API reference
The API is organized around five ideas: dimensions you filter on, metrics (computed ratios and amounts), asset quality (credit-quality series addressed by what they measure), data points (the raw Pillar 3 disclosures metrics are built from), and earnings (P&L, balance sheet, KPI and key-metrics figures from banks' own earnings releases). Every endpoint below returns JSON and sits under /v1/. All endpoints require an API key.
Every table below lists every field the endpoint accepts or returns, nothing is abbreviated.
- Dimensions, including banks and periods
- Metrics, including search, metrics metadata and metrics
- League tables: one ready-made table per topic, built for Excel
- Asset quality, including asset quality metadata and asset quality
- Data points, including search, tables and p3dh data
- Earnings, including items and data
- Currencies: every amount in EUR by default, or in a currency you choose
- Errors
Key dimensions
A metric or a data point is always the intersection of a bank, a reference period, and (for data points) a template cell. These two endpoints let you discover what values those dimensions can take before you filter on them.
Banks
GET /v1/banks/
Every reporting institution in the dataset, one row each, regardless of how many periods or templates it has submitted. That includes banks that took part in the EBA Transparency Exercise but have never filed Pillar 3 data with the EBA (record_source: "EBA_TE"), typically because they have since merged or left the exercise. Only institutions with a complete identity record are returned. A row missing its name, country, LEI, entity type, or institution type is excluded automatically, rather than returned with nulls. The one exception is the entity type and institution type of EBA_TE banks: the Transparency Exercise doesn't publish either, so both are null for them.
Request parameters: all optional, combine freely.
| Parameter | Match | Description |
|---|---|---|
entity_name | partial, case-insensitive | Institution name, e.g. entity_name=deutsche. |
entity_lei | exact | Legal Entity Identifier, .CON form (see below). |
slug | exact, ignores case | The bank's slug, e.g. slug=ERSTE_sub_croatia. Every endpoint that takes entity_lei also takes slug. |
bank | exact, ignores case, accents and punctuation | Any name the bank is known by: its slug, abbreviation, short_name or one of its aliases, e.g. bank=socgen or bank=HSH Nordbank. An alias can match more than one bank. |
entity_type | exact | One of Banking Group, CRD Credit Institution, PILLAR3 Sub Group. |
institution_type | exact | One of Large highest EEA, Other highest EEA, Large subsidiaries. |
country | exact | Full country name, e.g. Germany, not an ISO code. |
size | exact | One of very_small, small, medium, large, very_large (see size below). |
min_exposure_bn / max_exposure_bn | numeric range | Filter on leverage_exposure_eur_bn, in EUR billions. min_exposure_bn=100 returns institutions with at least €100bn; combine both for a range. |
systemic_relevance | exact | One of Not systemic, O-SII, G-SIB. |
ssm | exact | One of SI, LSI, False (the literal string, not a boolean; see ssm below). |
stoxx | true/false | STOXX Europe 600 Banks index membership. |
record_source | exact | P3DH (has filed Pillar 3 data) or EBA_TE (known only from the EBA Transparency Exercise). See record_source below. |
short_name, abbreviation, slug and aliases are null ([] for aliases) for a bank that hasn't been given them yet, typically one that only just started filing. The banks dictionary lists every bank with its abbreviation and slug.
Every filter based on the six enrichment fields above (size, min_exposure_bn/max_exposure_bn, systemic_relevance, ssm, stoxx) only ever matches institutions those fields have been computed for; see below.
Response fields: one object per institution.
| Field | Type | Description |
|---|---|---|
entity_lei | string | Legal Entity Identifier, with a .CON suffix for consolidated group entities. An external LEI registry returns the bare 20-character LEI without this suffix; always use the .CON form when querying this API. |
short_name | string or null | The name the bank is known by, e.g. BNP Paribas Fortis, Bank Austria. |
abbreviation | string or null | Its shorthand: the exchange ticker where the market uses it as the bank's name (BNPP, UCG, DBK), otherwise the brand (BAWAG, ERSTE, SOCGEN). Returned, with short_name and slug, next to entity_lei by /v1/p3dh-data/, /v1/metrics/, /v1/asset-quality/, /v1/earnings/, the search endpoints and the league tables. |
slug | string or null | A unique, stable key for the bank; filter on it with slug on any endpoint that takes entity_lei. A group or standalone bank uses its abbreviation (BNPP); a subsidiary reads <GROUP>_sub_<qualifier> (BNPP_sub_fortis, ERSTE_sub_croatia). Returned wherever abbreviation is. |
aliases | list of strings | Other names that identify the bank: tickers, former names, common acronyms (["SG", "GLE"] for Société Générale). Not unique across banks. |
entity_name | string | Institution's reporting name. |
entity_type | string or null | Banking Group, CRD Credit Institution, or PILLAR3 Sub Group. null for EBA_TE banks. |
institution_type | string or null | Large highest EEA, Other highest EEA, or Large subsidiaries. null for EBA_TE banks. |
country | string | Country of establishment, full name. |
record_source | string | P3DH if the institution has filed Pillar 3 data with the EBA, otherwise EBA_TE: known only from the EBA Transparency Exercise, with history in /v1/metrics/ and /v1/asset-quality/ but no cells in /v1/p3dh-data/. |
last_period | string (YYYY-MM-DD) | The institution's latest reference period in any source. For an EBA_TE bank, the last exercise it took part in. |
leverage_exposure_eur_bn | number or null | The LR2 leverage ratio's total exposure measure, converted to EUR and expressed in EUR billions (rounded to 2 decimals). Sourced from leverage_exposure_ref_period: 2025-12-31 wherever available (the best-covered P3DH period so far), otherwise the nearest period the institution actually reported one for. For an EBA_TE bank, the leverage exposure from its last Transparency Exercise period. |
leverage_exposure_currency | string or null | The currency the exposure was originally reported in, before conversion. |
leverage_exposure_ref_period | string (YYYY-MM-DD) or null | Which reference period leverage_exposure_eur_bn was sourced from. |
size | string or null | A baseline size classification derived from leverage_exposure_eur_bn: very_small (\< €10bn), small (€10-30bn), medium (€30-100bn), large (€100-500bn), or very_large (€500bn+). |
systemic_relevance | string or null | G-SIB (global systemically important bank, FSB list), O-SII (other systemically important institution, EBA/ESRB list), or Not systemic. |
ssm | string or null | Status under the ECB's Single Supervisory Mechanism: SI (significant institution, directly ECB-supervised), LSI (less-significant institution, NCA-supervised under the SSM), or the *string* "False" (not part of the SSM, outside the euro area). |
stoxx | boolean or null | Whether the institution is a STOXX Europe 600 Banks index constituent. |
leverage_exposure_eur_bn/_currency/_ref_period, size, systemic_relevance, ssm, and stoxx are computed separately from the raw P3DH disclosures, matched against the FSB's G-SIB list, EBA/ESRB's O-SII list, the ECB's SSM supervised-entities list, and the STOXX Europe 600 Banks index. All seven are null for a PILLAR3 Sub Group entity, which this isn't computed for at all; leverage_exposure_eur_bn/_currency/_ref_period/size can additionally be null for a consolidated group that hasn't reported a leverage exposure value in any period, even though its systemic_relevance/ssm/stoxx are populated.
systemic_relevance, ssm, and stoxx describe an institution's status today, matched by LEI against the current lists. An EBA_TE bank that has since been absorbed into another group therefore shows Not systemic, "False", and false, even if it was systemic while it took part in the exercise.
The coverage summary (Coverage) and the league tables only count institutions that file Pillar 3 data, so EBA_TE banks don't appear there.
Use this endpoint to discover a bank's entity_lei: the stable key you'll pass to /v1/metrics/ or /v1/p3dh-data/. The banks dictionary is the same list, searchable, without a request.
Periods
GET /v1/periods/
Every reference period (reporting date) present in the dataset, as YYYY-MM-DD.
Request parameters: all optional, combine freely.
| Parameter | Match | Description |
|---|---|---|
entity_name | partial, case-insensitive | Restrict to periods in which this institution has submitted data. |
table_code | exact | Restrict to periods covered by this template. Spaces are optional: K61.00 and K 61.00 resolve to the same table. |
table_name | partial, case-insensitive | Same as table_code, matched by template name instead. |
Response fields: one object per distinct period.
| Field | Type | Description |
|---|---|---|
ref_period | string (YYYY-MM-DD) | End date of the reporting period, e.g. 2025-12-31 for year-end 2025. |
This endpoint lists the periods that have Pillar 3 data points, so it is the source of truth for how far back /v1/p3dh-data/ goes. /v1/metrics/ and /v1/asset-quality/ reach further back, to the EBA Transparency Exercise: see History back to 2012 and Asset quality metadata. Your own key's entitlements may further restrict how far back each endpoint will actually return.
Metrics
Computed ratios and nominal amounts derived from Pillar 3 disclosures. Think CET1 ratio, leverage ratio, NSFR, and hundreds of others, organized into categories like capital_adequacy and liquidity. Start here if you want a comparable figure across banks without reshaping raw template cells yourself.
Search
GET /v1/search/?keyword=...
Finds banks and catalog metrics that match every word in keyword, in any order. Use this when you know roughly what you're looking for ("leverage ratio") but not its exact metric_name.
Request parameters
| Parameter | Required | Description |
|---|---|---|
keyword | yes | One or more words, space-separated. Every word must appear somewhere in a record's name/label/definition/category (or bank name) for it to match. Order doesn't matter, but this isn't typo-tolerant fuzzy matching. |
Response fields: a single object with two lists.
| Field | Type | Description |
|---|---|---|
banks | array | Institutions whose name matches every keyword token. Each entry has the same fields as Banks' response. |
metrics | array | Catalog entries whose name, label, definition, or category matches every token. |
metrics[].metric_name | string | Machine-readable identifier, e.g. cet1_ratio. Pass to /v1/metrics/'s metric_name filter. |
metrics[].category | string | Metric category, e.g. capital_adequacy. |
metrics[].label | string | Human-readable name. |
metrics[].definition | string | The compiled formula: which DPM cells feed this metric and how they combine (see Metrics metadata). |
metrics[].unit_kind | string | ratio, currency, years, or boolean. |
Metrics metadata
GET /v1/metadata-metrics/
The full metric catalog: names, labels, formulas, and which transformations each metric supports. This is a static, checked-in catalog, not a live query: cheap to call repeatedly. The metrics dictionary is the same catalog, searchable, grouped by category.
Request parameters
| Parameter | Match | Description |
|---|---|---|
category | exact, optional | Restrict to one metric group, e.g. capital_adequacy. Omit to return the whole catalog. |
Response fields: one object per metric.
| Field | Type | Description |
|---|---|---|
metric_name | string | Machine-readable identifier, e.g. cet1_ratio. Pass to /v1/metrics/'s metric_name filter. |
category | string | Metric group, e.g. capital_adequacy. Pass to /v1/metrics/'s category filter. |
label | string | Human-readable name. |
definition | string | The compiled formula: which DPM cells feed this metric and how they combine, given once as codes (e.g. [K 91.00.a & c0010 & r0260] / [...]) and once with the matching human-readable labels. Generated straight from the metric's own definition, so it can't drift from what /v1/metrics/ actually computes. For a metric built from earnings releases it names the earnings lines instead, e.g. [pnl: Total income, year-to-date], with annualised(...), average(...) (mean of the quarter-ends since the prior year-end) and prior_year(...). |
unit_kind | string | ratio, currency, years, or boolean; for earnings metrics also bps, currency_per_share (converted like currency), or count. |
available_transformations | array of strings | Which transformation values /v1/metrics/ accepts for this metric. Always ["none"] for boolean metrics; ["none", "delta_qoq", "delta_yoy", "growth_qoq", "growth_yoy"] for everything else. |
source_kind | string | Where the inputs come from: p3dh (the Pillar 3 filing), earnings (the bank's earnings release), or mixed (earnings figures over Pillar 3 RWA or CET1 capital). See Metrics from earnings releases. |
bank_defined | boolean | true for a bank's own KPI (reported_*), calculated the bank's own way rather than by one formula. |
See Getting started with the metrics endpoint for the full list of categories that currently exist.
Metrics
GET /v1/metrics/
The values themselves: one record per metric, bank, period, and transformation.
Request parameters: all optional, combine as a single query.
| Parameter | Match | Description |
|---|---|---|
category | exact | Metric group, e.g. capital_adequacy. |
metric_name | exact, repeatable | One or more machine-readable metric names, from Metrics metadata. |
metric_label | partial, multi-word | Matches against the catalog's label, e.g. metric_label=leverage+ratio. |
entity_lei | exact, repeatable | One or more banks, from Banks. |
entity_name | partial, case-insensitive | Institution name. |
entity_type | exact | Banking Group, CRD Credit Institution, or PILLAR3 Sub Group. |
institution_type | exact | Large highest EEA, Other highest EEA, or Large subsidiaries. |
country | exact, repeatable | Full country name(s), e.g. country=France&country=Germany. |
ref_period | date (YYYY-MM-DD), repeatable | One or more reference periods, from Periods. |
transformation | one of five values | none (default), delta_qoq, delta_yoy, growth_qoq, growth_yoy. An unrecognized value returns 422. |
include_null_values | true/false | Default false. Rows where the metric couldn't be computed for that bank/period are excluded unless set to true. |
only_latest | true/false | Default true. Collapses restatements: for each (entity_lei, ref_period, metric_name, transformation), only the row from the most-recently-submitted module is returned. Where sources overlap, a Pillar 3 filing wins over the EBA Transparency Exercise, which wins over an earnings release. Set false to get every submission's row back instead. |
source | exact | P3DH (Pillar 3 filings), EBA_TE (EBA Transparency Exercise history), or EARNINGS (banks' earnings releases). Omit for all. See History back to 2012 and Metrics from earnings releases. |
currency | one of 13 codes | Default EUR. The currency amount metrics are returned in. See Currencies for the codes and how conversion works. Any other value returns 422. |
limit | integer, ≥ 1 | Default 1000, silently clamped to your key's per-request maximum. |
offset | integer, ≥ 0 | Number of rows to skip, not a page number. Default 0. |
Response fields: one object per row.
| Field | Type | Description |
|---|---|---|
entity_lei | string | Bank's LEI, .CON form. |
entity_name | string | Institution name. |
entity_type | string | Banking Group, CRD Credit Institution, or PILLAR3 Sub Group. |
institution_type | string | Large highest EEA, Other highest EEA, or Large subsidiaries. |
country | string | Full country name. |
ref_period | string (YYYY-MM-DD) | Reporting period this value applies to. |
category | string | Metric group, e.g. capital_adequacy. |
metric_name | string | Machine-readable metric identifier. |
metric_label | string | Human-readable name. |
definition | string | The compiled formula behind this metric (see Metrics metadata). |
transformation | string | Which of the five kinds this row's value represents. |
value | number or null | The computed value, rounded to 4 decimal places. Amount metrics are in currency. Excluded (row absent) when the metric couldn't be computed, unless include_null_values=true. Also null when no exchange rate is available for this period; reported_value still holds the figure then. |
unit | string | ratio, years, boolean, or for amount metrics the currency of value (your currency). |
reported_value | number or null | The value in the bank's own reporting currency, before conversion. Equal to value when no conversion was needed. |
reported_unit | string | The unit of reported_value: the bank's reporting currency for amount metrics. |
fx_rate | number or null | The exchange rate applied to this period's value, so value = reported_value × fx_rate for plain values. 1.0 when already in currency. null for ratio, years and boolean metrics, and when no rate is available. |
fx_rate_date | string (YYYY-MM-DD) or null | Publication date of the ECB rate used. null unless a rate was applied. |
source | string | P3DH if the value was computed from the bank's Pillar 3 filing, EBA_TE if it comes from the EBA Transparency Exercise, EARNINGS if it comes from the bank's earnings release. |
quality | string | How closely the value matches the metric's definition: exact for every P3DH row; exact, derived, or approx for EBA_TE rows (see below); exact, partial, or bank_defined for EARNINGS rows. |
quality_note | string or null | For derived, approx and partial rows, what was done and which inputs it was built from; for EARNINGS rows also how the value was assembled (e.g. year-to-date summed from quarters). null otherwise. |
Response headers
| Header | Description |
|---|---|
X-Result-Count | Records in this response. |
X-Has-More | true if additional rows exist beyond this page. |
X-Limit | Effective limit applied (may be lower than requested). |
X-Offset | Offset applied. |
X-Plan-Capped | true when your key's per-request cap, not your own limit, determined the page size. |
History back to 2012
Pillar 3 filings to the EBA start at 2025-06-30. Values for earlier periods, back to 2012, come from the EBA Transparency Exercise (source: "EBA_TE"). They are the same catalog metrics, built from the exercise's figures mapped onto the same template cells, with amounts in EUR. The two sources overlap at 2025-06-30. There, with only_latest=true, the filed value is returned, and a Transparency Exercise value only fills a metric the filing doesn't carry. Quarter-on-quarter and year-on-year transformations bridge the two sources, so a series runs continuously from the exercise into the filings.
quality tells you how each Transparency Exercise value was obtained:
exact: the exercise reports the same concept the metric uses.derived: computed from exercise figures, for example performing exposures as total minus non-performing.approx: close but not identical in scope.
quality_note gives the detail for the last two. Filter with source=P3DH to work with filed values only.
Your key's entitlements may limit how far back this endpoint returns, in which case the Transparency Exercise history is cut off with everything else.
Metrics from earnings releases
Banks publish their earnings release weeks before their Pillar 3 filing reaches the EBA. Rows with source: "EARNINGS" come from those releases, the same figures /v1/earnings/ serves, in three kinds:
- Computed metrics in the
profitability,balance_sheetandshareholder_valuecategories: RoTE, RoE, net interest margin, cost/income, jaws, cost of risk, loan-to-deposit, organic capital generation, tangible book value per share, and more. Each is one formula applied to every bank alike, on year-to-date figures, annualised where a flow is divided by a balance, with balances averaged over the quarter-ends since the prior year-end. A missing input leaves the value out; it is never filled with zero. - The banks' own KPIs in
reported_kpis, e.g.reported_rotenext torote, withquality: "bank_defined". Each bank defines these its own way, so the gap between a computed metric and the bank's own figure is often worth a look. - Capital and liquidity figures from the release (CET1 ratio, RWA, leverage ratio, LCR, NSFR, MREL), stored under the Pillar 3 metric's own name, e.g.
cet1_ratio. Only figures that match the metric's definition exactly are used. They fill the latest quarters until the Pillar 3 filing arrives; from then on, withonly_latest=true, the filed value is returned for that quarter. Withonly_latest=falseyou get both.
Filter with source=P3DH to work with filed values only.
transformation defaults to none. Pass delta_qoq, delta_yoy, growth_qoq, or growth_yoy to get a change instead of the plain value. A delta_* value is in the same unit as the base metric; a growth_* value is always a dimensionless ratio. Restatements are deduplicated by default (only_latest=true, see above); set only_latest=false to see every submission's row. See Getting started with the metrics endpoint for a full walkthrough.
League tables
GET /v1/league_table/{table}/
A league table is a ready-made peer comparison: one row per bank, one column per metric, for a fixed set of metrics on one topic. /v1/metrics/ returns one row per metric, which you then pivot; a league table arrives already wide and complete in a single response. That makes it the quickest way to get a peer table into Excel, where one Power Query call loads it straight into a sheet (see Getting started with Excel and Power Query).
There are six tables. Each is a fixed list of metrics from Metrics metadata, and the column names are the metric names:
| Table | Path | Columns |
|---|---|---|
| Key metrics (EU KM1) | /v1/league_table/key_metrics/ | cet1_amount, total_capital_amount, rwa_total, cet1_ratio, t1_ratio, total_capital_ratio, cet1_ratio_unfloored, p1_p2r_requirement, combined_buffer_requirement, overall_capital_requirement, total_srep_cet1_requirement, headroom_cet1, leverage_exposure, leverage_ratio, hqla_total, net_cash_outflows, lcr, nsfr_asf_total, nsfr_rsf_total, nsfr |
| Capital | /v1/league_table/capital/ | cet1_ratio, t1_ratio, total_capital_ratio, cet1_ratio_unfloored, t1_ratio_unfloored, total_capital_ratio_unfloored, cet1_ratio_floor_drag, cet1_amount, at1_amount, t2_amount, total_capital_amount, total_srep_cet1_requirement, total_srep_t1_requirement, p1_p2r_requirement, overall_capital_requirement, headroom_cet1, headroom_t1, headroom_total_capital, leverage_ratio, leverage_ratio_ex_temp_exemptions, leverage_ratio_headroom, leverage_exposure, rwa_total, rwa_total_prefloor, output_floor_binding_pct, asset_density, credit_rwa_share, ccr_rwa_share, oprisk_rwa_share, irb_share_credit_rwa, sa_share_credit_rwa |
| Liquidity, funding and asset encumbrance | /v1/league_table/liquidity/ | lcr, hqla_total, net_cash_outflows, nsfr, nsfr_asf_total, nsfr_rsf_total, hqla_to_wholesale_funding, derivative_outflow_share, retail_deposit_outflow_base, wholesale_funding_outflow_base, wholesale_funding_share, stable_deposit_share, operational_deposit_share, runnable_deposit_share, encumbrance_ratio, hqla_encumbrance_ratio, unencumbered_hqla_carrying_amount, covered_bond_asset_encumbrance |
| Capital and MREL requirements | /v1/league_table/capital_requirements/ | p1_p2r_requirement, total_srep_cet1_requirement, total_srep_t1_requirement, capital_conservation_buffer_req, countercyclical_buffer_req, gsii_osii_buffer_req, systemic_risk_buffer_req, combined_buffer_requirement, overall_capital_requirement, mrel_requirement_rwa, mrel_requirement_lre, mrel_rwa, mrel_lre, mrel_headroom_rwa_excl_cbr, mrel_headroom_rwa_incl_cbr, mrel_headroom_lre_excl_cbr, sub_mrel_rwa, sub_mrel_lre |
| Asset quality | /v1/league_table/asset_quality/ | npl_ratio, npl_coverage_ratio, stage2_ratio, stage2_coverage_ratio, forbearance_ratio, forborne_npl_share, npl_inflow_rate, npl_outflow_rate, npl_net_formation_rate, npl_writeoff_rate, past_due_90_180_share, past_due_over_2y_share, households_npl_ratio, nfc_npl_ratio, nfc_sme_npl_ratio, households_coverage_ratio, nfc_coverage_ratio, cre_exposure_share |
| ESG (climate) | /v1/league_table/ESG/ | carbon_intensive_exposure_share, financed_emissions_intensity_total, financed_emissions_intensity_power_generation, financed_emissions_intensity_shipping, financed_emissions_scope3_share, emissions_data_company_specific_coverage, npe_share_carbon_intensive, stage2_share_carbon_intensive, avg_weighted_maturity_carbon_intensive, sector_concentration_top1_carbon_intensive, physical_risk_exposure_share, mortgage_epc_a_b_share, mortgage_epc_f_g_share, mortgage_epc_unlabelled_share, cre_epc_a_b_share |
The asset quality table leaves out the per-industry breakdowns, which would add dozens of columns; get those from Metrics. The tables will grow as new metrics are added. A breakdown of the countercyclical buffer by country, for example, is not yet available.
Choosing a period
period decides which values you get, and each of its three forms answers a different question:
period | Rows | Use it when |
|---|---|---|
latest (default) | One per bank. Each cell holds that metric's latest available value. | You want to know where every bank stands today. |
A quarter-end, e.g. 2026-06-30 | One per bank, every value from exactly that quarter. | You need a like-for-like ranking at one reporting date. |
A year, e.g. 2026, or a range, e.g. 2025-2026 | One per bank *per quarter* within those years. | You want trends, or to build a pivot table or a chart over time. |
With latest, a row can combine periods. Banks disclose some metrics only semi-annually or annually, so a bank's latest NSFR or asset encumbrance can be older than its latest CET1 ratio. The ref_period_min and ref_period_max columns show the oldest and newest period used in the row; when they are equal, the row is a single consistent snapshot. If mixed periods matter for your comparison, pass a quarter-end instead. A bank that didn't disclose a metric for that quarter then shows null for it, rather than an older value.
A date that isn't a quarter-end (03-31, 06-30, 09-30 or 12-31), or any other unrecognised value, returns 400 with a message listing the accepted forms.
Request parameters: all optional, combine freely. The filters work exactly as in Banks.
| Parameter | Match | Description |
|---|---|---|
period | see above | latest (default), a quarter-end date, a year, or a year range. |
entity_type | exact | One of Banking Group, CRD Credit Institution, PILLAR3 Sub Group. |
institution_type | exact | One of Large highest EEA, Other highest EEA, Large subsidiaries. |
country | exact | Full country name, e.g. Germany, not an ISO code. |
size | exact | One of very_small, small, medium, large, very_large. |
min_exposure_bn / max_exposure_bn | numeric range | Leverage exposure in EUR billions. |
systemic_relevance | exact | One of Not systemic, O-SII, G-SIB. |
ssm | exact | One of SI, LSI, False (the literal string). |
stoxx | true/false | STOXX Europe 600 Banks index membership. |
Response fields: one object per bank (or per bank and quarter, for a year or range). Every object carries every field, in the same order, with null where a bank has no value, so a spreadsheet sees the same columns on every row.
| Field | Type | Description |
|---|---|---|
entity_lei | string | Bank's LEI, .CON form. |
entity_name | string | Institution name. |
country | string | Full country name. |
entity_type | string | Banking Group, CRD Credit Institution, or PILLAR3 Sub Group. |
institution_type | string | Large highest EEA, Other highest EEA, or Large subsidiaries. |
ref_period_min / ref_period_max | string (YYYY-MM-DD) | period=latest only. The oldest and newest period of the values in this row. |
ref_period | string (YYYY-MM-DD) | Quarter-end, year, or range only. The period every value in this row comes from. |
currency | string | Tables with amount columns only (key metrics, capital, liquidity). The bank's reporting currency. Amounts are as reported, not converted to EUR. |
| *one field per metric* | number | The table's metrics, in the order listed above, rounded to 4 decimal places. Ratios are decimal fractions: 0.153 is 15.3%. |
Response headers
| Header | Description |
|---|---|
X-Result-Count | Rows in this response. |
No pagination, at most 200 banks. A league table always arrives whole: there is no limit or offset, because a table split across pages is unusable in a spreadsheet. To keep responses manageable, one request returns at most 200 banks. The count is taken after your filters and period are applied. If more banks match, you get 400 with a message saying how many matched, e.g. *"This request matches 347 banks; league tables return at most 200. Narrow it with country, size, systemic_relevance, …"*. Add a filter such as country, ssm=SI or min_exposure_bn and try again.
Like every endpoint, league tables respect your plan's history depth: on the free plan, values from before the start of the previous calendar year are excluded. The free plan's per-request data-point limit doesn't apply here; the 200-bank cap bounds these responses instead. Values follow the same rules as Metrics: restatements are collapsed to the latest submission, and only values computed with the current metric definition are served.
Let's say you're comparing CET1 headroom across banks directly supervised by the ECB, as of mid-2026:
GET /v1/league_table/key_metrics/?ssm=SI&period=2026-06-30[
{
"entity_lei": "R0MUWSFPU8MPRO8K5P83.CON",
"entity_name": "BNP Paribas",
"country": "France",
"entity_type": "Banking Group",
"institution_type": "Large highest EEA",
"ref_period": "2026-06-30",
"currency": "EUR",
"cet1_amount": 93512000000.0,
"cet1_ratio": 0.1263,
"headroom_cet1": 0.0237,
"lcr": 1.4,
"nsfr": null
}
]*(Illustrative values, and columns shortened for readability; the real response includes every key-metrics column.)*
In Python, the response is already a table:
import pandas as pd
import requests
rows = requests.get(
"https://app-2b3bd7f4-0899-420d-be54-0a0cfb9b0879.cleverapps.io/v1/league_table/key_metrics/",
params={"ssm": "SI", "period": "2026-06-30"},
headers={"X-API-Key": "eu_bank_data_..."},
).json()
peers = pd.DataFrame(rows).sort_values("headroom_cet1", ascending=False)In Excel, see Getting started with Excel and Power Query for a query that loads a league table and shows the 200-bank message as a readable row instead of a generic "Bad Request".
Asset quality
The credit quality of banks' exposures as long time series: performing and non-performing exposures, IFRS 9 stages, past-due buckets, forbearance, NACE sectors, and collateral. The values come from the Pillar 3 templates EU CR1, CQ1, CQ3, CQ5, and CQ6, but each value is addressed by what it measures (a set of dimensions) rather than by template row and column. A series for one bank runs from the EBA Transparency Exercise, starting with 2014, into its latest Pillar 3 filing.
Asset quality metadata
GET /v1/metadata-asset-quality/
The data dictionary for /v1/asset-quality/: every dimension with its allowed values and labels, plus how far back each source and template goes. No parameters.
Response fields: a single object with two lists.
| Field | Type | Description |
|---|---|---|
dimensions | array | One entry per dimension, in the order records list them. |
dimensions[].name | string | Dimension name, e.g. counterparty. Also the filter parameter's name on /v1/asset-quality/. |
dimensions[].values | array of {value, label} | Every allowed value, e.g. {"value": "non_financial_corporations", "label": "Non-financial corporations"}. all always means "not broken down by this dimension". |
coverage | array | One entry per source and template, reflecting the data loaded right now. |
coverage[].source | string | P3DH or EBA_TE. |
coverage[].template | string | CR1, CQ1, CQ3, CQ5, or CQ6. |
coverage[].first_period / coverage[].last_period | string (YYYY-MM-DD) | The earliest and latest reference periods this source has for the template. Transparency Exercise history for CQ5 starts in 2019 and for CQ6 in 2020; the others go back further. |
coverage[].banks | integer | Distinct institutions with values. |
coverage[].values | integer | Stored values. |
Asset quality
GET /v1/asset-quality/
The values themselves: one record per bank, period, source, and combination of dimensions.
Request parameters: all optional, combine as a single query. Every dimension filter is exact and repeatable, and accepts only the values below (anything else returns 422).
| Parameter | Match | Description |
|---|---|---|
entity_lei | exact, repeatable | One or more banks, from Banks. |
entity_name | partial, case-insensitive | Institution name. |
entity_type | exact | Banking Group, CRD Credit Institution, or PILLAR3 Sub Group. |
institution_type | exact | Large highest EEA, Other highest EEA, or Large subsidiaries. |
country | exact, repeatable | Full country name(s). |
ref_period | date (YYYY-MM-DD), repeatable | One or more reference periods. |
template | dimension | CR1 (performing and non-performing exposures and provisions), CQ1 (forborne exposures), CQ3 (by past-due days), CQ5 (loans to non-financial corporations by industry), CQ6 (collateral valuation). |
measure | dimension | gross_carrying_amount (the nominal amount for off-balance-sheet exposures and loan commitments), accumulated_impairment_and_provisions, accumulated_impairment, accumulated_negative_fv_changes, accumulated_partial_write_off, collateral_and_guarantees_received, collateral_received_capped, collateral_received_above_cap, financial_guarantees_received. |
instrument | dimension | cash_balances, loans_and_advances, debt_securities, off_balance_sheet, loan_commitments, all. |
counterparty | dimension | central_banks, general_governments, credit_institutions, other_financial_corporations, non_financial_corporations, households, all. |
performing_status | dimension | performing, non_performing, non_performing_defaulted, non_performing_impaired, all. |
stage | dimension | IFRS 9 stage: stage_1, stage_2, stage_3, all. |
past_due | dimension | not_past_due_or_le_30d, gt_30d_le_90d, utp_not_past_due_or_le_90d (unlikely to pay, not past due or up to 90 days), gt_90d, gt_90d_le_180d, gt_180d_le_1y, gt_1y_le_2y, gt_2y_le_5y, gt_5y_le_7y, gt_7y, all. |
forborne | dimension | forborne, all. |
nace_sector | dimension | NACE section A to S (CQ5 only), all. |
of_which | dimension | sme, subject_to_impairment, secured, secured_immovable_property, ltv_gt_60_le_80, ltv_gt_80_le_100, ltv_gt_100, immovable_property, all. |
source | exact | P3DH or EBA_TE. Omit for both. |
quality | exact, repeatable | exact, derived, approx. |
only_latest | true/false | Default true: where both sources have a value for the same bank, period, and template cell (currently 2025-06-30), only the P3DH value is returned. Set false to get both and compare. |
currency | one of 13 codes | Default EUR. See Currencies. |
limit | integer, ≥ 1 | Default 1000, silently clamped to your key's per-request maximum. |
offset | integer, ≥ 0 | Number of rows to skip, not a page number. Default 0. |
Response fields: one object per row.
| Field | Type | Description |
|---|---|---|
entity_lei | string | Bank's LEI, .CON form. |
entity_name | string | Institution name. |
entity_type | string or null | Banking Group, CRD Credit Institution, or PILLAR3 Sub Group. null for banks known only from the Transparency Exercise. |
institution_type | string or null | Large highest EEA, Other highest EEA, or Large subsidiaries. null for banks known only from the Transparency Exercise. |
country | string | Full country name. |
ref_period | string (YYYY-MM-DD) | Reporting period this value applies to. |
template, measure, instrument, counterparty, performing_status, stage, past_due, forborne, nace_sector, of_which | string | The dimensions, always all ten, each one of the values listed above. |
value | number or null | The amount in currency, rounded to 4 decimal places. null when no exchange rate is available for this period; reported_value still holds the figure then. |
unit | string | The currency of value (your currency). |
reported_value | number | The amount before conversion. |
reported_unit | string | The currency of reported_value: the bank's reporting currency for P3DH rows, EUR for EBA_TE rows. |
fx_rate | number or null | The exchange rate applied: value = reported_value × fx_rate. 1.0 when already in currency, null when no rate is available. |
fx_rate_date | string (YYYY-MM-DD) or null | Publication date of the ECB rate used. null unless a rate was applied. |
source | string | P3DH (the bank's own Pillar 3 filing, latest submission per cell) or EBA_TE (EBA Transparency Exercise). |
quality | string | exact for every P3DH row. For EBA_TE rows: exact (the same concept), derived (computed from exercise figures, e.g. performing as total minus non-performing), or approx (close but not identical scope, or a zero where an older exercise didn't collect the item). |
quality_note | string or null | For EBA_TE rows, how the value was matched and the exercise expression it came from. null for P3DH rows. |
table_code / row_code / col_code | string | The template cell the value belongs to. For P3DH rows, the same coordinates as in /v1/p3dh-data/. |
Response headers: the same as Metrics: X-Result-Count, X-Has-More, X-Limit, X-Offset, X-Plan-Capped.
The same concept can appear in more than one template. Performing loans to non-financial corporations, for example, are in both CR1 and CQ3. Filter on template before summing or comparing values.
Loans to non-financial corporations, performing and non-performing, for one bank across all periods:
GET /v1/asset-quality/?entity_lei=7LTWFZYICNSX8D621K86.CON&template=CR1&measure=gross_carrying_amount&instrument=loans_and_advances&counterparty=non_financial_corporations&stage=all&of_which=allNon-performing loans to non-financial corporations by NACE sector, all banks, one period:
GET /v1/asset-quality/?template=CQ5&measure=gross_carrying_amount&performing_status=non_performing&ref_period=2024-12-31Your key's entitlements may limit how far back this endpoint returns, in which case the Transparency Exercise history is cut off with everything else.
Metrics from earnings releases
Banks publish their earnings release weeks before their Pillar 3 filing reaches the EBA. Rows with source: "EARNINGS" come from those releases, the same figures /v1/earnings/ serves, in three kinds:
- Computed metrics in the
profitability,balance_sheetandshareholder_valuecategories: RoTE, RoE, net interest margin, cost/income, jaws, cost of risk, loan-to-deposit, organic capital generation, tangible book value per share, and more. Each is one formula applied to every bank alike, on year-to-date figures, annualised where a flow is divided by a balance, with balances averaged over the quarter-ends since the prior year-end. A missing input leaves the value out; it is never filled with zero. - The banks' own KPIs in
reported_kpis, e.g.reported_rotenext torote, withquality: "bank_defined". Each bank defines these its own way, so the gap between a computed metric and the bank's own figure is often worth a look. - Capital and liquidity figures from the release (CET1 ratio, RWA, leverage ratio, LCR, NSFR, MREL), stored under the Pillar 3 metric's own name, e.g.
cet1_ratio. Only figures that match the metric's definition exactly are used. They fill the latest quarters until the Pillar 3 filing arrives; from then on, withonly_latest=true, the filed value is returned for that quarter. Withonly_latest=falseyou get both.
Filter with source=P3DH to work with filed values only.
Data points
The raw cells behind every metric: one row per bank, period, template, row, and column. Use this when you want to build your own metric, run analytics on nominal amounts directly, or load the full dataset into a warehouse.
Search
GET /v1/search-p3dh/?keyword=...
Finds banks and template cells (table/row/column) whose name or label contains keyword. The raw-data counterpart to /v1/search/ above.
Request parameters
| Parameter | Required | Description |
|---|---|---|
keyword | yes | Matched case-insensitively as a substring, not split into tokens the way /v1/search/'s keyword is. |
Response fields: a single object with two lists.
| Field | Type | Description |
|---|---|---|
banks | array | Institutions whose name contains keyword. Same fields as Banks' response. |
tables | array | Distinct table/row/column combinations where the template name, row label, or column label contains keyword. Same fields as Tables' response. |
Tables
GET /v1/tables/
The schema map: one record per unique table_code/row_code/col_code combination, with human-readable labels. It's a map of the data schema before you query actual values.
Request parameters: all optional, combine freely.
| Parameter | Match | Description |
|---|---|---|
table_code | exact | Short EBA template identifier, e.g. K 61.00. Spaces optional. |
table_name | partial, case-insensitive | Human-readable template name. |
row_code | exact | DPM row identifier within the template, e.g. 0010. |
row_label | partial, case-insensitive | Plain-language row description, e.g. non-performing. |
col_code | exact | DPM column identifier, e.g. 0010. |
col_label | partial, case-insensitive | Plain-language column description. |
Response fields: one object per unique cell.
| Field | Type | Description |
|---|---|---|
table_code | string | Short EBA template identifier, e.g. K 61.00 (EU KM1: Key metrics), K 71.00 (EU LR2: Leverage ratio). |
table_name | string | Human-readable template name. |
row_code | string | DPM row identifier within the template. |
row_label | string | Plain-language row description, e.g. *"17. Liquidity coverage ratio (%)"*. |
col_code | string | DPM column identifier within the template. |
col_label | string | Plain-language column description, e.g. *"a. T"* (current period) or *"b. T-1"* (prior period). |
P3DH data
GET /v1/p3dh-data/
Individual data points: the primary raw-data endpoint. Each record is a single cell value, the intersection of one institution, one reference period, one EBA template, one row, and one column.
Request parameters: all optional, combine as a single query.
| Parameter | Match | Description |
|---|---|---|
entity_lei | exact, repeatable | One or more banks. |
entity_name | partial, case-insensitive | Institution name. |
entity_type | exact | Banking Group, CRD Credit Institution, or PILLAR3 Sub Group. |
institution_type | exact | Large highest EEA, Other highest EEA, or Large subsidiaries. |
country | exact, repeatable | Full country name(s). |
table_code | exact, repeatable | One or more templates, e.g. K61.00 (spaces optional). |
table_name | partial, case-insensitive | Template name. |
row_code | exact | DPM row identifier. |
row_label | partial, case-insensitive | Row description. |
col_code | exact | DPM column identifier. |
col_label | partial, case-insensitive | Column description. |
axis_dim | exact, case-insensitive | Axis dimension code for templates that repeat a cell once per country, sector or exposure class, e.g. qOOR (country in K 83.01.c/.d), RIO (country in K 67.01.a), GPR (country in K 45.00.a/.b), NAC (NACE sector), qEEA (IRB exposure class). Used alone, returns only rows that carry that dimension. See Filtering by axis. |
axis_member | exact, case-insensitive, repeatable | Member(s) of axis_dim to keep, e.g. FR or eba_GA:FR. Requires axis_dim (422 otherwise). Multiple values are OR'd. |
ref_period | date (YYYY-MM-DD), repeatable | One or more reference periods. |
submission_date | date (YYYY-MM-DD) | Restrict to data points from a submission filed on this exact date; useful to pin one vintage when a period has been resubmitted. |
only_latest | true/false | Default true. Collapses restatements: for each cell (entity_lei, ref_period, table_code, row_code, col_code), only the row from the most-recently-submitted module is returned. Set false to get every submission's row back instead. Ignored when submission_date is set: pinning an exact date is already more specific than "the latest one." |
currency | one of 13 codes | Default EUR. The currency monetary cells are returned in. See Currencies for the codes and how conversion works. Any other value returns 422. |
limit | integer, ≥ 1 | Default 1000, silently clamped to your key's per-request maximum. |
offset | integer, ≥ 0 | Number of rows to skip, not a page number. Default 0. |
Response fields: one object per data point.
| Field | Type | Description |
|---|---|---|
dp_id | string | Opaque unique identifier. Safe for equality comparison and deduplication only, not meaningful on its own. |
entity_lei | string | Bank's LEI, .CON form. Use this as the join key across endpoints. |
entity_name | string | Institution name. |
entity_type | string | Banking Group, CRD Credit Institution, or PILLAR3 Sub Group. |
institution_type | string | Large highest EEA, Other highest EEA, or Large subsidiaries. |
country | string | Full country name. |
ref_period | string (YYYY-MM-DD) | End date of the reporting period. |
submission_date | string (YYYY-MM-DD) or null | Date this submission was filed. |
base_currency | string | The currency the bank reports in, e.g. SEK, whatever currency you asked for. |
table_code | string | EBA template identifier. |
table_name | string | Human-readable template name. |
row_code | string | DPM row identifier. |
row_label | string | Plain-language row description. |
col_code | string | DPM column identifier. |
col_label | string | Plain-language column description. |
value | number or null | Numerical value, rounded to 4 decimal places (the stored value keeps full precision; only the response is rounded). Monetary cells are in currency; every other cell is as reported. null for text or empty cells, and for a monetary cell when no exchange rate is available for its period (reported_value still holds it). |
string_value | string or null | Text value. null for numerical cells. |
value_type | string | "numeric" or "string": check this before reading value or string_value; exactly one of the two is populated. |
data_type | string or null | The cell's type in the EBA Data Point Model: monetary, percentage, integer, decimal, boolean, date, enumeration, or string. Only monetary cells are converted. null in the rare case the cell isn't in any DPM release we have loaded; it is then returned as stored, unconverted. |
unit | string or null | For monetary cells, the currency of value (your currency). null for other cells. |
reported_value | number or null | The value exactly as filed, before conversion. Equal to value when no conversion was needed. |
reported_unit | string or null | For monetary cells, the currency the bank filed in (base_currency). null for other cells. |
fx_rate | number or null | The exchange rate applied: value = reported_value × fx_rate. 1.0 when the cell was already in currency. null for non-monetary cells, and when no rate is available. |
fx_rate_date | string (YYYY-MM-DD) or null | Publication date of the ECB rate used. null unless a rate was applied. |
axis_value | string or null | Raw extra-axis marker, only on templates that repeat the same row/column once per axis member (row_code *). One or more DIM=member parts separated by a vertical bar, e.g. qOOR=eba_GA:FR (country), or two parts for country and NACE sector (GPR=eba_GA:DE and NAC=eba_NC:I). Not resolved to a label. null on all other rows. Filter with axis_dim/axis_member. |
Response headers
| Header | Description |
|---|---|
X-Result-Count | Records in this response. |
X-Has-More | true if additional rows exist beyond this page. |
X-Limit | Effective limit applied (may be lower than requested). |
X-Offset | Offset applied. |
X-Plan-Capped | true when your key's per-request cap, not your own limit, determined the page size. |
See Getting started with the p3dh-data endpoint for dimensions, pagination, and a worked example of turning data points into your own metric
Coverage
GET /v1/coverage/
Summary of the full dataset's scope. No parameters.
Response fields
| Field | Type | Description |
|---|---|---|
bank_count | integer | Distinct institutions with Pillar 3 data. Banks known only from the EBA Transparency Exercise aren't counted. |
total_datapoints | integer | Total individual cell values in the dataset. |
period_range.earliest / period_range.latest | string (YYYY-MM-DD) | Earliest and latest reference periods available. |
template_count | integer | Number of distinct EBA templates present. |
templates | array of {table_code, table_name} | Every template in the dataset. |
last_submission_date | string (YYYY-MM-DD) or null | Date of the most recently ingested submission. |
metric_count | integer | Distinct computed metric names currently available (see /v1/metadata-metrics/). |
metric_group_count | integer | Number of metric categories currently available. |
metric_groups | array of strings | Every metric category currently available. |
last_metrics_computed_at | string (ISO 8601) or null | When the metrics pipeline last finished computing values; may be more recent than last_submission_date, which tracks the raw EBA data only. |
Earnings
P&L, balance sheet, KPI and key-metrics figures from banks' own earnings releases: the bank's data pack, or its PDF report where the data pack doesn't carry a figure. This is a different source from Metrics and Data points above: published by the bank alongside its own earnings release, not regulator-published EBA XBRL, and not cross-validated against it. entity_lei is the same identifier used everywhere else in this API, so earnings data joins directly onto Banks, Metrics, or Data points.
Earnings items
GET /v1/metadata-earnings/
The catalog of line items available on /v1/earnings/: the earnings equivalent of Tables. It is a live query over the data, one record per line item. The earnings dictionary is the same catalog, searchable.
Request parameters: all optional, combine freely.
| Parameter | Match | Description |
|---|---|---|
category | exact | pnl, asset, equity_liability, kpi or km12. |
schema_item | partial, case-insensitive | Normalised line-item name. |
Response fields: one object per distinct (schema_item, category).
| Field | Type | Description |
|---|---|---|
schema_item | string | Normalised line-item name, comparable across banks and publications, e.g. Net interest income. Pass to /v1/earnings/'s schema_item filter. |
category | string | pnl, asset, equity_liability, kpi or km12. |
unit | string | The unit its values are in, see unit below. |
bank_count | integer | How many banks report it. |
Earnings
GET /v1/earnings/
The values themselves. A data point is a bank, category, schema_item and period. Each publication reports its own quarter and restates comparatives for earlier ones, so one data point has a value in every publication that prints it. The view parameter picks which of those values you get:
view | Returns | Use it for |
|---|---|---|
time_series (default) | The newest value of each data point: the latest restated figure. | Comparable history. Has look-ahead: a 2019 figure may come from a 2021 publication. |
as_of | With as_of_date=YYYY-MM-DD: for each data point, the value from the latest publication out on that day. Publications with an unknown publication_date are left out. | Backtests: the data exactly as it stood on that day, free of look-ahead. |
release | With publication_period, e.g. 2025-Q3: the values that release stands by, comparatives included. | What one release said. |
first_published | Each data point's value as first published. | Measuring restatements against time_series. |
all | Every value of every publication, including ones a later parse of the same document superseded. | Audit. |
Every other parameter works the same in every view. Filters on the data point (entity_lei, slug, bank_name, category, schema_item, period) are applied before the view picks its value. Filters on the value (publication_period, section, reporting_basis, source) are applied after, so they narrow the result and never swap in an older value.
Request parameters: all optional except where a view needs one.
| Parameter | Match | Description |
|---|---|---|
view | exact | time_series (default), as_of, release, first_published or all. Anything else returns 422. |
as_of_date | YYYY-MM-DD | Required with view=as_of, rejected with any other view. |
entity_lei | exact, repeatable | One or more banks, from Banks. |
slug | exact, ignores case, repeatable | The bank's slug, e.g. slug=DBK, as an alternative to entity_lei. |
bank_name | partial, case-insensitive | The bank's name as in its publications; may differ slightly from the canonical entity_name on Banks. Join on entity_lei, not name. |
category | exact | pnl, asset, equity_liability, kpi or km12. |
schema_item | partial, case-insensitive | Normalised line-item name, from Earnings items. |
period | exact, repeatable | The period the value is for: 2025-Q2, 2025-6M, 2025-9M, 2025-FY. A label, not a date. |
publication_period | exact, repeatable | The quarter of the publication the value comes from, e.g. 2025-Q3. Required with view=release. |
section | exact, repeatable | The part of the statement, e.g. pnl, assets, liabilities, equity, kpi, capital. |
reporting_basis | exact | reported, underlying or pro forma. |
source | exact | The data pack, e.g. abn_factsheet, or ai_enrichment for values read from a PDF. |
limit | integer, ≥ 1 | Default 1000, silently clamped to your key's per-request maximum. |
offset | integer, ≥ 0 | Number of rows to skip, not a page number. Default 0. |
Response fields: one object per value.
| Field | Type | Description |
|---|---|---|
row_id | string | Stable id of this value in this publication. |
entity_lei | string | Bank's LEI, .CON form. Use this as the join key across endpoints. |
short_name, abbreviation, slug | string or null | The name the bank is known by, its shorthand and its lookup key, as on Banks. |
bank_id | string | Short bank key used by the dataset, e.g. abn. |
bank_name | string | Bank's name as in its publications. |
category | string | pnl, asset, equity_liability, kpi or km12. |
section | string or null | The part of the statement within the category. |
schema_item | string | Normalised line-item name. |
period | string | The period the value is for, e.g. 2025-Q2. Flows are for the period; stocks are at its end. |
period_end_date | string (YYYY-MM-DD) | period as a date, for sorting. |
value | number | The value, in unit, as published. |
unit | string | millions, fraction (0.123 = 12.3%), bps, count, currency_per_share or millions_shares. |
currency | string or null | ISO 4217 code for monetary units (millions, currency_per_share); null for ratios and counts. The bank's reporting currency; earnings values are never converted. |
value_type | string | reported, or derived: not printed, but computed from the publication's own parts (the formula is in note). |
reporting_basis | string | The bank's own basis: reported, underlying or pro forma. |
publication_period | string | The quarter of the publication the value comes from, e.g. 2025-Q3. |
publication_date | string (YYYY-MM-DD) or null | The day that publication came out; null where unknown. |
source | string | The data pack, or ai_enrichment for a value read from a PDF. |
source_file, source_file_name | string | The document the value comes from. |
sheet | string or null | The data-pack tab the value is on; null for a PDF value. |
source_page | integer or null | 1-based PDF page; null for a data-pack value. |
source_text | string | The text the value was read from: the row label in the data pack, or the quoted PDF line. |
basis_match | string | exact, or partial when the bank's line doesn't fully match the item's definition (note says why). |
note | string or null | Why a match is partial, how a value was derived, or how it was corrected. |
data_quality_score | string | high, medium or low. |
human_verification | boolean | A person checked this value against its source. |
date_stamped_view | boolean | The value this publication stands by (one per data point per publication). |
time_series_view | boolean | The newest value of the data point (one per data point). |
creation_date | string (ISO 8601) | When the value was produced. |
Examples
# Deutsche Bank's P&L, latest restated figures
GET /v1/earnings/?entity_lei=7LTWFZYICNSX8D621K86.CON&category=pnl
# Net interest income across banks, as known on 1 March 2025
GET /v1/earnings/?schema_item=Net+interest+income&view=as_of&as_of_date=2025-03-01
# Everything ABN AMRO's 2025-Q3 release reported
GET /v1/earnings/?entity_lei=BFXS5XCH7N0Y05NIXW11.CON&view=release&publication_period=2025-Q3Response headers
| Header | Description |
|---|---|
X-Result-Count | Records in this response. |
X-Has-More | true if additional rows exist beyond this page. |
X-Limit | Effective limit applied (may be lower than requested). |
X-Offset | Offset applied. |
X-Plan-Capped | true when your key's per-request cap, not your own limit, determined the page size. |
See Getting started with the earnings endpoint for a full walkthrough.
Currencies
Banks report in their own currency: a Swedish bank in SEK, a Polish bank in PLN. So that amounts compare across banks, Metrics and P3DH data return every amount in EUR by default. Pass currency to get them in another currency instead:
GET /v1/metrics/?metric_name=cet1_amount¤cy=GBP
GET /v1/p3dh-data/?table_code=K61.00&ref_period=2025-12-31¤cy=SEKcurrency | Currency |
|---|---|
EUR | Euro (default) |
BGN | Bulgarian lev |
CHF | Swiss franc |
CZK | Czech koruna |
DKK | Danish krone |
GBP | Pound sterling |
HUF | Hungarian forint |
ISK | Icelandic krona |
NOK | Norwegian krone |
PLN | Polish zloty |
RON | Romanian leu |
SEK | Swedish krona |
USD | US dollar |
These are the currencies of the European Economic Area, plus the pound sterling and the US dollar. Codes are case-sensitive; any other value returns 422.
How amounts are converted
- Each amount converts at the European Central Bank's euro reference rate for its own reference period. If that day was a weekend or bank holiday, the last rate published before it is used. A bank's 2025-12-31 figures use the 2025-12-31 rate, and its 2025-09-30 figures use the 2025-09-30 rate.
- Conversions between two non-euro currencies, such as SEK to NOK, go through the euro.
- Only amounts are converted: data points whose
data_typeismonetary, and metrics whoseunit_kindiscurrency. Ratios, percentages, counts, years, dates, flags, and text never change. - The Bulgarian lev was replaced by the euro on 2026-01-01. From then on it converts at its fixed rate of 1.95583 per euro.
- For
delta_*andgrowth_*transformations, both periods are converted at their own rate before the change is computed. Growth in EUR terms therefore includes exchange-rate moves. To see growth in the bank's own currency, request that currency.
Getting the figures as filed
Every row also carries reported_value and reported_unit: the figure exactly as the bank filed it, in its own currency. fx_rate is the rate that was applied, so value = reported_value × fx_rate. If no rate is available yet for a period, value is null and reported_value still holds the figure. Rates are added within a few hours of a new reporting period first appearing in the data.
Errors
All errors use the shape {"detail": "..."}, except 422 validation errors, which use {"detail": [{"loc": [...], "msg": "...", "type": "..."}]}.
| Status | Meaning |
|---|---|
400 | League tables: an unrecognised period, or more than 200 banks match. The detail message says which, and how to fix it. |
401 | Missing, unrecognized, or revoked API key. |
404 | Unknown route. |
409 | POST /v1/signup/: an active key already exists for this email. |
422 | An invalid parameter value, most often an unrecognized transformation or currency. |
There's no 400 for exceeding your key's per-request row cap on /v1/metrics/, /v1/p3dh-data/, or /v1/earnings/: limit is silently clamped instead, and X-Plan-Capped tells you it happened. Likewise there's no error for requesting data older than your key's history depth. It's just excluded from the response.