Effects of capital inflows on industrialization in low- and middle-income countries in sub-Saharan Africa

Achieving sustainable industrial development remains one of the main objectives of decision-makers. International capital inflows have become crucial to the global economy in the era of economic globalization and trade liberalization. According to Lormeau (1992), this capital ^[1] is defined as funds, investments or financial resources originating in one country and injected into the economy of another. These flows are widely recognized for promoting development, particularly in Asian economies (Bahodurov et al., 2025). However, several crises around the world have been linked to abnormal international capital movements (Roy and Kemme, 2020; Kaminsky and Reinhart, 1999). This raises concerns about the potential threat of this capital to the dynamics of industrial development (Ali and al., 2024; Nyang'oro, 2017; Orji et al., 2014). As globalization merges global capital markets, the volume and frequency of these flows have increased significantly in developing economies (Fang et al., 2025; Park and Yang, 2021). As a result, understanding the effects of international capital on industrial development has attracted the attention of researchers.

However, numerous studies provide mixed evidence on how capital inflows impact economic growth. On the one hand, many studies suggest that capital inflows can stimulate growth dynamics (Haskel et al., 2007; Acemoglu and Zilibotti, 1997; Grossman and Helpman, 1991). Studies by Grossman and Helpman (1991) argue that increased foreign direct investment (FDI) can facilitate the diffusion of technology and stimulate demand for skilled labour. Similarly, Acemoglu and Zilibotti (1997) find that improved access to foreign capital can ease credit constraints and enable local firms to undertake more productive investments.

On the other hand, some studies indicate that the volatility inherent in certain external capital flows could lead to instability and uncertainty (Reinhart and Reinhart, 2009; Kaminsky et al., 2005; Kaminsky and Reinhart, 1999). Studies by Adejumo (2013) and Kaya (2010) on African countries suggest that international capital can have negative effects on industrialization by creating dependency and hindering structural transformation.

Furthermore, some recent studies find no clear relationship between these capital inflows and industrial development (Ongo Nkoa, 2016; Gui-Diby and Renard, 2015). Most research suggests that international capital is not neutral: It can be a powerful catalyst for industrialization if properly regulated but can also exacerbate existing problems if poorly managed. This debate prompts our investigation into the relationship between foreign capital inflows and industrialization in sub-Saharan Africa (SSA). There are two reasons for focussing on this part of the African continent. First, previous work in East Asia concludes that foreign capital has a positive impact on industrialization. Specifically, this work shows that the economic transformation of Asian countries was made possible by the inflow of foreign capital (Bahodurov et al., 2025; You and Solomon, 2015; Dahlman, 2009). Second, for many decades, SSA has faced the paradox of increasing capital inflows and stagnant industrial development. Indeed, the shares of FDI and remittances from international migrants as a percentage of the gross domestic product (GDP) rose from 0.40 and 0.43% in 1981 to 2.2 and 3% in 2024, respectively. During the same period, however, the manufacturing value added fell from 18% to 10% (WDI, 2025).

While economists have emphasized the importance of international capital inflows, a related and equally critical question has received less attention in the literature: Do capital inflows promote industrial development in low- and middle-income countries in SSA, or do they exacerbate deindustrialization? This is particularly relevant when exploring the heterogeneity of income levels in the region. Understanding these issues is crucial for governments. Drawing on the literature, we hypothesize that the effect of foreign capital may differ between low-income and middle-income countries.

Our main finding is that these two groups of countries cannot sufficiently finance their domestic production and investment due to a shortage of savings. This implies that reducing this deficit will require additional sources of financing, particularly foreign capital, which is considered an alternative (Baye and Jansen, 2006). Governments should take advantage of this insight to manage these capital inflows more effectively. In middle-income countries with underdeveloped financial systems, policies should focus on optimizing capital inflows to boost industrial development. Conversely, for low-income countries with less developed financial systems, it is crucial to prioritize capital attractiveness in order to capitalize on its beneficial effects.

This article examines how different types of capital inflows influence industrialization in SSA. It contributes to the literature in three ways. First, it provides a comprehensive assessment by jointly analyzing four broad categories of external financing – aid, FDI, portfolio flows and remittances – thus capturing the direct and indirect impact of capital inflows. Second, it examines whether the effects of these flows depend on countries' structural characteristics, highlighting the role of absorption capacity and institutional conditions. Third, it explores the heterogeneity of income levels in the region, revealing whether the impact of external financing on industrialization varies according to the level of development. These contributions help to clarify why increased capital flows have not led to stronger industrial development in SSA.

The rest of the article is structured as follows: Section 2 reviews the literature. Section 3 describes the methodological elements used to highlight these effects. Section 4 presents the results, and Section 5 concludes.

Theoretically, the link between capital flows and economic growth is based on the Harrod–Domar model. This model has been used as a theoretical framework for assessing capital requirements in poor countries (Gillis et al., 1998). The Harrod–Domar model shows that the growth rate of total output is equal to the savings rate on the capital coefficient. Consequently, a one-point increase in the savings rate leads to an increase in the growth rate. Assuming that the capital coefficient is constant, the growth rate of total production is an increasing function of the savings rate (Bastidon et al., 2010). In order to achieve the growth target in a context of insufficient national savings, a calculation of the volume of foreign capital required to achieve the target growth rate could be considered. Given that a low savings rate leads to a low growth rate, all other things being equal, foreign capital must therefore be used to stimulate economic growth.

According to this model, countries with low capital stocks have low income levels and therefore low savings rates. As a result, their growth rates are doomed to remain low: This is the poverty trap. To escape this vicious circle, these countries must resort to foreign capital, including remittances from international migrants, official development assistance, FDI and other forms of investment. However, the Harrod–Domar model has been widely criticized. Among the criticisms levelled at it is the rigidity of the capital coefficient. To correct these shortcomings, the Solow model (1956) incorporates savings and investment, with the particular feature that the capital coefficient is no longer constant. In other words, the return on factors is decreasing. Indeed, Solow's model stipulates that poor countries will catch up with rich countries if the savings rate is high. From this perspective, international capital inflows to developing countries can promote growth.

On empirical ground, numerous studies have examined the impact of capital inflows on industrial development. On the one hand, recent studies have highlighted this relationship using three indicators, such as FDI (Sule, 2019; Ongo Nkoa, 2016; Combes et al., 2016; Orji et al., 2014), official development assistance (Combes et al., 2016; Orji et al., 2014; Awemoa, 2008) and remittances from international migrants (Sawadogo et al., 2024; Combes et al., 2016; Gbenou, 2015). Table 1 summarizes these studies.

On the other hand, numerous studies have examined the impact of capital inflows on industrialization using various indicators. For example, based on a system generalized method of moments (GMM), Mfere and Makosso (2023) established a strong negative relationship between FDI and industrial performance between 2000 and 2020 using various indicators (such as industrial value added, employment in the industrial sector, manufacturing value added and the production capacity index) in the countries of the Central African Economic and Monetary Community (CAEMC). Similarly, using external loans, FDI and remittances as indicators of foreign capital and the ARDL technique, Sule (2019) found that FDI had an aggravating impact, independently of other forms of capital, on industrial growth in Nigeria for the period 1985–2018. Similarly, Müller (2021) studied the impact of FDI in relation to the penetration of information and communication technologies (ICT) in 47 African countries over the period 1996–2017. Using the generalized least squares method, the results show that FDI has a negative impact on the industrialization process in SSA. More importantly, this relationship is exacerbated by the high level of ICT.

In contrast, Orji et al. (2014) studied the impact of FDI in countries in the West African monetary zone and found that FDI contributes significantly to output growth in Nigeria and The Gambia only if these countries exceed their low-income economic development level for the period between 1981 and 2010. Also in West Africa, Akorsu and Okyere (2023), using the augmented autoregressive distributed lag (ARDL) and nonlinear augmented autoregressive distributed lag (NARDL) cointegration approaches, found that FDI was an important lever for ensuring industrialization in Ghana during the period from 1983 to 2019. In the same year, Eje-Ojeka et al. (2023), using the fully modified ordinary least square method, assessed the impact of capital inflows and trade openness on industrial production in sub-Saharan African countries. Over the period 2002–2021, the authors highlighted the principal component analysis (PCA) technique to generate a single index that measures capital inflows. However, the study also revealed that supplementing capital inflows ^[2] with trade openness significantly reduces industrial production in the region. While FDI generally promotes industrialization in developing economies, its effect is less clear in low-income African countries due to limited absorptive capacity and weak financial institutions.

With regard to remittances, official development assistance and portfolio investment, Combes et al. (2016) found that, in 77 low- and middle-income countries with a financing gap, foreign capital inflows into these countries were not sufficiently effective in stimulating growth between 1980 and 2012. During the same period, however, Nyang'oro (2017) studied the interaction between debt flows and equity portfolios in 26 sub-Saharan African countries and found that debt was not effective in stimulating industrial growth, using the system GMM for the period between 1980 and 2011. Furthermore, using remittances, portfolio investment and official development assistance as indicators of foreign capital for 35 countries between 2000 and 2015, Ouedraogo et al. (2018) found using different techniques (fixed effects and system GMM) that remittances and official development assistance were important factors in changing patterns of transformation from primary products to manufactured goods.

With regard to portfolio investment, the same study shows that this type of capital has a detrimental effect on diversification. Although there is little literature on the effect of portfolio investment on industrial development, studies by Baharumshah and Thanoon (2006) show that, unlike other capital inflows, portfolio investment can be counterproductive because it may hinder economic growth through externalities arising from both booms and sudden reversals. This type of capital is often short-term and speculative, reflecting investors' desire for instant liquidity in risky markets. Similarly, other studies, such as those by Combes et al. (2012), have examined the financial impact of foreign capital (captured by portfolio investment, transfers and FDI) in 42 developing countries for the period 1980–2006. Using a pooled mean estimation technique, the authors found that among private capital inflows, portfolio investment has the greatest impact on real exchange rate appreciation, almost seven times more than FDI or bank loans, while private capital inflows have the smallest effect.

In all, although conclusions on the impact of capital inflows on industrialization are mixed depending on the type of indicators and methodology used, this study, like those by Bako (2024), Efobi et al. (2019) and Sule (2019), explores the role of foreign capital in industrial development in certain African countries. Unlike these studies, this article broadens the scope of research by simultaneously using four indicators of foreign capital. Furthermore, none of the previous studies on the relationship between capital inflows and industrialization have focussed on low- and middle-income African countries. This study tends to add an empirical gap to the body of knowledge.

We study a panel of 46 sub-Saharan African countries over the period 1995–2023 with data mainly from the World Development Indicators (WDI), the International Monetary Fund (IMF) Balance of Payments (2024) database and the IMF World Economic Outlook (WEO, 2024). The sample size of this study is primarily determined by the availability of data for all the variables of interest. The dependent variable is industrialization. It is part of a dynamic process involving the production or purchase of raw materials, their transformation into semi-finished or finished products and their sale on domestic or foreign markets (Agon, 2017).

However, to measure this, two indicators are commonly used in the literature. These are the share of manufacturing value added in GDP at constant prices and the share of manufacturing employment in total employment (UNIDO, 2013). Similar to Guidiby and Renard (2015) and Egunjobi and Akam (2025), due to the unavailability of data on manufacturing employment in African countries, the present study uses the share of manufacturing value added in GDP.

Capital inflows are measured by four different indicators, namely, remittances from international migrants as a percentage of GDP, official development assistance as a percentage of GNP and FDI, net flows as a percentage of GDP and portfolio investment as a percentage of GDP, although other flows, such as bank loans and external debt, also exist. The choice of these four indicators is consistent not only with the literature (see Ouedraogo et al., 2018) but also with the industrialization objectives in the context of sub-Saharan African countries. Indeed, FDI promotes technology transfer (Borensztein et al., 1998), official development assistance finances support infrastructure, remittances act as a lever for local private investment and portfolio investment acts as an alternative financing channel, enabling industrial companies to access long-term capital through regional stock markets while imposing governance discipline that promotes productive efficiency (Levine, 2005), unlike bank loans and external debt, which are often pro-cyclical and more closely linked to financing consumption or public deficits than to the structural transformation of the manufacturing sector in SSA (Sandow et al., 2022; Mbondo, 2013; Thornton, 2008; Frimpong and Oteng-Abayie, 2006).

The control variables are cell phone subscriptions, fixed-line subscriptions, internet users, human capital, trade openness, natural resource endowment, financial development and primary energy intensity level. The definitions of the variables and their statistical properties are presented in Appendix Table A1 and Table A2, respectively. Table A3 in the Appendix lists the countries included in the study, and Table A4 (in the Appendix) presents the correlation matrix.

Figure 1 shows the relationship between international capital flows and industrialization in SSA. The shape of the scatter plots and especially the regression line reveal a positive relationship between international capital and industrialization in African countries. What's more, this relationship is stronger when external capital flows are measured by official development assistance and remittances from international migrants. Furthermore, observation of these different graphs also reveals that the magnitude of the effects of international capital flows on structural transformation is weaker for African economies in the low-income bracket. In fact, whatever the type of external capital, the regression line for low-income countries has a positive slope, but of low magnitude. However, the magnitude of this effect depends not only on the type of foreign capital involved but also on each country's level of development. However, the conclusions drawn from this statistical analysis are only presumptions, given that the examination of the effect of international capital on structural transformation cannot be limited to a statistical analysis. As a result, econometric analysis of the effects of external capital on structural change is essential. More importantly, these econometric analyses will go some way to controlling the other determinants of structural transformation and the endogeneity problems that would ensue.

This section outlines the econometric specifications that will be used to estimate the effect of capital inflows on the industrialization of sub-Saharan African countries. The present study refers to the literature on industrialization (Mahamat, 2021; ACET, 2017 ^[3]; Mbondo and Bouwawe, 2023). In line with the literature, the additive function of the model is as follows:

where $i$ and $t$ are country and time, respectively. $MVAit−1$ is the dependent variable of one-period lagged manufacturing value added. $REM$⁠, $FAD$⁠, $FDI$ and $PFI$ represent international migrant remittances, official development assistance, FDI and portfolio investment, respectively. $Zk$ is the vector of control variables presented in the previous section; $β$ and $λ$ are the parameters to be estimated. $δt$ captures the country-specific unobserved effect, while $γi$ is the time-invariant fixed effect. $εit$ is the error term. However, Equation (1) introduces the interaction terms between the variables of interest and the control variables, illustrating the channels through which net capital inflows influence the process of industrialization dynamics of African countries. This dynamic involves Equation (2).

with $z$ representing some of the country's structural characteristics.

The estimation strategy adopted in this work is the system GMM. This is in line with the economic literature on industrialization (Mbondo and Bouwawe, 2023; Nyang'oro, 2017; Moshi, 2014). In what follows, the present study outlines four essential elements that preside over the choice of this empirical strategy: (1) the number of periods in each cross section (28 years) is substantially smaller than the number of countries (46); (2) the inclusion of the lagged dependent variable in the empirical model implies that there is a correlation between the regressors and the error term since historical evidence shows that the movement of resources towards manufacturing is a persistent and unidirectional process. Indeed, the dynamic dimension implies to include the lagged dependent variable among the explanatory variables in the model, given that the dependent variable is correlated with the error term, so is its lagged value, all of which consequently raises a problem of endogeneity; (3) In addition to these, this estimation strategy takes endogeneity into account by considering the simultaneity of explanatory variables through an instrumental variables approach on the one hand and by controlling unobserved heterogeneity through time-invariant indicators on the other; (4) the inherent biases that characterize the difference estimator are corrected by the system estimator. According to the work of Arellano and Bond (1991), unlike other estimation strategies, this method offers numerous advantages, such as correction of the persistence of the dependent variable, correction of endogeneity, consideration of unobserved country-specific effects, correction of autocorrelation in panel data models and the inclusion of time-invariant variables as explanatory variables that would be erased in fixed-effect GMM estimates or by difference. Due to these numerous advantages and in line with recent literature, compared to the classical empirical approach (method of difference GMM), this method tends to reduce instrument proliferation (over-identification) by taking cross-sectional dependence into account (Asongu et al., 2020).

The empirical results are shown in Tables 2 and 3, respectively. In addition, four fundamental information criteria are used to verify the validity of system GMM. According to the two criteria (AR (1) and AR (2) first- and second-order autocorrelation, respectively), the first-order autocorrelation test is accepted in the tables. The AR (2) test for the absence of autocorrelation in the residuals is not rejected, confirming the validity of the instruments used across alltables and columns. Also, the p-values of the Hansen test are above the 0.1 significance level, indicating the absence of over-identification of the instrumental variables. More importantly, the study ensured that the instruments were smaller than the value of the instrumental variable, in order to limit instrument proliferation. Table 2 presents the overall results of the system GMM. Column (1) shows the direct relationship between manufacturing value added and all the international capital variables, while columns (2) to (5) successively integrate the external capital variables in the presence of the control variables. Columns (6) to (13) repeat the work of column (1), integrating all the control variables one by one.

Indeed, the positive, non-significant effect of international remittances on industrialization may suggest that the large proportion of remittances is not destined for the investment project, but rather towards household consumption. This type of result was found by Ouédraogo et al. (2018) and is opposed to those of Sule (2019). Hence the need for African governments to encourage households receiving remittances to use these funds for profitable, income-generating investments. FDI has a negative and significant effect at the 1% threshold. More precisely, its coefficient is −0.042. Thus, a 1% increase in FDI is associated with a 4.2% drop in manufacturing value added. This result is consistent with dependency theory, which maintains that excessive dependence on foreign investment can compromise the growth and development of recipient countries.

According to this theory, local industries may find it difficult to compete with multinational corporations (MNCs), leading to deindustrialization and monopolization by these external entities. Also, SSA is a resource-rich region, and most FDI seems to go into the extractive rather than manufacturing sectors. These findings are similar to those of Sule (2019), Müller (2021) and Ali et al. (2024) and are opposed to those of Saba and Ngepah (2023) and Gui-Diby and Renard (2015).

Official development assistance has a negative and statistically significant effect at the 1% threshold. In fact, a 1% increase in public development aid is correlated with a 7% drop in manufacturing value added. The negative impact of public development aid can be justified by the fact that in this region, there is not only a high level of corruption but also a high level of embezzlement of public funds. It appears that aid received for development projects is being diverted for personal and selfish interests. This may also be caused by the inability of the appropriate institutions or entities responsible for directing these flows towards industrial development. This kind of result was found by Ouédraogo et al. (2018) and contradicts the findings of Bako (2024) and Eje-Ojeka et al. (2023), who found a positive and statistically significant effect on industrialization.

Similarly, the coefficients associated with portfolio investment are negative and highly significant at 1%. A 1% increase in portfolio investment leads to a 171.9% decrease in manufacturing value added. As argued by Baharumshah and Thanoon (2006), portfolio investment can be counterproductive as it can hinder economic growth through externalities resulting from both sudden booms and busts. This result is similar to those of Ndugbu et al. (2021) and in contrario to those of Ouédraogo et al. (2018).

Concerning the control variables, the results in columns (2) to (13) of Table 2 are similar to those in the existing literature on the determinants of industrialization. Indeed, the regression coefficients of fixed-line subscriptions, trade openness $(TRA)$ and electric power intensity $(EE)$ are positive to manufacturing value added (Sawadogo et al., 2024; Adanlawo and Vezi-Magigaba, 2021; Tahir et al., 2016), while cell phone subscriptions $(ATM)$ and natural resource endowment $(NNR)$ have a statistically negative effect (Sawadogo et al., 2024; Saba and Ngepah, 2023). Internet users $(IPOP)$⁠, human capital $(ESUP)$ and financial development $(DF)$ are insignificant. However, the sign associated with the coefficient of the human capital variable deserves particular attention. Indeed, the results show that industrialization is not directly linked to higher education enrolment. This result may have at least two explanations: Firstly, it may mean that the accumulation of human capital needed to stimulate industrialization is insufficient. Secondly, it may be due to the poor quality of institutions, which muzzle individuals endowed with human capital, preventing them from applying their know-how (Bouwawe, 2023).

In addition, the insignificance of certain variables led us to investigate the transmission channels through which international capital can stimulate industrialization. We have therefore decomposed our sample into low-income and middle-income countries. Eight interaction terms are introduced into the model. Tables 3 and 4 present the results of these estimations. Table 3 shows the results of the direct effect of these capitals according to the country's level of development, and Table 4 takes these interactions or potential effects into account.

Columns (1) to (4) of Table 3 present the regression results for linear effects. We find a significant and mixed relationship between foreign capital and industrialization in low-income countries in SSA. These results show that remittances (0.452) and portfolio investments (0.130) have a positive and statistically significant effect at the 5 and 1% levels, respectively. Conversely, official development assistance and FDI have a significant negative effect at the 5% threshold (−0.203 and −0.159, respectively). Our results corroborate those of Kaboré and Kangoye (2021), who showed that remittances promote manufacturing job creation, particularly in countries where access to bank credit is limited. These flows act as a substitute for financial sector development in stimulating productive investment (Babatunde and Martinetti, 2010). Similarly, Kurniawan et al. (2023) identify a positive link between portfolio investments and industrial expansion through market capitalization. These authors note that although volatile, portfolio investments complement domestic savings to finance industrial infrastructure in South Asia. However, the negative effect of official development assistance and FDI has also been noted in the work of Loko et al. (2024), Iqbal et al. (2022), Fagerberg and Verspagen (2021) and Sarr (2020). One explanation for this negative effect is that FDI is concentrated in extractive sectors (oil, mining) and causes currency appreciation, which undermines the competitiveness of manufacturing industries. For low-income countries, the vitality of this capital does not lie in the amount received but in the ability of institutions to channel it towards the local manufacturing sector, which creates sustainable jobs.

In middle-income countries, we also find a significant and mixed relationship between foreign capital and industrialization. Indeed, the results in columns (5) to (8) show that remittances and portfolio investments have a negative and significant effect on industrialization of (−0.543) and (−7.642), respectively. On the other hand, official development assistance (0.822) and FDI (0.313) have a positive and significant effect on manufacturing value added. Our results reveal a complex duality: While these financial flows provide liquidity, they also have a crowding-out effect or lead to early deindustrialization.

These results corroborate those of Le and Nguyen (2024), who identify a mechanism in the literature that explains this effect (the Dutch Disease paradox). Similarly, Khan et al. (2025) find a threshold effect beyond which remittances cease to support consumption and become a driver of structural deindustrialization in Pakistan. Similarly, in Latin America, Gomez-Ramirez (2023) points out that the dynamics of portfolio investment have accelerated deindustrialization in Mexico and Brazil by creating macroeconomic uncertainty that is incompatible with the long investment cycles in the industrial sector. Our results, using more recent data specific to the African region, confirm the studies by Dandume et al. (2024) on developing countries. The authors used a balanced panel data set for the period between 1990 and 2021 for 90 countries and found that aid targeted at transport and energy infrastructure has a positive impact on manufacturing value added. Furthermore, in emerging markets, FDI generates positive spillover effects. Local firms industrialize more quickly by imitating the technologies and production standards of multinationals (Yılmaz and Kiliç, 2024). Thus, the varied effect of this foreign capital on the industrial development of low- and middle-income countries in SSA highlights the existence of a development threshold effect in the effectiveness of financial flows in the region.

Table 4 presents the results of the indirect effect of foreign capital on industrialization. Columns (1) to (4) present the results for SSA. Columns (5) and (6) present the results for low-income African countries, and columns (7) and (8) present the results for middle-income African countries. Studies argue that capital inflows could lead to different returns depending on the level of economic development (Dandume et al., 2024; Gómez-Ramírez, 2023). In this research, we test the idea that the effect of foreign capital on industrialization depends on certain structural characteristics of the country. To do this, we introduce interaction variables between foreign capital and the following variables: education, financial development, energy infrastructure, foreign trade, natural resource endowment and mobile telephony. The results with interaction terms are presented in columns (1) to (8).

In SSA, the results of the interaction variables show that remittances are negatively associated with industrial development in countries with low higher education enrolment rates. On the other hand, aid is positively correlated with industrial development in countries with strong electricity infrastructure. At the same time, FDI is positively correlated with industrialization in countries with significant natural resources. Studies have shown that foreign capital can have an effect on enrolment levels in developing countries. Khan et al. (2025) and Nassem (2024) argue that remittances do not systematically promote economic complexity. In countries with low higher education enrolment rates, these financial flows are negatively associated with industrial development. In the absence of academic skills to channel these funds into entrepreneurship, remittances are used to consume imported goods. Similarly, our results corroborate those of Dadume et al. (2024), which show that electricity supply acts as a multiplier for productive aid. By reducing factory operating costs, energy enables infrastructure-targeted aid to translate into a real increase in manufacturing value added. Similarly, abundant natural resources act as a magnet for FDI, which is no longer limited to extraction activities but now promotes industrial development through local processing (mining to manufacturing) (UNCTAD, 2025).

In low-income countries, the results show that foreign capital (remittances and official development assistance) is negatively associated with industrialization in countries where higher education enrolment rates are low. In economic literature, foreign capital inflows are often seen as a remedy for the lack of domestic savings in developing countries. However, the effectiveness of this capital depends on the level of human capital. Numerous studies have highlighted this link (Loko et al., 2024; Nassem, 2024; Asongu and Nwachukwu, 2023). Although our results corroborate these recent studies, they suggest that industrialization cannot be bought with foreign capital. For these low-income African countries, the priority must be to strengthen higher education geared towards the industrial sector. Without a critical mass of engineers, researchers and technicians, transfers and aid will continue to fuel consumption rather than supporting industrial structures that create long-term added value.

In middle-income countries, the results show that FDI is positively correlated with industrialization in countries with a high trade openness ratio. Indeed, the coefficient of the FDI*TRA interaction variable is positive and significant. This result is similar to that of Yılmaz and Kiliç (2024), who show that the impact of FDI on industrial development is doubled in countries with a high trade openness ratio. Through this channel for disseminating knowledge and skills beyond national borders, multinational firms can import technological inputs and export finished products, thereby integrating local industry into global value chains. For these middle-income countries, attracting FDI without an active trade policy could create industrial enclaves, limited by their capacity to create added value. However, trade openness transforms FDI into a powerful vector for structural transformation (Seck, 2014). Following the financial sector, the results show that portfolio investments are negatively correlated with industrialization in countries with low levels of financial development. Our results confirm the work of Baharumshah et al. (2017) on 80 countries around the world. Indeed, the authors find that while portfolio equities significantly stimulate growth in financially developed countries, they have a negative effect in less financially developed countries. Other studies show that although portfolio investments can increase market activity, their effectiveness depends on the absorption capacity of the local financial system (Gómez-Ramírez, 2023; Kuvshinov and Zimmermann, 2022).

In order to verify that the results obtained are robust and consistent with those in the economic literature, this study uses additional estimates by replacing the dependent variable with another, modifying the specification. When the specification is modified, the study is guided by maximizingthe number of observations relative to that of the main specification. This involves either increasing the number of control variables likely to affect the dependent variable or removing other control variables in order to maximizethe number of observations. Under these conditions, we removed the number of control variables at the cross-sectional level (countries according to the level of development) in order to limit the proliferation of instruments. This was done with a view to ensuring that the instruments were fewer than the number of cross sections in all specifications. As shown in Table 5, for the dependent variable of industrialization, instead of using manufacturing value added (% of GDP), we use industrial value added (% of GDP) to verify the robustness of our estimation. This approach ensures that the differentiated impacts are not sensitive to the narrow definition of the manufacturing sector and verifies the consistency of our results t the level of the entire industrial sector. Thus, the results achieved in Table 5 are irrefutable since we have the same sign and the same significant variables as in the main results discussed above. The results remain unchanged in terms of the sign and level of significance of the coefficients. More importantly, our results remain unchanged when countries are stratified according to their level of development.

The idea that industrialization is the driving force behind economic growth dates back to Kaldor's (1967) laws of growth (UNIDO, 2015). Today, it remains the goal of structural transformation in economies at the industrialization stage. Until now, little attention has been paid in the literature to understanding the role of international capital flows in achieving this goal. The few studies that do exist focus either on developed economies or on developing economies. To address these limitations, this study examined the influence of international capital flows on industrialization in 46 sub-Saharan African countries over the period 1995–2023. Using a dynamic panel model with the system GMM, two main sets of results emerge. First, overall, the results show that the effect of international capital on industrialization depends on the type of capital. Indeed, these results showed that official development assistance, FDI and portfolio investment are negatively associated with manufacturing value added in SSA, while migrant remittances are positively associated. This implies that African governments can bring about a financial paradigm shift through targeted policies. The aim is to channel these flows towards productive sectors. Second, using a cross-sectional analysis, the study showed that the effect of external capital flows on manufacturing sector growth depends on a country's level of development. In low-income countries, the results showed that foreign capital (remittances and official development assistance) is negatively associated with industrialization in countries where higher education enrolment rates are low. This implies that in these countries, the priority must be to strengthen higher education geared towards the industrial sector. In the absence of a critical mass of engineers, technicians and managers, transfers and aid could continue to fuel consumption rather than supporting productive structures that create long-term value. In middle-income countries, the results showed that FDI is positively correlated with industrialization in countries with a high trade openness ratio. For these countries, attracting FDI without an active trade policy could create industrial enclaves, limited by their capacity to create added value. It is imperative that governments channel the attractiveness of this capital through a strategic trade policy. This policy must be designed to act as a lever, ensuring that FDI translates into real structural transformation and an upgrade of local industries.

However, this study has some shortcomings. Instead of using remittances from international migrants, foreign aid, FDI and portfolio investments separately, it would be more relevant to construct a composite index of these capital flows using principal component analysis. This would capture the overall effect of foreign capital. Future research could also take institutional quality into account in order to more accurately reflect economic reality. This approach would make it possible to determine whether this foreign capital serves as an engine of growth or a factor of instability in this part of the African continent.

We would like to thank the anonymous reviewers for their valuable comments.

The supplementary material for this article can be found online.