Consolidated Banking Data (CBD2), delivered by Datadory

Consolidated Banking Data (CBD2) is the Eurosystem's statistical view of EU banking groups on a consolidated basis — the numbers that treat a bank the way its own consolidated accounts do, parent group included, rather than the country-by-country slices that dominate bank statistics. The dataset carries 66,139 series across three layers (a Key-data layer of 6 headline series, a Detailed layer of 15,484, and an All-data layer of 66,139), split between 40,227 annual and 25,912 quarterly series, covering 30 reference areas: every EU member state, the euro-area and EU aggregates, and the United Kingdom.

The concept dimensions read like a macroprudential syllabus: credit and credit exposure, debt, balance sheet, Finrep, P&L, capital, asset quality, funding, liquidity, Corep, profitability, interconnectedness, commercial real estate exposures, lending, leverage and banking structure. Series are keyed by an 18-part dimension string — CBD2.Q.U2.W0.11._Z._Z.A.A.I2003._Z._Z._Z._Z._Z._Z.PC is euro-area return on equity of domestic banking groups, percent units — so one key pins down frequency, reference area, counterpart area, reporting sector, framework, item and unit at once.

Datadory delivers diversified banks data covering this dataset as clean, analysis-ready extracts: the dimension string parsed into typed columns, labels resolved against the concept definitions, and the whole grid shaped for your warehouse instead of your patience.

Three observations exactly as they land in an extract — the same euro-area return-on-equity series for domestic banking groups, three consecutive quarters:

```text key | freq | ref_area | count_area | cb_rep_sector | cb_item | unit_measure | time_period | obs_value | obs_status CBD2.Q.U2.W0.11._Z._Z.A.A.I2003._Z._Z._Z._Z._Z._Z.PC | Q | U2 | W0 | 11 | I2003 | PC | 2024-Q1 | 2.3735 | E CBD2.Q.U2.W0.11._Z._Z.A.A.I2003._Z._Z._Z._Z._Z._Z.PC | - | - | - | - | - | - | 2024-Q2 | 4.8864 | - CBD2.Q.U2.W0.11._Z._Z.A.A.I2003._Z._Z._Z._Z._Z._Z.PC | - | - | - | - | - | - | 2024-Q3 | 7.3301 | -

Every column above is one of the parsed dimensions or the observation itself: freq splits quarterly from annual, ref_area is U2 for the euro area aggregate (DE, FR, ES... each get their own), cb_rep_sector 11 means domestic banking groups and stand-alone banks, cb_item I2003 is return on equity in percent, and obs_status E flags an estimated value. The full untruncated series key rides along with each row, so nothing about provenance is lost by flattening.

Ask for any other item — capital ratios, loan loss rates, CRE exposures — and your sample arrives cut to those concepts instead.

Nine fields anchor every extract we cut from CBD2. They carry the full spine of each observation — when, where, which kind of banking group, which reporting framework, which indicator and the value itself — which is enough to pivot straight into equity analysis, funding studies or capital comparisons without touching the raw dimension grammar. The example column shows real values from a verified euro-area return-on-equity cut, so you can see how types land before anything ships.

Additional fields on request: the observation-quality flag (OBS_STATUS), the per-observation footnote (COMMENT_OBS), plus the deeper sector selectors behind the 18-part series key — counterparty breakdowns and the balance-sheet counting sector. Tell us the analysis you are running and we map those onto your sample rather than shipping columns nobody reads.

Three chips summarize the footprint:

• Geography: 30 reference areas — all EU member states, the euro area aggregate (U2), the EU aggregate (B0) and the UK — with a separate counterpart-area axis (COUNT_AREA) that distinguishes domestic exposures from worldwide ones.
• Time: quarterly series observed back to Q4 2007, annual series reaching further, with new quarters arriving on the ECB's statistical calendar. Long-run users should note the reporting-framework evolution across the sample periods; we flag the seams in any multi-year extract.
• Granularity: per reference area per period, split by reporting sector (domestic banking groups and stand-alone banks, branches of foreign institutions, subsidiaries), by reporting framework, and by consolidated-banking-data item — 66,139 series in the All-data layer, 40,227 annual and 25,912 quarterly.

That breadth is why CBD2 anchors European bank analysis: it is the only place where profitability, capital, funding, asset quality and commercial real estate exposure line up under one consolidated definition across the whole Union.

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

You pick the channel and the cadence; parsing SDMX keys, tracking revisions and holding the schema steady are our problem. When the source publishes a new quarter or restates a vintage, the feed you consume stays normalized — same columns, same types, same join keys. Bulk pulls land as files, continuous consumption runs through the API, and warehouse-native loads write straight into your own storage. A sample ships first either way, sized to test in your pipelines the same day.

Ranked by how directly the consolidated view answers their day job:

1. Investors & quants. Pan-European bank equity and credit analysis needs group-level numbers, not branch-level ones — consolidated profitability (the verified RoE run shows 2.37% for euro-area domestic groups in Q1 2024, 4.89% in Q2, 7.33% in Q3), capital and leverage items keyed identically across 30 areas. Deeper material on the investors & quants page.
2. Competitive intelligence & product teams. Benchmark a target bank's balance-sheet growth, funding mix and asset quality against the sector aggregate for its home market, without reconciling eleven national disclosure formats first. More in the competitive intelligence hub.
3. Data scientists & ML engineers. Two decades of quarterly panel data with a documented dimension grammar is rare training stock for stress-testing proxies, early-warning models and clustering of European banking systems. See the data scientists hub.
4. Market researchers & consultants. Market-structure chapters — banking structure, interconnectedness, lending — quantify consolidation and concentration for entry and sizing work.
5. Journalists & academics. One citable series per claim: euro-area capital ratios, EU-wide CRE exposure growth, national profitability gaps — attributed to the Eurosystem's consolidated statistics rather than a press release.

For contrast inside the same industry: BIS Locational Banking Statistics tracks cross-border positions residence-by-residence, while the World Bank Global Financial Development Database works at country-year breadth across 200+ economies — neither gives you consolidated group financials at EU-member depth.

Cards worth reading next: the neighboring datasets in this industry, the comparison that sets CBD2's consolidated group view against the World Bank's country-year breadth, and the primers that decode the standards behind the rows.