
	(1)	(2)	(3)	(4)	(5)	(6)
P25	0.147 ***	0.085 **	0.191 ***	0.270 ***	0.329 ***	0.409 ***
	(0.039)	(0.039)	(0.069)	(0.070)	(0.069)	(0.070)
P50	0.074 *	0.067 *	0.163 ***	0.219 ***	0.236 ***	0.278 ***
	(0.041)	(0.037)	(0.062)	(0.063)	(0.059)	(0.065)
P75	0.038	0.057 *	0.107 *	0.188 ***	0.131 **	0.184 ***
	(0.043)	(0.036)	(0.062)	(0.058)	(0.063)	(0.064)

Notes: P25, P50, and P75 are the 25th, 50th and 75th percentiles respectively. ***, **, * show 1%, 5%, 10% level of significance respectively.

reveal that the marginal effect of the shadow economy on N₂O emissions decreases with an increase in the level of CORC, RLAW, and RQ. These results are in line with the estimate given in Table 7.

Several control variables were included in the analysis to account for other factors that influence air pollution. Renewable energy consumption has a significant negative impact on both CO₂ and N₂O emissions. Countries that invest in renewable energy experience lower levels of pollution, supporting the findings of Chen et al. [54]. The results show that GDP growth is positively correlated with CO₂ and N₂O emissions, suggesting that economic expansion, particularly industrialization, contributes to higher levels of pollution. This finding aligns with Ang [55], who demonstrated that economic growth often leads to increased emissions in developing economies. Higher population density is associated with increased emissions, reflecting the greater energy demand and higher production levels in densely populated areas. This is consistent with the findings of Raheem and Ogebe [56], who noted that population growth in developing countries exacerbates environmental degradation. Industrialization has a positive and significant effect on both CO₂ and N₂O emissions. As countries industrialize, they typically increase their reliance on energy-intensive activities, leading to higher emissions. This supports the findings of the studies [56,57].

6. Discussion

The findings of this study provide critical insights into the link between the shadow economy (SECO) and air pollution (AIRP), emphasizing the moderating role of governance indicators (GOVI) in developing countries. The study confirmed that the SECO significantly contributes to increased levels of CO₂ and N₂O emissions. This finding is consistent with previous studies [5,6,25], which highlighted how unregulated activities in the SECO lead to environmental degradation. Specifically, the positive correlation between the SECO and AIRP in our results aligns with Blackman et al. [23], who observed similar trends in Mexico’s SECO, where industries evade environmental regulations and increase AIRP.

This study expands on the foundational work by Elgin and Oztunali [27] and Wang et al. [28], who explored the non-linear dynamics between SECO and AIRP. The current results show a robust and consistent positive relationship between the SECO and AIRP, indicating that without proper regulatory oversight, the SECO exacerbates environmental harm. The increase in emissions of CO₂ and N₂O aligns with Huynh [18], who documented how shadow sectors in Asian economies contribute significantly to CO₂ emissions due to the use of outdated technology and unregulated practices. Furthermore, the study’s results support Chen et al. [24], who found that the SECO in 30 Chinese provinces contributed to increased CO₂ emissions. However, our research adds to this by showing that the harmful environmental effects of the SECO can be mitigated by improvements in governance frameworks. This echoes the findings of Ahmad and Hussain [5], who argued that corruption control plays a crucial role in reducing environmental damage from shadow economic activities.

A critical finding of this study is the significant moderating role played by governance indicators such as CORC, RLAW, and RQ in reducing the SECO impact on AIRP. Our results are consistent with the studies [30,31], which demonstrated that improvements in governance reduce CO₂ emissions. The negative and significant interaction terms between the SECO and governance indicators suggest a substitutability effect, where stronger governance reduces the negative environmental impacts of shadow economic activities. This indicates that governance frameworks, through their ability to enforce regulations, control corruption, and ensure compliance with environmental standards, can significantly mitigate the adverse effects of the shadow economy.

These results reinforce the findings of Sultana et al. [34], who found that improvements in governance quality, particularly corruption control, led to lower CO₂ emissions. Similarly, Huynh and Ho [12] highlighted that strong governance significantly reduces the SECO size and its associated pollution emissions. The current study contributes to this body of work by showing that governance indicators not only reduce the size of the SECO but also limit its harmful environmental effects. Better regulatory quality (RQ), effective corruption controls (CORC), and strong rule of law (RLAW) are vital in ensuring that shadow activities transition to formal operations that comply with environmental standards. Furthermore, the positive relationship between SECO growth and AIRP, as revealed by the findings, supports the theory proposed by Biswas et al. [17], who modeled how environmental degradation is exacerbated by shadow economies operating under weak governance structures.

By combining shadow economy dynamics with governance quality, our study highlights a key policy lever: governance improvement. Governance indicators act as a critical moderating variable, as shown by the significant interaction terms in our regression models. Feld and Schneider [13] argued that a strong RLAW and CORC are necessary for formalizing the SECO, which the current study confirmed. This formalization leads to better compliance with environmental regulations, reducing pollution levels, as firms are held accountable for their environmental impact. These theoretical insights are supported by empirical evidence, as illustrated by the significant interaction between the shadow economy and governance indicators, demonstrating the substitutability effect proposed by Chaudhuri and Mukhopadhyay [19].

7. Conclusions and Policy Implication

Air pollution has been a challenging issue all over the world, especially in developing countries. Poor governance quality facilitates the growth of the shadow economy, and the unofficial and unregulated nature of this sector can lead to environmentally detrimental practices. From unsustainable resource extraction to illegal waste disposal, the lack of oversight and accountability in the shadow economy can contribute to air pollution. Addressing the role of governance indicators and formalizing economic activities are crucial steps toward promoting environmental sustainability and mitigating the negative impact of the shadow economy on air pollution. Therefore, this study offered significant insights into the nexus between the shadow economy and air pollution in developing economies, emphasizing the moderating role of governance indicators.

The study findings revealed that the shadow economy plays a damaging role in increasing levels of air pollutants such as CO₂ and N₂O emissions, which are major contributors to climate change. The absence of regulatory supervision in shadow sectors allows for environmentally harmful practices, including the use of outdated technologies, unsustainable production methods, and inefficient fuel consumption. These practices lead to a rise in the levels of air pollution, which damages global efforts to combat climate change and environmental degradation. The study also indicated the critical importance of governance indicators in reducing the adverse effects of the shadow economy. Governance quality, measured through corruption control, the rule of law, and regulatory quality, plays a moderating role in the relationship between the shadow economy and air pollution. The study found that when governance frameworks are robust, the negative environmental impact of the shadow economy is significantly reduced. Specifically, stronger corruption control helps to reduce opportunities for businesses to bypass environmental regulations, while the enforcement of the rule of law ensures compliance with environmental standards. Regulatory quality further discourages environmentally harmful practices by creating transparent and predictable legal environments where businesses are held accountable for their environmental impact.

The study findings confirmed a substitution effect between the shadow economy and governance indicators. In the time of good governance, the size of the shadow economy decreases, resulting in lower levels of air pollution. In contrast, poor governance allows the shadow economy to grow and worsen environmental degradation. This finding underscores the need for policymakers to prioritize governance reforms as part of broader environmental sustainability strategies. By improving governance quality, governments can both shrink the shadow economy and ensure that economic activities adhere to environmental regulations.

The inclusion of control variables in this study, such as renewable energy, growth of GDP, population density, and industrialization, further refines the analysis. These variables help isolate the specific impact of the shadow economy and governance indicators on air pollution. Renewable energy, for instance, was shown to reduce CO₂ and N₂O emissions, highlighting its role in promoting environmental sustainability. On the other hand, GDP growth, population density, and industrialization were associated with increased emissions, which suggests that economic expansion often comes at the expense of environmental health unless appropriate regulatory measures are implemented. By accounting for these control variables, the study provides a more comprehensive understanding of the factors influencing air pollution. Figure 2 also illustrates the effects of explanatory variables on air pollution (AIRP).

7.1. Policy Implication

Environmental sustainability in developing countries cannot be achieved without addressing the governance deficits that enable the shadow economy to operate without accountability. Governments must focus on building strong governance frameworks that include enhanced corruption control, stronger rule of law, and improved regulatory quality. These measures not only foster economic formalization but also ensure that businesses comply with environmental standards, thereby reducing the levels of air pollution. Moreover, integrating renewable energy into the economic structure is essential for reducing emissions and promoting a sustainable economic model.

7.2. Future Recommendations

Building on the findings of this study, future research could enhance the analysis of the shadow economy’s environmental impact by incorporating several additional factors, including public awareness of environmental protection, economic structures, natural resource economics, waste management practices, and enterprise-level investments in sustainability. Investigating public awareness could reveal how societal values shape both the occurrence and impact of shadow economic activities on pollution. Analyzing the economic structure of transition economies, such as the size and type of industry sectors, would offer insights into sector-specific governance challenges. Examining natural resource dependency and management could highlight sustainable resource governance’s role in mitigating pollution. Additionally, a focus on waste management practices may expose how ineffective disposal and recycling exacerbate pollution associated with the shadow economy while exploring corporate investments in environmental protection could identify opportunities for public-private collaborations to advance sustainable practices. Together, these dimensions would allow for a more comprehensive approach to addressing the environmental implications of the shadow economy.