Environmental Sustainability and Foreign Direct Investment in East Africa: Institutional and Policy Benefits for Environmental Sustainability

Abstract

Persistent drought is not a unique phenomenon in East African countries, different research findings cite different reasons for it, but the environmental problem is currently a major concern worldwide and in East African country (EAC) the problem is not an exception to this phenomenon; policymakers and researchers are interested in knowing the cause in order to mitigate environmental degradation and support policies and institutions for environmental sustainability. Therefore, this study examines the environmental sustainability laws and institutions in 18 EAC to determine the relationship between foreign direct investment and environmental quality. Using the generalized method of moments for analysis, the results show, among other things, that foreign direct investment, when associated with environmental sustainability policies and institutions, improves environmental quality in the long run while degrading it in the short run. Long- and short-run environmental improvements in EAC are also enabled by domestic investment, environmental sustainability institutions, and policies. The study, therefore, concludes that environmental sustainability institutions and policies are critical in EAC because they improve environmental quality and interact with foreign direct investment in the long run. Therefore, the study recommends that policymakers and other stakeholders in EAC take action to improve environmental quality and sustainable economies.

1. Introduction

Environmental concerns, such as air and water pollution, poor sanitation, and the loss of natural resources and forest reserves, have raised serious concerns in both developed and developing nations in recent years. Health, food security, access to clean water and land, as well as natural and physical capital, are all impacted by changes in the environment [1]. These environmental worries have generated a worldwide campaign to tackle environmental change through the Paris Agreement and the Kyoto Protocol. The primary goal of these global movements is to reduce the harmful effects of carbon emissions on the environment. The Global Energy Report 2021, which states that carbon emissions from energy combustion and industrial processes climbed again in 2021 and reached the highest yearly level ever, provides the foundation for the efforts to lower carbon emissions. Emissions increased to 36.3 gigatons, up 6% from 2020. In Africa, particularly East African, carbon emissions (CO₂ in kilotons) increased from 66,300 in 2012 to 101,420 in 2019 (World Bank 2020). Carbon emissions are mainly influenced by energy consumption, and the industrial structure in the context of economic growth determines energy consumption in many ways. For a sustainable and healthy development of the global economic society, new alternative energy sources must be developed to reduce carbon emissions [2,3,4]. Given the increase in CO₂ emissions in East Africa, there is no doubt that carbon emissions are affecting the quality of the environment in East Africa, negatively impacting the well-being of citizens, and therefore, need to be urgently addressed. According to the Institute for Economics and Peace, of the 39 African countries surveyed in the Lloyd’s Register Foundation World Risk Poll 2019, the highest level of concern about climate change was registered in Southern African countries. Lesotho (77.9%) and Malawi (74.6%), ranked sixth and ninth globally, respectively, were the countries with the highest percentage of the population concerned about climate change, followed by Eswatini (69.6%), Namibia (65.3%), and Zambia (64.4%), with most of these countries located in East Africa. The negative impacts of environmental change on human livelihoods, economic development, climate, and ecosystems will worsen if less attention is paid to them.

In recent years, air and water pollution, inadequate sanitation, and exhaustion of natural resources and forest reserves have caused great concern in industrialized and non-industrialized countries. A poor environmental situation adversely affects the health and economic well-being of citizens. According to [2], climate change has an impact on vital human livelihoods, such as wellbeing, ordinary and corporal capital, and a way into food, water, and land. The Paris Agreement and the Kyoto Protocol were established in response to these environmental issues in order to tackle climate change. The primary objectives of these international actions are to lessen the negative environmental effects of carbon emissions. According to a report from the International Energy Agency published in February 2020, worldwide energy-related carbon emissions remained constant at 33 Gt in 2019. However, in 2018, the report states that global carbon dioxide emissions increased by 1.7%.

Global carbon dioxide emissions grew 1.7% in 2017, a growth rate that was the greatest since 2013 and 70% greater than the average increase since 2010. From 66,300 kilotons of CO₂ in 2012 to 101,420 kilotons in 2019, East Africa’s carbon emissions grew. There is no question that carbon emissions are hurting the quality of the environment in East African countries, which has a detrimental influence on the well-being of inhabitants and needs to be quickly addressed given the growth in CO₂ emissions there even though there was a modest decline in 2015. Due to a lack of attention, climate change has had a severe influence on human existence, economic growth, the environment, and climate [2,3]. Foreign direct investment is generally viewed (empirically and theoretically) as complementary to domestic capital in promoting economic growth and development, as it offers advantages [4,5,6].

East Africa welcomed USD9.3 billion in FDI in 2019, an increase of 12.7% from USD8.12 billion in 2018 [7]. Based on the flow of foreign direct investment and its benefits, one can infer that East African countries are on the periphery of development. This interpretation, which is supported by the neoclassical theory (capital flows), asserts that foreign assets match domestic capital and foster economic development [8]. Ref. [9] Has been proven that foreign direct investment flows lead to economic revival through the sharing of managerial and skill-related knowledge, the transfer of technology, and the creation of new jobs. FDI has a favorable effect on economies such as those in east African countries, but it also interacts with the environment and in some way degrades environmental quality. Researchers and policymakers are equally interested in learning whether FDI is a factor in CO₂ emissions, the loss of forest reserves, and the depletion of natural resources. Numerous studies [4,10,11,12] have demonstrated that FDI worsens environmental quality (increases CO₂ emissions). However, others [4,13,14,15] contend that FDI enhances environmental quality (reduces CO₂ emissions), making the discussion of the issue still vital for aiding policymakers and other stakeholders in deciding on environmental improvement measures.

When attempting to determine the environmental impact of foreign direct investment, a number of earlier studies, particularly in Africa [11,12,16,17,18], neglected to take into account the benefits of institutions and policies for maintaining environmental sustainability. Specifically, it does not reflect how environmental policies are intended to protect natural resources and manage pollution, i.e., how well they protect and sustain them. FDI and environmental quality are intertwined, but environmental sustainability policies and institutions are an important variable to consider when examining the relationship. The implementation of environmental quality improvement strategies is more likely in economies with effective environmental sustainability policies and institutions. For instance, governments with strong environmental sustainability institutions and policies are more likely to put them into practice and so improve the environmental quality of their nations. As a result, the relationship between FDI and environmental quality depends heavily on institutions and policies for environmental sustainability. We also consider FDI and investigate their relationships with environmental institutions and policies in order to add to the corpus of knowledge on the nations of East Africa. Previous literature has largely used the quality of political institutions variable [2,10,16,17,18,19,20]. The quality of the environment is evaluated in this study using an index of policy and institutions. It is a departure, making it a valuable contribution. This inclusion allows us to determine how much environmental quality is impacted by institutions and policies for environmental sustainability. Using the interaction variable, it is feasible to evaluate the effect of institutions and policy on environmental sustainability in conjunction with foreign direct investment in environmental quality. Furthermore, FDI flows into various economic sectors, including mining, manufacturing, and pulling out activities, and could have a variety of effects on environmental quality, mostly pollution and diminution of forest reserves. The majority of research on FDI has focused on how it affects CO₂ emissions, generally ignoring the degradation of natural resources such as forest reserves. Carbon emissions have been stressed as a gauge of environmental quality in the majority of studies, especially those carried out in African countries [11,16,17,18]. This does not adequately capture the effect of foreign direct investment on environmental quality. We fill this gap by examining how FDI impacts carbon emissions, the natural resources depletion, and the forest reserves destruction. We also developed an EQ index by combining natural resource depletion, carbon dioxide emissions, and forest reserve depletion metrics in order to measure the overall effect of foreign direct investment on EQ. The major objective of the study is to investigate how FDI affects environmental quality in East African countries (See Appendix A,

Table A1. List of East African Countries.

Burundi	Madagascar	Somalia
Comoros	Malawi	South Sudan
Djibouti	Mauritius	Tanzania
Eritrea	Mozambique	Uganda
Ethiopia	Rwanda	Zambia
Kenya	Seychelles	Zimbabwe

), while taking into account institutions, policies, and interactions with FDI that are important to environmental sustainability. The remainder of the article is structured as follows: the theoretical and empirical overviews are covered in the following section, and the material and methodological framework, the analysis of the empirical data, and discussion are covered in Section 3 and Section 4, respectively. Section 5 of the essay discusses conclusions and recommendations.

2. Literature Review

Theoretical and Empirical Review

Research on the relationship between FDI, energy consumption, carbon emissions, trade openness, policy institution for environmental sustainability and international tourism has gained value with the deepening of human industrialization. The comprehensive and systematic study of the coordination relationship between these variables has been important and urgent since the 20th century/second-generation environmental issue in the 1980s. The academic literature on this issue has been abundant in recent years, but obtaining a clear conclusion has been difficult. There is also a lack of positive and negative evidence on the impact of FDI on East Africa’s environment. There are various theoretical and empirical perspectives on how FDI affects environmental quality. This section compiles a few of the relevant theoretical frameworks and empirical investigations. In terms of theoretical approaches, we investigate three hypotheses: scale effect, pollution haven, and pollution halo hypothesis. Ref. [21], proposed the pollution haven hypothesis, later on verified and approved as a perfect theory by [22]. According to the pollution haven hypothesis, because sophisticated economies frequently have stringent environmental regulations, businesses, particularly those from other countries, are severely penalized for pollutant emissions. Multinational corporations relocate to nations with loose environmental restrictions in an effort to increase profits, which increases pollution in developing nations. Basing the report from the World Investment, FDI flows to Africa show that although FDI flows to Africa decreased in 2019, previous FDI flows to African countries increased by 41.6% from 2017 to 2018, which is directly related to pollution [11]. Hymer, the first independent researcher of FDI theory, proposed monopolistic advantage theory in his doctoral thesis in 1960. According to Hymer, the motivation for FDI was that, compared with the host nation, the investor had more favorable knowledge advantages, including production technologies, management and organizational skills, sales skills and other intangible assets, as well as enterprise-scale and other monopolistic advantages, thereby obtaining more benefits from production abroad. Another reason for FDI is that strict energy environmental management and control policies are generally implemented in developed countries. Therefore, some polluting enterprises are more inclined to transfer their factories to developing countries, such as east Africa, with low environmental standards and lower production costs. According to the pollution haven hypothesis, foreign businesses that receive FDI harm the environment. Concerning the pollution haven hypothesis, ref. [23] demonstrated that foreign businesses are drawn to economies with laxer environmental regulations, which has a negative effect on EQ. According to the scale effect hypothesis, while foreign or multinational firms increase industrial sector output, they also consume more energy, which pollutes the receiving or host country [24]. According to the pollution halo hypothesis, foreign companies from developed countries should invest in developing countries, such as countries in Africa, to improve their environmental quality. Theorists believe that foreign businesses with forward-thinking managing ideas and innovative know-how benefit the environment. It is noted in [25] that foreign companies use hygienic energy, which ensures effective energy use and does not harm the environment.

This study offers a brief overview of selected literature about Africa and other parts of the globe based on empirical studies. Ref. [14] investigated the relationships between FDI, economic growth, energy consumption, and CO₂ emissions in six African nations between 1971 and 2009, including Congo, Kenya, and Zambia. After using the Granger causality test and autoregressive distributed lag for the investigation, they discovered that FDI has a significant positive effect on CO₂ emissions in Zimbabwe and a negative and significant effect on emissions in the Democratic Republic of the Congo, South Africa, and Kenya, but an insignificant effect on emissions in all other countries.

The Democratic Republic of the Congo had a significant negative impact on CO₂ emissions, while South Africa and Zimbabwe had a positive and significant effect, while the remaining countries had no impact. In terms of causal connections, the findings revealed that FDI, energy use, and economic growth, all contribute to CO₂ emissions in the Democratic Republic of the Congo, South Africa, and Kenya. Furthermore, ref. [18] documented the dynamic interplay between environmental quality, FDI, economic development as measured by CO₂ emissions in selected SSA countries, including Ghana, Angola, and Benin between 1980 and 2013. According to the study, which used panel vector error correction as its estimating method, FDI and economic expansion in SSA had considerable favorable effects on CO₂ emissions in SSA. Further, FDI and per capita income shocks were found to reduce environmental quality and increase carbon dioxide emissions. Furthermore, ref. [12] investigated the relationship between environmental quality, economic growth and foreign direct investment in six African countries between 1980 and 2014. According to the study, while gross domestic product and foreign direct investment have no effect on environmental quality, energy consumption and GDP do. The Pool Mean Group Estimator is used to demonstrate this. In terms of causality, in the short run, the finding shows that CO₂ emissions and energy consumption are bidirectional, but in the long run, only energy consumption causes carbon emissions.

Using the dynamic generalized method of moment model and the fixed effect as estimation methodologies, the relationship between transportation, energy usage, urbanization, and environmental quality, as measured by CO₂ emissions, was explored over the period 1980–2011 in 19 chosen African nations [11]. According to the study, while foreign direct investment and regulatory quality have a negative significant effect on environmental quality, transportation energy consumption, urbanization, electricity consumption, and the relations between regulation and energy consumption for transportation have a positive significant effect on carbon emissions. Panel data from 2000 to 2018, as well as the generalized method of moments as an estimation approach, were used to examine the CO₂ emissions thresholds from trade and FDI for a green economy using 45 African nations [17]. According to the findings, FDI has a positive significant effect on CO₂ emissions, while trade has a negative significant effect.

Researchers examined the relationship between growth determinants and environmental quality (as measured by CO₂ emissions) in the selected 44 African countries from 1990 to 2014 by the Dumitrescu and Hurlin’s Granger causality test and the pooled mean group estimator [11]. According to the estimated results, while economic growth has a positive significant effect on environmental quality, the renewable energy consumption and the financial development have a negative significant effect. The research found a two-way association between carbon emissions, industrial practices, economic growth and, renewable energy consumption. Ref. [4] conducted a separate study that looked at environmental quality and financial development in 35 SSA countries from 1985 to 2014. The study discovered that population, wealth, and skill have a positive significant effect on environmental quality, but economic development has a negative significant effect on carbon emissions using the Dumitrescu and Hurlin’s Granger causality test, and the augmented mean group estimator. According to findings, there is a two-way relationship between financial growth and carbon emissions. Furthermore, the interaction of technological progress and economic development has a downbeat effect on EQ.

Previous research conducted outside of Sub-Saharan Africa between 1975 and 2012 [12], examined whether foreign direct investment affects EQ in 99 low-, high-, and middle-income countries. The study revealed that both FDI and FDI squared improve environmental quality in all countries, whereas energy consumption and economic growth have a positive significant effect on CO₂ emissions. FDI improves EQ in high-income countries and worsens it in low- and middle-income countries (with a positive significant effect), depending on income levels. Dumitrescu and Hurlin’s findings on causality, FDI, and energy expenditure have a bidirectional causal relationship with carbon emissions in the selected 99 countries. While high-income countries have a bidirectional relationship between energy consumption and CO₂ emissions, high-income countries have a one-way relationship between FDI and CO₂ emissions. Ref. [7] uses the static panel estimation method to examine the relationship between FDI, pollution, and environmental quality in 28 Chinese provinces from 1997 to 2012, the finding documented that FDI, economic growth, and human capital all have a positive, significant impact on pollution emissions. Between 1990 and 2012 in the Middle East and North Africa (MENA), ref. [13] examined the causal relationships between EQ, economic growth and FDI in 17 nations, including Algeria, Morocco, Jordan, and Egypt. The results of the simultaneous analysis show that FDI has a significant positive impact on CO₂ emissions in all 17 countries and that there is a bidirectional relationship between FDI, economic growth, and CO₂ emissions. For the period from 1990 to 2013, robust OLS models, fixed panel corrected standard error models, and random effects models were used as estimation approaches to investigate and document the effect of FDI, institutions, and control quality on ecological sustainability in Africa. The findings showed that FDI and institutional quality had a significant positive (worsening) and negative (improving) impact on ecological sustainability, respectively. Additionally, it was discovered that the link between institutional quality and FDI enhanced environmental sustainability. In other words, it has a lot of detrimental impacts.

Seker, Fhah et al. [26], the triangular relationship between FDI, GDP, global tourist arrivals, energy consumption, and EQ in the Chinese economy was studied using periodical data beginning 1995Q1 to 2016Q4, and autoregressive distributed lag. The study discovered that FDI, GDP, and tourism arrivals have a positive short- and long-run effect on carbon emissions. Using the Dumitrescu–Hurlin panel causality test and the generalized method of moment as estimation approaches, ref. [7] examined the relationship between FDI and carbon emissions in MENA countries from 1990 to 2015. In contrast to MENA countries, the findings shown that FDI and economic growth have a positive and significant effect on carbon emissions. The square of foreign direct investment, economic growth, and biomass energy has a significant negative impact on CO₂ emissions. Ref. [18] examined the relationship between renewable energy consumption, economic growth, and foreign direct investment on CO₂ emissions in Pakistan from 1975 to 2016. While renewable energy consumption has a significant negative impact on carbon emissions, GDP, FDI, and the interaction between FDI and FDI all have a significant positive impact. Furthermore, the causality test revealed that the use of renewable energy has the opposite effect of increasing carbon emissions. More recently, ref. [27] used the Driscoll–Kraay standard error pooled ordinary least squares technique and the Dumitrescu and Hurlin panel causality test to examine the impact of foreign direct investment and financial development on environmental quality in 90 Belt and Road countries from 1990 to 2017. While FDI and trade openness have a significant negative impact on CO₂ emissions, financial development, energy consumption, and economic growth have a positive impact on environmental quality, according to the findings. The results of the causality test show a bidirectional causal relationship between foreign direct investment, financial development, energy consumption, openness, economic growth, and carbon emissions. Ref. [28] also looked at the relationship between economic growth, foreign direct investment, and carbon emissions in Asian countries between 1970 and 2014. The ARDL (autoregressive distributed lag) estimation method revealed that in the short and long run, energy consumption and economic growth had a significant positive impact on carbon emissions. The findings, however, revealed that foreign direct investment has little effect on carbon emissions in Asian countries. Ref. [10] used the GMM to assess the effects of globalization and foreign direct investment on environmental quality in OIC countries from 1991 to 2017. Industrialization, globalization, institutional quality, and foreign direct investment all have a significant positive impact on carbon emissions, according to the findings. However, the interaction of globalization and foreign direct investment has a negative effect on carbon emissions. In ref. [29], researchers examined the impact of FDI, renewable energy, and health spending on environmental degradation from 1995 to 2016 in 58 countries participating in the Belt and Road Initiative. While income and health spending have positive and significant effects on CO₂ emissions in two estimators, FDI and renewable energy use do not, according to FMOLS and the two-stage generalized moment approach empirical findings.

Gunarto, Toto [20], used spatial econometric models to examine the effects of FDI and technological innovation on environmental quality in 30 Chinese provinces from 1998 to 2016. They discovered that foreign direct investment and technological innovation have a significant negative impact on pollution. While there are numerous cross-national studies on erstwhile parts of the world, only a few [11,12,14,17,18] focus specifically on Sub-Saharan Africa. In order to help policymakers in East African countries implement their policies effectively, additional research is required to supplement earlier studies. It was also noted that earlier studies, particularly those on Africa, had downplayed the significance of a nation’s policies and institutions for environmental quality, environmental sustainability and had not discussed how FDI and those policies and institutions worked in tandem to affect environmental quality. Although some studies [10,17,19,27,30,31] attempted to use political institutions, which may not directly affect environmental quality given their broad scope, the results of these studies may prove policy implementation ineffective in comparison to environmental sustainability policies and institutions that center exclusively on environmental protection. Furthermore, considering that FDI primarily spreads across various economic sectors (extraction activities such as mining) that may have different environmental effects, most studies have actually used carbon emissions to measure EQ when analyzing the effect of FDI on EQ. This study measures environmental quality in addition to carbon emissions by looking at forest reserves and the depletion of natural resources in order to compare the diverse effects of FDI on environmental quality. According to the author’s knowledge and the empirical literature reviewed, no research has yet developed a merged index to fully evaluate the effect of FDI on EQ in East Africa. As a result, there is little information about policy implications. Against this backdrop, this study adds to the body of literature by developing an environmental quality index that provides policymakers in East African nations with the information they need to make environmental quality-based decisions. Therefore, this research hypothesis is advanced in this study: H1; FDI has a significantly positive impact on the intensity of CO₂ emissions in East African countries. H2; policies and institutions for environmental sustainability have a significantly positive impact on the intensity of CO₂ emissions in East African countries. H3; trade openness has a significantly positive impact on the intensity of CO₂ emissions in East African countries. H4; in East African countries, domestic investment significantly impacts CO₂ emission intensity. H5; forest reserve depletion has a significantly positive impact on the intensity of CO₂ emissions in East African countries. H6; natural resource depletion has a significantly positive impact on the intensity of CO₂ emissions in East African countries. H7; in East African countries, international tourism significantly impacts CO₂ emission intensity. H8; urbanization has a significantly positive impact on the intensity of CO₂ emissions in East African countries. H9; natural resource depletion has a significantly positive impact on the intensity of CO₂ emissions in East African countries.

3. Methodology

This section addresses the specification of the model variable descriptions and estimation procedures used.

3.1. Empirical and Theoretical Model Specification

STIRPAT model is used in this study to investigate how foreign direct investment, environmental sustainability policies and institutions (PIES), and their interactions affect environmental quality, as [21] documented. The model is based on Ehrlich and Holdren’s IPAT model [23], which has the main disadvantage of being a mathematical identity equation that frequently invalidates assumptions [24]. As a result, the IPAT model was converted into the STIRPAT (stochastic) model in [20], which resolves this issue. The STIRPAT model depicts the effects of technology, population growth, and wealth on the environment. Equation (1) below describes the STIRPAT model.

$$I_{it}=a{P}_{it}^{\beta }{A}_{it}^{\gamma }{T}_{it}^{\tau }{\epsilon }_{it}$$

(1)

I, P, A, and T stand for, respectively, population (as measured with urbanization), wealth (as measured with economic growth), and technical advancement. The constant term, the elasticity of population, wealth, and technological advancement on environmental quality, respectively, are denoted by $a, \beta , \gamma \mathrm{and} \tau$. $i, t \mathrm{and} \epsilon$ stand for cross-sectional units, time trend, and the stochastic error term.

It is assumed that FDI, domestic investment, and trade openness all have an impact on technological progress; therefore, $T=f\left(FDI,DI,TOP\right)$. The first equation can be transformed to equation two, as shown below:

$$I_{it}=a{P}_{it}^{\beta }{A}_{it}^{\gamma }FD{I}_{it}^{\tau }D{I}_{it}^{\phi }TO{P}_{it}^{\sigma }{\epsilon }_{it}$$

(2)

FDI, DI, and TOP represent foreign direct investment, domestic investment and trade openness, and $\tau , \phi \mathrm{and} \sigma$ are their respective elasticities. Equation (2) linear zed to Equation (3) are presented below:

$$ln{I}_{it}=a+\beta ln{P}_{it}+\gamma ln{A}_{it}+\tau lnFD{I}_{it}+\phi lnD{I}_{it}+\sigma lnTO{P}_{it}+{\epsilon }_{it}$$

(3)

The above equations, (1) and (2), elaborate on all DI, FDI, and TOP. According to Equation (3), domestic investment, foreign direct investment, urbanization, economic growth, and trade openness all have an impact on environmental quality. The general specification of Equation (3) is presented in Equation (4), which is presented as follows:

$$EQ=f\left(FDI,DI,TOP,URB, EG\right)$$

(4)

where: environmental quality, economic growth, and urbanization are represented, respectively, by EQ, EG, and URB. Equation was changed to include additional factors, including international tourism (IT) and PIES (4). Every nation implements laws and establishes organizations to safeguard the environment. To assess the degree to which these policies and institutions impact environmental quality, the policy and institution for environmental sustainability variables should be included in the environmental quality model. International travel has a favorable impact on EQ, and subsequently, deteriorates it, as reported by [25]. In light of the fact that there are many popular tourist destinations in African nations, we added international tourism to the model as well to investigate how it affects the environment in East African nations. In accordance with [4,19,23], this study additionally proposed an interaction term (PIES × FDI) between FDI and environmental sustainability policies and institutions to assess how FDI influence EQ while policies and institutions for environmental sustainability are enhanced. Equation (5) below gives the complete specification of the generalized model that includes this outcome:

$$EQ=f\left(FDI,PIESDI,IT,TOP,URB,EG,FDI \times PIES\right)$$

(5)

where: PIES, IT, and FDI × PIES are defined in Equation (5) seen above. This equation also transformed to a panel dynamic estimable form, as stated in Equations (6) and (7)

$$lnE{Q}_{it}=a_1+\lambda lnE{Q}_{it-1}+{\beta }_{1}PIE{S}_{it}+{\beta }_{2}FD{I}_{it}+{\beta }_{3}lnD{I}_{it}+{\beta }_{4}lnI{T}_{it}+{\beta }_{5}lnTO{P}_{it}+{\beta }_{6}lnUR{B}_{it}+{\beta }_{7}lnE{G}_{it}+{\gamma }_1{\left(FDI\times PIES\right)}_{it}+{\varrho }_i+{\eta }_t+{\epsilon }_{it}$$

(6)

$$E{Q}_{it}=a_2+\lambda E{Q}_{it-1}+{\delta }_{1}PIE{S}_{it}+{\delta }_{2}FD{I}_{it}+{\delta }_{3}lnD{I}_{it}+{\delta }_{4}lnI{T}_{it}+{\delta }_{5}lnTO{P}_{it}+{\delta }_{6}lnUR{B}_{it}+{\delta }_{7}lnE{G}_{it}+{\gamma }_2{\left(FDI\times PIES\right)}_{it}+{\varrho }_i+{\eta }_t+{\epsilon }_{it}$$

(7)

where:$a_1 \mathrm{and} a_2$ are the constant terms, $\left({\beta }_1,{\beta }_{2,} {\beta }_{3,}\dots ,{\beta }_7\right) \mathrm{and} \left({\delta }_{1,}{\delta }_{2,}{\delta }_{3,}\dots ,{\delta }_7\right)$ represent the variables’ estimated coefficients in Equations (6) and (7), respectively. ${\varrho }_i, ln,\epsilon \mathrm{and} {\eta }_t$ Represent the stochastic error term, which is normally distributed with amean of zero and constant variance, logarithm country-specific effect and time-specific effect respectively. If environmental policy and institutions improve, the lagged dependent variable coefficients ($\lambda , {\gamma }_1, \mathrm{and} {\gamma }_2$) and the interaction term capture the effect of foreign direct investment on environmental quality. It is worth noting that Equations (6) and (7) are estimated twice, with Models 2 and 1 in Equation (6) and 3, and 4 in Equation (7), respectively (7). Models 1, 2, and 3 estimate carbon dioxide emissions, forest reserve depletion, and natural resource depletion. Using principal component analysis, the estimation of the environmental quality index of Model 4 is produced. Since carbon dioxide emissions and the depletion of natural resources are discussed in logarithmic form in Equations (6) and (7), the difference between Model 1 and 2, and Model 3 and 4 is the reason for this equation.

3.2. Description of Variables and Data

This study analyzed balanced panel data from 18 East African nations from 2005 to 2020. East Africa was chosen because the influx of foreign direct investment into East African countries increased from time to time while the region was being plagued by chronic droughts at various times. The reason we chose 2005 is that there are insufficient data for the variable environmental sustainability policies and institutions prior to 2005, so we chose 2005 as the baseline year for all variables because this variable is critical to this study. The data for the variables used in this study, ${\mathrm{CO}}_2$ emissions, overexploitation of natural resources and forest reserves, environmental sustainability policies and institutions, foreign direct investment, domestic investment, international tourism, trade openness, economic growth, and urbanization, were collected from the World Bank following [32,33,34,35,36]. The following

Table 1. Description of variables and their sources.

Variable	Measurements	Definition	Notation	Sources
Carbon dioxide emission	Metric tons per capita	Emissions from the combustion of fossil fuels and the manufacture of cement. Emissions from the use of solid, liquid, and gaseous fuels, as well as gas flaring are also included.	CO2	WB-World Development Indicators (2022)
Forest reserve depletion	Net forest depletion (% of GNI)	The product of resource rent per unit and the excess of round wood harvest over natural increment.	FRD	WB-World Development Indicators (2022)
Natural resource depletion	Natural resource depletion (% of GNI)	The sum of net forest consumption, energy consumption, and mineral consumption.	NRD	WB-World Development Indicators (2022)
Foreign direct investment	Foreign direct investment net flows (% of GDP)	The net flow of investment to acquire a permanent interest in an enterprise operating in a different economy than the investor’s.	FDI	WB-World Development Indicators (2022)
Domestic investment	Gross fined capital formation (% of GDP)	Formerly gross domestic fixed capital formation. They include land improvements, the purchase of machinery, plants, and equipment, and the road, railroads construction, and other facilities.	DI	WB-World Development Indicators (2022)
International tourism	Number of arrival	The number of tourists travelling for a period not exceeding 12 months to a country other than that of their usual residence and outside their usual environment.	IT	WB-World Development Indicators (2022)
Trade openness	Trade (% of GDP)	The sum of goods and services exported and imported as a percentage of GDP.	TOP	WB-World Development Indicators (2022)
Urbanization	Urban population	Urban population defined by national statistical offices as people who live in urban areas.	URB	WB-World Development Indicators (2022)
Economic growth	GDP per capita (Constant 2010 USD)	Gross domestic product divided by midyear population.	EC	WB-World Development Indicators (2022)
Policy and institution for environmental sustainability	Policy and institution for environmental sustainability rating (1 = low to 6 = high)	Assessments of the extent to which environmental policies promote the conservation and sustainable use of natural resources, as well as pollution management.	PIES	WB-World Development Indicators (2022)

provides a brief description of the variables used in this study.

3.3. Estimation Technique

This study employs the estimation technique of the dynamic system’s generalized method of moments to investigate how FDI, environmental policy, and institutions, as well as their interactions, affect EQ. In our cross-sectional scenario, there were 16 years and 18 countries; therefore, the system-GMM is a suitable estimator for this study. This estimation technique, introduced by [14], was devised to achieve accurate estimates when T is small and N is large. Static panel estimators, such as fixed effect, random effect, pooled ordinary least squares, and the remainder, produce inaccurate and inconsistent results due to the dynamic nature of Equations (6) and (7) and the lag dependent variables ($lnE{Q}_{it-1} \mathrm{and} E{Q}_{it-1})$.

While Equations (6) and (7) are altered into their variation form, these static panel estimators fail to produce accurate estimates in a dynamic model due to an endogeneity problem caused by the inclusion of the lag dependent variable that is correlated with the stochastic error term (${\epsilon }_{it}$). The system-GMM estimating approach, which is an extension of the differenced-GMM [26,27], was chosen because it can be used to avoid this problem. Equation (8) specifies the generic system-GMM arrangement of our model, which is to be estimated as follows:

$$E{Q}_{it}-E{Q}_{it-1}=\lambda \left(E{Q}_{it-1}-E{Q}_{it-2}\right)+{\beta }^{\prime }\left(X_{it}-X_{it-1}\right)+({\epsilon }_{it}-{\epsilon }_{it-1})$$

(8)

By performing the Arellano–Bond test [12,20] for second-order serial correlations [AR (2)] and the test for the reliability of instruments, the correctness of the system-GMM estimate is evaluated (over-identification restriction). The instruments are correct and do not show second-order serial correlation; therefore, the estimates are consistent in the cases where the [12,20] tests do not succeed in rejecting the null hypothesis and the AR (2) test does not succeed either. We use the Delta-Method [${\beta }_k/\left(1-\lambda \right)]$ to determine the long-run coefficient from the short parameters since the long-run effect is more significant for policy implications [28]. To achieve this, we divided the lagged coefficients of the dependent variable by one less than the short-run coefficients.

Before using the system-GMM, a few preliminary tests, such as the cross-sectional dependence test, unit root test, and cointegration test, must be performed. This will make it easier to assess if there are dependencies in the data from the 18 nations under study without bias, false results, or conflicting findings [29]. In this test, the alternative hypothesis of cross-sectional dependence is contrasted with the null hypothesis of cross-sectional independence. The test follows the standard normal distribution with unit variation [i.e., CD$~N(0,1)$] and a mean of zero for a large cross-section and a small time dimension. The CD test is shown below in Equation (9):

$$CD=\sqrt{2T}/n\left(n-1\right)\left[{{\displaystyle \sum }}_{i=1}^{N-1}{{\displaystyle \sum }}_{j=1+1}^{N}pij\right]$$

(9)

Following the cross-sectional dependence test, the study continues on to determine the variables’ stationary properties. We use non-parametric and parametric unit root tests based on [37] to settle on the unit root of the variables. Due to statistical weakness, we use the Fisher unit root test in [29,38,39], and to relate to cross-sectional dependence we use the CIPS unit root test proposed by [29]. In these tests, the rejection of the null hypothesis that the panel contains a unit root (non-stationary) implies that the panels are homogeneous. However, rejection of the null hypothesis of non-homogeneous stationarity implies heterogeneous stationarity for the CIPS unit root test. To determine whether there is a long-run panel association with the sample variables, this study applies the panel cointegration test described in [40] and Kao’s cointegration as a robustness check test. The null hypothesis in both the Padroni and Kao tests is that there is no cointegration (i.e., no long-run relationship between variables) as opposed to the alternative hypothesis, which is that there is a long-term association (cointegration). Whether the null hypothesis is rejected or not, there is a long-run relationship between the variables (no cointegration).

Ref. [17] claims that it is crucial to identify the marginal effect and its importance when seeking to account for the interplay between FDI and PIES. Nevertheless, we compute the partial derivative of EQ in relation to the foreign direct relationship and calculate the standard error of the marginal effects to assess the relevance of the FDI marginal effect in estimable models. The following equation illustrates the generic specification of the partial derivatives for our estimable model:

$$\frac{\delta E{Q}_{it}}{\delta FD{I}_{it}}=\beta +\gamma PIES$$

(10)

Rather than focusing on individual environmental aspects, this study utilized panel principal component analysis to develop an index based on the three environmental quality methods in order to measure the overall effect of FDI, the impact of policies and institutions, and their interaction on the environment (forest reserve extraction, carbon emissions, and natural resource extraction). The normalized formula established by [41] is used to calculate the environmental quality index. It is not necessary to add that the weight for the sub-indices is determined by the relative standard deviation using the following formula:

$$EQ{I}_t={\mathcal{Q}}_1\left(\frac{C{O}_{2t}-minC{O}_2}{maxC{O}_2-minC{O}_2}\right)+{\mathcal{Q}}_2\left(\frac{NR{D}_t-minNRD}{maxNRD-minNRD}\right)+{\mathcal{Q}}_3\left(\frac{FR{D}_t-\min FRD}{maxFRD-minFRD}\right)$$

(11)

EQI stands for environmental quality index, CO₂ is carbon dioxide emissions, NRD is natural resource depletion, FRD is forest reserve depletion, and Q stands for relative standard deviation and environmental quality index. The EQI is normalized to a number between 0 and 1, with a value closer to 1 indicating higher environmental quality and a value closer to 0 indicating lower environmental quality.

4. Result and Discussion

In this section, we go over and evaluate the outcomes of the empirical assessment. We begin by discussing descriptive statistics and variable correlations. Following that, the preliminary test results are discussed. The long-run and short-run effects of the interaction between foreign direct investment and environmental policy and institutions, as well as the marginal effect, are then examined. The report on principal component analysis is presented and discussed in the final section.

4.1. Descriptive Statistics

Table 2. Descriptive statistics.

Variable(s)	Mean	Standard Deviation	Minimum	Maximum
lnCO2	−1.5251	1.2984	−3.8938	1.8323
lnNRD	0.8786	2.1412	−7.3907	3.4967
FRD	6.8014	7.0999	0.0006	33.0086
EQI	−8.3800	1.0000	−4.5440	1.6155
PIES	3.0052	0.6723	1.0000	4.5000
FDI	4.6566	6.5039	−4.3035	57.8773
lnDI	2.9550	0.4977	0.6933	3.7591
lnTOP	4.1422	0.6187	3.1019	5.8521
lnIT	13.0078	1.2517	8.8536	14.8698
lnEG	6.9518	1.0247	5.6008	9.7403
lnURB	14.7304	1.5405	10.6646	17.0320

Source: authors’ evaluation.

below depicts that the average values for CO₂ emissions are 1.52, natural resource depletion is 0.87, and forest depletion is 6.80; it is observed that, on average, there is a relatively small dispersion around the mean values of the variables. The standard deviation of the variables does not deviate significantly from the mean values because there is less dispersion around the mean values. A maximum and a minimum value 57.87 and −7.39 are also found in the sample data.

4.2. Correlation Analysis

Table 3. Correlation matrix.

	lnCO2	lnNRD	FRD	EQI	PIES	FDI	lnDI	lnTOP	lnIT	lnEG	lnURB
lnCO2	1
lnNRD	−0.775	1
FRD	−0.630	0.675	1
EQI	0.199	−0.252	−0.269	1
PIES	−0.088	0.057	−0.142	0.462	1
FDI	0.254	−0.226	−0.217	0.354	0.065	1
lnDI	0.199	−0.252	−0.269	0.495	0.462	0.354	1
lnTOP	0.586	−0.470	−0.283	0.189	−0.253	0.483	0.189	1
lnIT	0.010	0.004	−0.049	0.028	0.418	0.022	0.028	−0.147	1
lnEG	0.931	−0.777	−0.604	0.253	−0.150	0.210	0.253	0.636	−0.065	1
lnURB	−0.522	0.486	0.254	0.034	0.328	−0.145	0.034	−0.430	0.561	−0.626	1

Source: authors’ evaluation.

shows the correlation matrix between variables. We come across the fact that foreign direct investment and environmental sustainability policy and institution (PIES) are positively correlated with measures of environmental quality, except for foreign direct investment, which tends to be negatively correlated with measures of forest reserve destruction and natural resource depletion. The correlation between domestic investment and CO₂ emission and the environmental quality index tends to be positive, while the correlation between domestic investment and forest reserve depletion tends to be negative. On average, the correlation matrix shows that, with few exceptions, the correlation between variables is less than 0.50, indicating that there is less likelihood for multicollinearity to exist in the dataset employed.

4.3. The Panel Cross-Sectional Dependencies Test

Table 4. Test result for cross-sectional dependencies.

Variables	CD Test	p-Values
lnCO2	1.6864	0.045
lnNRD	1.4355	0.035
FRD	1.6530	0.049
EQI	3.1780	0.000
PIES	0.9178	0.179
FDI	4.4229	0.000
lnDI	3.1780	0.000
lnTOP	0.3526	0.362
lnIT	1.1736	0.879
lnEG	5.3179	0.000
lnURB	5.1331	0.000

Source: authors’ evaluation.

below documented that of the variables we used in this study, only three variables, namely environmental sustainability policies and institutions, trade openness, and international tourism, exhibit cross-sectional independencies across the 18 East African countries. This is because the probability values (0.179, 0.362, and 0.879) of these variables exceed the conventional 5% significance level, implying a non-rejection of the null hypothesis of cross-sectional independencies. Apart from these variables, all other variables give an indication of cross-sectional dependencies, given that the probability values lead us to rejection of the null hypothesis at a high significant level (1%). With that we can say that there exist cross-sectional dependencies across the 18 East African nations as most of the variables indicated cross-sectional dependencies, and thus, policymakers should consider cross-sectional dependencies among countries when implementing policies.

4.4. Panel Unit Root Test

The results of the unit root test are shown in

Table 5. Unit root test result.

	IPS		Fisher		HT		CIPS
Variable(s)	I(0)	I(1)	I(0)	I(1)	I(0)	I(1)	I(0)	I(1)
lnCO2	1.973	−6.908 ***	19.520	164.696 ***	0.913	−0.013 ***	−2.689	−3.736 ***
lnNRD	0.161	−6.610 ***	37.216	134.421 ***	0.619	0.046 ***	−1.406	−3.006 ***
FRD	0.355	−7.110 ***	33.030	151.135 ***	0.789	−0.043 ***	−1.777	−3.710 *
EQI	−0.671	−6.831 ***	8.667	153.192 ***	0.702	−0.051 ***	−2.008 *	−3.637 *
PIES	1.634	−7.453 ***	17.903	109.267 ***	0.795	−0.171 ***	−1.852 *	−3.270
FDI	−2.765	−7.937 ***	1.399	178.988 ***	0.345	−0.375 ***	−2.882	−4.549
lnDI	−0.671	−6.831 ***	8.667	53.192 ***	0.702	−0.051 ***	−2.008 *	−3.637
lnTOP	0.833	−4.981 ***	6.885	02.010 ***	0.793	−0.102 ***	−1.330	−3.396 *
lnIT	0.122	−4.744 ***	46.280	195.034 ***	0.577	−0.148 ***	−2.029	−3.211
lnEG	−1.535	−1.079	65.899	69.460 ***	0.883	0.074 ***	−1.707 *	−1.980 *
lnURB	2.576 *	−13.602 ***	27.813 *	46.457 ***	1.005 *	0.384 ***	−1.007 *	−1.952 *

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

below Despite cross-sectional dependencies between variables, we account for them using Pesaran’s CIPS unit root test. A different unit root test is used for consistency and robustness.

From Table 5 above, it is observed that all sample variables are found to be stationary, with all tests (IPS, H-T, Fisher, and CIPS) indicating that carbon dioxide emissions, natural resource depletion, forest reserve depletion, trade openness, environmental quality index, and urbanization are stationary at the first difference [I(1)], whereas the H-T, Fisher, and IPS tests show that international tourism is stationary at the first difference [I(1)]. IPS, Fisher and IPS tests also show that foreign direct investment, domestic investment, and policy and institution for environmental sustainability are stationary at the first difference level.

4.5. The Panel Cointegration Test

After valid confirmation of panel stationarity among variables, the study proceeds to establish whether there exists a long-run relationship between the variables using the cointegration test of Pedroni’s and Kao’s tests [22,42,43].

Table 6. Result of panel cointegration test.

Pedroni Cointegration
	Model One	Model Two	Model Three	Model Four
Panel V	−1.637 ***	−2.243 ***	−2.330 ***	−2.278 ***
Panel rho	2.491 ***	2.746 ***	2.711 ***	1.039 ***
Panel P-P	1.366 ***	2.005 ***	1.904 ***	−1.083 ***
Panel ADF	5.893 ***	4.589 ***	3.944 ***	1.331 ***
Group rho	3.646 ***	4.521 ***	4.519 ***	2.998 ***
Group P-P	3.182 ***	3.318 ***	3.302 ***	3.389 ***
Group ADF	4.447 ***	4.713 ***	3.593 ***	−3.109 ***
Kao Cointegration
ADF	1.994	−2.714	1.479	−1.991

Note: *** represents the significance at the 1% level.

shows that there is the validation of a cointegration relationship among the dependent variable (environmental quality) and the explanatory variables in all estimable models, such as 1, 2, and 3, as shown by Pedroni’s and Kao’s cointegration test. This is because the test statistics in both tests indicate significance at 1% and 5% error level. This implies a rejection of the no cointegration null hypothesis. Given the presence of a long-run relationship among the variables, the study continues to estimate the long-run and short-run results, as well as the marginal effect of FDI on environmental quality using the generalized methods of moments (System-GMM) techniques of estimation.

4.6. Marginal Effect and Their Estimation Results

The long-run results and their marginal effects are reported in

Table 7. The long-run estimate-GMM.

	Model One	Model Two	Model Three	Model Four
PIES	0.4131 **	−1.003 ***	−1.640 ***	1.2006 ***
	(0.1216)	(0.232)	(0.237)	(3.7008)
FDI	−0.124 ***	0.751 ***	−0.755 ***	2.7909
	(0.00894)	(0.0957)	(0.0975)	(1.5408)
FDI*PIES	−0.4341	0.314 ***	0.320 ***	7.1810
	(0.0286)	(0.0306)	(0.0311)	(5.1009)
lnIT	0.350 ***	−0.126 ***	1.251 ***	2.8307 ***
	(0.0118)	(0.027)	(0.129)	(4.5908)
lnDI	−0.411 ***	1.569 ***	1.637 ***	2.009 ***
	(0.0269)	(0.289)	(0.294)	(1.0407)
lnTOP	0.0861 ***	2.077 ***	2.709 ***	3.3908
	(0.0235)	(0.252)	(0.257)	(5.8008)
lnURB	0.453 ***	1.241 ***	1.361 ***	1.3306 ***
	(0.0115)	(0.124)	(0.126)	(1.2407)
lnEG	0.591 ***	−6.249 ***	−6.843 ***	4.4208
	(0.0175)	(0.188)	(0.191)	(1.7407)

Note: *** and ** represent the significance at the 1% and 5% levels, respectively.

and

Table 8. The long-run marginal impact of foreign direct investment on environmental quality as institution for environmental sustainability improves.

Percentiles	Percentile Values	Coefficient	Standard Error	Confidence Level (95%)
Model 1
5%	0.0446	0.3425	0.2342	0.0394	0.1049
25%	0.0446	0.3425	0.2342	0.0394	0.1049
50%	0.1638	−1.4532 ***	0.2657	−0.0452	0.1950
75%	0.3774	−1.2341 ***	0.4586	−0.0833	0.4868
95%	0.3774	−1.2341 ***	0.4586	−0.0833	0.4868
Model 2
5%	0.0080	0.3017 **	0.5147	0.0030	0.1110
25%	0.0080	0.3017 **	0.5147	0.0030	0.1110
50%	4.8561	1.4532	0.3478	3.7684	5.6393
75%	9.6317	1.2341	0.4384	7.9068	6.1848
95%	9.6317	1.2341	0.4384	7.9068	6.1848
Model 3
5%	0.0080	−0.2142 ***	0.0152	−0.0030	0.1110
25%	0.0080	−0.2142 ***	0.0152	−0.0030	0.1110
50%	4.6692	−1.4532 ***	0.0954	−0.1886	1.6871
75%	2.7716	−1.2532 ***	0.0454	−0.0386	1.6871
95%	2.7716	−1.2532 ***	0.0454	−0.0386	1.6871
Model 4
5%	−1.7426	−0.2135	0.3452	−0.7441	−0.5182
25%	−1.7426	−0.2135	0.3452	−0.7441	−0.5182
50%	0.0747	0.2532	0.2854	0.0533	0.2639
75%	0.7030	1.2341	1.1217	0.5816	0.9077
95%	0.7030	1.2341	1.1217	0.5816	0.9077

Note: *** and ** represent the significance at the 1% and 5% levels, respectively.

, respectively, while the short-run result and their marginal effects are shown, respectively in

Table 9. The GMM short-run results.

	Model 1	Model 2	Model 3	Model 4
L.lnEQI(CO2)	1.0908 **
	(0.1907)
L.lnEQI(NRD)		0.1310
		(0.1709)
L.lnEQI(FRD)			1.1209 ***
			(0.1909)
L.lnEQI(EQI)				1.2406 ***
				(0.1308)
FDI	0.1506 ***	0.2906 ***	0.2306 ***	1.1809 ***
	(0.9808)	(0.6908)	(0.9508)	(0.7210)
FDI*PIES	−0.0309	1.2108	−1.2209	1.2406 ***
	(0.7709)	(1.9408)	(0.7609)	(0.3308)
PIES	−0.3909 ***	−0.2306 ***	0.6109 **	1.2406 ***
	(0.0610)	(0.1808)	(2.0609)	(0.3308)
lnIT	0.1307 ***	0.3207 ***	1.2307 ***	0.2907 ***
	(0.2209)	(0.3108)	(0.4308)	(0.2908)
lnDI	0.039 ***	−1.129 ***	−1.349 ***	1.1039 ***
	(0.7308)	(1.4407)	(0.9908)	(0.7708)
lnTOP	1.7608 ***	0.5708	0.9208	0.2408
	(0.0508)	(0.0108)	(0.6808)	(0.5808)
lnURB	−1.1306 ***	−1.1806 ***	−1.1706 ***	−1.1506 ***
	(0.8608)	(0.9108)	(0.3308)	(0.2608)
lnEG	1.4008	−0.5908	−0.2208	−1.6608
	(1.6007)	(0.9108)	(0.4708)	(0.6808)
Constant	0.379 ***	−0.9026	−0.9120 ***	−0.937 ***
	(0.6307)	(0.8018)	(0.586)	(0.6307)
No. of observations	288	288	288	288
No. of groups	18	18	18	18
No. of instruments	17	17	17	17
AR(2) p-Value	0.928	0.086	0.855	0.759
Hansen p-Value	0.854	0.075	0.276	0.296

Note: *** and ** represent the significance at the 1% and 5% levels, respectively.

and

Table 10. The short-run marginal effects of FDI on EQ as policy and institution for environmental sustainability look up (is at its mean).

Percentiles	Percentile Values	Coefficient	Standard Error	Confidence Level (95%)
Model1
5%	0.0446	0.2653	0.2351	0.0394	0.0504
25%	0.0446	0.2653	0. 2351	0.0394	0.0504
50%	0.1638	0.4532 **	0.0636	0.0452	0.1950
75%	0.3774	0.2341 **	0.0535	0.0833	0.4868
95%	0.6425	0.3420 ***	0.0535	0.0833	0.4868
Model 2
5%	0.0080	0.1243 ***	0.0123	0.0030	0.1110
25%	0.0080	0.1243 ***	0.0123	0.0030	0.1110
50%	0.8561	1.4532	0.3434	0.7684	1.6393
75%	0.6317	1.2341	0.2219	0.6716	1.0146
95%	0.6317	1.2341	0.2219	0.6716	1.0146
Model 3
5%	0.7426	0.2142 ***	0.0232	0.0030	1.6627
25%	0.7426	0.2142 ***	0.0232	0.0030	1.6627
50%	0.0747	1.4532 ***	0.0236	−0.6386	1.6871
75%	0.4205	1.2341 *	0.0147	0.5350	1.0146
95%	0.4205	1.2341 *	0.0147	0.5350	1.0146
Model 4
5%	−0.7426	−0.2221	0.2342	−0.2125	−0.3792
25%	−0.7426	−0.2221	0.2342	−0.2125	−0.3792
50%	0.0174	0.0234 *	0.2854	−0.0533	0.2639
75%	0.5030	0.0343 *	1.1217	0.3244	0.5045
95%	0.4205	0.2125	1.1217	0.3244	0.5045

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

. We first analyze the long-run results and their marginal effect, and then followed with the short-run result and their marginal effects. In this analysis, it is important to note that a positive and a negative effect on carbon emission, natural resource depletion, forest reserve depletion and environmental quality index means deterioration of environmental quality and improvement in environmental quality, respectively.

In Table 7 Model 1, where environmental quality is measured by carbon emission, it is observed that policy and institution for environmental sustainability are positively and significantly associated with CO₂ emissions, herby, deteriorating environmental quality. However, we find the effect of policy and institution for environmental sustainability on natural resource depletion (Model 2) and forest reserve depletion (Model 3) to be negative and significant, i.e., improving environmental quality, and policy and institution for environmental sustainability positively and significantly associated with CO₂ emissions, herby, deteriorating environmental quality (Model 4) in the long run.

In particular, Model 1 indicates that an increase in environmental sustainability policies and institutions worsens environmental quality (increase in carbon dioxide emissions) by 0.4 percent at the 1 percent error level. This outcome can be attributed to the fact that the attitude of policy and institutions towards carbon dioxide emission in East African countries are weak, and this could explain why CO₂ emissions continue to increase in East African countries. Additionally, the coefficients in model two and model three show that improving environmental sustainability policies and institutions improves environmental quality in East African countries by reducing the depletion of natural resource and the depletion of forest reserve by 1% and 0.03% at 1% and 10% significance level, respectively. These findings highlight the importance of policies and institutions in influencing EQ in East Africa, and policymakers should consider policies that improve environmental quality, which may have a positive impact on the region’s persistent drought. The negative impact of environmental policies and institutions is consistent with various authors, such as [9,19,27], while the positive impact is consistent with [10,17,23].

When it comes to FDI, we observe that unconditional impacts, i.e., the best strategy for improving environmental quality in Models 1 and 3, are to reduce forest depletion, and carbon emissions (Table 8). Environmental quality declines in Model 2 due to increased resource depletion, and the overall impact on the ecosystem also has its effect (Model 4). Consider the marginal effect of the interaction term to understand the precise effect of FDI on EQ. This is because the estimable model’s partial derivative shows how FDI affects environmental quality after environmental sustainability institutions and policies have reached their mean (see Equation (10)).

In Table 8, we estimate the long-run marginal effect and find that an increase in FDI improves environmental quality in Models 1, 2, and, 3. FDI, on the other hand, degrades environmental quality by depleting natural resources in Model 2. In particular, in Model 1, the coefficient shows that when environmental sustainability policies and institutions in East African countries are equal to the mean/improvement, FDI improves environmental quality and reduces carbon emissions by 1.45 and 1.23 percent at the 50th to 95th percentiles, respectively, when FDI increases by one point in East African countries. Model 3 (forest reserves) indicated in the coefficients that a particular increase in FDI improves ecological quality by 0.21, 1.45, and 1.25 points, respectively, at all percentile levels when environmental sustainability policies and institutions in East African countries are at their mean/improvement levels, and they are significant at the 1% level error when all other covariates are kept constant. These findings support the pollution halo hypothesis, which holds that foreign firms adopt modern technologies, improving environmental quality, and is supported by the finding of [15].

When environmental sustainability policies and institutions in East African countries reach/improve their mean, the coefficient in Model 2 shows that foreign direct investment worsens environmental quality by 0.30% at the 5th, and 25th percentiles for each supplementary increase in foreign direct investment, and it is significant at the 5% error level when all other variables are kept constant. This finding supports both the scale effect and the pollution hypothesis. When environmental sustainability policies and institutions in East African countries reach their mean or improvement, we found that FDI improves the environment at the 5th and 25th percentiles by 0.21 points, but it worsens the environment from the 50th to the 95th percentiles by nearly 0.25 and 1.23 points, when all covariates are held constant, which may be due to the different signs in Table 8.

The short-run result in Table 9 shows that all lag environmental quality metrics, such as carbon emissions, depletion of natural resource, depletion of forest reserve, and the index, have a positive significant effect on environmental quality. However, natural resource depletion has no significant effect, which is insignificant. The coefficient depict that a previous one percent increase in carbon emission, that is, forest depletion and environmental quality index, induces carbon emission (forest reserve depletion and the environmental quality index) to rise by about 1.09 percent, 1.12 percent, and 1.24 percent, respectively at a 1% error level for Model two and three and at a 5% error level for Model one when all other factors are constant. This estimate validates the findings of [9,10,16,17].

The short-run effects of foreign direct investment on environmental quality were evident in Models 1, 2, 3, and 4. These effects included a rise in carbon emissions, the depletion of natural resources, the loss of forest reserves, and a decline in the index value. The coefficient demonstrates that when other factors/variables are kept constant, the coefficient shows that an increase of one percent in environmental sustainability-related institutions and policies results in increases in carbon emissions, natural resource depletion, forest reserve depletion, and the index about 1.2, 1.3, 1.23, and 5.5 percent, respectively.

In both Models 1 and 2, we find that policies and institutions for environmental sustainability improve environmental quality, particularly by lowering carbon emissions and resource depletion. When all other variables remain constant, a 1% improvement in environmental institutional and policy sustainability reduces carbon emissions and natural resource depletion by about 0.39 and 0.23 percent, respectively. The impact on forest reserves, which is similar to the long-run result, and on the environmental quality index are also significant. In East African countries, policies and institutions aimed at environmental protection and the need to pay attention to them to improve environmental quality also play an important role in short-run environmental sustainability. This effect is in line with the findings of [9,15,16,19].

In all models (Models 1, 2, 3, and 4) at a 1% significance level, a percentage increase in international tourism is linked with 0.13, 0.32, 1.23, and 0.29 percentage points, respectively. The results of study [26] support this conclusion. In the long run, trade openness degrades environmental quality in Model 1, but not in Models 2, 3, or 4. In Model 1, a percentage increase in trade openness results in a 1.76 percent decrease in environmental quality, that is, a 1%significant level, this result is supported by the finding of [44]. Environmental quality in East African countries worsens with further economic growth (models 1, 2, 3, and 4). Environmental quality in East African nation is improved by urbanization in models 1, 2, 3, and 4. According to the coefficient, environmental quality in East African countries improves by about −1.13, −1.18, −1.17, and −1.15 percentage points when urbanization boost by 1% and it is significant at the 1% level for all models regardless of other variables.

It is possible that environmental policies and institutions in East African countries are less stringent in monitoring or auditing foreign direct investment flows and evaluating foreign companies using modern technology, resulting in FDI degrading the quality of the environment in East African countries. Therefore, the short-run result confirms the hypothesis that environmental degradation in developing countries is due to less stringent policies. These results are similar to the findings of [7,8,13,15,18].

Models 2 and 3 show that DI improves environmental quality in the short run. Owing to their impacts in Models 1 and 4 (carbon emissions), the environmental quality is worsened, which is different from the long-term results. As a result of an increase in DI, environmental quality improves by about 1.12 points in Model 2, and by 1.34 points in Model 3, and 1% is significant level. The coefficients in Models 1 and 4 demonstrate that environmental quality declines by 0.03 and 1.10 points/percent when DI rises by 1% while holding all covariates constant. The improvement in environmental quality in Models 2 and 3 may be explained by long-term findings that domestic investors who care about their wellbeing do not engage in activities that degrade environmental quality as much. The result registered in Models 1 and 4 supports the findings of [5,8,19], but the result in Models 2 and 3 contradicts the findings of [19,20,45].

The second-order {AR(2)} serial correlation and the over-identification test in Table 10 above show that there is lack of a second-order serial correlation and there exist valid instruments, in all the estimable models, i.e., Models 1, 2, 3, and 4. This is because, the probability value of the AR(2) [0.928, 0.086, 0.855 and 0.759] and the Hansen test [0.854, 0.075, 0.276, and 0.296] fail to reject the null hypothesis of no second-order serial correlation and instrument validity at a 5% level of significance. This reveals that the accuracy and consistency of the estimable models’ parameters.

Basing the finding of the short-run marginal effects of FDI on EQ as policy andinstitution for environmental sustainability, short-run result show that contrary to the long-run result, international tourism deteriorates, which has a positive effect on environmental quality, which is significant in all models. Model 1, 2, and 3 represent the estimated model CO₂ emission, natural resource depletion and forest reserve depletion as a measure of environmental quality, respectively; whereas Model 4 denotes the estimated model with the environmental quality index as a dependent variable.

Contrary to the long-run result we notice that the unconditional and conditional impacts of FDI deteriorate environmental quality in the short run in all models (Model 1, 2, 3, and 4) stated in Table 10 above. Specifically, the marginal effect, which is the conditional effect of FDI, in

Table 11. The analysis report of principal component.

Environmental Quality Index	Eigen Val.	Proportion Explain	Primary Var.	Eigen Vect.	Correlation Coef.	Bartlett’s p-Val.
Component 1	0.3455	0.1337	CO2	−0.2166	0.7286	0.000
Component 2	1.0678	0.2791	NRD	0.5587	0.4912
Component 3	1.2357	0.3935	FRD	0.3576	0.5635

Source: authors’ evaluation.

below reveals that if all other variables are held constant, a single increase in FDI will deteriorate environmental quality, i.e., increase carbon emission, natural resource depletion, and forest reserve depletion and decrease the index value closer to zero by 0.45, 0.23, and 0.34 point values at the 50th–95th percentile in Model 1; 0.12 point value at the 5th–50th percentile in Model 2; 0.21, 1.45, 1.23 values at the 5th–95th percentile in Model 3; and 0.02 and 0.03 point values at the 50th–75th percentile in Model 4, if policy and institutions for environmental sustainability are at their mean (improve) and are all significant at the 5% and 1%, 1% and 10% and 10% significance levels in Model 1, 2, 3, and 4, respectively.

4.7. Principal Component Analysis Report

As shown in the Table 11 above, all components are strongly positively related to the environmental quality index. Additionally, it was discovered that natural resource depletion and carbon emissions are anticipated index for the variables because their Eigen values (1.08470 and 1.4809) are bigger than one and they make up a higher share of the created environmental quality index (about 0.50 and 036). When adopting environmental regulations, it is important to keep in mind that carbon emissions and the depletion of natural resources are important environmental challenges in East African nations. We draw the conclusion that the three variables utilized to calculate the index are connected based on the Bartlett probability value of 1%.

5. Conclusions and Policy Suggestions

Persistent drought is not a unique phenomenon in East African countries, different research results cite different reasons for it, but the environmental problem is currently a major concern worldwide, and in East African countries the problem is not an exception to this phenomenon; therefore, using a balanced panel of 18 East African nations from 2005 to 2020, this study looks into the relationship between foreign direct investment and environmental policies. As an estimating method, using long-term GMM estimation, environmental quality is measured by carbon emissions, and environmental sustainability policies and institutions are found to be positively and significantly associated with CO₂ emissions, and thus, with environmental degradation. However, it was found that the effects of environmental sustainability policy and institution on natural resource depletion (Model 2) and forest reserve depletion (Model 3) are negative and significant, i.e., improve environmental quality, and that environmental sustainability policy and institution are positively and significantly associated with CO₂ emissions and environmental quality degradation in the long-run (model 4). Specifically, Model 1 shows that an increase in environmental sustainability policies and institutions worsens environmental quality (increase in carbon dioxide emissions) by 0.4 percent at the 1 percent error level. This result can be attributed to the fact that policy and institutional attitudes toward carbon dioxide emissions are weak in East African countries, which could explain why CO₂ emissions continue to increase in East African countries. The negative impact of environmental policies and institutions is consistent with various authors, such as [9,19,46,47,48,49], while the positive impact is consistent with [10,17,23,50,51].

Regarding FDI, we note that the unconditional impact, i.e., the best strategy to improve environmental quality is to reduce forest depletion and carbon emissions (in Models 1 and 3, Table 8). Environmental quality decreases due to increasing resource depletion, and overall ecosystem impacts also have their effects (in Model 2 and 4). Consider the marginal effect of the interaction term to understand the exact impact of FDI on environmental quality. On the other hand, an increase in FDI improves environmental quality in Models 1, 2, and 3, while FDI worsens environmental quality in Model 2 due to natural resource depletion. Specifically, in Model 1, the coefficient shows that when environmental sustainability policies and institutions in East African countries are equal to the mean/improvement, FDI improves environmental quality and reduces carbon emissions by 1.45 and 1.23 percent at the 50th to 95th percentile, respectively, when FDI increases by one point in East African countries, this finding supported by the findings of [45,50,51,52,53,54,55]. In Model 3 (forest reserves), the coefficients suggest that a given increase in FDI improves ecological quality by 0.21, 1.45, and 1.25 points, respectively, at all percentile levels when environmental sustainability policies and institutions in East African countries are at their mean/improvement levels. These results support the pollution halo hypothesis that foreign companies adopt modern technologies, and thereby, improve environmental quality, and are supported by the results of [15,56,57,58,59,60,61,62].

Moreover, while all models show a short-run worsening of environmental quality due to foreign travel, only Model 2 shows a long-term improvement. Trade openness worsens environmental quality in all scenarios, both in the long and short run. While urbanization worsens environmental quality in the long run and improves it in the short run in all models, economic growth improves environmental quality in the long run (models 2 and 3). Considering that they improve environmental quality and FDI is balanced to improve environmental quality over time, the study concludes that policies and institutions for environmental sustainability are worthwhile in East African countries. The study also concludes that domestic investment is also important for improving environmental quality in East African countries.

Based on how policies and institutions affect environmental sustainability, the study makes policy recommendations for governments, policymakers, and other stakeholders in East African countries to enact stronger environmental laws. This is possible if institutions and policies that prioritize environmental quality are free from opinion bias and dishonesty. As a result, companies, especially international companies, will be forced to adopt technologies that do not harm the environment in East African countries. According to the study, domestic investments have a positive impact on environmental quality. Therefore, governments and other stakeholders in East African countries should develop policies that encourage domestic investment to improve environmental quality. This is possible if East African governments and stakeholders work together to create a friendlier business environment.