Bilateral trade corridors classified by structural trade potential, intervention tractability, and development objective. Each point represents one corridor. Point size indicates the estimated trade gap. Only corridors meeting the three-product classification criteria are shown.
The horizontal axis measures economic scale — the log of the gravity model's prediction of how much two countries should trade, normalised across all corridors. The vertical axis measures tractability — how amenable the identified barrier is to policy intervention. Tractability is higher for logistics and policy barriers than for political or structural constraints.
Gateway Efficiency corridors are active trade corridors with at least one coastal endpoint where efficiency improvements unlock large immediate gains.
Trade Bottleneck corridors have high predicted potential but low realisation due to logistics or policy barriers amenable to targeted intervention.
Inclusion Corridors connect landlocked and structurally excluded economies. The objective is development connectivity, not trade volume maximisation.
Corridors that do not qualify for any of the three categories are not shown. They are retained in the underlying dataset but are not presented as diagnostic targets.
Top 50 active trade corridors where efficiency improvements unlock immediate gains. At least one endpoint has sea access in every Gateway corridor. Dark bars show observed trade; light extensions show remaining gap.
These corridors are already trading at meaningful scale. The dark bar shows observed bilateral trade (BACI 2022). The light extension shows the remaining gap to the model's structural prediction. The instrument here is efficiency — not access infrastructure.
Algeria–Egypt and Egypt–Tunisia are the North African frontier: active trade relationships where customs friction and documentary requirements prevent fuller realisation. Ghana–Nigeria is West Africa's most commercially significant corridor — both countries have functioning ports, complementary economic structures, and existing trade relationships. The Seme–Krake border crossing is the identified friction point.
Corridors where the model predicts substantially more trade than is observed. Light bars show the structural prediction; dark bars show observed trade. The gap is not explained by geography or economic incompatibility.
The gravity model estimates a structural level of trade based on patterns observed across 2,862 African bilateral pairs from 2015 to 2019. When observed trade is far below this prediction, and when the barrier is identified as logistics or policy rather than political conflict, the corridor enters this category.
Nigeria appears as the underperforming partner across more corridors than any other country in this analysis. Guinea–Nigeria, Angola–Nigeria, DR Congo–Nigeria, Ethiopia–Nigeria — in each case the model predicts significant bilateral trade that is not occurring. This convergence across independent bilateral pairs suggests the constraint is specific to Nigeria's border administration rather than to any individual corridor. The intervention is customs reform, not roads.
Corridors connecting landlocked and structurally excluded economies, ranked by inclusion score — a composite of landlocked access deprivation, network isolation, economic vulnerability, and trade concentration.
The inclusion score combines four components: landlocked status (40% weight), network isolation (25% — how few trade relationships each country maintains relative to its economic size), economic vulnerability (20% — inverse of GDP per capita), and trade concentration (15% — Herfindahl index of partner concentration).
These corridors should not be evaluated on the same criteria as Gateway or Bottleneck corridors. The appropriate metric is development connectivity — how many people gain access to regional markets — not bilateral trade value. Niger–Chad and Burundi–Rwanda score highest because both endpoints are landlocked and both countries maintain very few existing trade relationships. They require physical access infrastructure, not customs reform.
Several inclusion corridors span multiple sovereign borders. They are analytically valid as bilateral pairs but operationally correspond to named multimodal transport routes. Phase 3 will map them onto physical corridors.
Total bilateral goods trade between African countries 2000–2024. Source: CEPII BACI V202601.
Standard bilateral reporting significantly undercounts intra-African trade because many African countries do not consistently report their bilateral flows. BACI resolves this by reconciling both sides of every bilateral relationship. The result is approximately three times more recorded trade than standard IMF data — not because trade increased, but because the measurement is more complete.
Trade grew from approximately $20 billion in 2000 to $114 billion in 2023. The agreement establishing AfCFTA officially entered into force on May 30, 2019. The 2020 dip reflects COVID-19. Trade recovered in 2021, though attributing this recovery to AfCFTA specifically requires caution given the simultaneous global trade rebound.
Each point is one bilateral trade corridor. The diagonal line shows where a corridor would sit if it traded exactly as much as the model predicts. Points below the line trade less than expected.
The diagonal line represents full realisation — corridors on this line trade exactly as much as the model predicts. Corridors above the line are overperforming: South Africa–Mozambique and South Africa–Botswana sit above because SADC integration is working and trade exceeds structural predictions.
Corridors far below the line are the diagnostic targets. The colour indicates the probable constraint type: rust for logistics barriers, red for political or institutional barriers, navy for policy or information gaps.
The model was estimated on 2015–2019 data across 2,862 African bilateral pairs. It controls for GDP, distance, shared borders, common language, colonial ties, FTA membership, and landlocked status. Predictions for 2024 apply the estimated structural relationships to current GDP with bilateral characteristics treated as stable.
Average number of bilateral trade relationships above $10 million annually, 2021–2024. Dark bars: African partners. Light bars: non-African partners.
Several of Africa's largest economies by population and GDP maintain more trade relationships with non-African partners than with other African countries. Ethiopia — 126 million people — averaged six intra-African relationships against thirty-four outside Africa. Nigeria averaged roughly three times as many non-African as African relationships.
This asymmetry reflects the historical structure of African trade: economies oriented toward commodity exports to Europe and Asia, with comparatively weak intra-regional exchange. The pattern holds consistently across all four years (2021–2024), confirming it is structural rather than cyclical.
World Bank Logistics Performance Index scores for African countries. Scale of 1 (poor) to 5 (excellent). Sorted by overall score.
The LPI measures customs procedures, infrastructure quality, logistics competence, and international shipment ease. Customs is the most directly actionable dimension — it is a software and institutional reform, not a capital infrastructure investment.
South Africa scores 3.7 overall — the continental benchmark. Angola and Cameroon, which appear repeatedly in the Trade Bottleneck corridors, score 2.1 on overall logistics. The correlation between low customs scores and high bottleneck frequency is the empirical basis for the customs modernisation recommendation in this analysis.
African countries trade with each other far less than their economic size and geography suggest they could. This analysis asks a precise question: if two African countries faced no avoidable barriers — only geography, economic size, and structural realities — how much trade would we expect between them? We then compare that expected level with what is actually observed. The difference is the trade opportunity.
Illustrative example. Hover over charts for actual corridor figures.
For every pair of African countries, the model evaluates six structural drivers of trade — characteristics that shape how much two economies are likely to exchange regardless of policy choices. The model learns the strength of each driver from five years of African trade data (2015–2019), a period before COVID-19 and supply-chain disruption introduced temporary distortions.
The analysis classifies corridors into three categories, each corresponding to a different type of barrier and a different intervention instrument.
Trade is already happening. The opportunity comes from making border processes faster, simpler, and more predictable. These corridors have at least one coastal country and active trade relationships. The instrument is efficiency, not access.
The model predicts strong trade potential, but very little trade is occurring. These corridors are not structurally incompatible — something specific is blocking them. The barrier is typically border administration or policy misalignment rather than missing roads.
These corridors connect landlocked and structurally isolated economies. The objective here is not trade value alone — it is access, resilience, and economic inclusion for populations with limited alternatives. These corridors correspond to concessional finance.
Africa's largest economy — $363 billion GDP, 223 million people — maintains meaningful trade with fewer than 20 African partners. The model predicts large bilateral flows between Nigeria and its neighbours; actual trade is a fraction of that. Nigeria is not disconnected from the world: it trades meaningfully with 49 non-African partners. It is specifically disconnected within Africa. The analysis points to border administration — not road infrastructure — as the binding constraint, suggesting customs reform as the highest-return intervention for the West African trade network.
Algeria and Morocco share a 1,559-kilometre land border, a common language, a common colonial history, and combined GDP approaching $400 billion. The model predicts they should be among Africa's most active bilateral traders. Observed trade is approximately $250 million — around 12 percent of the model's prediction. The border has been closed since 1994 following a political dispute. This single decision represents the largest bilateral trade gap in North Africa, and cannot be addressed through trade facilitation instruments alone.
South Africa trades meaningfully with 42 African partners — the highest intra-African connectivity on the continent — and several of its SADC corridors (Botswana, Mozambique, Namibia) trade above the model's prediction. These are the only corridors in this analysis where observed trade consistently exceeds structural expectations. South Africa's SADC corridors function as the benchmark — demonstrating that the barriers identified elsewhere are not inherent features of intra-African trade but outcomes of specific, addressable institutional conditions.
This analysis identifies where structural trade opportunities exist, what kind of constraint is most likely binding, and which policy objective each corridor serves. It does not assume every gap should be closed, does not replace feasibility studies or engineering assessments, and does not model physical infrastructure costs or returns.
Barrier classifications — logistics, political, policy — are probabilistic inferences from observable data, not direct observations of the underlying mechanism. Investment decisions based on this analysis should be accompanied by corridor-level operational assessments. Phase 3 will map identified corridors onto named physical transport routes.