Loading workbench page
Fetching primary parquet sources and computing exhibits.
Loading chart...
Fetching primary parquet sources and computing exhibits.
In 2024, the world median ratio of merchandise trade to GDP is 57% and of commercial-bank credit to GDP is 40%. The bottom-right quadrant of Figure 1 is the underserved zone: open, trade-heavy economies with shallow banking. The Asian Development Bank's 2023 Trade Finance Gaps, Growth and Jobs Survey (ADB Brief No. 256, September 2023) puts the global trade-finance gap in 2022 at $2.50T, with rejection rates concentrated in developing-economy SMEs. This page maps where private-credit funds, development finance institutions, and factoring firms should look first.
Each edition of the ADB Trade Finance Gaps, Growth and Jobs Survey reports a single global estimate of unmet but bankable demand for trade finance. Read in sequence, the editions show the gap holding near USD 1.5 trillion through the late 2010s, then stepping up after the pandemic to $2.50T in 2022 (about 10% of merchandise trade), where the 2025 edition leaves it. The 2017 to 2023 points are taken directly from the ADB brief PDFs (Briefs No. 83, 113, 192 and 256); the 2016 and 2025 points are from the ADB publication page and news release.
SELECT reference_year, gap_usd_trillion, pct_of_global_trade FROM 'trade_finance_gap.parquet' ORDER BY edition_year;
The gap is not evenly distributed across applicants. In the 2019 survey (2018 data, ADB Brief No. 113), the rejection rate on small and medium-enterprise proposals was 45%, against 39% for mid-sized and large firms and 17% for multinationals. Women-owned firms were rejected more often than men-owned firms (44% versus 38%). SMEs were about 37% of trade-finance demand but absorbed a disproportionate share of the rejections, which is why development-finance and factoring capacity aimed at the SME segment closes more of the gap per dollar than capacity aimed at large corporates.
SELECT label, value_pct FROM 'trade_finance_breakdowns.parquet' WHERE edition_year=2019 AND metric_key LIKE '%rejection_rate' ORDER BY value_pct DESC;
The x-axis is commercial-bank outstanding loans as a share of GDP (IMF Financial Access Survey, equivalent in concept to the World Bank's FS.AST.PRVT.GD.ZS domestic credit to private sector). The y-axis is merchandise trade (exports plus imports, BACI, in current USD) divided by GDP in current USD (World Bank WDI, NY.GDP.MKTP.CD). Countries in the lower-right of the plot are where trade openness is high but bank intermediation is thin, so a rising share of each invoice travels without bank financing. Auboin (2009, WTO Staff Working Paper ERSD-2009-16) and Niepmann and Schmidt-Eisenlohr (2017, Journal of International Economics, 107: 111-126) document that bank-intermediated trade finance is concentrated in advanced economies, so the underserved zone is precisely where specialised trade-finance providers face the least competition from domestic banks.
WITH trade AS (
SELECT country_code, (total_exports+total_imports)*1000 AS trade_usd
FROM 'country_year_totals.parquet' WHERE year=2024
), gdp AS (
SELECT iso3, value AS gdp_usd FROM 'wdi_data.parquet'
WHERE indicator='NY.GDP.MKTP.CD' AND year=2024
), credit AS (
SELECT iso3, value AS credit_gdp FROM 'imf/financial_access.parquet'
WHERE indicator='Outstanding loans, Commercial banks' AND year=2024
)
SELECT c.iso3, credit.credit_gdp AS x,
(trade.trade_usd/gdp.gdp_usd)*100 AS y
FROM 'countries.parquet' c
JOIN trade ON trade.country_code=c.code
JOIN gdp ON gdp.iso3=c.iso3 JOIN credit ON credit.iso3=c.iso3;The Asian Development Bank's 2023 Trade Finance Gaps, Growth and Jobs Survey (ADB 2023, Brief No. 256) reports a global unmet trade-finance need of $2.50T in 2022, up from USD 1.7 trillion in 2020. Aggregated across surveyed banks, roughly 10% of proposed trade-finance transactions are rejected, with the rejection rate substantially higher for SMEs and for applicants in developing economies. We project that 10% coefficient onto each underserved country's BACI-measured total trade to obtain an order-of-magnitude country-level gap. This is an upper-bound estimate in countries with deep offshore bank finance (e.g. Hong Kong trade routed through Singapore) and a lower-bound estimate in countries where the ADB survey rejection rate runs above its global average (Sub-Saharan Africa, parts of South Asia).
The IMF Financial Access Survey reports outstanding commercial-bank loans to small and medium enterprises as a percentage of GDP, alongside total outstanding commercial-bank loans. Their ratio is the share of the loan book committed to SMEs, a proxy for how much of a country's credit stock is reaching the firms that drive the bulk of traded-goods production (Beck, Demirguç-Kunt and Maksimovic 2005, Journal of Finance 60(1): 137-177 on SME financing constraints). Economies with a low SME-loan share and a high trade-to-GDP ratio are where working-capital demand from exporters and importers most exceeds domestic banking supply, which is the operating thesis of the ADB 2023 Survey. Displayed year is 2024, the latest FAS vintage with coverage across the underserved set.
Trade finance is extended at the corridor level: a bank in Singapore finances a letter of credit for a Vietnamese exporter shipping to a Bangladeshi buyer. If either leg sits in the underserved quadrant, the corridor is a candidate for specialist intervention. We list the top-20 bilateral flows (BACI 2024, current USD) where at least one leg is an underserved economy, annotated with each leg's credit/GDP ratio. Corridor-level gap = bilateral flow × 0.1, the same ADB rejection-rate coefficient applied in Figure 2. Head and Mayer (2014, Handbook of International Economics, vol. 4, ch. 3) formalise the corridor as the unit of analysis.
Ahn, Amiti & Weinstein (2011, American Economic Review Papers & Proceedings101(3): 298-302) show that letter-of-credit usage rises with distance, with counterparty risk, and with the novelty of the buyer-seller relationship. Niepmann & Schmidt-Eisenlohr (2017, Journal of International Economics 107: 111-126) document the same on US bank-level data and add that when correspondent banks cut LC lines during the 2008-09 crisis, the most exposed exporters were those with the most-concentrated destination books. Partner-HHI on export shares is therefore a clean proxy for concentration risk of trade-finance demand: high HHI + underserved banking = a country where a single corridor withdrawal can collapse the trade-finance pipeline. BIS consolidated banking statistics (see Table B4, 'Foreign claims of BIS reporting banks') show the same pattern from the lender side. The bars below rank the underserved set by partner-HHI in 2024.
The Bank for International Settlements' CPMI correspondent-banking reports (BIS-CPMI 2016, 'Correspondent banking'; BIS-CPMI 2020, 'New correspondent banking data'; Rice, von Peter & Boar 2020, BIS Quarterly Review) document a sharp long-run contraction in active correspondent-bank corridors: roughly a 20% decline in active correspondents between 2011 and 2018, concentrated on small economies in the Caribbean, Pacific, and Sub-Saharan Africa. A correspondent-bank corridor withdrawal cuts precisely the rail that letter-of-credit and documentary-collection trade finance rides on. We do not have BIS SWIFT correspondent-relationship microdata in this workbench, so as a proxy we plot the top-20 bilateral trade corridors with at least one underserved leg (BACI 2024), on the principle that a correspondent-banking map approximately follows the merchandise- trade map (Head & Mayer 2014). Nodes are sized by each economy's total trade value in the set; underserved legs are flagged in orange; edges are the BACI bilateral flows (thicker = higher flow, same corridors as Figure 4).
The OECD Arrangement on Officially Supported Export Credits (first 1978 Consensus; current 2025 revised text) sets minimum pricing, tenor, and premium floors for the 11 participants' export-credit agencies, backing roughly USD 100 billion of medium- and long-term official support each year (Berne Union Yearbook). Private non-life insurers cover the bulk of short-term export-credit insurance outside that perimeter. IMF FAS does not publish premia directly; the insurer density (per 100,000 adults) per USD billion of trade is the closest open-data proxy for where private capacity is thinnest.
The most-cited long-run change in trade-finance structure is the shift out of letters of credit (LCs) into open-account payment as a share of world merchandise trade. The IMF & Bank of Korea (2009) Survey on Trade Financeestablished the foundational read: in the early 2000s, bank-intermediated instruments (LCs plus documentary collections) accounted for roughly 40% of trade by value; by the GFC it was in the mid-30s, sliding into the low 20s by the mid-2010s as corporates shifted to open account on the back of deepening bank relationships, credit insurance, and supply-chain-finance platforms. Subsequent ICC Trade Registerreports (ICC Banking Commission, annual) and the Boston Consulting Group & SWIFT Global Trade: Securing Future Growth (2017) refine the series. The anchors below are published benchmark points from those sources; they are not a continuous survey of all banks (no such survey exists) but they bracket the direction of travel.
The Figure 1 cross-section is a snapshot. Whether the credit gap is widening, stable, or closing matters for the trade-finance-gap forecast: Chor & Manova (2012, Journal of International Economics 87(1): 117-133) show that credit-supply contractions in source countries depress exports of financially-dependent sectors more than the trade-credit channel alone predicts, and the symmetric implication is that credit deepening loosens that constraint over time. Below we plot the IMF FAS commercial-bank-loans-to-GDP series across the 10-year window ending in 2024 for the top-5 underserved economies by the Figure 2 gap ranking. Rising lines mean the gap should narrow organically; flat or falling lines indicate the gap is structural and waiting for external supply.
SELECT iso3, year, value AS credit_gdp
FROM 'imf/financial_access.parquet'
WHERE indicator = 'Outstanding loans, Commercial banks'
AND value IS NOT NULL
AND year BETWEEN 2014 AND 2024
AND iso3 IN ('MHL','GUY','ZMB','GMB','GIN')
ORDER BY iso3, year;