Environmental degradation and female economic inclusion in sub-Saharan Africa: Effort towards Sustainable Development Goal 5

ABSTRACT

The paper examines how environmental degradation affects female economic inclusion. Using Generalised Least Squares (GLS) and Instrumental Variable Approach in Two-Stage Least Squares (IV-2SLS), a panel of 22 countries in sub-Saharan Africa was analysed for the period 1990–2013. We observe that CO₂ emission (metric per capita), CO₂ emissions from electricity and heat, CO₂ emissions from liquid and fuel consumption, and CO₂ intensity (emission from solid fuels such as wood, charcoal, and coal) appear to negatively affect female economic inclusion. This implies that environmental deterioration disproportionally affects the labour force participation of women. The paper recommends that policies that are designed to mitigate environmental pollution should also incorporate measures to address gender exclusion. This effort will not only reduce environmental vulnerability but will also create a window for women empowerment in vulnerable situations. This focus holds the potential for the attainment of SDG 5 in SSA.

KEYWORDS:

• Environmental degradation
• female economic inclusion
• sustainable development
• SSA

1. Introduction

In recent years, persistent gender inequality has become a topical issue on the global development agenda. It features prominently in the Sustainable Development Goal (SDG) 5 and urges national governments to implement policies that are aimed at achieving ‘gender equality and empower all women and girls’ by 2030. Although SDG Goal 5 is the only dedicated and stand-alone gender equality goal, it is emphasised that no economic, social, ecological, or political issue is gender-neutral (Koehler, 2016:56). This reinforces the view that gender inequality is ‘intersectional’ – which is intensified and exacerbated by a constellation of factors (Kabeer, 2010). Given the recent surge in unsustainable development policies that prioritise economic growth beyond what the planet can cope with (Raworth, 2012), it is believed that women and children would be the most hurt by the evils of climate change. From this underpinning flows the principle that gender equality needs to be linked to climate justice. This paper examines the relationship between environmental degradation and women economic inclusion in sub-Saharan Africa (SSA).

Scholars of African renaissance have consistently maintained that gender exclusion is a fundamental setback to post-2015 Africa sustainable development agenda (Asongu & Kodila-Tedika, 2017). This assumption is confirmed by the recent policy trends. First, the 2018 World Economic Forum’s report on global gender gap suggests that with the step-back score of −0.6%, SSA countries have the next 135 years to achieve gender parity. Second, the transition from the Millennium Development Goals (MDGs) to the SDGs shows that SSA remains a region with the worst human development outcomes (World Bank, 2016). Third, environmental sustainability is one of the key themes that underpin the Post-2015 development agenda (Akpan et al., 2015). However, the consequences of climate change and global warming are predicted to be dire for developing countries (Asongu & Odhiambo, 2019). For instance, the World Bank reports that household air pollution from cooking with solid fuels was the major cause of welfare losses in sub-Saharan Africa in 2013 (World Bank, 2016: xii).

The literature on the triggers of gender inequality has long been a debate between political, economic, and social forces that marginalise women and girls (see Seguino, 2000; Quaranta & Dotti, 2018). These studies shed light on the systematic disadvantage of women, both in the sense of deliberate exclusion from access to assets and in the sense of persistent attitudes and beliefs that women are weak, or inferior, to men (Eastin, 2018). However, the missing link in the existing scholarship is how the entrenched power relations and social structures create gender-specific set of risks that pose disproportionate hardships for women by exposing them to the vagaries of climate change (Davis et al., 2012). Unfortunately, evidence on the impact of the environmental degradation on gender norms is largely limited (Denton, 2002), and has been particularly scarce with regards to how it influences women’s labour force inclusion particularly in emerging and transitional economies (Montt, 2018). Although it is tempting to assume that climate change equally influences the lives of women and men because the most visible effects occur on social scales, the literature suggests that already marginalised and vulnerable segments of the population bears the most brunt (Eastin, 2018).

This paper adopts the Organization for Economic Cooperation and Development’s (OECD) fit-for-purpose conceptualisation of female economic inclusion as the ‘capacity of women (ages 15+) to participate, contribute, and benefit from growth processes in ways that recognise the value of their contributions, respect their dignity and make it possible for them to negotiate for fairer distribution of benefits of economic development’ (OECD, 2011:6). Its menu includes but not limited to equal access to and control over critical economic resources and opportunities, and the elimination of structural gender inequalities in the labour market (Tornqvist & Schmitz, 2009). Labour force participation is critical in this regard because it is linked to female empowerment in the family as well as the broader society. Active labour force participation by women increases their real power over economic decisions that influence their lives and priorities in society. The main question is: to what extent does environmental degradation affect female economic inclusion in sub-Saharan Africa? The paper contributes to the scholarship on environment-women nexus in developing countries. The paper proceeds as follows: the related literature is discussed, the methodology is presented, the empirical findings are discussed, and the policy implications are offered.

2. Literature review

2.1. Linking environmental degradation to female labour force participation

In this paper, environmental degradation captures carbon (CO₂) emissions, land degradation, deforestation, and organic pollution in water (Li & Reuveny, 2006), and its consequential impacts on global warming, climate change, biodiversity, air quality, and natural resources (Donohoe, 2003). Therefore, environmental degradation and environmental pollution are used interchangeably in this paper. The environment and the world of work are inextricably intertwined. There are three pathways through which environmental degradation (in)directly impacts the levels of economic activities of women. First, the ecosystem and other infrastructure that sustain jobs often get deteriorated by environmental degradation (Montt, 2018). In most developing countries, the majority of women are engaged in climate-sensitive sectors such as agriculture, forestry, and fisheries for their livelihoods (Terry, 2009). However, since the consequences of climate change and global warming are caused by the emission of greenhouse gases, female labour participation may be disproportionately affected by environmental degradation and depletion of natural resources. Studies have established that the majority of small-holder farmers on marginal lands are women, and these lands are subject to floods, landslides, droughts, and other climate hazards (Koehler, 2016). Second, evidence suggests that environmental pollution directly influences hours worked through workers’ health status (ILO, 2018). Thus, in a situation where environmental degradation reaches critical levels, workers may be ill and call in sick (Montt, 2018:2). Therefore, Kim et al. (2017) observe that workers’ own health tends to contribute to the reduced labour supply in the long-run. The third channel by which environmental pollution affects labour supply is through the care of the illness of their dependents (Aragón et al., 2017). Evidence shows that children are uniquely vulnerable to the damaging health effect of air pollution. For instance, the World Health Organization (WHO) estimated that in 2016, about 600 000 children died from acute lower respiratory infections caused by polluted air (WHO, 2018). Hence, if care for sick children and other dependents (e.g. elderly) relies on working adults, they will likely skip work to take care of their dependents or stop job entirely if the illness becomes severe (Aragón et al., 2017). Since women are generally responsible for cooking, and as children often spend time with their mothers, household air pollution disproportionately affects women and children. It becomes apparent, for instance, when children cannot go to school due to illness and have to stay at home. Eriksson & Nermo (2010) argue that it is usually the mother, even in developed societies where gender inequality is not pervasive. In most developing countries including sub-Saharan Africa where care for dependents is chiefly the responsibility of women within the family, labour force participation of women may be affected as a result of children falling ill due to air pollution. This resonates with the extant literature that shows that labour force participation of women is negatively affected by the presence of young children (Afridi et al., 2012). It should be noted that environmental degradation in the form of ambient (outside) and household air pollution as well as water and soil pollution affect women’s ability to work through their own illness or, when the care of dependents (children or the elderly) relies on working women. Nevertheless, it may also limit the possibility of other group of females waiting to enter the labour market if the same burden of care for dependents relies on them.

2.2. Existing gender inequality and women’s environmental vulnerability

The Intergovernmental Panel on Climate Change (IPCC) articulates that climate change perpetuates existing gender inequalities (Fröhlich & Gioli, 2015). In developing countries, environmental degradation may accentuate existing gender inequality due to power relations that are intertwined within other socio-cultural structures (Djoudi & Brockhaus, 2011). According to Eastin (2018), women’s vulnerability is hugely amplified by disproportionate gendered divisions of household labour. For instance, in Uganda, Kenya, Nigeria, Mali, Benin, Zimbabwe, and in other SSA countries, women and girls particularly in rural areas are responsible for household-based work such as gathering fuel (firewood), fodder, and water for household consumption (Mubila, et al., 2011). Eastin (2018) argues that these socially constructed household responsibilities make women disproportionately vulnerable to the environmental deterioration. Similarly, women are also disproportionately discriminated against in accessing and owning lands (Agarwal, 2003). This argument needs a more nuanced explanation particularly in sub-Saharan Africa. In fact, Tiondi (2000) argues that the marginalisation of women in sub-Saharan Africa has colonial undertones. Scholars have argued that colonial government’s policies and the institutional arrangement were aimed at facilitating agricultural production and the extraction of raw materials (Sachs, 1997). Thus, the colonial policies were such that it allowed land to be consolidated in the name of men with the goal of stimulating cash crop production (Sachs, 1997). Indeed, Stichter & Hay (1995) have argued that women’s unequal access to land in sub-Saharan Africa can be attributed to European settlers who prioritised and institutionalised indigenous male authority in land administration and ownership on condition that men were superior to women in the art of farming. Unfortunately, since most post-colonial nations in sub-Saharan have inherited such institutionalised male-dominated institutions and laws, women’s low socioeconomic status has persisted (Tiondi 2000).

Analogous to this argument, Hallward-Driemeier & Hasan (2012) posit that property regimes, family codes, and inherited laws and institutions across sub-Saharan Africa often give men primacy of place over women. Because men often as the automatic heads of household provision, it restricts women’s ownership and control over property. Hallward-Driemeier & Hasan (2012) argue that head of household in nearly half of household in sub-Saharan Africa are men, which restricts women’s access to land titling and redistributive programmes. For example, while in Kenya, land is overwhelmingly titled in men’s name (Hallward-Driemeier & Hasan 2012), Ethiopia reformed the situation mandating that land be titled jointly (Deininger et al., 2008), and Botswana recently revised its 2015 land policy to allow women’s access to land. In Nigeria, evidence suggests that banks are required to contact male spouses when a married female tries to open a bank account with funds from her husband (World Bank & IFC, 2013; O’Sullivan, 2017). In other five sub-Saharan African countries: Niger, Gabon, Togo, Mali, and Democratic Republic of Congo, Hallward-Driemeier & Hasan (2012) observe that a married woman is legally required to obtain the consent of the husband before opening a bank account or withdraw funds obtained from her husband. This wanton disparity generally results in acute hardships for women when climatic change undermines agricultural livelihoods (Nuryartono, 2005). Consistent with this view, McKinney & Fulkerson (2015) and Eastin (2018) argue that environmental deterioration disproportionately exacerbates and undermines women’s economic opportunities.

Analysts contend that a consistent trend of economic growth logically has a positive effect on the emission of greenhouse gases which pose great challenges to human welfare development. The rapid economic growth and industrialisation require greater use natural resources, dirty energy sources (fossil-based), and obsolete technology, which leads to the emission of pollutants into the environment (Dinda, 2004). Consistent with this view Akpan & Akpan (2012) argue that carbon dioxide (CO₂) emissions due to economic activities constitute about 75% of global greenhouse gas emissions. Therefore, the available evidence suggests that environmental deterioration due to economic activities tend to affect the three dimensions of human development: health, education, and well-being. From an education perspective, atmospheric pollution negatively affects the ability of parents to send their children to school and this affects the ability of students to study effectively in class (Clark et al., 2012). From a heath standpoint, air pollution tends to harm peoples’ health and reduces life expectancy (Boogaard et al., 2017), and overall well-being (Graff & Neidell, 2012). Taken together, Eastin (2018) asserts that environmental pollution disproportionally hurts girls’ school attendance, undermines women’s ability to generate independent revenue, and ultimately affect gender equality. Further, Björkman-Nyqvist (2013) observes that during episodes of income shocks due to environmental threats, households in Uganda are most likely to use girls as supplementary domestic labour, which undermine their ability to stay in school. Consequently, female labour force participation is likely to be affected since education is found to have a positive relationship with employment status (Cameron et al., 2001).

Assessing the ecofeminists view that women are more likely to support environmental protection policies than men, Ergas & York (2012) revealed that carbon dioxide emissions per capita were lower in countries where women are given higher political status compared to those countries where women are accorded low political rights. Similarly, McKinney & Fulkerson (2015) reported that nations with greater female representation in governing bodies have a lower predisposition to climate footprints. These preceding discussions suggest that women and the environment are interconnected, and ecological losses tend to weaken women’s status. To this end, we hypothesise that environmental degradation exerts an independent negative impact on female economic inclusion in sub-Saharan Africa.

2.3. Data and methodology

2.3.1. Data

The paper assesses a panel of 22¹ countries in sub-Saharan Africa with the recent dataset (1990–2013). The selection of the countries and the time period was subject to data availability. The data were taken from two sources: International Labour Organization (ILO) and World Development Indicators (WDI) of the World Bank. Our dependent variable is women inclusion which is proxied by female labour force participation (ILO modelled estimates) aged 15 years and above. It measures the proportion of female population (ages 15+) that is economically active. The independent variables and our variables of interest are: carbon dioxide emissions (metric tons per capita), carbon dioxide emissions from electricity and heat production (% of total fuel combustion), carbon dioxide emission from liquid fuel consumption (% of total), and carbon intensity (kg per kg of oil equivalent energy use). Based on the preceding discussion on the environment-women nexus, the measures of environmental degradation are expected to exert negative effect on female labour force participation. The data were taken from the WDI. Five control variables were used in order to avoid omitted variable bias: gross domestic product per capita (GDP per capita), foreign direct investment (FDI), urbanisation, foreign aid, and British colonial origin (dummy). Asongu (2015) argues that English common law countries are more associated with effective institutions compared to their French Civil countries. Analogous to this historical view, Asongu et al. (2019) posit that English common law countries are more likely to adapt more rapidly to the impact of environment deterioration compared to their French counterparts. Detailed description of the variables is shown in Appendix 1.

2.3.2. Estimation techniques

It should be noted that previous studies on environment-women nexus are largely based on one estimation technique such as pooled/Ordinary Least Square (OLS), which relies on the parameter estimates at conditional mean. However, OLS may not offer optimal results when there are serious issues with regards to autocorrelation, first-order serial correlation, and endogeneity. As result, Generalised Least Squares (GLS) and the instrumental variable in Two-Staged Least square IV(2SLS) estimations were employed to control for specific characteristics. The GLS with fixed effect for both the country and year was used to control for unobserved country level effects such as historical differences and structural peculiarities. According to Greene (2000), by controlling for country fixed effect, the problem of omitted variable bias in cross-sectional studies is reduced. By estimating the weighted least squares through fixed effects, the GLS model is superior to OLS that assigns an equal weight to each observation (Reed & Ye, 2011). Of course, the GLS model is efficient in dealing with heteroscedasticity and serial correlation in the panel datasets. Then, we employed IV(2SLS) to deal with the problem of endogeneity that is likely to persist despite the use of GLS. It should be noted that the efficiency of IV(2SLS) depends on the validity of the instruments. Nevertheless, the appropriate instruments often used in the literature include but not limited to ethno-linguistic fractionalisation (elf), common law, latitude, and corruption. These instruments have been extensively used as policy variables in the growth literature (Alesina et al., 2003). The baseline specification assumes the following model: (1) $\text{FC}{\text{L}}_{it}={\theta }_0+{\text{β}}_1\text{ENVIRONMENTD}+\sum _{h=1}^5{w}_h{W}_{hit-t}+{\eta }_i+{ϵ}_{it}$(1) where fcl_i,t is our dependent variable (female labour force participation) for country i at period t, ${\theta }_o$ is a constant, ENVIRONMENTD is environmental degradation involving four pollutants: carbon dioxide emissions – metric tons per capita (CO₂mtpc), carbon dioxide emissions from electricity and heat (CO₂elc), carbon dioxide emission from liquid fuel consumption (CO₂lfc), and carbon intensity (CO₂itn), W is the vector of control variables, η_i is the country-specific effect and ϵ_i,t is the error term.

2.4. Empirical results

The GLS and IV(2SLS) techniques were used to estimate the impact of environmental degradation on female economic inclusion. Columns (1) to (6) in Table 1 present the GLS estimates for the effect of environmental degradation on female economic inclusion. The rationale is that the measures of environmental degradation come from different sources and it is important to estimate their independent effect on female economic inclusion. From columns (3), (4), and (5), it can be seen that carbon dioxide emission from electricity and heat, carbon dioxide emission from liquid and fuel consumption, and carbon dioxide intensity (emission from solid fuels) have significant and negative effect on female labour force participation. However, in column (6) which happens to be the combined model, carbon dioxide emission (metric per capita), emission from electricity and heat, and emission from liquid and fuel consumption are significant and shows a negative relationship with female labour force participation. These findings signal that environmental degradation with regards to carbon dioxide emission (metric per capita), carbon dioxide emission from electricity and heat, emission from liquid and fuel consumption, carbon intensity due to coal consumption appear to impair female economic inclusion in sub-Saharan Africa.

Table 1. Dependent variable – female economic inclusion: GLS estimations.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
GDP Per Capita	0.00133*** (0.000221)	0.00132*** (0.000212)	0.00154*** (0.000204)	0.00133*** (0.000200)	0.00130*** (0.000208)	0.00142*** (0.000216)
Urbanisation	−1.60e-06*** (5.02e-07)	−1.43e-06*** (5.21e-07)	−1.77e06*** (4.93e-07)	−1.43e06*** (5.06e-07)	−1.65e-06*** (5.14e-07)	−1.26e-06** (5.33e-07)
Foreign Direct Investment	0.441** (0.174)	0.447*** (0.170)	0.548*** (0.171)	0.386** (0.172)	0.440** (0.175)	0.390** (0.166)
Foreign Aid	0.495*** (0.0965)	0.510*** (0.0939)	0.564*** (0.0965)	0.499*** (0.0979)	0.537*** (0.0987)	0.487*** (0.0998)
British Colonial Origin	6.618*** (1.644)	7.123*** (1.703)	5.494*** (1.614)	6.945*** (1.620)	6.400*** (1.646)	5.170*** (1.761)
CO2 Metric Per Capita	−0.548 (0.476)					−5.790*** (1.264)
CO2 kt		−1.68e-05 (1.05e-05)				6.96e-05*** (2.17e-05)
CO2 from electricity & heat				−0.133*** (0.0390)		−0.400*** (0.0619)
CO2 from liquid & fuel C.			−0.0594* (0.0330)			−0.392*** (0.0555)
CO2 Intensity					−0.963** (0.660)	3.351*** (0.967)
Constant	62.87*** (2.106)	61.98*** (2.151)	67.00*** (3.172)	64.76*** (2.353)	63.01*** (2.388)	99.24*** (5.670)
Observations	474	474	474	449	447	447

Note: Standard errors in parentheses.

***p < 0.01.

**p < 0.05.

*p < 0.1.

However, we performed a number of tests to determine whether the problem of endogeneity exists between environmental degradation and the female economic inclusion variable. The results from the Hausman test in Table 2 indicate that the measures of environmental degradation are endogenous to female inclusion, suggesting a reverse causality. This is consistent with the ecological feminists view that both women and the environment are mutually reinforcing and interconnected (Kabeer, 2010). Empirical studies have also demonstrated that women are more concerned about the environment compared to men and having more women in the decision-making position will engender environmental quality (McKinney & Fulkerson, 2015). To deal with the problem of endogeneity, we instrumented the measures of environmental degradation with ethnic fractionalisation. Further, we used Anderson Canon LM test to ascertain whether the instrument used is associated with the endogenous variables. The assumption is that the instrument should be associated with the endogenous variable. The Anderson Canon LM correlation statistics were significant in columns (1) to (5), indicating that ethnic fractionalisation is associated with endogenous variables.

Table 2. Dependent variable – female economic inclusion: IV(2SLS) estimations.

Variables	(1)	(2)	(3)	(4)	(5)	(6)
GDP per Capita (PPP $)	−0.00117*** (0.000225)	−0.00104* (0.000541)	−0.00133*** (0.000201)	−0.00130*** (0.000272)	−0.00105** (0.000531)	−0.0192 (0.106)
Urbanisation	−1.82e-06*** (5.28e-07)	−1.32e-06* (6.94e-07)	−1.43e-06*** (5.09e-07)	−1.61e-06*** (5.18e-07)	−8.43e-07 (1.18e-06)	−2.20e-05 (0.000145)
Foreign Direct Investment	0.344* (0.186)	0.316 (0.273)	0.385** (0.177)	0.399* (0.206)	0.354 (0.239)	3.262 (16.99)
Foreign Aid	0.540*** (0.0995)	0.428*** (0.149)	0.499*** (0.0985)	0.454*** (0.127)	0.486*** (0.104)	−1.804 (20.98)
British Colonial Origin	6.103*** (1.668)	7.942*** (2.777)	6.946*** (1.622)	7.048*** (2.000)	9.335** (4.308)	−97.71 (619.2)
CO2 emission per capita		−1.948 (2.403)				−52.95 (472.7)
CO2 kt					−5.33e-05 (6.58e-05)	0.00297 (0.0171)
CO2 from Electricity & Heat			−0.134** (0.0641)			−7.119 (61.93)
CO2 from liquid & fuel C.				0.207** (0.116)		0.869 (19.17)
CO2 Intensity	−2.628** (1.216)					126.3 (1,117)
Constant	65.30*** (2.785)	63.23*** (2.209)	64.77*** (2.618)	56.03*** (8.622)	60.36*** (3.620)	149.5 (2.198)
Observations	447	474	449	474	474	447
Year Dummy	NO	NO	NO	NO	NO	NO
Country Dummy	NO	NO	NO	NO	NO	NO
R-squared	0.273	0.270	0.298	0.253	0.266	–
Anderson Canon LM Statistics	133.421 (0.0000)	18.948 (0.0000)	165.713 (0.0000)	40.177 (0.0000)	12.348 (0.0004)	0.017 (0.8955)
Cragg-Donald Wald F	187.211	109.445	258.554	113.249	122.491	0.003
Hausman Test	0.07911	0.04823	0.00144	0.13912	0.00192	–
Sargan	0.542	0.134	0.324	0.050	0.123	0.013
Endogeneity Test	0.0998	0.0729	0.0011	0.003	0.0043	0.1302

Notes: Standard errors in parentheses. From column (1), Stock & Yogo (2005) weak identification test critical values for 10%, 15%, 20%, and 25% are: 16.38%, 8.96%, 6.66%, and 5.53%, respectively. From Column (3), the weak identification test critical values for 10%, 15%, 20%, and 25% are: 16.38%, 8.96% 6.665%, 5.53%, respectively. We reported the Stock & Yogo (2005) weak identification for only consistent estimations.

***p < 0.01.

**p < 0.05.

*p < 0.1.

Additionally, Sargan test was performed to determine the validity of the instrument. The statistical significance of the Sargan test indicates that valid instrumental variable was used. It must be noted that the power of instrumental variables in IV(2SLS) is determined by the Cragg-Donald F-statistic. According to Stock & Yogo (2005), the instrumental variable should be larger than the maximum IV size of 10%. The Cragg-Donald F statistic shows that ‘powerful’ instruments were used. From Table 2, columns (1) and (3) show that carbon dioxide intensity (emission from coal consumption) and emission from electricity and heat are associated with female economic inclusion. Thus, both carbon Intensity and carbon dioxide emission from electricity and heat are negative and significant at 5% level respectively. The result from column 6 (Table 2) is not different from column 6 (Table 1). However, one key point must be flagged. Thus, carbon dioxide emissions from electricity and heat and carbon dioxide intensity (emission from coal) appear to individually have the most consistent and negative impact on female economic inclusion in sub-Saharan Africa because the findings hold in both GLS and IV(SLS) estimations.

2.5. Discussion

Some of our findings are consistent with the theoretical and empirical literature. The findings resonate with the hegemonic school of thought which holds the view that the continued emphasis on economic activities such as production, capital accumulation, and overexploitation of natural resources amidst disrespect for environmental justice tend to promote cynicism and self-interest to the detriment of inclusive human welfare outcomes (Tsai, 2007). The negative statistical significance of CO₂ emissions (metric per capita) CO₂ emissions from electricity and heat consumption, CO₂ emissions from liquid and fuel consumption, and CO₂ intensity have serious implications for post-2015 African development agenda. For instance, the World Bank report on SDGs shows that household air pollution from cooking with solid fuels such as wood, charcoal, and coal was the major cause of welfare losses in sub-Saharan Africa (World Bank, 2016: xii). The 2019 report on SDGs also shows that 44% of people in sub-Saharan Africa lack access to electricity (UNDP SGD Report, 2019:36). Despite the electricity gap in the sub-region, our findings suggest that the current level of energy consumption appear to pose a serious threat to efforts aimed at women empowerment. This means that greenhouse emissions due to unsustainable use of energy sources and the corresponding impact of climate change are likely to have a devastating impact on women. This is important because majority of women in sub-Saharan Africa depends on rain-fed agriculture (Davis et al., 2012), suggesting that the impact of climate change due to CO₂ emissions will have a disproportionate impact on women (Djoudi & Brockhaus, 2011). However, it is not clear why most of the environmental degradation variables have negative signs but insignificant in IV(2SLS) estimations. The plausible reason could be that majority of sub-Saharan African countries are still depended on primary products and are not fully industrialised. Hence, their economies may not be currently emitting huge pollutants from various sources to have an effect. Nevertheless, the negative signs suggest that environmental degradation could have a devastating effect on gender norms for which reason sustainable practices should be incorporated into development policies as most countries in the sub-region plan to industrialise their economies. Furthermore, our findings indicate that GDP per capita tends to have a negative impact on female economic inclusion. The plausible reason is that the current level of economic development in sub-Saharan Africa appears not to support female economic empowerment. Also, urbanisation is significant and exert a negative influence on female inclusion and resonates very well with the extant literature. However, foreign direct investment and foreign aid have proven to relevant for female economic advancement in the sub-region.

2.6. Conclusion and policy implication

The paper examines the relationship between environmental degradation and female economic inclusion in sub-Sahara Africa. The GLS and IV(2SLS) estimations were employed. Generally, the paper finds that CO₂ emission (metric per capita), CO₂ emissions from electricity and heat, CO₂ emissions from liquid and fuel consumption, and CO₂ intensity (emission from solid fuels such as wood, charcoal, and coal) appear to negatively affect female economic inclusion. Furthermore, economic growth, foreign direct investment, and foreign aid are important ingredients of women economic empowerment in the region. Our results also signal countries with English common law background are most likely to promote women economic inclusion in sub-Saharan Africa. However, urbanisation was observed to negatively impact on women economic empowerment. Within the framework of gender inequality, the findings suggest that environmental degradation is likely to hinder the achievement of the SDG 5 in sub-Saharan Africa. The SDG 5 seeks to achieve ‘gender equality and empower all women and girls’ by 2030. The paper challenges cultural beliefs and national laws that have denied equal opportunities and rights for women in sub-Saharan Africa.

Based on the above results, the paper recommends the following: First, it is important to integrate sustainable practices into the development policy. Thus, environmental standards should be strictly enforced to ensure that economic activities and businesses operate within environmental justice and ethical governance. Second, policies designed to mitigate environmental degradation should also incorporate measures to address gender exclusion. This will mean that in addition to reforming socio-cultural norms and values as well as national laws that make women extremely vulnerable to adverse environmental concerns, policies that aim at improving women’s rights should be vigorously pursued since available evidence suggests that having more women at decision-making positions has the potential to improve environmental quality (Ergas & York, 2012). Furthermore, since women are disproportionately hurt by the vagaries of environmental degradation, efforts aim at advancing women’s social status would not only reduce environmental vulnerability but will also create another window for women empowerment in vulnerable situations. Future studies can improve on the extant scholarship by investigating how various components of environmental pollution affect gender development based on case studies reflecting country-level realities. Also, the measures of environmental degradation used in this paper do not include water pollution and deforestation damages, future studies should explore such dynamics.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 The 22 countries are: Benin, Botswana, Burundi, Cameroon, Congo Republic, Cote d’Voire, Democratic Republic of Congo, Ethiopia, Gabon, Ghana, Kenya, Mauritius, Mozambique, Namibia, Niger, South Africa, Senegal, Sudan, Tanzania, Togo, Zambia, and Zimbabwe.

Appendices

Appendix 1. Variables, description, and sources

Table

Variable	Description	Source
Female economic inclusion (female labour force participation)	Percentage of female population, ages 15+ (modelled ILO estimate).	International Labour Organization (ILO)
CO2 emissions per capita	CO2 emissions (metric tons per capita)	World Bank (WDI)
CO2 emissions (kt)	CO2 emissions (kt)	World Bank (WDI)
CO2 emissions from electricity and heat	CO2 emissions from electricity and heat production, total (% of total fuel combustion)	World Bank (WDI)
CO2 emissions from liquid and fuel consumption	CO2 emissions from liquid and fuel consumption (% of total)	World Bank (WDI)
CO2 Intensity	CO2 Intensity (kg per kg of oil energy equivalent use)	World Bank (WDI)
GDP per capita	GDP per capita (Annual %)	World Bank (WDI)
Foreign direct investment	Foreign direct investment, net inflows (% of GDP)	World Bank (WDI)
Urbanisation	Population living in the largest metropolitan area	World Bank (WDI)
Foreign aid	Net Official Development Assistance (% of GNI)	World Bank (WDI)
British Colonial origin	British Colonial origin (dummy variable)	World Flag Database

Appendix 2. Descriptive statistics

Table

Variable	Obs	Mean	Std.Dev.	Min	Max
fcl	528	61.371	19.469	15.8	94.2
co2kt	528	22 822.33	85 091.65	40.337	503 000
co2mtpc	528	1.209	2.013	.021	9.979
co2lfc	528	68.782	25.746	0	100
co2itn	500	1.377	1.284	.085	10.39
co2elec	502	24.823	21.099	16.2	96.296
fdi	525	2.721	4.449	−8.589	41.81
urb	480	1 850 000	1 640 000	139 000	1.07e+07
gdppc	528	4127.121	4647.815	247.766	18 282.23
elf	528	.685	.159	.295	.875
bcolony	528	.456	.499	0	1
faid	525	8.434	8.927	–.251	67.735
bcolony	528	.456	.499	0	1

Source: Authors’ Computation.

Appendix 3. Correlation matrix

Table

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)
(1) fncl	1.000
(2) co2mtpc	−0.275	1.000
(3) co2kt	−0.195	0.843	1.000
(4) co2lfc	0.092	−0.750	−0.502	1.000
(5) co2elec	−0.227	0.514	0.494	−0.555	1.000
(6) co2itn	−0.177	0.538	0.365	−0.387	0.641	1.000
(7) gdppc	−0.362	0.440	0.286	−0.291	0.077	0.286	1.000
(8) urb	−0.119	0.092	0.227	−0.046	0.091	−0.171	−0.244	1.000
(9) fdi	0.127	−0.146	−0.070	0.138	−0.187	−0.177	0.104	−0.094	1.000
(10) faid	0.390	−0.377	−0.222	0.325	−0.134	−0.073	−0.327	−0.152	0.038	1.000
(11) bcolony	0.265	0.120	0.267	−0.089	−0.022	−0.110	−0.114	−0.097	0.120	0.158	1.000

Source: Authors’ Computation.