Does environmental quality react asymmetrically to unemployment and inflation rates? African OPEC countries’ perspective

Environmental degradation, inflation, and unemployment are unquestionably among the current global issues. However, there has not been an in-depth investigation of how unemployment and inflation rates affect environmental quality, particularly when considering the asymmetric scenario in oil-producing countries. This gap in the literature motivated this study to investigate how the environment (proxied by CO2 emissions) reacts to asymmetric shocks in inflation and unemployment rates using the panel NARDL model methodology. This study also examines whether the environmental Phillips curve (EPC) hypothesis holds true in the context of African OPEC countries over the period 1990 to 2019. The study presents three interesting findings. First, CO2 emissions are adversely associated with unemployment and inflation rates, meaning that protecting a healthy environment would have to come at the expense of two undesirable outcomes: losing employment and a decline in purchasing power. Second, the asymmetry analysis demonstrates that both negative unemployment and positive inflation shocks have a larger effect on CO2 emissions than the opposite scenario. Finally, long-term evidence exists to support the presence of the EPC in these countries.


Introduction
One fact that all countries across the world can agree on is that environmental degradation is more than just a concern and a hot topic. The underlying cause of this issue is a sustained increase in greenhouse gas emissions, with carbon dioxide accounting for a substantial portion and being the single most important of those emissions that trap heat and remain in the atmosphere for hundreds to thousands of years, eventually harming environmental quality (Global Carbon Atlas 2023).
As long as nonrenewable energy remains the lifeblood of the economy, there is no trade-off between sustainable economic growth and higher CO 2 emissions. According to the environmental Kuznets curve (EKC) hypothesis proposed by Grossman and Krueger (1991), both environmental deterioration and economic expansion are initially inextricably linked, but at certain high income levels, environmental quality starts to improve and emissions start to go down. The change in direction is due to the fact that pollution may be combated through the use of clean technology or abatement initiatives. This hypothesis, however, has been extensively investigated; some studies support it (Apergis and Ozturk 2015;Rafindadi and Usman 2019;Pham et al. 2020;Malik et al. 2020;Wang et al. 2023a), while others do not (Kang et al. 2016;Adu and Denkyirah 2019;Nasir et al. 2021;Nguyen et al. 2021), and other investigations discover mixed findings across the selected sample (Liu and Lai 2021;Isik et al. 2021;Bibi and Jamil 2021;Wang et al. 2023b).
Countless empirical studies have been conducted to explore the drivers of carbon emissions (as a best proxy for environmental degradation and global warming), which may be split into two categories: the first category consists of conventional drivers including GDP, energy consumption, industry, tourism, FDI, trade openness, urbanization, and import and export Zhang and Zhang 2021;Li et al. 2021;Wang and Zhang 2021;Wang et al. 2022;Li et al. 2022a;Li et al. 2022b;Farooq et al. 2023). The second includes nonconventional drivers such as education level, innovation, and female employers (Chen and Lee 2020;Alkhateeb et al. 2020;uz Zaman et al. 2021;Mehmood 2022). The latter set of factors, however, has received less attention.
The social and economic factors mentioned above contribute significantly to the existing empirical literature and offer various perspectives on reducing carbon emissions and improving environmental policy frameworks. However, further examination of other relevant drivers is still necessary. This study focuses on two ambiguous and indirect drivers: unemployment and inflation rates.
According to Bhowmik et al. (2021), unemployment can impact CO 2 emissions and the environment through two channels: the "growth channel" and the "preference channel." The growth channel suggests that unemployment hinders economic growth, leading to lower energy use and ultimately lower carbon emissions (enhancing environmental quality). The preference channel argues that unemployment negatively affects the environment because it reduces consumer income, limiting individuals and households' ability to fulfill their desires and live a healthy lifestyle that promotes environmental quality (see Veisten et al. 2004;Meyer 2016;Mulderij et al. 2021). Kashem and Rahman (2020) propose the environmental Phillips curve (EPC) hypothesis as a novel approach to investigating the relationship between unemployment and environmental pollution. The EPC hypothesis provides strong evidence supporting the negative relationship between unemployment and environmental degradation. This suggests that unemployment can lead to lower carbon emissions and enhance environmental quality.
The environmental impact of inflation has not been widely recognized. Alam et al. (2015) show that high inflation has been associated with lower CO 2 emissions over the long term. It is argued that inflation harms the economy by reducing aggregate demand, resulting in less environmental degradation . Rising inflation can affect the environment through two channels: cutting back on consumption and preventing investment plans, leading to a slowdown in growth. Similarly, Ullah et al. (2020) argue that consumers choose to participate in more economic activities that produce more carbon dioxide in an area of stable inflation.
This study selected the African OPEC members (Algeria, Angola, Equatorial Guinea, Gabon, Libya, Nigeria, and the Republic of the Congo) as a sample not only because there is a dearth of literature on them but also because they are roughly net oil-producing countries (Algeria, Nigeria, Angola, and Libya are frequently listed among the top 20 oil-producing countries in the world) that collectively contributed more than 29% of all emissions in Africa in 2021, with three of them (Algeria, Nigeria, and Libya) being among the top five emitters on the continent (Global Carbon Atlas 2023). Except for Equatorial Guinea, which averaged inflation and unemployment rates below 10% for the previous decade (2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018)(2019), all other members failed to attain single-digit rates on at least one indicator during the same decade (World Bank 2023).
This paper makes several significant literary contributions. First, unlike previous studies that have primarily focused on well-known CO 2 drivers (e.g., GDP, energy consumption, FDI, and tourism), this paper sheds light on two ambiguous drivers that have yet to be fully investigated, namely, unemployment and inflation rates, both of which have an indirect relationship with CO 2 . Since environmental concerns are given the utmost attention by policymakers, every prospective factor has a reason to be concerned. Second, with the exception of Ullah et al. (2020), who investigated the asymmetric impacts of inflation instability on environmental quality, no previous research has focused on how CO 2 reacts to asymmetric shocks in both inflation and unemployment rates, as positive fluctuations in both indicators are expected to have a different impact on the ecosystem than negative fluctuations. Third, as a novel concept, the EPC should be kept in check. As far as we are aware, no research has been undertaken to ascertain whether or not the EPC exists for the African OPEC region.
The remaining sections of the paper are organized as follows: the "Literature review" section discusses the existing research and relevant literature on the topic, the "Econometric framework" section presents the econometric framework. the "Empirical results and discussions" section discusses the empirical analysis findings and interpretations. and the "Conclusion and policy implications" section concludes the paper.

Literature review
The relationship between inflation and unemployment rates and environmental deterioration has received little attention in recent research, although it has been briefly discussed in conjunction with other indicators.

Inflation rate and environmental quality
Previous research has rarely addressed the topic of how inflation affects environmental quality. In general, inflation slows economic activity and, as a result, relieves strain on the environment. Alola et al. (2019) run a dynamic panel ARDL approach to study the dynamic long-run link between CO 2 emissions, food production, and inflation rates for the 16 Coastline Mediterranean Countries (CMCs) from 1995 to 2014. Ahmad et al. (2021) use panel FMOLS estimation to look into the connection between inflation instability and pollution emissions in 40 Asian economies from 1990 to 2018. A common finding of both studies is that high inflation (instability or uncertainty) is related to low CO 2 emissions, which enhances environmental performance in the long run. This result is in line with Tamazian and Rao (2010), who employ the standard reduced-form modeling technique and GMM estimation for 24 transition economies using panel data from 1993 to 2004 to explore the relationship between economic development, environmental quality, financial development, and institutional quality. They conclude that inflation helps to reduce environmental deterioration, despite the fact that the impact of inflation on CO 2 emissions has coefficients close to zero. However, a few other studies confirm the negative (positive) link between inflation and CO 2 emissions (environmental quality) (Musarat et al. 2021;Setyadharma et al. 2021). In contrast, Khan (2019) discovers a positive relationship between inflation uncertainty and CO 2 emissions in the case of the Pakistan economy, implying that inflation instability is deleterious to environmental quality. Moreover, Ullah et al. (2020) use the NARDL model to examine the asymmetric effects of inflation instability on environmental quality in Pakistan from 1975 to 2018. Asymmetric findings reveal that negative shocks from inflation instability have a positive impact on CO 2 emissions in the long run, whereas positive shocks have no effect.

Unemployment rate and environmental quality
The relationship between environmental quality and unemployment has also received little attention in the literature. For the first time, Adesina and Mwamba (2019) consider the unemployment rate as part of the economic circumstances influencing CO 2 emissions in 24 African countries classified into three groups: low, lower-middle, and uppermiddle income. They conclude that while unemployment has no significant negative effects on CO 2 emissions for the entire sample or in lower-and upper-middle-income countries, it significantly reduces CO 2 emissions in lowincome countries. Liu and Feng (2022) use the STIRPAT model to assess the possible impacts of nurturing pressure and unemployment on global CO 2 emissions for 77 countries grouped into five sub-panels: Africa, the Americas, Asia-Pacific, Europe, and the Middle East. At the aggregate level, the results show that unemployment has a negative effect on CO 2 emissions. Nonetheless, outcomes differ at the regional level. Unemployment reduces CO 2 emissions in Africa, the Americas, and Europe, but it increases them in the Middle East, and there is scant evidence that unemployment in the Asia-Pacific region specifically affects CO 2 emissions. However, the negative association reported in both studies is supported by a few further investigations that look into the relationship between CO 2 and unemployment (Granados and Spash 2019;Wang and Li 2021) or employment (Gyamfi et al. 2020;Joshua and Alola 2020). Kashem and Rahman (2020) use a static panel model to examine the relationship between environmental pollution and unemployment in 30 countries from 1990 to 2016. They coined the term "environmental Phillips curve" to describe the inverse relationship between the two indicators. This curve is reexamined for BRICST economies by Anser et al. (2021) and for OECD countries by Ng et al. (2022). They confirm the EPC by demonstrating a significant trade-off between unemployment and environmental damage. Furthermore, evidence from the USA (Bhowmik et al. 2021) and South Asian countries (Tariq et al. 2022) also supports the EPC.
Although previous research looked at a variety of renowned CO 2 drivers in developed and developing countries, more research is needed to thoroughly investigate the determinants of environmental damage, using unemployment and inflation rates as two neglected drivers that have not yet been thoroughly examined. This study aims to bridge this gap, as no previous study has been undertaken to demonstrate the asymmetric influence of these determinants on CO 2 emissions. It also tests the EPC hypothesis in African OPEC oil-producing countries for the first time.

Data description
In this empirical study, we mainly focus on the asymmetric impacts of unemployment and inflation rates on CO 2 as a proxy for environmental quality. We use annual data from 1990 to 2019 for the seven OPEC African countries (Algeria, Angola, Congo, Equatorial Guinea, Gabon, Libya, and Nigeria). The data was obtained from World Development Indicators. A thorough explanation of the variables is shown in Table 1.

Model specification
To investigate the link between environmental quality (as represented by carbon dioxide) and both unemployment and inflation rates, we construct the following equation: where CO 2 denotes carbon dioxide, Unem denotes unemployment rate, and Infl denotes inflation rate. (1) For empirical analysis, we then express the equation in the following function form: where lnCO 2 , lnUnem, and lnInlf are the natural logarithms of CO 2 , Unem, and Inlf, respectively. Changing variables to their natural logarithmic form reduces data skewness. i indicates the cross section of countries used, t represents time, and μ it specifies the error term.
The goal of this analysis is to look at the probable asymmetric influence of Unem and Inlf on CO 2 , which may be more visible than a symmetric analysis. Thus, the nonlinear model will be estimated by integrating the cumulative sums of positive and negative Unem and Infl changes. The estimated nonlinear model is as follows: where lnUnem_pos it and lnInlf_pos it (lnUnem_neg it and lnInlf_neg it ) represent the accumulated sums of positive (negative) changes of lnUnem and lnInlf, respectively.
Equation (3) is therefore the most suitable framework for analyzing the asymmetric effects of inflation and unemployment rates on CO 2 emissions as well as verifying the existence of the EPC. The anticipated signs of all coefficients may be negative, especially for positive shocks.

Estimation method
As the impact of positive and negative shocks differs when looking at macroeconomic performance indicators, asymmetric analysis has recently received a lot of attention. Asymmetric behavior has also been observed in unemployment and inflation rates (see Mirbagherijam 2014;Meo et al. 2018;Ullah et al. 2020). To address the research issue empirically, the study employs a panel of nonlinear autoregressive distributed lag (PNARDL) analysis, developed by Shin et al. (2014) as a new procedure in the ARDL model, to capture the asymmetric impact by including both positive and negative partial sums of independent variables. It also has the advantage of being flexible in terms of integration order, allowing the time series to be I(0), I(1), or a combination of both (Pesaran and Smith 1995;Pesaran et al. 1996).
(2) ln CO2it = 0 + 1 ln Unemit + 2 ln Inlfit + it (3) ln CO2it = 0 + 1 ln Unem_posit + 2 ln Unem_negit To prevent the strong heterogeneity bias that might arise if we use a dynamic fixed effect (DFE) model, we employ the pooled mean group (PMG) estimator as described by Pesaran et al. (1999). Additionally, the PMG panel model restricts the long-run coefficients to be the same and covers the unobserved cross-sectional characteristics while permitting the variation in short-run coefficients across panels.
For the reasons outlined above, the PMG-NARDL model is effective for precisely detecting the asymmetric effects of changes in unemployment and inflation rates on CO 2 , as well as for small samples like the one we are examining, which spans 29 years and includes seven countries. Equation (4) expresses the PMG-NARDL model.
The independent variables are divided as follows: This paper conducts initial tests and chooses appropriate methodologies to ensure reliable results and achieve its objective. The selected methodologies are effective in min ΔInfl − J , 0 handling the diverse characteristics of different time series. Additionally, the paper explores the application of modern methodologies proposed in the field and specifically runs the PNARDL model using the pooled mean group (PMG) approach. As a diagnostic test, the paper checks for the existence of asymmetries between negative and positive shocks to inflation and unemployment rates. The long-run symmetry is verified using the Wald test of the null hypothesis in Eq. (4), assuming β 2i = β 3i and β 4i = β 5i . Asymmetry exists if the null hypothesis is not accepted.

Empirical results and discussions
A complete description of the data set utilized in this inquiry is provided in Table 2. Based on the standard deviation, CO 2 is more volatile than the other variables under consideration. Yet each variable deviates slightly from its mean value, demonstrating steady variance during the sample period. It should also be noted that both unemployment and CO 2 are negatively skewed.

Cross-section dependence test
Cross-sectional dependence (CSD) is a common problem in panel data, especially when there is considerable interaction across national economies, as in the case of the countries under consideration, all of which are members of the same organization (OPEC) and have extensive oil extraction activities. Hence, ignoring the potential existence of CSD in the panel data can cause irrelevant panel econometric methodologies and subsequently inaccurate test statistics (Chudik and Pesaran 2013). For this reason, we used the Breusch-Pagan (LM), Pesaran scaled (LM), and Pesaran (CD) tests to check for CSD in our data. The results in Table 3 demonstrate that in two out of three tests, the null hypothesis of no CSD between the selected countries is rejected. These findings are expected given that the African OPEC countries share a number of economic characteristics, and any shock affecting a variable in one country would likewise affect that same variable in the other.

Panel unit root tests
If CSD is detected in the series, it is recommended to use second-generation unit root tests, such as the CIPS and CADF tests proposed by Pesaran (2007), as well as the Herwartz and Siedenburg (2008) test (HS), to determine whether either series has a unit root or is stationary. According to the findings in Table 4, unit roots exist at levels for both lnCO 2 and lnUnem, but they become stationary after the first difference. However, even at level, lnInfl is stationary, demonstrating that the primary requirement for the PNARDL model has been met.

Panel nonlinear ARDL results
We proceed to estimate the PNARDL model using the partial sum decomposition of both unemployment and inflation rates in order to determine the effects of negative and positive shocks in each variable on CO 2 . The optimum lags for the PNARDL model according to the Akaike information criteria (AIC) are (4, 4, 4). The asymmetric long-run and short-run estimates for the entire sample are shown in Table 5. Table 5 shows that changes in unemployment and inflation rates, both positive and negative, have significant longterm effects on CO 2 emissions, and all coefficients have negative signs, indicating that shocks to each variable have  ------------a negative association with CO 2 emissions, which is in principle consistent with the relevant literature. The coefficient of the positive (negative) unemployment shock is −0.25 (−0.47). More specifically, a 1% rise (reduction) in unemployment results in a 25% (47%) decrease (increase) in CO 2 emissions. This outcome implies two crucial points. First, a negative unemployment shock has a higher positive impact on carbon emissions as compared to the negative impact of a positive unemployment shock. In other words, carbon dioxide reacts asymmetrically to unemployment shocks. Although several studies have supported this negative association (Adesina and Mwamba 2019;Liu and Feng 2022), the asymmetric scenario has yet to be examined. According to our asymmetric analysis (especially negative shocks), these rentier countries are unaware of environmental issues when implementing their employment policies. Second, it appears that reducing pollution in these countries can only be achieved at the expense of human jobs. This means that an increase in the unemployment rate induces an improvement in environmental quality, providing credence to the existence of the EPC for the panel of African OPEC economies. This result supplements the findings of Kashem and Rahman (2020) and Anser et al. (2021).
Possible explanations could help clarify this result. During times of temporary prosperity, rent redistribution may prioritize unemployment reduction (negative shocks). This unsustainable policy would inevitably result in a shift in consumption and production in which neither the government nor individuals consider environmental quality because this behavior necessitates cultural patterns that can be acquired over time and do not appear in patchwork policies that frequently emerge to meet social needs. In contrast, in these oil-producing countries known for their inelastic production systems and notable socioeconomic support, energy-related consumption and activities would fall as unemployment rose, thereby enhancing environmental quality. As a result, these governments must create green jobs in nonhydrocarbon sectors such as sustainable agriculture, waste management, and renewable energy to protect people from unemployment risks while simultaneously improving public awareness of environmental preservation. Ullah et al. (2023) further asserted that by fostering innovation and creativity, human development may create opportunities for sustainable employment, allowing individuals and communities to actively participate in the transition to renewable energy.
Regarding the inflation rate, both positive and negative shocks are statistically significant at 0.01 and have a negative sign, revealing a negative (positive) correlation with CO 2 (environmental quality). While a 1% decrease in inflation only causes an 8% growth in CO 2 emissions, a 1% increase cuts CO 2 emissions by 23%, indicating that inflation rates influence carbon dioxide asymmetrically; however, unlike unemployment shocks, CO 2 reacts more strongly to positive inflation shocks than negative ones. Ullah et al. (2020) also estimated asymmetric findings and claimed that only negative shocks caused by inflation instability have a positive long-run influence on CO 2 emissions. A few other studies, however, found a negative linear relationship between inflation and CO 2 emissions (Alola et al. 2019;Ahmad et al. 2021;Musarat et al. 2021). OPEC members in Africa rely heavily on oil revenues. Therefore, during the welfare period, a time of high oil prices and low inflation rates (negative shocks), governments sought to redistribute rents to meet social needs by increasing basic commodity subsidies, assisting poor socioeconomic categories, and providing funds for local investment initiatives. As a result, consumers' purchasing power rises while their consumption costs fall, encouraging consumers (both individuals and institutions) to use more subsidized nonrenewable energy and engage in other forms of environmentally harmful behaviors that would ultimately drive CO 2 emissions to new highs. Musarat et al. (2021) also found that when the inflation rate falls, it can lead to increased demand due to cheaper raw material costs. This, in turn, motivates end-users to engage in more economic activity before prices rise again. While this development in economic activities may be favorable in terms of economic and social sustainability, it is important to note that it is significantly increasing CO 2 emissions. Policymakers in these rentier countries should be aware that low inflation is unsuitable for protecting the environment. They should play a critical role in preserving inflation stability and carefully examine raising the inflation goal to a level that is not detrimental to the economy. Subsidies for basic commodities, such as fuel, can be phased out gradually, resulting in higher prices and lower energy-related CO 2 emissions without harming the economy. Moreover, Djedaiet and Ayad (2022) argued that controlling subsidy policy, the real effective exchange rate, and the balance-of-payment situation is critical to determining inflation in a rentier country.
In the opposite situation, if oil prices fell, these governments would intentionally generate certain inflation waves (positive shocks) through exchange rate depreciation to raise oil revenues in nominal terms (i.e., in local currency terms) in order to meet spending requests. They may also impose austerity measures and progressively phase out essential commodity subsidies. This would diminish purchasing power and lead to a reduction in all types of consumption, particularly nonrenewable energy use, resulting in a reduction in CO 2 emissions. In general, the high cost of inflation is a decline in consumer purchasing power, which disrupts aggregate demand and slows economic activity, ultimately leading to a slowdown and alleviating pressure on the environment (Setyadharma et al. 2021).
Concerning short-run dynamics, the findings reveal that either positive or negative shocks in unemployment and inflation rates have no statistically significant influence on CO 2 emissions, indicating that the EPC is not supported in the short run. However, our findings align with those of Bhowmik et al. (2021), who concluded that the environmental Phillips curve (EPC) hypothesis holds true only over the long term in the case of the USA. The negative and significant error correction term (ECT) provides further evidence of a long-term link between the variables. Since ECT is equal to −0.24, any deviation from long-run equilibrium is rectified annually by 24%.

Wald test
To ascertain if the coefficients of inflation and unemployment shocks are indeed asymmetric, the Wald tests of long-run symmetry are carried out. The null hypothesis states that both indicators, positive and negative shocks, have the same effect on CO 2 in African OPEC countries (i.e., C(1) = C(2) and C(3) = C(4)). The estimated Wald test findings demonstrate that positive and negative shocks in both indicators have asymmetric long-term effects because the null hypothesis of symmetric effects is rejected at the 1% significance level for unemployment (Table 6, Part A) and inflation coefficients (Table 6, Part B). This lends further credence to the presence of a nonlinear connection between variables.

Conclusion
This paper investigates the asymmetric influence of inflation and unemployment rates on carbon dioxide emissions for African OPEC countries between 1990 and 2019, using the PNARDL simulation approach. Our paper contributes to previous research by examining the asymmetric scenario of two neglected CO 2 emission drivers and testing the validity of the EPC for the first time in these rentier countries. The study found three interesting results. First, CO 2 emissions in these countries are negatively related to unemployment and inflation rates. This appears to imply that preserving a sustainable environment would have to come at the expense of two undesirable phenomena: loss of employment and purchasing power erosion. Second, the asymmetric analysis shows that both negative unemployment and positive inflation shocks have a larger impact on CO 2 emissions than the inverse scenario. This suggests that improved environmental quality may coexist with lower employment and purchasing power levels. This is a complex issue that lacks a clear solution and instead creates a new alarming situation where three undesirable events coincide. Rising socioeconomic instability caused by unemployment and inflation may eventually have a negative impact on the environment due to people's poor quality of life. Finally, there is evidence to support the long-term existence of the EPC for African OPEC economies since CO 2 emissions and both unemployment shocks are negatively linked. This implies that in these oil-producing countries, there is no trade-off between enhanced environmental quality and unemployment.

Policy recommendations
The current research has policy implications. Policymakers should figure out how to increase employment through high levels of economic productivity while minimizing emissions. This could be achieved by boosting investments in clean technology and renewable energy industries, along with developing green business possibilities that result in steady employment growth. Moreover, the discovered link between inflation rates and CO 2 emissions in these countries may aid policymakers in implementing a consistent policy that can preserve financial stability while ensuring better environmental quality by encouraging environmentally friendly consumption behavior and imposing taxes on polluting goods and services to monitor CO 2 emissions. Furthermore, government subsidies for private expenditures on public goods, particularly fuel, may be gradually phased out, resulting in greater costs and a reduction in CO 2 emissions connected to energy. Lastly, African OPEC governments should shift from nongreen to green economic growth, and environmental concerns should not be overlooked as economic issues gain prominence.

Limitations and future directions
The study had two primary limitations: a lack of recent data, which limited the study period from 1990 to 2019, and a limited number of relevant previous studies due to the research area not being adequately studied. However, the research can be extended to include other current environmental indicators, such as the ecological footprint and the load capacity factor, while maintaining the same goals and possibly extending the data. Future studies can also consider the augmented NARDL analysis of the asymmetric effects of unemployment and inflation rates on the carbon footprint in the same region or in developed and emerging countries.