The African Continental Free Trade Area (AfCFTA) constitutes one of the central initiatives within Africa’s strategic development agenda. It holds significant importance for Africa’s developmental objectives as it establishes a cohesive, extensive market that enhances intra-African commerce, stimulates industrialisation and fosters economic self-reliance. By removing trade constraints and unifying heterogeneous economies, it advances economic growth and situates Africa as a more significant and cohesive player in the global economic framework (Tadelle 2025).

The initiative was launched in January 2012 during the 18th Ordinary Session of the Assembly of Heads of State and Government in Ethiopia (Marinov 2021; Mhonyera & Meyer 2023). The AfCFTA agreement was signed in March 2018 in Kigali, Rwanda and is considered one of Africa’s most important legal instruments to promote continent-wide economic integration (United Nations Development Programme [UNDP] 2022). As of today, there are 54 member states participating in the AfCFTA, and it has an estimated total market size of over $3 trillion based on the combined gross domestic product (GDP) of its member states (Maliszewska et al. 2020). It is designed to increase intra-African trade by creating a larger single market for goods and services through reducing or eliminating tariffs and non-tariff trade restrictions among member states. African Continental Free Trade Area will also help to drive the process of regional industrialisation, which could lead to increased job creation, poverty reduction and ultimately improved living standards across the entire region (Ajewumi, Afolabi & Joe-Akunne 2024; Oloruntoba & Nshimbi 2023). Despite these ambitions, intra-African trade remains limited, accounting for approximately 15% of total African trade flows. This share remains substantially lower than in other regional blocs, such as the European Union and ASEAN, where intra-regional trade ranges between 60% and 70% (Kassa & Sawadogo 2021; Omoju 2021).

The academic relevance of this study derives from the limited availability of empirical evidence on the realised macroeconomic outcomes of the AfCFTA, as the existing literature is dominated by ex-ante simulation-based analyses and forward-looking projections (Abrego et al. 2019; Maliszewska et al. 2020; Mangeni & Mold 2024). However, the computable general equilibrium (CGE) models project an increase in welfare and GDP growth (Abrego et al. 2019; Maliszewska et al. 2020). Nevertheless, substantial debate persists regarding the distribution and magnitude of benefits across individual member states. The larger economy countries, including South Africa and Nigeria, are likely to be the ones to benefit from the greatest proportional gains because of their market size and exporting capacity (Mhonyera & Meyer 2023); on the other hand, small and less diverse economies could experience either limited or unevenly distributed growth outcomes (Mengistu, Fan & Feleke 2024). Additionally, both empirical and institutional analyses have found that tariff reductions alone cannot produce broad-based benefits to achieve this, complementary policies and reforms would be required (Fontagné et al. 2023; Saygili, Peters & Knebel 2018). Specifically, investments in infrastructure development and trade facilitations are needed to transform policy commitments into quantifiable macroeconomic results (Ajewumi et al. 2024; Onuwa & Adedire 2023). This highlights the need for empirical studies based on observed macroeconomic data that complement existing ex ante simulation evidence.

To address this gap in the literature, the present study provides an empirical panel-based assessment of GDP dynamics among seven early AfCFTA ratifiers (Ghana, Kenya, Rwanda, Niger, Chad, Eswatini and Côte d’Ivoire). These countries were selected because they deposited their instruments of ratification during the initial phase of the Agreement in 2018 and provide a consistent observation window for examining the early implementation period using a balanced panel of annual data. The analysis focuses on whether the post-2018 period is associated with statistically and economically significant changes in GDP across participating economies. The year 2018 is selected as the structural breakpoint, as it corresponds to the signing and initial ratification phase of the AfCFTA agreement, representing the transition from negotiation to formal policy commitment. Although the implementation of the Agreement has been gradual, 2018 represents the earliest common policy milestone from which potential macroeconomic adjustments associated with the AfCFTA can be examined. The empirical methodology employs an ordinary least squares (OLS) framework with country-specific fixed effects to control for unobserved heterogeneity, combined with cluster-robust inference, Wild Cluster Bootstrap procedures and a counterfactual trend analysis. The study contributes to the literature by offering evidence based on observed macroeconomic data during the early implementation phase of the AfCFTA. The results are interpreted as empirical associations rather than definitive causal effects, given the evolving nature of the agreement and the limited post-ratification time horizon. This positioning allows the study to complement existing simulation-based evidence by providing an early empirical perspective on macroeconomic performance under regional trade integration in Africa.

Based on the theoretical mechanisms linking trade integration and macroeconomic performance, and in line with the empirical strategy adopted in this study, the following hypotheses are formulated:

The examination of the relationship between the AfCFTA and GDP in African countries represents a critical area of inquiry, as regional integration may alter the structural composition of the continent’s economies (Ajewumi et al. 2024; Maliszewska et al. 2020).

Tariff liberalisation under AfCFTA is effective to the degree that it reduces transactional trade costs, expands intra-African trade and increases overall economic efficiency for member states through specialisation based on comparative advantages (Abrego et al. 2020). While long-term implications of a free trade area include higher productivity levels, increased foreign direct investment and ultimately improved GDP growth rates for member states, success will depend heavily upon the existence of complementary measures, including physical infrastructure development, institutional capacity building and trade readiness (Abrego et al. 2020; Adeleke, Osakede & Ajeigbe 2021). However, these expected gains are not automatic, as the effectiveness of tariff liberalisation in driving economic growth remains uncertain when non-tariff barriers and infrastructure constraints are not addressed simultaneously, since such limitations can significantly restrict the full benefits of trade integration (Fontagné et al. 2023; Saygili et al. 2018).

Several studies have questioned the extent to which the AfCFTA will translate into broad-based aggregate growth, pointing out that its benefits are likely to be unevenly distributed across countries (Ofori 2021; Ofori et al. 2022). Economies with more developed industrial bases, stronger institutions and better infrastructure are better positioned to take advantage of expanded market access, scale up production and attract investment, while less developed countries may struggle to compete, leading to asymmetric gains (Ezugwu & Duruji 2023). As a result, rather than uniformly boosting growth, AfCFTA could exacerbate income and development disparities between member states.

On the other hand, trade liberalisation reduces tariff and non-tariff barriers, which lowers trade costs and facilitates increased intra-regional exchange (Fontagné et al. 2023; Saygili et al. 2018). Similarly, this expansion in trade flows can generate scale effects, improve resource allocation and enhance productivity through increased competition and specialisation (Abrego et al. 2019; Maliszewska et al. 2020). At the same time, improved market access may stimulate investment and industrial development, particularly in sectors with comparative advantage (Ajewumi et al. 2024; Cooper & Sama 2024). However, the magnitude of these effects depends on structural conditions, including infrastructure quality, institutional capacity and the degree of economic diversification across countries (Ajewumi et al. 2024; Gbigbidje, Temishi & Asoro 2023).

The empirical literature identifies multiple transmission channels through which AfCFTA is associated with GDP outcomes, primarily through tariff reduction, removal of non-tariff barriers and infrastructure modernisation (Abrego et al. 2020). These mechanisms could have an impact on trade flows and overall economic welfare. For example, empirical findings reported by Ajewumi et al. (2024) indicate a positive relationship between AfCFTA implementation and GDP growth. Similarly, CGE-based simulations found a substantial expansion of intra-African trade under model-based assumptions, with a potential doubling of trade volumes, reductions in poverty and increases in real income of approximately 7% (Maliszewska et al. 2020). Several studies further demonstrate that the removal of non-tariff barriers constitutes a key determinant of economic gains (Fontagné et al. 2023; Omoju 2021; Saygili et al. 2018). In contrast, other analyses highlight potential risks associated with reduced tariff revenues and unequal distribution of gains, which may result in net welfare losses for certain countries or regional groupings (Mureverwi 2016; Saygili et al. 2018).

At the sectoral level, simulation-based projections suggest that services, manufacturing and transport may capture the largest gains from the agreement (Cooper & Sama 2024; Mhonyera & Meyer 2023). As mentioned previously, these sectoral dynamics suggest that larger economies, such as Nigeria, South Africa and Ethiopia, may realise greater gains because of their export capacity and domestic market size (Ajewumi et al. 2024; Mengistu et al. 2024; Mhonyera & Meyer 2023). The heterogeneous effects of the AfCFTA across countries indicate that while larger economies may experience stronger GDP growth, other countries may face adjustment costs associated with reduced fiscal revenues or limited economic diversification (Cooper & Sama 2024; Mengistu et al. 2024). Empirical and simulation-based estimates of GDP and income effects vary significantly, ranging from approximately 0.6% to over 7% increases in real income (Fontagné et al. 2023; Maliszewska et al. 2020; Mengistu et al. 2024).

A fundamental question in the trade and development literature is whether trade liberalisation alone is sufficient to generate sustained economic growth. The dominant evidence indicates that infrastructure deficits, persistent non-tariff barriers and disparities in economic development levels continue to constrain the effectiveness of the AfCFTA (Ajewumi et al. 2024; Gbigbidje et al. 2023; Saygili et al. 2018). Governance capacity and policy coordination are also important factors that affect development outcomes (Adamu, Jazbhay & Benyera 2023; Onuwa & Adedire 2023). Some existing studies rely on CGE models, which, while useful in simulating policy scenarios, are grounded in equilibrium assumptions that may not fully capture the dynamic adjustments, structural transformations and heterogeneous responses associated with evolving trade policies and economic systems (Cooper & Sama 2024; Fontagné et al. 2023; Omoju 2021). This methodological limitation reduces external validity and simplifies complex economic interactions (Abrego et al. 2019; Cooper & Sama 2024; Ekobena et al. 2021). Given that the AfCFTA remains a relatively recent agreement, a large share of the existing literature relies on ex ante projections or pre-implementation data. The limited temporal horizon also constrains the ability to identify realised effects and affects the robustness of conclusions regarding long-term impacts on GDP and welfare (Alechenu 2021; Fofack, Dzene & Hussein 2021; Mangeni & Mold 2024; Mengistu et al. 2024).

In summary, the literature indicates that the AfCFTA is expected to support economic growth and improve living standards across Africa, although the magnitude and distribution of these outcomes remain uncertain. The current body of evidence is largely informed by simulation models, with comparatively few studies examining the actual effects of implementation using observed macroeconomic data, especially during the initial implementation period. This study contributes to the literature by providing a panel-based empirical assessment of GDP dynamics among early AfCFTA ratifiers, offering complementary evidence to the existing simulation-based findings.

This study uses a longitudinal panel data approach to measure the relationship between GDP and AfCFTA participation among selected countries that ratified the agreement during its early adoption phase. This analytical framework enables the evaluation of both the scale and dynamics of economic activity, as well as the effectiveness of economic intervention policies within the national economy. In the context of international trade, GDP is widely used as an indicator of how free trade agreements, such as the AfCFTA, impact economic development, productivity and competitiveness among economies. The panel structure allows the analysis to capture relationships between variables across multiple countries over an extended time period (Marinov 2014).

The empirical strategy adopts an interrupted panel framework, which evaluates changes in GDP dynamics associated with the AfCFTA ratification period beginning in 2018. The choice of 2018 as the intervention point is based on institutional and methodological considerations. Institutionally, it corresponds to the signing of the AfCFTA Agreement and the initial ratification by the countries included in the sample, representing the first common policy milestone across all observations (Abrego et al. 2019). Methodologically, the use of a common breakpoint ensures a consistent pre- and post-ratification comparison while recognising that the estimated associations reflect the early implementation phase of the Agreement rather than its fully realised long-term effects. This approach allows for the identification of structural shifts in the data over time and provides an estimate of whether the post-ratification period is associated with statistically significant changes in the dependent variable. The identification relies on within-country temporal variation and should be interpreted as capturing associations conditional on the model specification rather than a strictly causal treatment effect.

The analytical sample was defined on the basis of the official AfCFTA ratification timeline published by the African Union (AU 2023). Specifically, the study focuses on countries that deposited their instruments of ratification during the initial ratification phase between May 2018 and November 2018, thereby constituting the first cohort of early AfCFTA ratifiers. Restricting the analysis to this sample provides the longest common post-ratification observation window currently available for empirical assessment of the Agreement, while reducing heterogeneity arising from differences in the timing of ratification across member states. The sample consists of seven countries (Ghana, Kenya, Rwanda, Niger, Chad, Eswatini, and Côte d’Ivoire) for which complete and consistent annual GDP observations were available throughout the observed period. Consequently, the final sample was determined by the intersection of two methodological criteria: (1) membership in the earliest ratification cohort and (2) the availability of a complete longitudinal dataset suitable for panel estimation. While this sampling strategy strengthens temporal comparability across countries, it does not provide a formal control group; therefore, the estimated coefficients should be interpreted as within-sample associations rather than causal treatment effects. The analysis covers the period from 2012 to 2024 and utilises data from the World Bank’s World Development Indicators, which provide standardised and internationally comparable macroeconomic data (World Bank 2026). The dataset was constructed by extracting annual observations for the selected variables over the period 2012–2024 for the seven countries included in the sample. Data cleaning involved consistency checks for missing values and alignment of country-year observations to ensure comparability across the panel. Two dependent variables are specified: (1) the natural logarithm of nominal GDP in current US dollars, and (2) nominal GDP in current US dollars. The logarithmic transformation allows for the interpretation of estimated coefficients in proportional terms and mitigates heteroskedasticity, while the level specification preserves absolute economic dynamics and captures price-related effects.

The empirical strategy consists of three complementary stages. Firstly, a fixed-effects panel regression is estimated to evaluate the association between the post-2018 ratification period and GDP while controlling for country-specific heterogeneity, common time trends and the COVID-19 shock. Secondly, a series of robustness analyses including cluster-robust inference, Wild Cluster Bootstrap estimation, a lagged dependent variable specification, a placebo test and a supplementary segmented analysis of the pre- and post-2018 GDP trajectory are conducted to assess the stability of the findings. Finally, a counterfactual analysis based on pre-2020 trends is employed to compare observed GDP trajectories with projected outcomes.

The baseline regression specification is defined as follows (Equation 1):

where yit denotes the macroeconomic outcome for country i in year t. The model was estimated using an OLS framework with country dummy variables αi, which is mathematically equivalent to a conventional country fixed-effects estimator. Accordingly, the reported country coefficients represent country-specific intercepts relative to the omitted reference category and are included solely to control for unobserved, time-invariant heterogeneity rather than for substantive interpretation. The coefficient β captures the association between the post-ratification period and GDP levels. The coefficient δ represents the long-term growth trend through the centred time variable YearCenteredt. The parameters θ2020 and θ2021 capture the shocks associated with the COVID-19 pandemic for 2020 and 2021, respectively, through dummy variables D2020,t and D2021,t. εit denotes the stochastic error component. The coefficient on the Post2018 variable captures the average within-country change in GDP levels associated with the post-ratification period, conditional on country fixed effects, the common time trend and pandemic controls. Given the absence of a comparison group, this coefficient should be interpreted as an association reflecting shifts in GDP dynamics rather than a causal estimate of the AfCFTA effect.

The estimation relies on an OLS framework with country-specific fixed effects. These fixed effects absorb all time-invariant characteristics, including economic structure, geographical conditions and infrastructure development, thereby eliminating cross-country heterogeneity and allowing for consistent estimation of within-country changes associated with the post-ratification period. Statistical inference is based on cluster-robust standard errors at the country level, which correct for both heteroskedasticity and serial correlation. However, given that the sample contains only seven country clusters, statistical inference should be interpreted with caution, as cluster-robust standard errors may be less reliable in panels with a very small number of clusters.

The analysis further incorporates a counterfactual framework implemented in Python, which projects log-linear GDP trends based on the pre-2020 period. This projection provides a descriptive benchmark against which observed post-2020 GDP levels can be compared. The year 2020 is selected as the counterfactual baseline, as it represents a structural break in the macroeconomic trajectories of most African economies as a result of the COVID-19 pandemic, which disrupted previously stable growth patterns.

This approach provides an additional perspective on GDP dynamics associated with the post-ratification period, allowing deviations from pre-pandemic growth trajectories to be identified and compared with observed outcomes. The counterfactual estimates should be interpreted as descriptive benchmarks rather than as evidence isolating the effects of the AfCFTA from other contemporaneous economic influences. The counterfactual analysis is based on a log-linear model estimated separately for each country using pre-2020 data. The model is specified as follows (Equation 2):

where GDP_it denotes the gross domestic product of country i in year t, and ln(GDP_it) represents its natural logarithm, applied to linearise the exponential growth trajectory. The constant αi is a country-specific intercept of the individual time-series regression. The slope coefficient βi reflects the GDP growth trend over the pre-2020 period, while t denotes the time index. The error term ε_it captures idiosyncratic, country-specific fluctuations that are not explained by the model. Because the counterfactual model is estimated independently for each country, αi should be interpreted as a country-specific intercept rather than a fixed effect in the conventional panel-data sense.

The empirical design is subject to several limitations. Firstly, the absence of a formal control group restricts causal interpretation. Secondly, the relatively small sample size and short post-ratification period limit the ability to capture long-term effects. Thirdly, GDP represents a broad aggregate indicator and does not allow for sector-specific analysis. These limitations imply that the results should be interpreted as early empirical evidence on GDP dynamics associated with AfCFTA participation rather than definitive causal estimates.

Table 1 presents the results from the fixed-effects panel regression used to estimate the association of the AfCFTA with nominal GDP.

The coefficient of the main explanatory variable ‘Post2018’ (0.0692; p = 0.0014) indicates a positive and statistically significant association between the post-2018 period and nominal GDP. In the logarithmic specification, this coefficient corresponds to an approximate 7.17% higher level of nominal GDP relative to the pre-2018 trajectory, conditional on country fixed effects, the time trend and pandemic controls. This suggests that the estimated association is not only statistically significant but also economically meaningful within the sampled early-ratifying economies. The 95% confidence interval (2.71% – 11.83%) supports both statistical significance and economic relevance. This finding provides empirical support for Hypothesis 1, which posits a positive association between the post-2018 AfCFTA implementation period and GDP levels among early-ratifying economies. The coefficient of ‘YearCentered’ (0.0364; p = 0.0014) is positive and statistically significant, reflecting the underlying common growth trend across the sample.

The COVID-19 shock is captured through the pandemic dummy variables. The coefficient for 2020 (−0.0717; p = 0.0042) is negative and statistically significant, indicating a contraction in GDP consistent with the global economic downturn. The coefficient for 2021 (0.0056; p = 0.7966) is small and statistically insignificant, suggesting a partial and uneven recovery across countries.

The country-specific coefficients reflect differences in baseline GDP levels relative to the reference category. Kenya (1.6707), Ghana (1.3644) and Côte d’Ivoire (1.2525) exhibit positive coefficients, indicating higher average GDP levels relative to the reference category. In contrast, Niger (−0.2603), Rwanda (−0.4885) and Eswatini (−1.2992) display negative coefficients, reflecting relatively smaller economic scales. These coefficients are included to account for unobserved, time-invariant heterogeneity across countries and are not interpreted as evidence of the effects of the explanatory variables.

The model demonstrates a strong statistical fit to the data (Table 2), with an R² of 0.989 and an adjusted R² of 0.987. These values indicate that a large share of the variation in nominal GDP is explained by the model. However, the high explanatory power should be interpreted with caution, as it partly reflects the inclusion of country fixed effects and a common time trend capturing persistent growth dynamics. As such, the R² should not be interpreted as evidence of strong predictive performance, but rather as an indication of the model’s ability to account for deterministic components of GDP variation. The robustness of inference is enhanced through cluster-robust standard errors and bootstrap procedures (Table 4), although results should still be interpreted with caution given the small number of country clusters. The F-statistic (507.1; p < 0.001) indicates that the explanatory variables are jointly significant. The Akaike Information Criterion (AIC = −119.1) and Bayesian Information Criterion (BIC = −91.51) are reported as measures of model fit and may be used for comparison with alternative specifications.

The diagnostic tests reported in Table 3 provide additional insight into the properties of the residuals and the extent to which the assumptions underlying the OLS estimator are satisfied (Petkov 2017). The Durbin–Watson statistic (0.671) suggests the potential presence of positive serial correlation, which is commonly observed in macroeconomic panel data because of the persistence of GDP over time. While serial correlation does not affect the consistency of the fixed-effects coefficient estimates, it may lead to downward-biased conventional standard errors. In addition, the Omnibus test (15.138; p = 0.001) and the Jarque–Bera test (17.430; p = 0.000164) suggest that the residuals deviate from normality. To address these issues, the estimation employs cluster-robust standard errors at the country level, which account for heteroskedasticity and within-country serial correlation while reducing the sensitivity of statistical inference to departures from normality. Given the relatively small number of cross-sectional units, inference is further strengthened through the Wild Cluster Bootstrap procedure reported in Table 4, providing an additional robustness check for statistical significance, following Cameron, Gelbach and Miller (2008). This approach provides robust p-values in settings with a limited number of clusters. The estimated model includes country fixed effects, a time trend, and control variables for the pandemic years 2020 and 2021. The main independent variable is ‘Post2018’, which captures the period following the implementation of the AfCFTA.

The results indicate that the coefficient Post2018 is positive (0.069) and statistically significant under both the standard cluster robust regression (p = 0.0014) and the Wild Cluster Bootstrap procedure (p < 0.001). These findings suggest that the estimated post-2018 association remains robust and statistically significant, even after accounting for potential biases associated with the limited number of cross-sectional units.

As a robustness check, the model is re-estimated with the inclusion of a lagged dependent variable to account for GDP persistence (Table 5). The coefficient on the lagged term is positive and statistically significant (0.6029; p < 0.001), confirming the strong temporal dependence in GDP dynamics. Importantly, the Post2018 coefficient remains positive and statistically significant (0.0485; p = 0.0025), although reduced in magnitude, indicating that the main findings are not driven solely by persistence or trending behaviour. This result suggests that the positive post-2018 association is not solely attributed to GDP persistence and remains evident after controlling for dynamic adjustment effects. The reduction in the magnitude of the Post2018 coefficient from 0.6029 to 0.0485 after the inclusion of lagged GDP indicates that part of the observed association reflects underlying GDP persistence. Nevertheless, the coefficient remains positive and statistically significant, suggesting that the post-2018 association is not entirely explained by dynamic adjustment effects.

A placebo test was conducted using an artificial structural break in 2016 to examine whether a similar association would have been detected prior to the AfCFTA ratification period (Table 6). The coefficient of the placebo variable is negative and statistically insignificant (−0.0807; p = 0.1348), indicating the absence of a spurious structural shift prior to the AfCFTA ratification period. This finding provides evidence that the observed post-2018 association is unlikely to be attributable solely to an arbitrary choice of intervention timing. However, the placebo test should be interpreted with caution. As all countries in the sample belong to the treated group and no formal control group is available, the placebo specification does not constitute a conventional falsification test. Rather, it serves as an internal robustness check against spurious pre-existing structural changes within the analysed sample and supports the interpretation that the estimated post-2018 association is unlikely to be driven solely by the selected intervention date.

To complement the baseline regression, a supplementary segmented analysis was undertaken to examine whether the post-2018 GDP trajectory differs from the continuation of the pre-2018 trajectory (Table 7).

The supplementary segmented analysis indicates that the estimated annual GDP trend increased descriptively from approximately 3.03% before 2018 to 5.36% after 2018. However, neither the estimated post-2018 level change nor the change in the annual GDP trend is statistically significant, and the joint Wald test does not reject the null hypothesis of no statistically significant level or slope change after 2018 (p = 0.140). Consequently, while the supplementary analysis suggests stronger GDP dynamics during the post-ratification period, it does not provide conclusive statistical evidence of a structural change in the underlying growth pattern.

As a complementary analysis, a counterfactual approach was employed to compare observed GDP trajectories with projected GDP paths derived from pre-2020 log-linear growth trends. This approach offers a descriptive benchmark for assessing deviations from historical growth patterns during the post-ratification period. Figure 1 presents the observed and counterfactual GDP trajectories for the sample countries.

Larger economies, such as Kenya and Ghana, exhibit persistent negative deviations, with actual GDP in 2024 remaining below counterfactual projections. In contrast, Chad, Rwanda and Niger display substantial positive deviations, as their observed GDP levels in 2024 exceed expected values. Côte d’Ivoire also demonstrates a positive deviation, although more moderate in magnitude. Eswatini shows minimal deviation, indicating a relatively stable growth path. These observed deviations from projected pre-implementation trends provide descriptive evidence that GDP dynamics differed from historical trends during the post-implementation period. Consistent with Hypothesis 2, the results indicate the presence of structural changes in GDP trajectories during the post-implementation period.

Smaller economies, including Chad, Niger, Rwanda and Eswatini, generally exhibit stronger positive deviations from their counterfactual trajectories, suggesting greater resilience and faster post-2020 adjustment. Larger economies, such as Kenya, Ghana and Côte d’Ivoire, display more moderate recovery patterns and, in some cases, remain below their pre-existing growth trends. Taken together, the econometric and counterfactual analyses indicate a positive and statistically significant post-2018 association with GDP dynamics across the sample. However, recovery from the pandemic-induced downturn remains uneven across countries. These findings should be interpreted as descriptive empirical evidence, rather than as a definitive estimate of the policy’s causal effect.

The findings of this study provide empirical evidence of a positive and statistically significant association between the post-ratification period of the AfCFTA and GDP dynamics among the sampled member states. The estimated increase of 7.17% in nominal GDP indicates that the post-2018 period is associated with higher GDP levels relative to pre-ratification trajectories. This result remains statistically significant under both standard cluster-robust estimation and the Wild Cluster Bootstrap procedure, which supports the robustness of the estimated association. The robustness of the results to the inclusion of a lagged dependent variable further supports the interpretation that the observed post-2018 association is not solely driven by underlying GDP persistence. The absence of a statistically significant placebo effect provides additional support for the temporal specification adopted in the study. However, given that all countries in the sample belong to the treated group, the placebo test should be interpreted as an internal robustness check rather than a definitive falsification test.

Taken together, the results provide support for both hypotheses. The fixed-effects estimation identifies a statistically significant post-2018 association with GDP levels (H1), while the counterfactual analysis reveals deviations from pre-implementation trajectories (H2), suggesting the presence of structural shifts in macroeconomic performance during the early phase of AfCFTA implementation. However, these deviations should be interpreted as descriptive evidence rather than as definitive estimates of the causal effects of the AfCFTA.

The empirical findings align with simulation-based evidence suggesting positive macroeconomic gains from the AfCFTA (Abrego et al. 2019; Maliszewska et al. 2020) but differ in magnitude and interpretation. While CGE models project long-term welfare and income increases under full implementation scenarios, the present results reflect short-term realised dynamics under partial and evolving implementation conditions. This distinction highlights the gap between projected and observed outcomes and underscores the importance of empirical validation of simulation-based findings.

The counterfactual analysis further suggests that the macroeconomic response during the early implementation phase of the AfCFTA was not uniform across the sampled economies. While smaller economies such as Chad, Rwanda and Niger exceeded their projected pre-2020 GDP trajectories, larger economies including Kenya and Ghana remained below their historical trends, indicating greater divergence from their projected growth paths. The present study does not attempt to identify the underlying determinants of these differences. Nevertheless, the observed heterogeneity is consistent with the broader literature indicating that the macroeconomic outcomes of regional trade integration may vary according to country-specific structural characteristics, including differences in production structures, trade composition, trade costs and the pace of complementary policy reforms (Abrego et al. 2019; Baier, Bergstrand & Clance 2018; Maliszewska et al. 2020). These findings therefore should be interpreted within the broader national economic context of each participating country, while identifying the mechanisms underlying these heterogeneous responses represents an important avenue for future research.

The inclusion of the COVID-19 pandemic as a controlling factor is critical for the interpretation of the results. The economic downturn in 2020 that was followed by a partial rebound in 2021 illustrates how temporary external shocks can interfere with the normal transmission mechanism of integration policy. The incorporation of dummy variables into the model illustrates how GDP deviation can be distinguished from pre-pandemic growth trends and also from other factors, such as those resulting from the pandemic. Consequently, these dummy variables improve the reliability of the estimated post-2018 associations. Furthermore, this emphasises the need to consider external shocks when assessing macroeconomic factors within the framework of regional integration.

This study contributes to the existing literature by complementing simulation-based projections with empirical evidence derived from observed macroeconomic data. By focusing on early AfCFTA ratifiers and employing a panel-based framework with robust inference techniques and counterfactual trend comparison, the analysis provides an initial empirical perspective on GDP dynamics during the early implementation phase of the agreement.

The findings suggest that the post-2018 implementation period of the AfCFTA is positively associated with GDP dynamics among the sampled early ratifying economies, although the magnitude of these associations differs across countries. On one hand, the integration increases market size, economies of scale and resource allocation efficiencies, thereby enabling firms to augment production and engage in more specialised activities across national borders. On theother hand, the structural differences across countries in terms of industrial capabilities, infrastructure and institutional quality cause the magnitudes and the distributions of potential gains to vary across countries.

The observed heterogeneity indicates that a combination of complementary policies, related to infrastructure advancement and trade facilitation, will be necessary to maximise the potential benefits of regional integration. To better assess the economic consequences of the AfCFTA, future research should include longer time horizons, additional countries in the analysis, and investigate the effects on specific sectors.

Beyond providing empirical evidence on GDP dynamics during the early implementation phase of the AfCFTA, this study contributes to the existing literature by complementing the predominantly ex ante simulation-based evidence with estimates derived from observed macroeconomic data. While the findings should not be interpreted as establishing causal effects, they demonstrate the value of evaluating regional integration initiatives using realised macroeconomic outcomes as implementation progresses. From a policy perspective, the results indicate that the macroeconomic associations observed during the early implementation phase are not uniform across the sampled economies, highlighting the importance of continued empirical monitoring of realised macroeconomic outcomes as the Agreement matures.

Several limitations should be acknowledged. Firstly, the empirical design does not include a formal control group, which restricts the ability to establish causal inference. Secondly, the relatively small sample size and the short post-ratification period limit the generalisability of the findings and constrain the analysis of long-term effects. Thirdly, the use of aggregate GDP as the sole outcome variable does not allow for sector-specific or distributional analysis. Fourthly, the AfCFTA remains in an evolving implementation phase, and the timing and depth of policy application may differ across countries. Accordingly, the estimated post-2018 associations should be interpreted as reflecting GDP dynamics during the early implementation phase of the AfCFTA among the sampled economies, rather than as evidence of causal effects attributable exclusively to the Agreement.