For several decades, the desire to improve the living conditions of their populations has led African economies to establish trade partnerships with each other and with the rest of the world. Whether it concerns Free Trade Areas (FTAs), Regional Economic Communities (RECs) or even simple bilateral trade partnerships, the opening of these economies to foreign trade is not always without environmental implications (Copeland and Taylor, 2004; Frankel and Rose, 2005; Managi et al., 2009; Gozgor, 2017; Opoku-Mensah et al., 2021). Thus, the issue of climate change as well as its implications for African countries has attracted a lot of attention from researchers in recent years (Acheampong et al., 2019; Acheampong and Dzator, 2020). Moreover, the natural resource wealth of most of these economies has opened them up to foreign trade and foreign investment which contributes greatly to the growth of their respective economies (World Bank, 2020; Sun et al., 2020). We also cannot ignore the fact that intra-African trade has developed enough, particularly with the General Agreement on Tariffs and Trade (GATT) and the establishment of a Continental Free Trade Area (ACFTA) with the aim of facilitating trade between member countries with high expectations of gains in terms of trade in goods and services as well as in terms of social and environmental outcomes (Awad, 2019; Sun et al., 2020).
Regarding environmental protection, a considerable part of global funding for climate change research has been allocated towards the African continent (Overland et al., 2022), with a focus on poorer and less developed countries (Fonta et al., 2018) especially since the continent is home to the largest share of the population most vulnerable to climate change (Bond, 2014). Yet, paradoxically, Africa is the continent that has contributed the least to the causes of current climate change (Busby et al., 2014; Shukla et al., 2019). This is likely related to the trade relations that the continent has with the rest of the world. For example, the volume of inward foreign direct investment in sub-Saharan Africa has increased from only $1.16 billion in 1990 to over $73.65 billion in 2021 (World Bank, 2022), thus characterizing the strong growth dynamic in which are the African economies. However, Lindmark (2002) and Mahmood et al. (2020) point out that during the early stages of a country's development, there is usually a large increase in carbon emissions. One of the explanations for this high rate of emissions is often the fact that these economies tend to relax constraints and taxes against the entry of companies and goods that are sources of pollution, in favor of internal productivity growth and job. Several studies to date point out that the tax on carbon emissions is an effective instrument to fight against environmental degradation (see Baumol and Oates, 1971; Köppl and Schratzenstaller, 2022), especially since many African countries mainly depend on imported products such as electronic waste and second-hand products from developed countries (Dauda et al., 2021).
Moreover, some authors agree on the fact that the fight to reduce CO2 emissions does not necessarily involve increasing protectionist measures or import taxes, but also increasing the domestic price of fuels and the introduction of taxes to limit the consumption of fossil fuels in order to encourage the consumption of non-polluting energies. However, the latter approach could cost more in terms of productivity and may instead reduce economic activity (Hogan and Jorgenson, 1991; Grubb et al., 1993; Glanemann et al., 2020), given that African countries are not yet shining in terms of adoption of renewable energy or green technologies (UNCTAD, 2023). However, the challenges to date are clear: the decarbonization of African economies, which will have to be resolved sooner or later, implies acting quickly by revising upwards the competitiveness in terms of new energy technologies and by encouraging a green industry, which involves seizing the opportunities offered by green growth and minimizing the risks of CO2 emissions (Saidi and Omri, 2020).
Studies linking trade openness and CO2 emissions (see Hossain, 2011; Shahbaz et al., 2013, 2017; Acheampong and Dzator, 2020; Dauda et al., 2021; Udeagha and Ngepah, 2022) attempt to propose solutions that can keep emissions relatively low compared to the maximum acceptable industrial level of 2°C. However, efforts still need to be made because even if the industrial sector is still weak, the transport sector in African economies consumes even more energy and is very polluting (World Bank, 2020). In addition, production growth, urbanization and population growth have not helped in this fight because these latter factors tend to increase CO2 emissions in Africa considerably (IEA, 2019; Acheampong et al., 2021; Djeufack et al., 2023). Indeed, despite the abundance of literature on the subject, the results reached by the studies remain mixed. On the one hand, some point to the existence of a statistically significant effect of trade openness on CO2 emissions in Africa (Shahbaz et al., 2013; Adams and Opoku, 2020; Tawiah et al., 2021; Adams and Kaffo, 2022). On the other hand, authors find no causality between the two variables (Zerbo, 2017; Yameogo et al., 2021). Even when there is a significant influence as encountered in most cases, the signs obtained lead to a real maze. Attempts at explanation highlight the fact that this divergence in results depends on both the methods of analysis used and the measures adopted (Ho and Iyke, 2019; Mignamissi and Nguekeng, 2022). Moreover, the study periods considered by the authors are most often different as well as the samples, which also explains why the results obtained in short-run analyzes on a sample differ from those in the long run.
In this work, we reassess the relationship between trade openness and CO2 emissions in the African context, using a less common and more informative measure. It is the composite measure of trade openness calculated following the methodology proposed by Squalli and Wilson (2011). It is a two-dimensional measure that takes into account not only the contribution of individual countries to world trade4 (Trade Share), but also their interactions and interconnections with the rest of the world (World Trade Share). We then use robust estimation techniques for the empirical specifications, including the Ordinary Least Squares (OLS), then the Generalized Least Squares (GLS) estimator as well as the Driscoll-Kraay specification whose coefficients are more robust in the presence of heteroscedasticity, autocorrelation and any form of spatial and temporal dependence (Driscoll and Kraay, 1998; Hoechle, 2007). Based on the assumption that the relationship between trade openness and CO2 emissions could be endogenous (Frankel and Rose, 2005; Adams and Apoku, 2020; Mignamissi and Nguekeng, 2022), we apply the Two Stage Least Squares (2SLS) estimator with internal and external instrumentation. For sensitivity analyses, we first use alternative measures of trade openness. Subsequently, we assess the effect of the presence of natural resources and, given that historical and cultural factors could also influence environmental outcomes (Acemoglu et al., 2001; Álvarez-Díaz et al., 2011; Koehrsen, 2015; Acheampong et al., 2021; Wang and Luo, 2022), we perform a sensitivity with the aim of seeing if religion, ethnicity, and spoken language can explain environmental outcomes in African countries. Finally, we appreciate the heterogeneity of the relationship across the five main sub-regions of Africa. The study shows that, overall, trade openness increases CO2 emissions in Africa. These results confirm the pollution haven hypothesis. Our results are robust to changing environmental quality indicators, using different specifications of Quantile Regression (QR), and changing the instrumentation approach. Taken together, these results provide the basis for sound and specific policy recommendations for African economies.
The rest of the paper is structured around 5 sections. Section 2 presents a brief overview of the literature. Section 3 presents the methodological approach and the data. Section 4 highlights the results of the study and their discussion. Section 5 presents the sensitivity and robustness analyses. Finally, section 6 is devoted to the conclusion and recommendations.