World Bank Development Indicators (Financial Sector)

World Bank Development Indicators (Financial Sector) is the Diversified Financial Services shelf's window onto country-scale financial depth - the frame every cross-border benchmark eventually stands on. The World Bank organizes roughly 1,498 World Development Indicators into twenty topics, and the Financial Sector topic (topic 7) tags 72 of them, from market capitalization of listed domestic companies (CM.MKT.LCAP.CD) through domestic credit to private sector (FS.AST.PRVT.GD.ZS), bank capital to assets ratio (FB.BNK.CAPA.ZS) and total reserves including gold (FI.RES.TOTL.CD).

Around those flagships sit the supporting cast: broad money, interest rate spreads, nonperforming loans, commercial branch and ATM density, portfolio equity flows. The panel shape is deliberate - one observation per country per indicator per year, across roughly 217 economies plus regional and income aggregates, with coverage reaching to 1960 and most financial series starting in the mid-1970s. Other records on this shelf cut finance finer: FDIC Bank Data & Statistics goes institution by institution and the Federal Reserve Economic Data Portal (Board Releases) owns the rate complex. This one owns the denominator underneath them both - how deep credit runs, how capitalized the banks stand, how large listed markets are relative to each economy.

Three consecutive year-end prints for one flagship indicator, exactly as they arrive:

indicator.id     : CM.MKT.LCAP.CD
indicator.value  : Market capitalization of listed domestic companies (current US$)
country.id       : US
country.value    : United States
countryiso3code  : USA
date             : 2024
value            : 62185685320000

date             : 2023
value            : 48979397700000

date             : 2022
value            : 40297980320000

Two things worth noticing. First, the spine never moves: whichever of the 72 financial-sector indicators you pull, the same seven fields appear in the same roles, so stacking credit-to-GDP beneath market cap is a union, not a remapping exercise. Second, the code arrives paired with its full descriptive name - CM.MKT.LCAP.CD travels with 'Market capitalization of listed domestic companies (current US$)' - so pipelines join on codes while humans read labels.

The values themselves tell a story in three lines: US listed-market capitalization stepping from roughly 40.3 to 49.0 to 62.2 trillion current US dollars across 2022-2024. One flat row per country-year, no parsing required.

Seven fields carry the verified spine. Two address the series: indicator.id holds the WDI code (CM.MKT.LCAP.CD) and indicator.value its full descriptive name, so a result set is self-documenting without a lookup join.

Three address the geography. country.id carries the two-letter World Bank code or aggregate code, country.value the readable name, and countryiso3code the ISO 3166-1 alpha-3 standard (USA) that most internal reference tables already speak.

Two carry the observation: date, an annual year for this panel, and value, the number itself - null when an economy simply did not report, which is information rather than noise.

Three qualifier columns - unit of measure, observation-status flags and display decimals - ride on every record but populate differently indicator by indicator, so they fold under additional fields on request. Name the indicators your workflow touches when you request a sample and they arrive as populated columns rather than documentation.

Geography - roughly 217 economies contribute national observations, from the largest listed markets to economies with a handful of quoted companies. Regional and income aggregates sit alongside in the identical column layout, so one query can return domestic credit for Brazil, for Sub-Saharan Africa and for upper-middle-income economies side by side. The corollary: filter deliberately, because aggregate rows will happily pose as countries.

Temporal - the panel reaches to 1960, though most financial-sector series begin in the mid-1970s as national reporting standardized. One planning note with teeth: high-profile market series such as capitalization report with multi-year lags, so the latest printed year can trail the calendar by about two. Backtests anchored on the newest print should know that.

Granularity - one observation per country per indicator per year. Annual rhythm is the native beat of the panel; nothing is interpolated up to quarterly or monthly, and delivered tables preserve that honesty rather than manufacturing frequency that was never observed.

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

Files suit the classic move: load the full country-year panel once and join it to your own exposure tables. Feeds suit dashboards that surface a credit-depth reading or a reserve-cover ratio inside a product. Warehouse delivery suits teams running cross-country models on a schedule, where the panel refreshes beside first-party data without a human in the loop. Whichever channel you choose, the tables arrive decoded - code-and-name pairs resolved, aggregate rows labeled, nulls left honest - and field names hold constant across channels and cadences, so changing frequency is configuration rather than migration.

• Investors & quant researchers - frame every market call with the local denominator: listed-market depth against GDP, private-credit intensity, reserve cover, before deciding a geography deserves a position (investors & quants use cases).
• Strategy & market-entry teams - rank candidate operating geographies on credit depth, bank capitalization and branch/ATM density before committing capital or headcount.
• Risk & treasury teams - watch domestic-credit cycles and nonperforming-loan ratios across every country a book touches, with six decades of history to separate signal from season.
• Economists & policy researchers - work from the citation-grade country-year backbone whose definitions have been stable long enough to support longitudinal study.
• Data scientists & ML engineers - engineer macro-financial features from a rectangular panel with stable field names across 72 indicators, enough depth for regime models rather than snapshot correlations (data scientists use cases).

Economists and macro strategists come first - the panel exists precisely so nobody has to reassemble a country's financial depth from twenty statistical annexes, and the stable definitions make decade-spanning comparisons legitimate rather than heroic. Investors and quant researchers come next: valuation multiples travel badly without the local credit and market-depth context this record supplies, and the aggregate rows make peer-group construction a filter rather than a project. Strategy teams sizing a new market get the fastest honest read on how banked, how levered and how listed an economy actually is, while academics, journalists and students gain numbers they can cite with attribution and reproduce - the rare panel whose provenance outlives the news cycle.

The shelf splits by the question being asked, and this record owns the country-scale system view: 217 economies, six decades, one schema. Neighbors own different jobs. FDIC Bank Data & Statistics cuts US banks institution by institution; the Federal Reserve Economic Data Portal (Board Releases) owns the rate complex everyone prices against; SEC EDGAR Full-Text Search API (EFTS) reads issuer filings; the BIS Data Portal records who owes whom across borders. Against market-conduct data the contrast is starkest - this comparison with the FINRA Data Catalog shows two records sharing an industry and almost nothing else. Run them together and the macro-financial frame meets its micro evidence.

Three things worth having in hand.

First, codes arrive paired with names. Every record carries both CM.MKT.LCAP.CD and its full descriptive label, so joins happen on codes while dashboards render language - and delivered tables ship with both intact.

Second, depth is real but uneven. The panel reaches to 1960, yet most financial-series histories begin in the mid-1970s, and headline market series can trail the calendar by roughly two years because of national reporting lags. A sample cut to your countries and indicators maps exactly where each history begins and ends before anything joins it.

Third, units are per-indicator, not global - current US dollars here, percent of GDP there, percent elsewhere. Delivered tables carry the unit metadata alongside values, so a capitalization figure never quietly plots against an interest-rate spread.