World Bank Microdata Library - Global Findex Catalog

Datadory delivers World Bank Microdata Library Global Findex catalog data — the complete study catalog behind the Global Findex Database, the most-quoted measurement of how adults worldwide store, move and borrow money. The Microdata Library's Central Data Catalog holds 7,098 studies; the Global Financial Inclusion (Global Findex) collection inside it accounts for 717 study records, each carrying anonymized respondent-level microdata, the fielded questionnaire, and a browsable data dictionary.

Every wave ships as its own set of catalog entries: the 2025 edition (survey year 2024), 2021, 2017, 2014 and 2011, with country-specific entries sitting beside the world file. The flagship 2025 world file holds 144,090 cases described by 199 variables — roughly 145,000 adults surveyed across 141 economies by Gallup through its World Poll, fieldwork May to December 2024.

This is the row-level layer beneath the headline statistics: not '76 percent of adults have an account' but each adult's own answer, weighted, coded and joined to demographics. Get a sample of this dataset to see real records with the full dictionary before committing.

One adult, one row — outcome flags beside the demographics that segment them, exactly as the labelled dictionary keys them:

RESPONDENT RECORD — one row per surveyed adult
| wpid_random | economy   | female | age | inc_q | urbanicity | account_fin | account_mob |
| ----------- | --------- | ------ | --- | ----- | ---------- | ----------- | ----------- |
| (anon id)   | Albania   | 0      | 34  | 3     | 0          | 1           | 0           |
| (anon id)   | Zimbabwe  | 1      | 27  | 2     | 1          | 0           | 1           |

VARIABLE DICTIONARY — label join shipped with every extract
| VARIABLE       | LABEL                                        |
| -------------- | -------------------------------------------- |
| account_fin    | Has an account at a financial institution    |
| anydigpayment  | Made or received a digital payment           |
| urbanicity     | Respondent lives in rural area               |

Two properties worth noticing. Every respondent keeps an anonymized identifier (wpid_random) plus a sampling weight (wgt), so counts scale back to national adult populations instead of staying raw tallies. And outcomes arrive as explicit booleans rather than pre-aggregated shares — saved, borrowed, anydigpayment, merchantpay_dig — which means a segment you invent today ('digitally paid, no bank account, bottom two quintiles') is a filter, not a special request.

Twenty-seven verified dictionary fields anchor the 199-variable 2025 file: identity and design columns that key every record, demographic cuts that make segmentation possible, and the outcome battery covering accounts, saving, borrowing, payments, cards and identity.

• Geography: worldwide. The 2025 wave covers 141 economies, Albania to Zimbabwe; earlier editions cover roughly 140 economies each, with country-specific catalog entries alongside the pooled world file.
• Temporal: five waves — 2011, 2014, 2017, 2021 and 2025, the last fielded May to December 2024. Consistent core question wording across all five is what makes a fifteen-year trend line quotable rather than approximate.
• Granularity: individual respondents, anonymized, each carrying a sampling weight so results aggregate cleanly to economy level — while remaining disaggregated enough to cut by gender, age band, education, income quintile, workforce status or rural/urban residence without touching a published table.

Set against the wider Datadory catalog — average quality score 7.81 across all 1,744 datasets — this slice scores 8/10, carried by verified field documentation and coverage depth (717 study records) that private panels cannot replicate at this breadth.

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

You pick the channel and the cadence; the field dictionary above travels unchanged across all three. Bulk files suit analysts loading all five waves once and joining them to their own macro tables, structured payloads suit products surfacing a single indicator inside an app, and warehouse delivery suits teams running cross-country models in SQL against weighted microdata. Changing cadence is a settings conversation, not a re-integration project.

• Emerging-market sizing for fintech and payments — account ownership (account, account_fin, account_mob) filtered to the economies you serve, scaled by pop_adult, converts penetration headlines into addressable-adult counts per income quintile and region.
• Segment-level product research — crossing inc_q with urbanicity and female isolates the underserved cells (rural women outside the top quintile are the classic) that country averages hide entirely.
• Barriers-to-adoption analysis — fin11a and its sibling reason codes turn 'why don't the unbanked open accounts' from anecdote into a ranked distribution, market by market.
• Credit-access context models — borrowed, fin10 and fin2 give per-economy behavioral priors where bureau files are thin or absent.
• Digital-ID readiness work — the 2025 additions fin46 (ID ownership) and fin51 (digital ID used online) measure the KYC infrastructure under account growth before you build on it.
• Trend attribution since 2011 — same wording, five waves: durable inclusion gains separate cleanly from pandemic-era digital-payment jumps.

Market researchers and consultants get the citation-grade source beneath the most-quoted numbers in development economics, plus the ability to rebuild any headline statistic for exactly the segment their client sells into. Investors and quant researchers get economy-level adoption trajectories across five waves — the leading indicators behind payment-infrastructure and neobank theses in emerging markets. Data scientists and ML engineers get a weighted respondent-level file that joins onto macro panels by ISO code and year, ready for feature engineering across 199 documented columns. Journalists, academics and NGOs get respondent rows behind claims they need to defend, with clean demographic cuts for subgroup statements.

Notes and adjacent reading:

• World Bank Global Findex Database 2025 — the aggregated sibling: nearly 300 country-level indicators in a flat 438-column table when you want the share, not the rows beneath it.
• FDIC National Survey of Unbanked and Underbanked Households — the US household-level counterpart, biennial since 2009, for anchoring global comparisons domestically.
• UK FCA Financial Lives Survey — regulator-fielded UK consumer waves, 12,865-19,145 respondents each, for a high-income-market contrast.
• Federal Reserve Survey of Consumer Finances (SCF) — household balance sheets triennially since 1989 when the question moves from access to wealth.

Browse the whole vertical on the Consumer Finance data hub or the ranked shortlist of the best consumer finance datasets.