📚 IMF Database Directory

May 10, 2026 · View on GitHub

Purpose: Full enumeration of the SDMX 3.0 dataflows advertised by GET https://api.imf.org/external/sdmx/3.0/structure/dataflow, grouped by IMF publisher unit, with each database tagged for EP editorial relevance. Companion to indicator-catalog.md — this file enumerates the databases (dataflows); the catalog enumerates the indicators (SDMX codes) inside them.

📅 Last Updated: 2026-05-10 | 🏷️ Classification: Public | 🌀 Wave: 4

Agency note (post-Sept-2025 IMF Data Portal): the umbrella IMF agency was retired and now returns 204 for every editorial dataflow. Each dataflow now lives under its publishing sub-department: IMF.RES (Research — WEO, PCPS, ITS), IMF.FAD (Fiscal Affairs — FM), IMF.STA (Statistics — everything else editorial: CPI, BOP, BOP_AGG, ER, IFS, DOT, CDIS, CPIS, GFS, GFSR, FSI, MFS). The TypeScript client (src/mcp/imf-mcp-client.ts) auto-resolves agency from dataflow id; agents may pass an explicit agencyId override on getParameterDefs / getParameterCodes / fetchData for vintage or non-editorial dataflows.

🔑 Relevance Legend

Tag	Meaning	EP editorial use
🟢	Actively wired	Cited by at least one article type today; mapped in `imf-indicator-mapping.md`. Indicators enumerated in `indicator-catalog.md §2`.
🟡	On-demand	Not wired to a recurring article-type mapping but may be queried ad-hoc via `imf-search-databases` + `imf-fetch-data` when an editorial brief requires it.
⚪	Out of scope	Not applicable to EU Parliament Monitor editorial surface (e.g. regional REOs covering non-European geographies, or specialised IMF-operational series).

Discovery is always possible: any agent can call imf-list-databases or imf-search-databases("keyword") to surface a database not pre-listed here. This file is the editorial relevance map, not a hard allowlist.

1. Flagship Publications (🟢 all wired)

Database ID	Full name	Cadence	Forecasts	Relevance	EP usage
`WEO`	World Economic Outlook	A (April + October)	+5y	🟢	ECON/BUDG/AFET macro + all forward-looking article types
`FM`	Fiscal Monitor	A (April + October)	+5y	🟢	BUDG debt/deficit, primary + structural balance
`GFSR`	Global Financial Stability Report	S (April + October)	—	🟢	ECON banking-sector stability; financial-stability risk commentary
`EREO` (alias `EUREO`)	Regional Economic Outlook — Europe	S (May + November)	+3y	🟢	ECON/AFET Europe-specific macro narrative + policy risk callouts
`IFS`	International Financial Statistics	M / Q	—	🟢	ECON policy-rate tracking; quarterly growth + monetary aggregates

Note on EREO vs EUREO: the SDMX dataflow ID is EREO in the Wave-1 REST surface; some legacy documents still use EUREO. The TypeScript client accepts either alias and normalises to EREO before issuing the request.

2. Statistics Department Series

Database ID	Full name	Cadence	Forecasts	Relevance	Notes
`CPI`	Consumer Price Index	M	—	🟢	Breaking-news inflation context
`BOP_AGG`	Balance of Payments — Aggregates	Q	—	🟢	INTA trade/capital flows, FDI
`BOP`	Balance of Payments — detailed	Q	—	🟡	Drill-down for components (goods, services, primary/secondary income)
`DOT`	Direction of Trade Statistics	M / Q	—	🟢	INTA bilateral trade flows EU ↔ RoW (US, CN, RU)
`ER`	Exchange Rates (REER / NEER)	M	—	🟢	ECON FX competitiveness framing
`PCPS`	Primary Commodity Price System	M	—	🟢	ITRE energy, AGRI food-price, ENVI raw-material context
`CDIS`	Coordinated Direct Investment Survey	A	—	🟡	INTA long-run bilateral FDI stock
`CPIS`	Coordinated Portfolio Investment Survey	S	—	🟡	ECON cross-border portfolio holdings
`IRFCL`	International Reserves & Foreign Currency Liquidity	M	—	🟡	ECON monetary-policy & FX-reserve commentary
`PGI`	Principal Global Indicators	M	—	🟡	G20-scoped snapshot for international comparisons

3. Fiscal Affairs Department Series

Database ID	Full name	Cadence	Forecasts	Relevance	Notes
`GFS`	Government Finance Statistics	A (+Q subset)	—	🟢	BUDG revenue structure, tax-to-GDP, expenditure functional breakdown
`GFSM`	GFS Manual reference dataset	A	—	⚪	Documentation dataset; use `GFS` for data
`FAD_EXPREV`	Expenditure & revenue by economic function	A	—	🟡	ECON/BUDG functional allocation
`TAXREV`	Tax Revenue (aggregate)	A	—	🟡	BUDG tax structure comparisons

4. Monetary & Capital Markets Department Series

Database ID	Full name	Cadence	Forecasts	Relevance	Notes
`FSI`	Financial Soundness Indicators	Q	—	🟢	ECON banking-sector health — NPLs, capital adequacy, ROA, ROE
`MCM_FSAP`	Financial Sector Assessment Program	Episodic	—	🟡	ECON country-level financial-sector stress snapshots
`SRF`	Standardized Report Forms (monetary statistics)	M	—	🟡	ECON central-bank balance sheet + monetary aggregates

5. Regional Economic Outlooks (REOs)

Database ID	Region covered	Cadence	Forecasts	Relevance
`EREO` / `EUREO`	Europe	S (May + November)	+3y	🟢
`MCDREO`	Middle East & Central Asia	S (May + October)	+3y	🟡 (cited in AFET Middle-East briefs)
`APDREO`	Asia & Pacific	S (April + October)	+3y	🟡 (cited in INTA China-trade + AFET Indo-Pacific briefs)
`WHDREO`	Western Hemisphere	S (April + October)	+3y	🟡 (AFET transatlantic briefs)
`AFRREO`	Sub-Saharan Africa	S (April + October)	+3y	⚪ (DEVE only — rarely cited)

6. Research & Analytical Datasets

Database ID	Full name	Cadence	Relevance	Notes
`HSG`	Historical Spending Dataset (incl. pre-1900)	Episodic	🟡	`historical-baseline.md` long-run fiscal context
`HPDD`	Historical Public Debt Database	Episodic	🟡	BUDG long-run debt-to-GDP comparisons
`HISTINF`	Historical Inflation Dataset	Episodic	🟡	`historical-baseline.md` long-run inflation context
`NMS`	New Member States convergence set	A	🟡	ECON convergence-criteria tracking for Bulgaria, Romania, etc.
`IMFDB`	Legacy all-indicator composite	—	⚪	Superseded by per-domain dataflows

7. Specialised / Operational (⚪ mostly)

Database ID	Full name	Relevance	Notes
`COFR`	Currency Composition of Foreign Reserves	🟡	ECON reserve-currency narrative
`IMFSDDS`	Special Data Dissemination Standard	⚪	Metadata about national data-publication compliance
`GCR`	Global Competitiveness Report dataset	⚪	Not an IMF core dataset; sometimes confused with World Economic Forum
`SDG`	Sustainable Development Goals (IMF surface)	🟡	ENVI/DEVE SDG monitoring — prefer UN SDG dataset for primary sourcing

8. Discovery Workflow (agent-usable)

When an editorial brief needs a database not listed above:

Call imf-search-databases("<keyword>") — returns matching dataflows.
Call imf-get-parameter-defs("<database_id>") — returns the SDMX data-structure definition listing dimensions.
Call imf-get-parameter-codes("<database_id>", "<dimension>", "<optional_search>") — returns the codelist for each dimension (e.g. the set of valid INDICATOR codes for WEO).
Call imf-fetch-data({ databaseId, startYear, endYear, filters }) with the discovered codes.

The five-step sequence is the canonical pattern for safe ad-hoc data fetching and is documented in chart-integration-guide.md §Discovery examples.

9. Drift Guard

The canonical list of "🟢 actively wired" databases is duplicated in:

This markdown table (source of truth for editors and agents).
The integration test test/integration/mcp/imf-mcp.test.js which exercises each 🟢 dataflow against the live SDMX 3.0 endpoint.
The indicator map in indicator-catalog.md.

The earlier IMF_ACTIVELY_WIRED_DATABASES / IMF_POLICY_INDICATORS TypeScript constants lived in src/utils/imf-data.ts and were purged in the April-2026 aggregator-pipeline migration — the agent now reads this document and the indicator catalogue directly at Stage A.

10. See also

indicator-catalog.md — per-database indicator codes
sdmx-dimensions-reference.md — SDMX dimension semantics
release-calendar.md — rolling vintage calendar
use-cases.md — article-type → database decision matrix
../methodologies/imf-indicator-mapping.md — validator wiring & policy