BIS Statistics Exchange and Derivatives Data

Central-bank statistics usually reach you one topic at a time - a credit table here, an exchange-rate sheet there, none of them agreeing on keys or units. The BIS Statistics Exchange and Derivatives Data record assembles the Bank for International Settlements' (BIS) statistical suite as one catalogue with one dimensional grammar. Credit to the non-financial sector is the spine: financing from all sources - domestic banks, other domestic financial corporations, non-financial corporations and non-residents - broken down by borrowing sector (general government; the private non-financial sector split into non-financial corporations and households plus NPISHs) and by lending sector, at market values per SNA 2008 with government debt additionally stated at nominal value. Currency and deposits, loans and debt securities together form what the BIS calls core debt.

Around that spine sit the market-facing families: exchange-traded derivatives, OTC derivatives turnover and outstanding positions, international debt securities, effective exchange rates, US dollar exchange rates and global liquidity indicators - 27-plus dataflows in all, spanning banking, derivatives, debt securities, FX and property prices. Get a sample of this dataset(/pricing) cut to the economies, sectors and window you model.

Three series from the total credit panel, exactly as the wide layout lands:

FREQ BORROWERS_CTY  TC_BORROWERS TC_LENDERS          VALUATION UNIT_TYPE               TC_ADJUST 2023-Q4  2025-Q4
Q    LU             N            A (All sectors)     M         770 (% of GDP)          A         402.8    358.8
Q    NZ             P            B (Banks, domestic) M         770 (% of GDP)          A         132.3    132.8
Q    SE             P            A (All sectors)     M         USD (Billions)          A         1463.297 1657.085

Read left to right, each row is a fully specified promise: quarterly frequency, a borrower country, a borrowing sector, a lending sector, a valuation basis, a unit and a break-adjustment flag - then one value per quarter. Luxembourg non-financial-corporation credit from all lenders fell 44 points of GDP in eight quarters, 402.8 down to 358.8. New Zealand's bank lending to its private non-financial sector barely moved, 132.3 to 132.8. Sweden's private-sector credit in US dollar terms climbed from $1,463.3 billion to $1,657.1 billion, a 13 percent rise that mixes borrowing growth with currency translation.

And every row keeps running left: 344 dated columns back to 1940-Q2, so the same cell that answers a 2025 question also carries the 1950s.

Eight documented fields define every series, verified during our August 2026 research pass against the published files rather than marketing copy. The design point is that identity lives in dimensions, not column names: nothing about Q.LU.N.A.M.770.A is hard-coded, so a new economy or a new sector cut extends the grid instead of reshaping it. Frequency comes first in the key, the unit travels as a three-part code (measure, multiplier, descriptor), and the break-adjustment flag tells you whether a series has been spliced across statistical breaks - the difference between a clean long-run read and a false step.

The dictionary below is the spine of the credit panel; the derivatives, debt-securities and exchange-rate dataflows carry their own dimension sets, shipped as additional fields on request alongside your sample.

Three chips, honestly drawn:

• Geography - more than 40 economies plus G20, advanced-economy and emerging-market aggregates, with a March 2026 reclassification moving Czechia, Hong Kong SAR, Israel, Korea and Singapore into the advanced group. Country grain throughout; no sub-national cut exists anywhere in the suite.
• Temporal - the total credit panel holds 344 quarterly periods from 1940-Q2 through 2025-Q4. Private non-financial sector series average more than 45 years of history, some starting in the 1940s and 1950s; government-sector series average about 20 years.
• Granularity - one series per country per borrowing sector per lending sector per valuation per unit. The derivatives families differ by construction: turnover arrives from a triennial survey while outstanding positions run as quarterly aggregates.

Scale check: 1,133 series by 344 quarters in the core credit file alone, against 27-plus dataflows across the wider portal - the benchmark aggregates against which exchange volumes get measured.

API, files, or your warehouse. Daily, weekly, or hourly.

The tables land however your stack wants them - pushed to storage, served over an endpoint, or synced straight into your database, as CSV, JSON, XML or SDMX-ML, in the wide one-column-per-quarter layout or the long one-observation-per-row shape. Name the economies, the sector cuts, the units and the cadence when you request the sample; the sample ships first either way, with the matching dictionary attached.

A record this deep earns its keep in specific jobs:

• Track credit-to-GDP cycles across economies - one definition of core debt applied consistently from 1940-Q2 onward means a leverage cycle reads as one curve, not a splice of national conventions. See data scientists use cases.
• Size derivative exposures - OTC turnover from the triennial survey and quarterly outstanding aggregates, with the exchange-traded family alongside, give the positioning picture the price tapes cannot show.
• Read currencies structurally - effective exchange-rate indices beside bilateral US dollar rates separate competitiveness moves from dollar-cycle noise.
• Feed macro and quant models - 85 years of continuous quarterly history needs no backcast stitching, which is why long-sample studies keep landing here. See investors and quants use cases.
• Frame country risk and financial depth - credit, debt-securities and exchange-rate tables across 40-plus economies give client decks a common yardstick for market sizing.
• Publish with archival continuity - teach or cite a series that predates most national statistical programs and keeps its identifiers.

Five of Datadory's eight personas tag this record, ranked by relevance:

• Data scientists & ML engineers (relevance 3/3) - quarterly credit, debt-securities, derivatives and FX panels that pivot into features without schema surgery; the work is in unit discipline, not cleanup.
• Investors & quant researchers (3/3) - global leverage and derivatives exposure on private-sector credit histories averaging 45-plus years, the long sample backtests starve for elsewhere.
• Journalists, academics & students (3/3) - citable official statistics with continuity back to 1940-Q2 and methodology documents behind every series.
• Market researchers & consultants (2/3) - country-risk and financial-depth sections framed from one internally consistent source rather than a collage of national sites.
• Developers & data-product builders (2/3) - a dimensional grid that extends cleanly into products, accepting an SDMX learning curve steeper than a plain JSON feed. See developers and builders use cases.

Four choices shape every extract, so make them upfront. First, unit: percent-of-GDP, US dollars or domestic currency - and for cross-country comparisons, whether market or PPP-weighted exchange rates apply. Mixing them mid-panel displaces every figure. Second, valuation basis: market value versus nominal face value, which diverge exactly when the price of credit is doing something interesting. Third, break adjustment: adjusted series splice statistical breaks for continuous history, unadjusted series preserve the raw prints, and a pipeline that grabs whichever loads faster gets both. Fourth, sector cut: borrowing sector, lending sector, or both crossed - the household-versus-corporate split and the bank-versus-bond-finance split are different questions wearing the same file.

One honest note from our August 2026 research pass: the credit panel's dictionary is verified end to end; each derivatives and exchange-rate dataflow carries its own dimension set, which we confirm against your requested cut before anything ships.

Notes worth reading next:

• Financial exchanges data hub - the pooled view of the industry slice, eleven primary datasets deep, from tick vendors to official statisticians.
• Economic Census NAICS 424 vs BIS Derivatives Data - American wholesale structure against global credit and derivatives aggregates; the scored trade-offs, compared honestly.
• Massive (formerly Polygon.io) Stock Market API - the nanosecond tape these macro aggregates are measured against; pair them when exposure needs both views.
• Tiingo Financial Markets API - corporate-action-adjusted end-of-day prices back to 1962 when the job is security-level history.
• Nasdaq Data Link (formerly Quandl) - Sharadar, Mergent and Zacks tables when fundamentals outgrow one vendor's statement set.
• Bank for International Settlements (BIS) source profile - who publishes the suite, how the record is cataloged, and what else ships from the same source.
• Best financial exchanges datasets - where this record ranks within the slice and what beats it for other jobs.
• SDMX API and OTC derivatives - the two terms this page leans on hardest, unpacked.