The purpose of this study is to analyze the impact of gold, oil and timber extractions on the environment, second, to test the Environmental Kuznets Curve (EKC) hypothesis, and finally, to compare the effect of institutional regulation and military intervention on natural resource extraction and its impact on the environment in Ghana.
The Autoregressive Distributed Lag and Seemingly Unrelated Regression estimators were used.
It was revealed that gold, timber and oil production lead to an increase in the ecological footprint in Ghana. It was also revealed in the EKC hypothesis that gold and oil production increase ecological footprint, while timber production exhibited the EKC hypothesis. Also, institutional interventions on gold and oil extraction had a positive relationship with ecological footprint.
Other natural resources, such as bauxite and diamond, can also be looked at. Spatial analysis can also be adopted to determine the effects of resource extractions at the district level.
The paper highlights the need for an effective institutional intervention to mitigate environmental pollution from natural resource extraction.
The study disaggregated natural resources into gold, oil and timber. Also, the production levels of resources were used. The ecological footprint was used as a measurement for environmental pollution.
Glossary
- Autoregressive Distributed Lag (ARDL)
This is an econometric approach to analyze the long-run relationships between variables in time series or panel data analysis
- Error correction mode (EC)
This is another ARDL econometric approach to assess both the long-run and short-run relationship between variables in time series or panel data analysis
- Cointegration
This is a statistical approach to test if the variables used for the study are suitable for a long-run relationship.
- Ecological footprint (ECF)
This is a measurement of environmental pollution that includes the effects of all human activities, including natural resources extraction, on the environment
- Environmental Kuznets Curve (EKC)
A hypothesis that suggests environmental pollution increases with economic growth, reaches a turning point and then pollution declines after a certain level of development
- STIRPAT (Stochastic Impacts by Regression on Population, Affluence and Technology)
This is an analytical model used to assess the environmental impacts of population, affluence and technology
Introduction
Every economy in Africa and the world at large has its means of improving the standard of living of its people, and Ghana is no exception (Omodero, 2018; Tang et al., 2022). Yet Ghana largely borrows from external and internal sources to fund its projects and improve the livelihoods of the citizens of the country. However, the overdependence on debt leads to an appalling crisis when it reaches its threshold (Prah, 2022). Hence, to limit the dependence of Ghana’s economy on foreign aid, there is also the need to supplement the revenue generated from natural resource extraction (Adabor et al., 2022; Brunnschweiler et al., 2021).
Hence, Ghana is so endowed with an abundance of natural resources that are found across the nation, including in its northern, southern, eastern, western and central regions. These resources are primarily found in the nation’s isolated settlements (Armah et al., 2014; Osumanu, 2020; Baddianaah and Baaweh, 2021). As a source of funding or support for the nation, these minerals are, nonetheless, mostly produced in enormous amounts. For example, minerals make up about 40% of all exports, although the mining industry in Ghana contributes just 5% of the country’s GDP (Ennin and Wiafe, 2023). Gold also accounts for almost 90% of all mineral exports (Besada and Golla, 2023). Because the majority of Ghana’s mining and minerals development industry is still focused on the production of gold, the country produced 117.6 tonnes of gold overall in 2021, ranking sixth in the world and first in Africa (Donkor et al., 2023). In addition, Ghana is a major producer of bauxite, manganese and diamonds. Ghana also has large-scale mining companies that produce bauxite, manganese, diamonds and gold, and more than 300 registered small-scale mining organizations and 90 mine support service enterprises (Hira and Busumtwi-Sam, 2018; Boaduo, 2022; Wireko–Gyebi et al., 2023). Other minerals produced in the country include salt, silver, petrol and natural gas.
Even though Ghana produces a lot of minerals and other natural resources, the country’s economic officials work to restrict how much of these resources are produced and used. This is due to their belief that excessive resource extraction harms both the environment and the next generation, as these resources are not produced sustainably (Kwakwa et al., 2020; Alhassan and Kwakwa, 2023; Adabor et al., 2022). Therefore, to lessen the influence on the environment, several steps have been implemented to limit or restrict the extraction of minerals, particularly illicit ones. However, these efforts have been in vain.
The central motivation behind this study is that Ghana is a country with a vast number of resources, including gold, oil, timber, bauxite, crude oil, lithium and magnesium (Danso, 2020; Worlanyo and Jiangfeng, 2021). The extraction of these resources yields a significantly large amount of revenue in the country, as well as fostering foreign exchange and international trade between Ghana and its trading partners (Yeboah et al., 2020; Ndikumana, 2023; Kindo et al., 2024). Because of the need to construct developmental projects in the country and the generation of income by individuals, there is always an excess production of these resources. However, the negative aspect of the over-extraction of these resources can also not be overlooked despite its non-negligible impact on the economy and standard of living of the people. To reduce its impact on the environment, there have been various measures by the government of the country and external sources, backed by other studies (Kwakwa et al., 2020; Adabor et al., 2020, 2022; Alhassan and Kwakwa, 2023) to limit its impact on the environment; however, there is still a continuous degradation of the environment, notably by the mining activities and mining sites.
Kwakwa et al. (2020) analyzed the impact of natural resource extraction on the environment in Ghana. They, however, examined natural resources in an aggregate manner without specifying which of these resources are detrimental to the environment. Hence, this study disaggregates natural resources in Ghana into their numerous types to determine if the extraction of most of these resources degrades the environment or not. This study specifically analyses the effect of gold, oil and timber extractions on the environment because these three resources are the most produced in the country. Furthermore, studies (Ahmed et al., 2020; Liu et al., 2022; Oteng–Abayie et al., 2022) that assessed natural resources extraction on the environment used natural resources rent as a measurement of resource extraction. This study, however, uses the production levels of gold, oil and timber instead of the rent-based data. This is because production levels well define the total quantities of resources extracted, instead of rent data that only describes the earnings from resource extraction. Again, these studies that analyzed natural resources on the environment only looked at a single part of the environment. Notably, Kwakwa et al. (2020) only looked at CO2 emissions; meanwhile, there are other parts of the environment, including water and land, which are also susceptible to natural resource extraction. Hence, this study uses the ecological footprint as a measure of environmental impact, which encompasses various aspects of land, water and CO2 emissions. Hence, it covers the entirety of the environment without only focusing on a single part of the environment. Even though other studies (Ahmed et al., 2020; Liu et al., 2022; Oteng–Abayie et al., 2022) have used ecological footprint as a measurement of environmental pollution outside Ghana and across the globe, studies conducted in Ghana have been silent on it when it comes to natural resources on the environment in Ghana. Lastly, Oteng–Abayie et al. (2022) have assessed the impact of laws and regulations on natural resource extraction and how these regulations either help to improve or deteriorate the environment. Other studies have also assessed the effect of military intervention on natural resource extraction, especially illegal mining, and how it affects the environment. Irrespective of the measures used by the government to control environmental pollution caused by these mining sites, they still have some consequences despite their ideology and positive impact. Hence, this study also assesses the impact of institutional regulation and military intervention on natural resource extraction to make a comparison between the two to determine which of these interventions will be suitable and will provide a better solution to reduce environmental pollution from natural resource extraction.
Due to these gaps, this paper addresses these objectives: First, to analyze the impact of gold, oil and timber extractions on the environment. Second, this study aims to test the EKC hypothesis by examining the relationship between gold, oil and timber production and ecological footprint. Lastly, to compare the effect of institutional regulation and military intervention on natural resources extraction and its impact on the environment. This study also covers both theoretical and empirical literature review, methodology, results discussion and conclusion.
Literature review
The theoretical framework surrounding the relationship between natural resource extraction and the environment evolves from the Structural Human Ecology (SHE) theory developed by Dietz and Rosa (1997). This theory provides an intricate relationship between society and the environment. Thus, it assesses the relationship between population, affluence and technology. The SHE theory comprises the STIRPAT (Stochastic Impacts of Regression on Population, Affluence and Technology) model. Thomas Dietz and Eugene Rosa developed the STIRPAT model. The STIRPAT model first started as an IPAT, which was initiated by Ehrlich and Holdren in 1971, considering environmental impact as a function of population, affluence and technology. However, Waggoner and Ausubel (2002) also expanded on the IPAT model and transformed the technology in the IPAT model into consumption per unit of GDP and impact per unit of consumption. The evolution of the STIRPAT model erupted when it was realized that both IPAT and ImPACT lacked some significant features. First, the STIRPAT model introduced the stochastic approach, which allows for hypothesis testing and statistical analysis. Unlike the IPAT and ImPACT analyses, they were just accounting equations, without accounting for hypothesis testing. Studies (Hasan, 2016; Li and Lin, 2015; Shahbaz et al., 2016; Kwakwa et al., 2020) have analyzed environmental impact using the STIRPAT model, and this study similarly adopts the STIRPAT model to evaluate its analysis.
Also, the EKC hypothesis by Simon Kuznets in the 1950s and 1960s is employed in this study. The EKC hypothesis postulates that there is an inverted U-shaped relationship between income and environmental degradation in an economy. That means environmental degradation changes as income increases over time. At the early stages of economic growth, pollution increases due to higher industrialization, resulting in higher pollution. However, after the economy reaches a certain level of income (the turning point), pollution starts to decrease because the country starts to develop strategies to curb pollution. Hence, this study also analyses the EKC relationship between gold production and ecological footprint, as well as oil production, and timber production EKC hypothesis with ecological footprint.
Numerous studies (Ahmed et al., 2020; Liu et al., 2022; Oteng–Abayie et al., 2022; Cai et al., 2023) have analyzed the impact of natural resource extraction on the environment around the globe and in Ghana as well. Cai et al. (2023) looked at how China’s natural resources, including GDP, technological innovation and forest exploitation, affected the country’s CO2 emissions between 1989 and 2021. It finds long-run cointegration among variables using time series econometrics (fully modified OLS (FMOLS), dynamic OLS (DOLS), and Canonical Cointegrating Regression (CCR), with coal, forest, gas and mineral rents positively influencing emissions. Also, Ahmed et al. (2020) analyzed the effects of urbanization, human capital and natural resources on China’s ecological footprint. It discovers that while human capital decreases the ecological footprint, natural resources, urbanization and economic expansion increase it through the use of cointegration and causality tests. Liu et al. (2022) examined how education and access to natural resources affect CO2 emissions in Latin America between 1990 and 2020, taking into account economic growth, remittances and green energy. It discovers a U-shaped relationship between natural resources and CO2 emissions. Additionally, education raises CO2 emissions; yet, remittances and green energy lower CO2 emissions and enhance environmental quality. Oteng–Abayie et al. (2022) analyzed the relationship between environmental sustainability and natural resources in 28 sub-Saharan African nations is examining this period from 2005 to 2017 with a focus on the effect of environmental regulatory quality. It is discovered through system-GMM estimation that although natural resources weaken environmental sustainability, environmental regulatory quality increases it. Furthermore, Environmental Regulatory Quality (ERQ) in conjunction with natural resources significantly diminishes sustainability. Additionally, Ibrahim et al. (2023) examined how natural resources such as coal, oil and gas affect CO2 emissions in 10 of the most resource-dependent nations between 1995 and 2019. It also takes structural change, green finance, renewable energy and technological advancement into account. It concluded that although natural resources have an impact on CO2 emissions, these emissions are mitigated by structural change, renewable energy, green finance and technology. This study looks at how natural resource dependence in resource-rich African nations relates to environmental degradation, income levels and life expectancy between 1980 and 2019. Among the tests and estimators, it employs are the FMOLS, DOLS, quantile regression (QR), Kao co-integration test and first-generation stationarity test. According to the research, longevity is positively impacted by income level but negatively by natural resources and environmental degradation. Their impacts are mitigated by the relationship between income and natural resources. To prolong life, the study recommends raising income levels and encouraging the use of cleaner energy. Omokanmi et al. (2022) analyzed how natural resource dependence in resource-rich African nations relates to environmental degradation and income levels between 1980 and 2019, employing the FMOLS, DOLS, QR, Kao co-integration test and first-generation stationarity test. According to the research, longevity is positively impacted by income level but negatively by natural resources and environmental degradation. Their impacts are mitigated by the relationship between income and natural resources.
Even though numerous studies have been conducted outside Ghana about natural resources on the environment. Studies conducted in Ghana about natural resources on the environment are mostly on a sample basis or at a particular mining site or neighborhood. Adam et al. (2021) examined how the decentralization of artisanal and small-scale gold mining might have paradoxical effects that are detrimental to sustainability through the lenses of political ecology and new institutionalism. According to two case studies (Bole and Talensi) conducted in Ghana, actors’ responses to decentralization frequently result in unfavorable outcomes from formalization initiatives. In the study of Attiogbe and Nkansah (2017), the effects of Newmont Akyem on water bodies in Akyem District are evaluated using both qualitative and quantitative approaches by the Department of Energy and Environmental Engineering at the University of Energy and Natural Resources. Although most bodies of water satisfy Environmental Protection Agency (EPA) regulations, the Pra River exhibits high levels of Total Suspended Solids, most likely as a result of illicit mining occurring upstream. Erdiaw-Kwasie et al. (2014) investigated a large-scale mining company’s operation in Prestea, Ghana. Surveys, SPSS analysis, in-depth interviews and content analysis were among the quantitative and qualitative techniques used in the case study. The results indicate that although the mining industry makes a substantial contribution to the local economy, the corporation does not sufficiently address social and environmental issues. Moreover, Adam et al. (2021) examined how two mining centers in Ghana, Bole and Talensi, are empowering local governments to handle natural resources. Rules are also being made for small-scale gold miners. Utilizing research and interviews, two locations were taken into consideration in Ghana. It was, however, discovered that increasing the central government’s power can negatively impact local people’s ability to manage resources and negatively impact the environment.
Other studies also looked at how natural resource extraction affects the whole country of Ghana. As such, Kwakwa et al. (2020) employed the STIRPAT model to examine the relationship between Ghana’s energy consumption and carbon emissions from the mining of natural resources. Urbanization and resource extraction raise carbon emissions, while development assistance contributes to a long-term reduction in emissions from 1971 to 2013. Ahakwa et al. (2023) likewise examined the relationship between Ghana’s natural resources and environmental deterioration from 1990 to 2020, taking into account industrialization, trade openness, economic growth and the use of renewable energy. Using the ARDL model, it concluded that while trade openness, economic expansion and natural resources all contribute to environmental deterioration over the long run, the use of more renewable energy reduces it. With an emphasis on Ghana, Adams et al. (2019) investigated the “curse of the natural resource” in emerging nations. Examining 222 cases from 18 different stakeholders, it concluded that although programs such as the Extractive Industries Transparency Initiative (EITI) are crucial, they are not enough to break the curse on their own. Crucial elements at the national level are governance, institutional quality and corruption prevention.
Few other studies have engaged in critical political ecology frameworks that examine militarization and resource conflicts, particularly given the focus on military intervention. For instance, Hilson and Maconachie (2020) assessed the effect of military intervention on artisanal and informal gold mining in Ghana. It was argued that projecting the ban on informal mining through military intervention has helped protect the environment in Ghana. Bansah et al. (2022) similarly assessed the effect of military intervention on artisanal gold mining in Ghana. It was, however, revealed in their results that military intervention leads to the burning of mining equipment such as excavators and injuries to the miners.
In terms of social impacts, the gendered consequences of resource extraction remain underexplored. Andrews et al. (2022) assessed how oil extraction leads to gender inequality in the Western region of Ghana. By employing the theoretical framework of feminist political ecology, this study revealed that local, cultural and social practices determine who has access to, manages and uses oil extractives in the Western region of Ghana. This limited women’s access to oil extraction, limited them in managing and leading extractive processes in Ghana. Likewise, Soliku (2021) concluded that men have greater economic impacts on resource extraction in Northern Ghana than women. Also, due to cultural and social norms, women suffer from economic and socio-psychological impacts. Finally, Adabor et al. (2022) assessed the causative relationship between oil rents and economic growth. It was concluded in their analysis that oil extraction has become a resource blessing to the growth of Ghana; however, it has become a curse to the country through its distraction on the environment.
Even though studies have analyzed natural resource extraction around the globe, most of the studies covered were outside Ghana and used an integrated measurement of the environment, thus, ecological footprint. Other studies have also done similar analyses in Ghana. However, it was only analyzed on one of the components of the environment, specifically, only on CO2 emissions. This study does a similar analysis to the studies outside Ghana by using the complex or integrated measurement of the environment that comprises various components of land, CO2 emissions and water. Again, most of the studies on Ghana were done on a sample basis. However, this study focuses on the entire country, like other studies such as Kwakwa et al. (2020) and Ahakwa et al. (2023), but disaggregates the natural resources into the major natural resources extracted in Ghana. Again, this study further analyzes the impact of military intervention and institutional regulation on natural resource extraction, to make a comparison and determine which of the measures would be effective in curtailing environmental degradation. Previous studies, such as Hilson and Maconachie (2020) and Bansah et al. (2022), assessed the military intervention effect on resource extraction in Ghana. It was, however, done on a sample basis and only focused on artisanal mining instead of focusing on the broader scale of mining.
The conceptual framework from Figure 1, developed in this study, illustrates how natural resource extraction, especially gold, oil and timber, contributes to environmental degradation as indicated by the ecological footprint. The ecological footprint measures various forms of pollution, including CO2 emissions, land degradation and water pollution. Additionally, control variables such as renewable energy, trade, urbanization and gross capital formation are included to evaluate their influence on the environment. Beyond the direct effects, this study incorporates interaction terms between military intervention and institutional regulation with each resource type to examine how government enforcement affects the environmental impact of resource extraction.
The flowchart shows two circles on the bottom left and bottom right. The bottom left circle is labeled “MILITARY INTERVENTION, INSTITUTIONAL REGULATION.” The bottom right circle is labeled “RENEWABLE ENERGY, TRADE, URBANIZATION, INDUSTRY, GROSS CAPITAL FORMATION.” An arrow from the bottom left circle points to another circle above labeled “GOLD, OIL, TIMBER.” An arrow from the bottom right circle points to another circle above labeled “ECOLOGICAL FOOTPRINT (C O 2, LAND, AND WATER POLLUTION).” An arrow from “GOLD, OIL, TIMBER” points to “ECOLOGICAL FOOTPRINT (C O 2, LAND, AND WATER POLLUTION).”Conceptual framework. Source: Author’s own creation
The flowchart shows two circles on the bottom left and bottom right. The bottom left circle is labeled “MILITARY INTERVENTION, INSTITUTIONAL REGULATION.” The bottom right circle is labeled “RENEWABLE ENERGY, TRADE, URBANIZATION, INDUSTRY, GROSS CAPITAL FORMATION.” An arrow from the bottom left circle points to another circle above labeled “GOLD, OIL, TIMBER.” An arrow from the bottom right circle points to another circle above labeled “ECOLOGICAL FOOTPRINT (C O 2, LAND, AND WATER POLLUTION).” An arrow from “GOLD, OIL, TIMBER” points to “ECOLOGICAL FOOTPRINT (C O 2, LAND, AND WATER POLLUTION).”Conceptual framework. Source: Author’s own creation
Materials and methods
Theoretical and empirical specification
The STIRPAT model indicated in Section 2 will be used to investigate the relationship between the specific natural resources’ extraction and ecological footprint. The STIRPAT model, as indicated above, analyzes the relationship between Population, Affluence, Technology and environment presented below.
The STIRPAT model is presented in the equation.
The natural log of each variable in Equation (1) is also taken to interpret the estimated parameters as elasticities, as shown in Equation (2).
From Equations (1) and (2), Represents ecological footprint, which is a proxy for environmental impact, C is the constant term, P refers to Population, T also is Technology and A represents Affluence. are the parameters to be estimated, is the stochastic term, and i and t refer to the number of county sets and periods, respectively.
In the context of this study, I symbolize environmental pollution as a proxy for environmental impact. The production levels of specific natural resources represent the affluence of each country, while urbanization consistent with previous studies such as Hassan (2016) serves as a proxy for population within this STIRPAT model framework. Additionally, industrialization, a key variable, is used to measure technology in the model.
From other empirical analyses, other key variables are also likely to influence environmental quality. The inclusion of other relevant variables helps to avoid any omitted variable bias and controls for some standard covariates. Based on the analysis of the equation, the STIRPAT model also allows for the inclusion of other relevant variables such as trade as a percentage of GDP (Trade), renewable energy (Renew) and gross capital formation (Gross). Hence, Equations (3), (4), (5), and (6) present an expanded model including all other relevant variables apart from the main variables (P, A, T) in the STIRPAT model.
From Equations (3), (4), (5), and (6), Gold, Oil and Timber refer to the affluence of the country, Ghana, since these resources contribute significantly to the country’s revenue and growth. Again, Urban represents urbanization, Trade represents trade as a percentage of GDP, Renew is also renewable energy and Industry refers to Industrialization, respectively, as shown in Table A1 in Appendix. Equation (4) also presents the EKC model, which analyses the non–linear relationship between gold, oil and timber extractions on ecological footprint. By including the linear and squared terms, the model presents a U-shaped relationship between resource production and ecological footprint as proposed by the EKC hypothesis. Also, from Equations (5) and (6), Goldmili, Oilmili and Timbermili refer to the influence of military intervention on gold, oil and timber extraction, respectively. Likewise, Goldins, Oilins and Timberins also refer to the influence of institutional regulation on gold, oil and timber production, respectively.
Estimation strategy
In Objective 1 from Equation (3), this study used the Augmented Dickey–Fuller test to ensure the variables have no unit root at order zero and first difference. Again, the bounds testing approach within the ARDL framework was also employed to test the level (cointegration) relationship between ecological footprint and natural resource extraction. The ARDL technique was also used to estimate the multipliers. Finally, the EC model was employed to estimate the long-run and short-run estimates.
Again, the Seemingly Unrelated Regressions (SURE) model was also used for objective 2, which is presented in Equations (4) and (5). The SURE model is being employed to analyze the effect of institutional regulation and military intervention on gold, oil and timber extractions in Ghana. The SURE model is therefore the ideal estimation technique when it comes to multiple regressions and simultaneous equations. This estimator enables simultaneous estimates of several regression models while taking into consideration any possible correlation between the error terms. The ecological footprint as the dependent variable in Equations (5) and (6) is likely to be influenced by various factors such as trade, gold, oil, timber extraction, urbanization, military intervention and institutional regulation. Hence, these factors may affect the ecological footprint in the two equations simultaneously, leading to correlated error terms. The SURE estimator accounts for this correlation, leading to efficient and accurate parameter estimates.
Results
Descriptive statistics
This study includes a number of variables that go into detail on how the ecosystem is affected by the extraction of natural resources. From Table 1, ECF, the dependent variable, has a standard deviation of 0.32 and an average of 1.449, with minimum and highest values of 0.94 and 1.88, respectively. On the other hand, the standard deviations of the extraction levels of gold, oil and lumber are larger (32.065, 21.517, 13,790,428), indicating a divergence from their respective averages (51.528, 11.553, 30,334,517).
Descriptive results
| Variable | Obs | Mean | Std. dev. | Min | Max |
|---|---|---|---|---|---|
| ECF | 42 | 1.449 | 0.32 | 0.94 | 1.88 |
| Gold | 36 | 51.528 | 32.065 | 9 | 98 |
| Oil | 37 | 11.553 | 21.517 | 0 | 66.744 |
| Timber | 43 | 30,334,517 | 13,790,428 | 11,315,500 | 55,222,506 |
| Urban | 42 | 3.974 | 0.424 | 3.02 | 4.59 |
| Renew | 31 | 61.576 | 14.606 | 40.25 | 82.93 |
| Trade | 42 | 64.116 | 26.439 | 6.32 | 116.05 |
| Gross | 40 | 18.619 | 6.255 | 3.75 | 29 |
| Industry | 42 | 2.424e+10 | 4.538e+10 | 540,100 | 1.950e+11 |
Diagnostic test results
To ascertain the presence of no autocorrelation in our study, an autocorrelation test using the Breusch–Godfrey test had to be performed in order to determine whether autocorrelation existed in our study. It is clear from Table A2 in Appendix that the null hypothesis, which holds that there is a serial correlation between the variables, is rejected. Therefore, we conclude that, at a 5% significance level, there is no correlation between the variables. Furthermore, it was shown by our correlation matrix in Table A3 in Appendix that there is a perfect linear relationship between the variables. As a result, Table A3 revealed that the correlation or relationship, between the variables was not equal to 1 or −1, indicating a non-perfect linear relationship and offering valuable insights into the interdependency of our variables.
Again, from Table A4 in Appendix, the White test (Cameron and Trivedi’s decomposition of the IM-test) was used to check for homoskedasticity in our study. The p-value generated was 0.4017, indicating the presence of homoskedasticity since the p-value is greater than the critical value of 0.05. In this study, we also applied a cointegration test to examine the long-run relationship between the variables. From Table 2, it was revealed that the calculated F-statistic (60.028) is greater than all the upper bound critical values (1.95, 3.06, 2.22, 3.39, 2.48, 3.70, 2.79, 4.10), and the t-statistic (−14.150) was also lower than all the lower bound critical values (−2.57 –4.40 –2.86 –4.72 –3.13 –5.02 –3.43 –5.37). Finally, there was also the need to test for stationarity of the variables. All of the variables exhibited stationarity at levels except the dependent variable ECF, which was stationary at first difference from Table A5 in Appendix.
ARDL bounds cointegration test and error correction term results
| F-bounds test (60.928) Ho: No cointegration | 1.95 | 3.06 | 2.22 | 3.39 | 2.48 | 3.70 | 2.79 | 4.10 | Cointegration confirmed |
| Test-statistic (−14.150) (EC term) | −2.57 | −4.40 | −2.86 | −4.72 | −3.13 | −5.02 | −3.43 | −5.37 | Long–run adjustment confirmed |
From Objective 1, this study used the ARDL estimator to analyze the long-run effect of gold, oil and timber extractions on the environment in Ghana. The optimal lag structure was obtained using the Akaike Information Criterion. From Table 3, it can be noticed that ECF and Gold had two optimal lag structures, with the other variables having only one optimal lag structure.
ARDL results
| Variables | Coefficients |
|---|---|
| L1. ECF | −1.3530*** (0.1190) |
| L2. ECF | −0.1860 (0.0993) |
| lnGold | 0.7130*** (0.0375) |
| L1. lnGold | 0.3830** (0.0576) |
| L2. lnGold | 0.2860** (0.0535) |
| lnOil | 0.1080*** (0.0074) |
| L1. lnOil | 0.0651** (0.0080) |
| lnTimber | 1.0980** (0.2460) |
| L1. lnTimber | −0.6900** (0.0860) |
| Urban | 0.5880* (0.1760) |
| L1. Urban | −0.4780 (0.1970) |
| Renew | −0.0138** (0.0024) |
| L1. Renew | −0.0384*** (0.0024) |
| Trade | −0.0050*** (0.0004) |
| L1. Trade | 0.0003 (0.0003) |
| Gross | 0.0275*** (0.0016) |
| L1. Gross | −0.0026 (0.0012) |
| lnIndustry | −0.3520** (0.0400) |
| L1. lnIndustry | 0.0956 (0.0373) |
| Constant | −0.7690 (3.9560) |
| Observations | 22 |
| R-squared | 1.000 |
Note(s): The dependent variable is ECF
Standard errors in parentheses, ***p < 0.01, **p < 0.05, *p < 0.1
The results in Table 3 revealed that the compounding effect of the ecological footprint from the previous year (L1.ECF) and two years ago (L2.ECF) had a negative relationship with the ecological footprint. This means that the compounding effect of environmental deterioration in Ghana rather leads to an improvement in the environment. The first previous year, which had a negative and significant relationship with ecological footprint, means that once the environment deteriorates based on human activities in Ghana, various EPAs and governmental organizations put in immediate measures to resolve the built-up pollution rather than allowing the already built-up pollution to escalate environmental pollution in Ghana. Again, gold production (InGold) in Ghana leads to an increase in the ecological footprint in Ghana in the current period. Thus, an increase in gold production leads to an increase in environmental pollution in Ghana. That is, a percentage increase in gold production leads to a 0.00713 increase in ecological footprint. Also, gold production in the previous years (L1 lnGold and L2 lnGold) leads to an increase in environmental pollution in Ghana. Thus, a percentage increase in gold production leads to an increase in ecological footprints in Ghana by 0.00383 and 0.00286, respectively. Gold mining and production increase environmental pollution because of the extensive degradation of the land during mining, which makes the land prone to degradation and soil erosion. Again, gold mining in Ghana also leads to water pollution because of the constant usage of chemicals like mercury and cyanide used in mining. Likewise, gold production also leads to an increase in CO2 emissions because of the constant use of fossil fuels for heavy tractors and excavators, which emit CO2 into the atmosphere. However, the continuous mining of gold in Ghana also leads to an increase in environmental pollution because Gold is the leading mineral mined in the country, including small-scale, large-scale or industrial ones, and illegal mining of gold in Ghana. Oil production likewise leads to an increase in the ecological footprint in Ghana in the current year of extraction. A percentage increase in oil production leads to a 0.00108 increase in ecological footprint in Ghana. This reflects the direct impact that oil production in the current period has on the environment. Oil production negatively affects the environment through the use of fossil fuels to power the drilling machines, which emit CO2 into the atmosphere, thereby contributing to an increase in global warming. Oil production also deteriorates the environment through oil spillages, leakages and the discharge of contaminated wastewater. Similarly, oil production in the previous period (L1. lnOil) leads to an increase in ecological footprint or environmental pollution. As such, a percentage increase in oil production in the previous period leads to a 0.000651 increase in environmental pollution. This means that a persistent increase in oil production continues to affect the environment negatively, as the negative effects of oil production continue to deteriorate the environment through continuous water pollution and aquatic habitat loss.
From Table 3, Timber production leads to an increase in ecological footprint in Ghana. Thus, a percentage increase in timber production in Ghana leads to a 0.01098 increase in environmental pollution. Timber production in Ghana, which involves the use of sawmill machines powered by fossil fuels, also emits CO2, thereby polluting the environment. Again, timber production in Ghana mostly leads to land degradation because of the constant cutting of timber, which makes the land bare, making it prone to soil erosion. Meanwhile, a percentage increase in timber production in the previous period leads to a 0.0069 decrease in the ecological footprint in Ghana. This is because Ghana has enacted laws and regulations regarding the forestry ministry. To the extent that, even when timber is cut down or after initial deforestation, the older ones are rather cut off, leaving behind the younger ones, which prevent the land from becoming bare and also the adoption of afforestation to replace the cut-down timber.
The ARDL EC model is also used as a robustness check on the ARDL results from Table 3. The EC model is best as a robustness check on the ARDL model because it explicitly tests for and corrects deviations from long-run equilibrium, as well as presents the results of both short-run and long-run equilibrium.
The results in Table 4 present the long-run relationship between gold, oil and timber extraction on ecological footprint. The results in Table 4 show that gold and oil production are positive and highly significant. That is, a percentage increase in gold production leads to an increase in ecological footprint by 0.00544 in the long run. Similarly, a percentage increase in oil production leads to an increase in ecological footprint by 0.0000683. Timber production likewise increases the ecological footprint by 0.00161 in the long run, though not significantly. Thus, the results from the EC model in the long run confirm the results from the ARDL. They both show that gold and oil production increase environmental pollution in the long run due to the energy-intensive processes involved in gold mining and oil drilling, which lead to CO2 emissions, land degradation and water pollution. Timber production’s results are not significant because, in the long run, there is the practice of afforestation to regenerate the timber species.
EC model; long-run effect
| Variables | Coefficients |
|---|---|
| ADJ. L1.ECF | −2.5390*** (0.1850) |
| lnGold | 0.5440*** (0.0224) |
| lnOil | 0.0683*** (0.0057) |
| lnTimber | 0.1610 (0.0728) |
| Urban | 0.0431 (0.0232) |
| Renew | −0.0206** (0.0021) |
| Trade | −0.0019** (0.0002) |
| Gross | 0.0098*** (0.0006) |
| lnIndustry | −0.1010** (0.0102) |
Note(s): The dependent variable is the ECF
Standard errors in parentheses, ***p < 0.01, **p < 0.05, *p < 0.1
In the short run from Table 5, the dynamics differ. Gold and oil production exhibit a negative and significant relationship with ecological footprint, while timber production shows a positive relationship with ecological footprint. In the short run, gold and oil output may be accompanied by investments in cleaner or more efficient extraction methods, regulatory compliance or localized containment measures that temporarily limit noticeable environmental degradation. Furthermore, the ecological footprint may lag behind production changes, especially if environmental effects (such as deforestation or pollution) appear gradually. It is also plausible that short-term production increases do not last long enough to cause acute ecological stress, as opposed to the cumulative pressures witnessed over time in gold and oil production. Timber production also showed a positive effect due to its immediate impact on the environment. Timber production initially leads to land degradation and deforestation, while there is the practice of afforestation and replanting of timber species afterwards, hence showing an insignificant relationship.
EC model; short-run effect
| Variables | Coefficients |
|---|---|
| L1.ECF | 0.1860 (0.0993) |
| L1.Gold | −0.6680** (0.0992) |
| L2.lnGold | −0.2860** (0.0535) |
| lnOil | −0.0651** (0.0080) |
| lnTimber | 0.6900** (0.0860) |
| Urban | 0.4780 (0.1970) |
| Renew | 0.0384*** (0.0024) |
| Trade | −0.0003 (0.0004) |
| Gross | 0.0026 (0.0012) |
| lnIndustry | −0.0956 (0.0373) |
Note(s): The dependent variable is ECF
Standard errors in parentheses, ***p < 0.01, **p < 0.05, *p < 0.1
Table 6 presents the EKC hypothesis results. According to the EKC hypothesis, the relationship between economic growth (resource extraction) and environmental pollution follows an inverted U-shape. This implies that the linear term of the resource variable should have a positive value, with the squared term or quadratic term having a negative coefficient. Such a pattern implies that pollution increases with natural resources extraction initially, reaches a turning point and then pollution starts to decline as development or income of the country increases.
EKC results
| Variables | Coefficients |
|---|---|
| lnGold | −3.4340 (4.0360) |
| (lngold)2 | 0.4520 (0.4870) |
| lnOil | 0.0131 (0.0308) |
| (lnoill)2 | 3.16e‐05 (3.15e‐05) |
| lnTimber | 46.4100 (33.9100) |
| (lntimber2) | −1.3540 (0.9900) |
| Urban | −0.0709 (0.2770) |
| Renew | −0.0161 (0.0091) |
| Trade | −0.0010 (0.0009) |
| Gross | 0.0046 (0.0053) |
| lnIndustry | −0.0132 (0.0746) |
| Constant | −388.0 (283.9) |
| Observations | 23 |
| R-squared | 0.9740 |
Note(s): The dependent variable is ECF
Standard error in parentheses, ***p < 0.01, **p < 0.05, *p < 0.1
From Table 6, the EKC results prove that gold and oil extraction continue to deteriorate the environment, while there is no turning point for controlling the harm to the environment. Specifically, for gold production, the linear relationship is negatively signed (−3.434) while the squared term is positive (0.4518584). This confirms a U-shaped rather than the EKC inverted U-shape. Similarly, oil production (lnOil) and its squared term (lnoil)2 have a positive relationship with ECF, also contradicting the EKC hypothesis. Hence, the results prove no statistically significant U-shaped relationship for gold and oil production. This outcome aligns with the status of Ghana’s economy as a developing country. This means Ghana has not reached its developed or maximum stage for the country to effectively find means to control pollution in gold and oil production. This comes from limited regulation, a lack of control over corruption and inefficient use of the military to resolve environmental pollution. Hence, Ghana is still in the early phase of the EKC-like pattern, where pollution increases with economic growth or resource extraction. However, timber production displayed an EKC hypothesis pattern. Pollution increased initially (lnTimber with coefficient 46.40505), then declined as production rose (lnTimber2 with coefficient −1.353884). Though not statistically strong, this may reflect gradual environmental progress in Ghana’s forestry sector, contrasting with the continued environmental pressures from mining and oil activities.
Table A6 from Appendix presents a comparison between the results of ARDL and EC models as well as the EKC hypothesis. The results from Table A6 prove that gold and oil production continually destroy the environment. While timber production destroys the environment, the results also prove that there are measures such as afforestation and replanting of timber species to maintain timber production.
Objective 2 of this study was to compare the effect of military intervention and institutional regulation on natural resource extraction in Ghana and how it affects the environment in the country. Table 7 presents the results of Objective 2. Starting with military intervention, the coefficient on lnGoldmili is positive (1.39420) but statistically insignificant. This means that even though there may be military intervention associated with gold mining, it might be ineffective and or inconsistently implemented to reduce environmental deterioration. Additionally, reverse causality comes into play where environmental effects due to gold mining drew the attention of military intervention, yet military presence was not felt in action to reduce environmental pollution. Conversely, Timbermili shows a negative but insignificant relationship effect on ecological footprint. This suggests that military efforts in controlling environmental pollution in timber production may be helpful, but not strong or consistent enough to make a significant impact. This could reflect weak or limited military involvement, irregular military intervention or challenges associated with preventing illegal logging activities. Similarly, military involvement in oil production shows a positive and significant relationship with ecological footprint at a 10% significance level. Even though it shows a significant relationship, it is, however, marginally significant, further suggesting a weak but potentially meaningful association that calls for deeper military intervention.
SURE model
| Variables | Coefficient | Variables | Coefficient |
|---|---|---|---|
| InGoldmili | 1.3942 (0.9030) | lnGoldins | −3.3870** (0.0318) |
| lnOilmili | 0.0578* (0.0613) | lnOilins | 0.0297*** (0.0018) |
| lnTimbermili | −0.3656 (0.2386) | lnTimberins | 0.0297*** (0.0081) |
| Gross | 0.0099*** (0.0064) | Gross | 0.0225*** (0.0002) |
| Industry | 0.0013*** (0.0025) | Industry | 0.0001*** (0.0001) |
| Urban | 0.3258 (0.1817) | Urban | 0.2678** (0.0101) |
| Renew | 0.0156*** (0.0057) | Renew | 0.0566*** (0.0004) |
| Constant | 1.1767 (0.6627) | Constant | 4.0520** (0.0301) |
| R-squared | 0.9019 | R-squared | 0.99990 |
Note(s): The dependent variable is ECF
Standard errors in parentheses, ***p < 0.01, **p < 0.05, *p < 0.1
From the second results of the SURE estimator, it was realized that institutional regulation is effective in improving the quality of the environment when it comes to gold and oil extractions. The institutional regulation effect on gold and oil extraction had a positive effect on the ecological footprint in Ghana. This is because of the strong regulations laid down by the government and the EPA in Ghana to prevent pollution from the mining site and also to reduce illegal mining activities in the country. However, institutional regulation effect on timber extraction in Ghana had a positive effect on the ecological footprint. This is due to weak enforcement, inadequate resources and prioritizing economic profits over environmental quality.
Comparing the military intervention and the institutional regulation effect on natural resources and how they affect the environment in Ghana has been something untold. Table 7 shows that institutional regulation is effective in reducing environmental pollution when it comes to gold and oil extraction in the country. This is because institutional regulation ensures stringent environmental standards to reduce waste and pollution and ensures the use of environmentally friendly equipment, which reduces pollution. However, military intervention’s main focus is on defense and the use of weapons to either seize mining equipment or threaten miners, especially those involved in illegal mining, to stop mining. The activities and jobs of the military concerning the prevention of environmental pollution on the mining site are always futile because they do not ensure that miners follow the stringent environmental policies to prevent degradation. They would rather use weapons to chase them out of the mining site and seize or burn their mining equipment. Even with the seizure of mining equipment, the workers still find means to acquire different ones to work on different mining sites and still pollute the environment because they lack the technical know-how to protect the environment, even as they mine to fend for themselves.
With timber production, there are confusing results because military intervention has no significant effect on timber extraction and institutional regulation likewise has a positive relationship with the ecological footprint. Both interventions help reduce environmental regulation in the country. This implied a clear lack of effectiveness in either strategy for minimizing the environmental impact of Ghana’s timber production.
Conclusion
This study delved into natural resource extraction on the environment. Specifically, the study examined gold, oil and timber extractions on the environment. From the results of our objective 1, in which ARDL was employed, it was concluded that gold production increases its ecological footprint in the current period and two periods ago. Also, oil production increased the ecological footprint in the current period and the previous period. These results align with the findings of Mantey et al. (2020), Besada and Golla (2023) and Ahakwa et al. (2023). However, timber production increased the ecological footprint in the current period but decreased the ecological footprint in the previous period. Also, the results of the EKC hypothesis revealed that gold and oil production do not follow the EKC inverted U-shaped pattern, while timber production does. Finally, using the SURE estimator revealed that military intervention has no significant impact on curtailing environmental pollution in gold, oil and timber production. However, institutional regulation significantly influenced gold and oil production by minimizing its environmental impact.
This study provides practical recommendations to tackle these environmental pollution problems in the study. The current study highlights that gold and oil production increase the ecological footprint and increase environmental pollution, while institutional regulation is a better remedy in controlling pollution as compared to military intervention. To address this problem, the labor force engaged in mining, legal and illicit, large- and small-scale mining should receive extensive training in mining operations to solve this issue, while military involvement in mining-related activities should be drastically reduced. However, institutional regulation should be strictly adhered to, especially in the areas of illegal mining popularly called “galmsey” in Ghana. This makes it highly commendable to use institutional regulation as it ensures orderliness, accountability, coordination, transparency and ultimately, protection of the environment. Institutional regulation should also be free from corruption, weak governance and political interference to be able to fully measure the effectiveness of institutional regulation in controlling and preventing environmental pollution.
Because the military often destroys mining equipment or occasionally kills or seriously hurts mining workers, which results in a loss of life and a reduction in labor. If the government authorities believe that excessive resource extraction is harming the environment, the labor force involved should be prepared for other lucrative jobs in the nation or properly trained. Also, mining should be concentrated at one location rather than being dispersed across the nation. When it comes to timber extraction, from this study, it was revealed that timber significantly increases the ecological footprint in the current period but reduces the ecological footprint in the previous period. However, both military intervention and institutional regulation did not prove to be successful in curtailing environmental pollution from timber production. However, this study recommends that various EPAs and governmental institutions should mainly focus on military intervention in reducing environmental pollution from the logging industry. From the results of Table 7, it was noticed that the interaction between timber and military intervention had a positive relationship with ecological footprint, even though it was not significant. This indicates that military intervention has an influence on reducing environmental pollution in the logging industry compared to the interaction between institutional regulation and timber production, which had a negative relationship with ECF but was significant.
It is better to impose institutional regulation on gold and oil production than timber production because gold and oil production are internationally traded among other foreign countries or neighboring countries unlike timber. From the report of World’s Top Exports in 2022, it stated Ghana’s top 10 exports and mineral fuels, including oil, the top export in Ghana, generating US$6 billion (37.1% of total exports) with gems and metals contributing $4.9 billion (30.1%) to Ghana’s total exports. Meanwhile, Timber was not even part of the list of the top 10 natural resources contributing to higher income in Ghana. This makes it easier for institutional regulation to work out for the two extractive industries, as other foreign bodies are mostly involved in ensuring compliance with environmental standards to ensure environmental sustainability. However, the timber industry is best suited to adopt military intervention because, even though traded internationally, but not as huge as gold and oil production, so the government of Ghana has greater control over it to use radical or military intervention to ensure that, even as they produce timber, they still maintain the environment.
This study, however, has some limitations. First, Ghana extracts numerous minerals, and future studies can both expand this study and also assess the impact of other natural resources like diamond, bauxite, aluminum and others on the ecological footprint. Future research can also highlight the need for disaggregated analyses (e.g. distinguishing artisanal vs. industrial mining) and longitudinal studies to track regulatory efficacy over time. Furthermore, future studies can propose cross-country comparisons to examine whether Ghana’s institutional dynamics generalize to other resource-dependent economies. Again, this study only explored the influence of military intervention and institutional regulation on mineral extraction and how it affects the environment. Future studies or scholars can also explore or generate other means, such as alternative regulatory mechanisms like community-based monitoring systems. EPAs or governmental bodies can also adapt to control pollution in the mining sector. Also, the variable for timber production had extreme values or outliers as compared to the other variables, which may affect the robustness and reliability of the estimated relationships. Future studies can also use other variables for timber production or convert the timber production variable to a different unit to achieve smaller values, like those of gold and oil production. The methodology applied in this study provides a solid framework by employing sector-specific effects on the environment. It, however, excludes spatial econometric analysis, due to limited time and resources to undergo such an experiment. Future studies can also incorporate spatial analysis and conduct a regional or district-based analysis since the extractive sectors are mostly centered in the rural communities or particular districts.
The supplementary material for this article can be found online.
