The “Africa Rising”: An Empirical Analysis of the Determinants of Per-Capita Growth

In developing economies, the race toward inclusive development has prompted researchers to reconsider the drivers of growth in view of achieving the Sustainable Development Goals (SDGs). This research note has the purpose to explore the determinants of African growth after analysing reference literature to select the explanatory variables. We examine growth in a panel of 54 African countries using the generalised method of moment system estimators (GMM-sys), by controlling idiosyncratic unobserved individual effects of countries with fixed-effects models. Using GMM-sys to estimate growth models is not novel, and many previous studies have used this appropriate approach for growth analyses. As far as we know, we have not found any studies analysing the determinants of growth in a panel of 54 African countries and/or using GMM estimators with data referring to the last decade. We have used the real per-capita GDP as a dependent variable. As a data source, the main international organisations (UN, WB, IMF) have been considered, and the analysed timeframe is 2010-2019, where we have more data completeness. Based on the time-series data collected and estimation methodology used, the findings show determinants of African growth, and our findings can have an implication for managers and decision-makers with an interest in African emerging markets. The real per-capita GDP with one order of lags is the variable with the highest magnitude in all models and it is expected of us. Furthermore, we have found confirmations for the variables of the macroeconomic framework highlighted in the recent reference literature on growth, and we have found new evidence for the variables of the African merchandise trade across developed and developing countries. In the conclusion, we highlight the importance of a sound governance and business environment in African countries for achieving the SDGs.


Introduction
Since the new millennium, Africa has had considerable GDP growth that has more than doubled.Over the last decade, six of the fastest-growing economies in the world were in Africa."Africa rising" has prompted academics and analysts to reconsider the issue of African economic growth.This is affecting firms' strategies, and those from developed countries are seeking to enter these emerging markets (Ferrucci et al., 2018;Scalamonti, 2022;Abdu et al., 2022).
African countries that gained independence from colonial rule in the sixties experienced their own model of cultural, social, and economic development.According to the theory of development, all societies advance through similar stages of development, and this means that underdeveloped countries today are in the same condition as developed countries were in the past (Marini, 2004;Schwab, 2014Schwab, , 2015)).This means that underdeveloped countries are not merely a primitive version of developed countries, but they are unique in their features and structures (De Jong, 2009;Hofstede et al., 2010).
Underdeveloped countries can accelerate their development, for instance, by implementing structural reforms, by attracting capital flows, and by enhancing technology transfers, as well as, they have increased the integration along global value chains.
Although the political and macroeconomic framework of African countries can be unstable, nonetheless, these emerging markets are often considered an opportunity, anyway with a given operative risk.In other words, labour markets may be inadequately regulated, the rule of law may be poor, and corruption may be high, logistical difficulties may depend on infrastructural deficiencies, or trade may be difficult due to the absence of codes of conduct and best practices.
Can African markets grow and break with their colonial trading past?On the other hand, can Africa successfully integrate into the global economy as it has in other areas of the world?Despite the difficulties, many African countries are encouraging drawing up action agendas and implementing structural reforms (Nafziger, 2012;Kuada, 2014;Mazrui and Wiafe-Amoako, 2015;Mol et al., 2017;Heshmati, 2017Heshmati, , 2018;;Oluwatayo and Ojo, 2018;Wiafe-Amoako, 2021).
In the last decade, Africa has experimented with high levels of growth, but there are still governance weaknesses.This means that debate on the determinants of African growth is central for economists and scholars.Economies in transition, such as emerging and developing, are experiencing an evident socio-economic dynamism, and, in the near future, they will have to face the challenge of change, as much as developed countries will have to do (Dallago and Casagrande, 2023).Globalisation has broken the consolidated production paradigms, and new business opportunities have emerged around the world.This has increased competition between firms to enter global value chains, and the pressure on emerging markets.Those African trade with developed countries primarily raw materials and commodities, while, if these are traded with developing countries, they are dutiable.
The state of crisis that began in 2008 has never stopped.Former, the sovereign debt crisis in the Eurozone, then, the uprisings in the Middle East and North Africa, and now, the global pandemic caused by Covid-19, the imbalances in the US-China relationship, and the Russia-Ukraine war are seriously damaging the world economy.Advanced economies are growing slowly, or risk a new recession, as they have begun to suffer from growing internal imbalances and income disparity risking compromising the long-run economic and social stability (IMF, 2020(IMF, , 2021(IMF, , 2022)).
Interest in African emerging markets is growing at least for three reasons: (i) governance in developed and developing countries, especially in Eastern and Southern Asia, is concerned about ensuring the supply of strategic raw materials to manufacturing industries; (ii) the so-called "African lions" -Ethiopia, Ghana, Kenya, Mozambique, Nigeria, and South Africa -have experienced fast growth (IMF, 2019); (iii) it expects that the African Continental Free Trade Area Agreement-AfCFTA will increase income in countries by at least 9% by 2035 (World Bank, 2020, 2022).Multilateral and free trade agreements will be important for the success of African countries, as a consequence of the economic and geopolitical processes that are affecting the globalised world.
Therefore, our research note has the purpose of exploring the determinants of African growth in a panel of 54 counties 1 after analysing reference literature to select the explanatory variables.The rest of the paper has been structured as follows: (i) the reference literature, (ii) the data analysis, (iii) the conclusions.

The reference literature
We have found the reference literature to select the explanatory variables of African growth from the online search engine discovered.ed.ac.uk developed by the University of Edinburgh, by inserting the following title key: Africa growth; filter: gross domestic product; time frame: 2011-2022, articles' type: business and economics.A clustering of the reference literature on African growth is shown at Table A.1 in the Annex.
Development economists have produced many empirical studies about the drivers of growth, however, their findings can change over time and based on the considered countries.As a result, authors can differently explain growth, based on the specific research questions or analysis methods they adopt.From the reference literature, it emerges that differences in economic policy among countries can explain their gaps in economic growth.
The neoclassical economic theory considers capital accumulation as a driver of growth, but endogenous growth models have also highlighted the key role of employment, productivity, human capital formation, and technology, as a result, unemployment and low knowledge capital cause slow growth (Solow, 1956(Solow, , 1957;;Lucas, 1988;Romer, 1990;Grossman andHelpman, 1990, 1991;Barro andSala-i-Martin, 1992,1997).In other words, development can depend on public and private investments, foreign direct investments, and international aid.The latter, if granted on the basis of the level of development reached by the recipient countries, has proven to be better.Moreover, in a globalised world, a part of the trade is along global value chains.For example, African trade with China has intensified as a result of diplomatic actions in countries rich in natural resources.
The neoclassical economic models consist of a set of assumptions useful for analysing market behaviour with an adequate mathematical formalisation.Under its assumptions, the appropriate unit of analysis is the individual consumer or firm rationally choosing among available alternatives according to preferences, by maximizing the utility function or profit.Therefore, preferences are exogenous to models, transaction costs are null, and information is perfectly available as the Coase theorem implies.
However, this can be deceptive, in that each individual or firm always knows the quality of offered goods and their prices.Nonetheless, the neoclassical framework is a fundament in economic analysis.
Alternatively, two main theories of economic development emerged in the second half of the XX century.On the one hand, there is the "modernisation theory" grasping from the cultural and technological differences among countries the explanation of their underdevelopment (Sadik-Zada, 2021).
The capitalistic way should be a solution for underdeveloped economies and technology is the most important factor in the development analysis, while cultural differences and traditionalist behaviours can be an obstacle to progress and innovation.
On the other hand, there is the "dependency theory", suggesting that underdevelopment can be explained in part by the exploitation of poor countries by the rich ones, due to the globalisation that privileges some countries at the expense of others (Sadik-Zada, 2023).
South countries rich in natural resources remain underdeveloped to advantage of developed ones exporting cheaply raw materials, which are then processed in developed countries, and then sold in developing countries as manufacturing goods with a higher value-added.In most ex-colonised countries exists a dependency on trade with ex-colonising countries, properly traced to this past experience shaped the economic relations.In the past, import substitution policies with locally manufactured products and inter-and intra-regional trade agreements were adopted as solutions to this order of issues.
In conclusion, from our clustering over the reference literature, the most questioned variable in empirical studies is the governance climate.The institutional level reached by countries affects their business environment.In other words, there is a positive nexus between the quality of the institutional and business environment with growth, but it requires sound governance (Acemoglu and Robinson, 2012;Acemoglu et al., 2019;Babajide et al., 2021;Festré, 2021;Glegg et al., 2021;Lin, 2021;Razin, 2022).Indeed, studies suggest that productivity, innovation, and a stable macroeconomic framework can affect growth based on the quality of the institutional and business environment in the country.
A clustering on the reference literature has been useful to determine which explanatory variables to include in econometric models.We develop our models starting from the macroeconomic determinants of growth most frequently used in empirical studies, such as openness, inflation, unemployment, external debt, net-ODA, FDI inflows, received remittances, natural resource rent, urbanisation, public expenditure, fixed and human capital formation, innovation, productivity, and last but not least the governance climate.Furthermore, we want to consider how the global interdependencies across markets can contribute to explaining African growth (Cooper and Barro, 1997).In other words, growth can create the conditions for a country to have a competitive advantage in the trade with other countries.Therefore, trade and growth can be dependent or independent, otherwise, when there is a negative relationship between them, this can depend on the imports being higher than exports.
As a result, it seem to us necessary to include in our framework of analysis proxy variables for African countries' merchandise trade across clusters of developed and developing countries as World Additionally, the academic and political debate on the trade-off between efficiency in resource allocation and public interventionism in the economy led us to separately consider in our models the variables related to public expenditure, fixed and human capital formation, innovation, and productivity.
Finally, we have found studies analysing growth with the generalised method of moment (GMM), but as far as we know, there are no studies analysing growth with a panel of 54 African countries and/or using GMM estimators over the period 2010-2019.
3. The data analysis

Dataset and econometric models
We have used a set of explanatory variables extracted from the UN-dataset (UNCTAD and UNDP), and the WB-dataset (World Development Indicators-WDI and World Governance Indicators-WGI) over the period 2010-2019 for all 54 African countries.Time-series have been integrated, when necessary, for a few missing values (2%) with data from the IMF (World Economic Outlook-WEO), otherwise from secondary sources (CIA-World Factbook's country surveys).A panel-dataset allows us to explore not only the cross-sectional dimension but also the time-variant one.As a result, the reliability of our panel-dataset based on its completeness is at 98%.
In the Annex at Table A.2, we show the main descriptive statistics and proxies for the variables used.An acceptable level of variability over the time dimension exists, while the cross-sectional dimension shows a higher level of this for some variables.In Table A.3, in the Annex always, we show the statistical associations between variables that are non-excessive for the dependent variables, particularly, and among the regressors, generally.
Our estimation strategy examines African growth using the Generalised Method of Moments estimator (GMM).We have computed over the time-series 2010-2019 the average value every two timeunits, then having a stationary time-series of five time-units.Using a GMM to estimate growth models is certainly nothing new, and many previous studies have utilised this approach for growth analyses, therefore, it is suitable for a study such as ours.In other words, this estimator allows to correct endogeneity when using a panel-dataset with variables that are potentially endogenously determined (Bond et al., 2001).By using the orthogonality conditions, the GMM estimators allow efficient estimation even in the presence of heteroscedasticity of unknown form.
We have adopted the "two-step" GMM-system estimator (Arellano and Bover, 1995;Blundell and Bond, 1998).This procedure is more efficient than the differencing, especially for a panel dataset like ours, where N is more than T. The GMM-sys extended the difference model by adding equations in levels to the regressions run in the first differences.The second equation allows the introduction of additional instruments.
In other words, for endogenous variables in levels, their own lagged differences serve as instruments, thus the additional moment conditions efficiency is increased.This means the modelling also takes care of finite sample bias whether variables are highly persistent and used as weak instruments for the first differences (Bond et al, 2001).We have also used the finite sample bias correction by Windmeijer (2005) for robust standard errors in the models.Finally, an unbiased GMM estimator depends on the validity of the instruments and maintaining the number of this below the number of groups is a good rule of thumb (Roodman, 2009a, b).
Instruments should be correlated with endogenous instrumented variables while conforming to the orthogonality condition to prevent errors (Baum, 2003).A high p-value for Sargan and Hansen is a confirmation of the correct specification of models under the null hypothesis of non-overidentification and instrumental validity.
On the one hand, Sargan relies on the assumption of homoscedastic errors, but this puts limitations on the strength of the test when the assumption is weak.On the other hand, the test is not exposed to the same instrumental proliferation weaknesses as the Hansen test.
Based on our heterogeneous dataset, there is a high probability of idiosyncratic shocks in each country, and there is a potential violation of the homoscedasticity assumption.To consider the Sargan test alone may be misleading, while considering both tests can be more convenient.However, the Hansen test is better in our case.
The dependent variable is real per-capita GDP, and the dynamic specification is given by the same dependent variable with one order of lags into models.We have estimated models with regressors at time t, and t-1 to consider the eventual lagged effect of macroeconomic policies on variables, or the persistent effect on trade.The models have been estimated with the open-source statistical software Gretl, as below [1]: where, Y i,t is the vector of the dependent variable; Y , −1 is the vector of the dependent variable with one order of lags on the right side of the equation; X , and X , −1 are vectors of time-variant explanatory variables -with one order of lags, these grasp the nearest and significant among the lagged effects of macroeconomic policies and the persistence of trade; is the vector of the idiosyncratic unobserved time-specific effects to prevent a contemporaneous correlation due to time-related shocks; finally, α, β, and δ are vectors of the coefficients that want to be estimated, , is the vector of the idiosyncratic individual and time-specific error terms in the regressions.
Finally, we have also estimated fixed-effects models for controlling idiosyncratic unobserved individual effects of countries, such as cultural, ethnic, and religious, that indirectly can influence growth, and without dependent variables with one order of lags in the right side of the equation to avoid a multicollinearity issue, but with time-dummies for controlling the time-related shocks.
We have selected the list of fitting regressors implementing a sequential selection algorithm minimizing the Akaike information criterion-AIC (Chakrabarti and Ghosh, 2011).Wald test for the joint significance of regressors, fixed-effects, and time dummies is shown, as well as the Welch test to select the fixed-effects model with respect to pulled-OLS is also shown.

Findings and interpretation
Estimated models with the significant variables are shown in Tables 2, 3, and 4. In Table 4, we show models with fixed effects to consider the cultural, ethnic, and religious features that can additionally contribute to promoting growth and explaining the possible gaps between countries.
The real per-capita GDP with one order of lags is the significant variable with the highest magnitude in all models.It is expected of us coherently to the literature on growth.This is clear confirmation that the dynamic approach is suitable for capturing the effects of past policies on growth.
Inflation could be expected to negatively impact growth.In fact, it is not uncommon to find it associated with a more unstable economic system (Kagochi et al., 2013;Asongu, 2014;Walle, 2014).
However, its significance and positive magnitude for the variable with one order of lags can be affected by competitive devaluations in the related foreign exchange markets implemented by policymakers to encourage import-export, otherwise, it can be related to short-term assessments of labour market efficiency.As well, external debt has been found to be significant, which means that African countries have had a need to finance their growth recurring to founds of international organisation financing development (Mbate, 2013;Kedir et al., 2017;Mensah et al., 2019;Ehigiamusoe and Lean, 2020;Idun, 2021).
Openness is significant and it is not uncommon for more liberalised economies to be better positioned along global value chains, for instance, benefiting from positive externalities on productivity by learning-by-doing in trade, or by collaborations and competition on international markets (Chang and Mendy, 2012; Elhiraika et al., 2014;Brueckner and Lederman, 2015;Koomson-Abekah and Nwaba, 2018;Osei et al., 2019;Udeagha and Ngepah, 2021;Abdu et al., 2021).However, a rapid openness can also damage development, if it does not occur in the appropriate way and at the right time, especially in developing countries, where they tend to specialise in traditional productions or in industries where innovation is not the core, thus, by becoming more vulnerable to external negative shocks.This means the openness effect on growth can be ambiguous in developing countries, and more openness can have a crowding-out effect on domestic growth and investments.
Nonetheless, more openness can increase productivity, facilitate the manufacturing industry's upgrading, promote technological and institutional advancement, and finally, increase capital accumulation, as a result, intermediate manufacturing imports and goods exports rise.
Productivity reflects the manufacturing industry's ability to add value to outputs.Higher productivity has important implications for growth.Although the variable shown in the below models is significant, it has a negative sign.This effect can depend on the positioning of countries' manufacturing systems along the global value chains, primarily on that of traditional industries.
Natural resource rent is significant.Countries rich in natural resources are usually characterised by their high dependency on them, their low economic diversification, and the volatility of their commodity prices and revenues.As a result, a negative sign for the variable with one order of lags refers to a crowding-out effect, i.e., the Dutch disease due to an abundance of natural resources and raw materials.
In developing and emerging economies, FDI inflows can contribute to growth in a different way, but proofs of their effect can be contrasting (Poku, 2016;Shittu et al., 2020;Hagan and Amoah, 2020).
Indeed, we have found evidence that FDI inflows positively affect growth, but controlling for the fixed-effects they negatively contribute.Nevertheless, this evidence is not too strong, and the magnitude of estimated coefficients is very poor.In other words, it can depend on the complementarity degree, or the substitution effect between FDIs and the other foreign capital -such as net-ODA and received remittance, or the countries' domestic policies on physical investments and human capital formation.Otherwise, it can be caused by increased competition in markets.As a result, their net effect can be positive even if the substitution effect has crowded-out domestic fixed investment.After removing the effect of FDI inflows and those of variables proxying innovationproductivity and ITC-diffusion, the effect of human capital formations and gross fixed investments has been found significant and with a positive sign in models.
Therefore, in line with our expectations, the gross fixed investments and also unemployment rate have been found significant (Seetanah and Rojid, 2011;Calderon and Boreux, 2016;Shittu et al., 2020).Gross fixed investment has been found with a positive magnitude, while the unemployment rate and it with one order of lags have been found with a negative and positive magnitude, respectively.Gross fixed investment and unemployment rate are variables related to income equation and labour market efficiency.
Human capital proxied a more training, or, a more educational level is usually associated with higher growth (Kagochi et al., 2013;Kayaoglu and Naval, 2017;Ibrahim, 2018;Anetor, 2020;Nwani, 2021), nonetheless, we would find their effects directly and indirectly impacting growth to be positive, negative, or neutral too.
Remittances can have an effect on the economy to which they are directed through the Keynesian multiplier.Even if all the income is consumed by the households that received the remittances, this would indirectly stimulate the exogenous component of demand, as there will be a general increase in aggregate income.This suggests that migrant workers' earnings proxied by received remittance with one order of lags are positively contributing to African growth (Adusah-Poku, 2016).
Received remittances are also related to population growth and the migration rate toward developed countries with better growth prospectives.Indeed, the urban population growth with one order of lags is also significant and with a positive sign in models in Table 2, although in the models in Table 4, by controlling for the countries' fixed effects, its magnitude is negative (Bruckner, 2012;Onjala and K'Akumu, 2016;Oluwatayo and Ojo, 2018).This may be related to the population growth rate, the extension of urban areas, and the development of global cities following an increase in income, or the opposite; in any case, sound country governance is required.In other words, institutional quality and stability positively influence the countries' growth.Improvements in the institutional and business environment can produce spillover effects on growth.As a result, the governance climate is an important indicator of the level of development reached by a country, and it has been found significant and with a positive sign.
The macroeconomic variables related to public spending, such as government, health and military expenditure, net-ODA, and ITC-diffusion have not been found significant in models in Table 2, but they are significant in models in Table 4, after having controlled for the countries' fixed-effects (Seetanah and Rojid, 2011;Pinkovskiy and Sala-i-Martin, 2014).
Government expenditure is a proxy for the governance's bureaucratic size, and it can be associated with a negative impact on growth due to the issue of available resource allocation (Arizala et al., 2020).
The health expenditure and military one with one order of lags have been found significant and positively and negatively impacting growth, respectively, although the second without the lag has had a positive impact (Ahmed, 2012;Akhmat et al., 2014;Shaaba and Ngepah, 2018).However, assessing these impacts should be difficult, especially in economies with a permanent or semipermanent war and riots, or because they could be related to other unconsidered exogenous variables (Mijiyawa, 2013;Nsiah et al., 2016;Boreux and Calderon, 2016;Franses and Welz, 2022).
Last but not least, the ICT-diffusion has been found significant in models in Table 4, after the control for the countries' fixed effects.Although the use of new technologies such as the internet and mobile devices has been found significant in studies of growth, their significance or lack thereof may depend on the variables used as a proxy (Batuo, 2015;Donou-Adonsou et al., 2016;David, 2019;David and Grobler, 2020;Ngameni et al., 2022).According to Haftu (2019), African societies still lag behind in the adoption of new information and communication technologies.In Table 3, we show models with the proxy variables for African countries' merchandise trade across clusters of developed and developing countries separately considered, and as World Bank Group researchers have classified them.In the models shown below, the significant proxy variables of the African merchandise trade across clusters of developed and developing countries are the lagged exports to the cluster of developed economies with a positive magnitude, as well as, the imports and exports from or to the cluster of developing economies in Eastern-Asia or Pacific -in which Asian giants China and India are the two fast-growing economies -and imports from the LMDCs in North Africa or the Middle East have been found significant.Then, other significant evidence has not been found in the remaining country clusters.
Meanwhile, in the models shown below in Table 4, we have found proof of a positive linkage with the cluster of developing economies in Latin America or the Caribbean and those in North Africa or the Middle East, as well as exports to developed economies with one order of lags.
Instead, imports and exports from or to Southern-and Eastern-Asia or Pacific have been found significant with a negative magnitude on growth, as well as, the same way, the imports from HDCs with one order of lags, imports from LMDCs in Sub-Saharan Africa, and exports to LMDCs cluster in Europe or Central-Asia -in which the Russian economy leads.
Finally, Sargan and Hansen tests, Wald tests on regressors and time-dummies, as well as respecting the rule of thumb to maintain the number of instruments less than the number of cross-sectional units, demonstrate the reliability of our estimations in models.As well, Wald's and Welch's tests computed on fixed-effects models confirm the reliability of estimations.Note: *** significance for α = 0.01 ** significance for α = 0.05 * significance for α = 0.10.

Concluding remarks
This study has analysed growth in a panel of 54 African countries using the generalised method of moment (GMM) estimators over the period 2010-2019.As far as we know, we have not found studies analysing growth with a panel of 54 African countries and/or using GMM-estimators over this period.
We have also controlled idiosyncratic unobserved individual effects of countries with fixed-effects models.Moreover, we introduce a novelty element in the analysis, by considering African growth in relation to specific openness degrees as proxied by merchandise imports and exports for the clusters of developed and developing countries such as defined by the World Bank Group researchers.
Therefore, based on the time-series data collected, and the estimation methodology used, our results show the determinants of African growth over the period 2010-2019.
Finally, over the question that we have highlighted in the introduction, where it asks whether Africa can break with its colonial commercial past, we have found proof of the existence of a trade dependency added to that with the cluster of developed countries, in which certainly the ex-colonising countries are situated, indeed, we have found evidence of a trade dependency with the cluster of developing countries.
A more developed institutional and business environment should lead to sustainable long-run African growth, but this depends on sound governance.Indeed, the governance climate has been found to be significant in models.
Generally, the significance found in import and export from or to LMDCs can depend on favourable linkages between developing economies that, for instance, have a similar institutional and business environment, or a similar technological gap with respect to developed countries.However, the significance found in the trade with HDCs highlights the linkage along global value chains between African markets and developed ones.
Therefore, both statistical significances can depend on the linkages found by the growth studies analysing the development in the North-South and South-South frameworks.This can explain the trade relationship existing along global value chains between African markets and developed and developing countries.
Global value chains have expanded in the new millennium, and low transport costs, low trade barriers, few embargoes, as well as technological and financial spillovers, have made this possible, but this has also meant greater uncertainty in the markets, which are then closely interconnected with each other (World Bank, 2020, 2022).This means that economies in transition have to face negative aspects related to globalisation.As a result, there is a trade-off between the lowering of trade barriers, and technological advancements deriving from an international specialisation of productions, and the exposure of countries and their firms to economic and political unbalances ad shocks (Togati and Visaggio, 2016).
Especially, in developing countries, trade in semi-finished products has intensified firms' activities along global value chains, however, these goods may escape from national accounting due to the absence of international accounting harmonisation (Wolf and Zedillo, 2015).It should be considered that the system of harmonisation of national accounts developed by the UN statistical commission is stuck in the fifth version since 2008 as an upgrade of the previous one released in the early Nineties.Therefore, causing trade intensification along the global value chains, products can transit from developing country to another after they have had an increase in value at least equal to the labor cost, and in turn, they can return to developed countries, but without having been properly accounted for.This means that a "country-factory" can show macroeconomic structures characterised by only consumed income, for instance, as small economies focusing on import-export activities with a dominant manufacturing production.

Policy implications
In the future, African growth could depend more on sound governance, but countries should improve their institutional and business environments in order to achieve more inclusive and sustainable growth (Acemoglu et al., 2019;Lin, 2021).African governance could lead the growth, both by pursuing incentive policies on exports rather than imports or by improving the opportunities for firms (Glegg et al., 2021).Therefore, sound institutions and forward-looking policies can lead firms toward progress, technological specialisation, and wellbeing (Kurtishi-Kastrati, 2013;Collier, 2014;Trebilcock, 2015;Kim and Heshmati, 2019;Farahane and Heshmati, 2020;Babajide et al., 2021).
However, changes in institutional structures are generally burdened: (i) by a heavy inertial mass to change in defence of the elites' interests; and (ii) by the slowness of adaptive responses typified by many societies.As a result, the acceptance of a new techno-economic paradigm as well as a new socio-institutional system is a difficult process, as the country will have to bear a greater sunk-cost due to the specificity of its historical development path and the variety characterising the capitalistic system as an expression of the institutional structure (Acemoglu and Robinson, 2012;Granovetter, 2017).
The socio-economic and institutional transition processes will inevitably lead to internal contradictions within capitalism and to paradigmatic fluctuations.Recurring crises are showing that the governance of globalisation is an important issue related to capitalism (Dallago and Casagrande, 2023).
In the capitalist system, there will be a certain selfish impulse to capital accumulation, such that intrinsic instability is not its failure, but constitutes its vital impulse (Razin, 2022).Therefore, capitalism is changing by its nature, and its ability to self-production does not contribute to making the socio-economic system stable for too long.
In other words, social progress depends on the choices made by the agents and the probability that an endogenous shock to the system occurs, such that it triggers a dynamic process of change (Hallett et al., 2010).This process would evolve the system for incremental leaps, and the alternative solutions prospected would be those near the optimal points of the Paretian-frontier.It is precisely the proximity of the socio-economic system to such Pareto-efficient points that triggers change, which makes mobile over time the steady state achieved by the system.This means that perturbations, triggered by the agents within the socio-economic system, push it toward a "natural" search for possible Nash's equilibriums, which would then be chosen among those sustaining and not among those responding to the maximisation logic (Festré, 2021).
The global pandemic and war in Eastern Europe are showing that global governance is an important issue at the current stage of globalisation (Cowling and Tomlinson, 2011;Autor et al., 2016;Eichengreen, 2018;Saccone, 2021).The evolution of the world economy has for too long been left solely to the regulatory automatisms of the markets, and this has increased social inequalities.
The trajectory followed by globalisation is progressively abrading the stability and social cohesion in the advanced and emerging economies, as it is not consciously governed.Globalisation, on the other hand, can foster convergence between countries while also increasing economic and political competition between them by causing a disruption in global balances (Heshmati and Lee, 2010;Valli, 2018;Obstfeld, 2020;Marelli and Signorelli, 2022).
For instance, difficulties in multilateral trade negotiations within the World Trade Organisation-WTO have resulted in a generalised focus shift toward regional agreements, which have grown in number and complexity over the last decade.African countries' signatories to the AfCFTA agreement have accepted to limit their governance's unilateral action in order to jointly improve their attractiveness.The AfCFTA agreement may be the biggest trade area in the world, with which the African countries could enhance the position of their manufacturing systems along the global value chains and reach sustainable development in the direction of the SDGs.Nonetheless, the AfCFTA agreement is burdened by significant lags in its agenda.
Greater attention has then been given to the growth-wellbeing relationship in recent years.Two commonly adopted indicators are per-capita GDP and the Human Development Index-HDI by the United Nations.The first is widely used, and it is annually available for all countries, although it measures only the economic dimension of development and suffers from some methodological issues, it remains a reliable growth index.The second is better, but data may not be available for all countries.
In other words, HDI is a composite and synthetic indicator measuring, on average, the country's performance based on three aspects: life expectancy at birth, schooling, and income.However, another indicator that considers the social impacts also is the Social Progress Index-SPI, developed by Porter et al. (2014) starting from the works by Sen, North, and Stiglitz.This index measures society's ability to satisfy basic human needs and improve people's quality of life, so that everyone can aspire to achieve the best possible personal fulfilment.Therefore, the HDI and SPI-index are well-being indicators prioritising social progress over economic progress.

Limitations and future lines of research
The primary limitation of our analysis is that African countries cannot share the same growth functions, and these cannot be stable over time.This can be the result of a different resource endowment and distinct stages of progress that could be separately considered (Marini, 2004;Schwab, 2014Schwab, , 2015)).Furthermore, it is known in the economic literature that structural reforms in a country need more time to manifest their effects on the economic system.As a result, this may explain why some lagged variables have not been found to be significant in this study.
Moreover, it could be a need to include other explicative variables in models, for instance, related to demographic characteristics, such as fertility rate and life expectancy, or, related to the quality of the institutional and business environment, which literature has shown to be significant for growth.As a result, it has greater data completeness.In addition, starting a backward time-series reconstruction project to extend the time dimension can be necessary.
Finally, a clustering based on the country's level of income could also be necessary to have control in models with reference to it.Four clusters could be used: low-income, lower middle-income, upper middle-income, and high-income countries, as provided by World Bank Group researchers.For instance, the gross national income-GNI with Atlas correction to reduce the impact of exchange rate fluctuations in the cross-country comparison, and the Gini coefficient for the income inequality could alternatively be used.They have found the effects of domestic and transnational terrorism on the per-capita income growth of 51 African countries from 1970 to 2007 by accounting for the cross-sectional spatial dependence of conflicts.The findings suggest that transnational terrorism has a modest marginal impact on per-capita income growth and that domestic terrorist events, surprisingly, do not affect it.According to the authors, the modest impact of transnational terrorism on African growth indicates that developing economies are more resilient to terrorism than is commonly assumed.

Narayan et al. (2011)
They examine the relationship between democracy and economic growth in 30 SSA countries, finding mixed support for the Lipset theory in the long run.
Bertocchi and Guerzoni ( They explore the empirical determinants of fragility in SSA over the period 1992-2007 by using a battery of development indicators and finding that institutions are the main cause of the fragility.The probability that a country will be fragile increases with the restrictions on civil liberties and with the increase in revolutions.In fact, the per-capita GDP growth and investments are significant explanatory variables, but the economic growth has an uncertain net impact as it reduces the country's fragility, while the investments increase it.

Jaunky (2013)
He studies the linkage between democracy and economic development in 28 SSA countries over the period 1980-2005 using the GMM model.He has found that economic growth precedes democracy in the short run, while bi-directional causality is found in the long run.At last, the effects on growth are positive.Fayissa and Nsiah (2013) They use fixed and random effects models, and GMM models for investigating the governance effect on African growth.They have found that governance contributes to the growth gap of African countries, which depends on the countries' income.

Ahlerup et al. (2016)
They examine how an impartial government toward ethnic groups can improve the growth of 20 SSA countries beginning in the late Nineties.They have found that countries with a governance perceived as impartial have a better chance of growth.

Akobeng (2016)
He investigates whether the linkage between growth and poverty reduction can be strengthened across the institutions in 41 SSA countries over the period 1981-2010 by using the GMM estimator.He finds that improvements in governance are significant for supporting the link between growth and poverty reduction in SSA.
Toh (2016) He investigates the long-run growth drivers of a group of SSA emerging economies.His findings indicate that the economies diverge more on economic characteristics, institutional quality, and governance than the slow-growth group.
Epaphra and Kombe (2017) They examine the impact of institutions on African growth using the GMM, fixed-and random-effects models over a sample of 48 countries from 1996 to 2016, discovering that political stability is the most important factor in explaining African per-capita GDP growth.Other significant explanatory variables are openness, gross fixed investments, human capital formation, and foreign direct investments.

Ogbuabor et al. (2020)
They examine the impact of governance on economic growth in Western Africa after the global economic recession using a panel of 13 countries and find a negative relationship between governance and growth.Specifically, corruption, government ineffectiveness, political instability, the weakness of the rule of law, and the absence of accountability are the main obstacles to growth, while the per-capita GDP, gross fixed investments, employment, and foreign direct investment are the other significant drivers of growth in the region.(2012) He explores the relationship between military expenditure, external debts, and growth in a sample of 25 SSA countries over the period 1988-2007, by finding that military expenditure has a positive impact on the external debt of African countries, and GDP growth negatively affects their total debt stock.

Kagochi et al. (2013)
They investigate the relationship between financial development and growth in a sample of SSA countries and find that stock-market development has a positive effect on growth.Instead, the other financial development indicators have an uncertain impact on the growth, while the control variables such as capital formation, schooling, and life expectancy have a positive effect on the growth.

Mbate (2013)
He investigates the impact of the domestic debt on growth and the private sector in 21 SSA countries over the period 1985-2010 by using GMM models.He has found that domestic debt crowds out the private sector and deters capital accumulation.

Asongu (2014)
He uses a VAR approach to examine the effects of monetary policy on African growth from 1987 to 2010, testing whether monetary policy variables affect growth in the short and long run, but with inconclusive results.

Walle (2014)
He examines the long-run relationship between the financial development and growth in 17 SSA countries over the period 1975-2005 by applying an error correction term based on the co-integration tests for considering the cross-sectional dependence between the countries.He has found that there is a long-run relationship between financial development and growth, although there is a weak reverse causal impact.Shaaba and Ngepah (2018) On a panel of 35 African countries from 1990 to 2015, they analyse the relationship between military expenditure, industrialization, and growth, by finding that industrialisation and growth precede military expenditure in the short-and long run, but that military power can be used to achieve industrialisation and growth under given conditions.

Mensah et al. (2019)
They used ADL models to exaggerate the impact of public debt on growth in 38 African countries from 1970 to 2015, discovering that public debt stifles growth when it exceeds 50% of the country's GDP.

Arizala et al. (2020)
They investigate the effects of government expenditures and revenues on growth in SSA from 1990 to 2016.They discovered that cutting off public investments has a greater impact on growth than cutting off public consumption or increasing revenues.Attempts to consolidate public finances, on the other hand, have had a negative impact on short-and medium-term growth, which has been mitigated by financial adjustments.

Ehigiamusoe and Lean (2020)
They examine the effects of public debt and deficit on growth in Western Africa by implementing empirical strategies that account for various econometric issues.They find that the impact of financial development on growth depends on the levels of debt and deficit.When debt and deficit levels exceed a certain threshold, the marginal effects of financial development on growth are negative.Idun (2021) He believes that the use of technology in financial systems can contribute to African growth in the long run, provided that other growth drivers such as human capital formation, openness, and infrastructural capital are present in the countries.However, financial development produces divergent responses to growth within African country clusters.Financial innovation in COMESA and ECCAS causes growth, while that in ECOWAS and ARABMAG has been found to be dangerous to growth.

Batuo (2015)
He finds that ICT infrastructures are positively related to the growth of a panel of 44 African countries over the period 1990-2010.A dynamic panel data approach has been employed.Findings show that additional ICT investments have a positive impact on growth.

Donou-Adonsou et al. (2016)
They examine the impact of the ICT infrastructures on the growth of 47 SSA countries over the period 1993-2012, by finding the positive impact of internet adoption and mobile technology.ICT ADVANCEMENT David (2019) Over the period 2000-2015, he investigated the impact of ICT infrastructure on growth as measured by the GDP and HDI index in 46 African countries.He uses a composite index as a proxy for the ICT depth finding and finds that it contributes to the growth.

Haftu (2019)
Using GMM models with internet and mobile telephone penetration rates as proxies for ICT depth, he discovered that an increase in mobile telephone penetration rate contributes to growth while an increase in internet penetration rate does not, as the countries remain in a relatively immature state in terms of technology use.David and Grobler (2020) They investigate the impact of ICT infrastructure on growth in African countries.They discovered that the depth of ICT has a positive impact on growth.

Ngameni et al. (2022)
They study the impact of the ICT infrastructure on the growth-gap between China and 30 African countries over the period 2000-2016, by using internet penetration and ICT-good exports as proxies.Their results suggest that the technological gap has a positive impact on African growth.The increase in Chinese ICT investments could benefit African economies through the positive externalities induced.

FOREIGN CAPITAL INFLOWS
Alemu and Lee ( 2015) For a panel of 20 middle-income economies and one for 19 low-income economies over the period 1995-2010, they used GMM models that found a positive relationship between foreign aid and growth only in the low-income countries, while the growth is subordinated to foreign investments and oil-export revenues in the middle-income countries.
Adusah-Poku (2016) He investigates the impact of foreign capital inflows -foreign aid, foreign direct investments, and personal remittances -on SSA growth from 1990 to 2010, concluding that all three inflows have a positive and significant impact on growth in the long run, while personal remittances are significant only in the short run.Cai et al. (2018) They investigate the effects of aid on African growth using panel data from 47 African countries from 1980 to 2013, discovering that international aid promotes growth, but its effectiveness is dependent on governance.
Hagan and Amoah ( 2019) Using an instrumental variable approach to panel data, they investigate whether the effect of foreign investments on African growth is dependent on the resilience of the financial system.They have found that when the financial markets are fragile, as they are in some African countries, the foreign investment inflows can have a small positive effect on growth.
Kumar and Saleh (2021) They use co-integrated vector autoregressive analysis to examine the output and prices of tradable and non-tradable sectors in SSA countries.They find that aids have a heterogeneous effect on sectoral output and prices.

Anoruo and Elike (2015)
They analyse the causal relationship between human capital formation and growth in a panel of 29 African countries.The results show a bidirectional causality between the two variables and reinforce the nexus between education and growth.

Kayaoglu and Naval (2017)
They simulate the trend for the formation of human capital, the urbanisation rate, and the per-capita GDP in African countries.They contend that in the short run, a low, or negative return on education investments can be attributed to systemic transitory adjustment or urbanisation costs.

Ibrahim (2018)
He examines the effect of human capital formation on the financial depth and growth in 29 SSA countries over the period 1980-2014 by using GMM models.They discovered that human capital formation and financial depth both promote growth in the short and long run, with financial depth stimulating human capital formation.

Anetor (2020)
He analyses the impact of human capital formation on foreign direct investment and growth in 28 SSA countries over the period 1999-2017.He finds that SSA countries do not have a sufficient, high-quality workforce for absorbing and transforming the FDI's spillover toward growth.

Nwani (2021)
He examines the role of human capital formation in relation to foreign aid and growth in SSA countries from 1985 to 2019.He has found that foreign aid and human capital formation have a negative impact on growth, nevertheless, this impact is mitigated by the interaction between human capital formation and foreign aid, which reduces the negative effect of foreign aid on growth.

Chang and Mendy (2012)
They examine the effects of openness on growth in 36 African countries over the period 1980-2009, by using fixed-effects models.Their results show that openness and investments positively impact growth, with North Africa being the best, while foreign aid, domestic savings, and gross fixed investments show a negative impact.
Brueckner and Lederman (2015) They use the instrumental variables approach to estimate the reciprocal effects of openness and growth in SSA discovering that growth has a negative effect on openness while having a positive effect on growth.

Osei et al. (2019)
They compare the influencing factors of openness in low-and low-middle-income African countries using the GMM approach.They have found that growth robustly enhances openness in low-income countries, while the impact is not robust and is largely negative in low-income countries.This suggests to them that higher growth is associated with less openness.Furthermore, the growth-openness relationship is non-linear and has an inverted Ushape in low-income countries.This means that an increase in the per-capita GDP improves openness, but beyond a given threshold, further increases penalise openness.

Udeagha and Ngepah (2021)
They use a non-linear ARDL approach for exploiting the relationship between openness and growth in South Africa over the period 1960-2016, by finding that there is a short-and long-run causality from the openness to the growth.

Doku et al. (2017)
They use fixed-effect models and Granger causality tests to examine the effects and causal nexus of Chinese FDIs on African growth over a sample of 20 countries from 2003 to 2012.They have found that Chinese FDIs increase the GDP growth rate in Africa, and all other things being equal, they have found that there is a unidirectional causality between GDP growth and Chinese FDIs in Africa.

Koomson-Abekah and Nwaba (2018)
They primarily examine the effects of Chinese FDIs on African growth using ADL models and Granger causality tests on data dating back to the millennium.They discovered that Chinese FDIs have a negative impact on African growth in both the short and long run because their inflows are directed toward capital-intensive activities with a lower impact on employment.They also discovered that FDIs from the United States and Chinese trade had little impact on African growth.The Granger-causality test has confirmed that there is a unidirectional relationship between growth and the other variables, except for human capital formation, which does not show causality.
More FDI inflows to labour-intensive activities will, according to the authors, boost African growth by lowering unemployment.

Bruckner (2012)
He analyses the effects of the value-added growth in the agricultural sector and per-capita GDP growth on the urbanisation rate in African countries over the period 1960-2007.He has found that an increase in the urbanisation rate has a negative effect on the per-capita GDP growth on average, but this does not affect the urbanisation rate.At last, he has found that a decrease in the value-added in agriculture leads to an increase in urbanisation.
Onjala and K'Akumu (2016) They found that the relationship between GDP and urbanisation in sub-Saharan African countries differs from that in developed economies.Their results indicate that the traditional thesis is still valid in the SSA countries, in fact, they urbanise without growth.However, new trends emerge when urbanization coexists with growth.

Seetanah and Rojid (2011)
They analyse the drivers of growth in the selected African COMESA member countries and find that gross fixed investments, openness, and human capital formation are the most important drivers of growth, as well as governance, financial depth, international aid, and spillover effects from foreign capital inflows.

Mijiyawa (2013)
He explores the drivers of African growth over the period 1995-2005, by finding that investments, access to finance, governance improvements, exports, and the share of value-added from agriculture have positively contributed to the growth.

Akhmat et al. (2014)
They investigate the relationship between the public health indicators and growth in Africa from 1975 to 2011, by establishing that there exists a moderately bidirectional causality between the variables.

Elhiraika et al. (2014)
They investigate the role of manufacturing transformations along the global value chains in 50 growing African countries.By using GMM models, they find that GDP increases when human capital formation drives the output growth in manufacturing, at last, this has a positive impact on the GDP growth rate, reducing the volatility.
Pinkovskiy and Sala-i-Martin ( 2014) They look at the recent growth in Africa in relation to poverty.They estimate the income distribution, the poverty rate, and the inequality index in African countries over the period 1990-2011.They show that African poverty is falling rapidly, and the growth that began in the second half of the Nineties has decreased income inequality, even in countries with geographical or historical disadvantages.Addison et al. (2016) They investigate the commodity price shocks in SSA countries dependent on agricultural commodities, by finding that there are inconclusive proofs of unanticipated price variations as responses to variations in per-capita GDP.

Calderon and Boreux (2016)
They investigate if African growth was accompanied by improved structural and macroeconomic indicators, if African countries had liquidity, and if governments implemented countercyclical policies following the global economic crisis between 1995 and 2008.They have found that improvements in the macroeconomic framework have allowed some African countries to better resist the global crisis.

Nsiah et al. (2016)
They examine the determinants of growth in 48 African countries from 1980 to 2011, by taking into account the economic impacts of neighbouring countries.They control for some drivers of growth, such as the gross fixed capital investment, openness, aids, and inflation, by finding a significant level for the gross fixed investments and education, as well as, for the spatial linkages across countries.When recessions occur, neighbouring SSA countries with similar growth compete for resources.Kedir et al. (2017) They estimate the additional investments required to achieve the SDGs and reduce poverty in Africa by 2030.They have found that estimates of the required growth rates vary widely across African subregions and countries.
Countries and subregions with low initial poverty levels and higher responsiveness to the poverty contrast will need less development assistance.

Oluwatayo and Ojo (2018)
They examine growth drivers and poverty reduction in African countries, by finding that African development is advancing inequality and poverty.In other words, this is manifested through persistent inequality, poverty, armed conflict, and indiscriminate young people's migration toward developed countries in search of better living conditions.

Shittu et al. (2020)
They study the impacts of FDIs, globalisation, and governance on the growth of Western Africa over the period of 1996-2016 using ADL models.They discover a positive relationship between globalization, governance, and growth.Even if the findings on the relationship between FDI and growth are inconclusive, governance has a positive impact on FDIs and growth.The other considered drivers of growth are employment, gross fixed capital investment, and government expenditure, whose effects on growth are negative on the first two and positive on the last.

Table 2 .
The GMM-system models with the variables of macroeconomic framework.

Table 3 .
The GMM-sys models with the merchandise trade variables.

Table 4 .
The models controlling for the fixed-effects, robust standard errors.

Table A .
1.The clustering of the reference literature, authors, and summary.