For the Treasury bill datasets and the Sovereign Bill Index, the full citation record — definitions market by market, stable URLs, the timeliness log and the corrections log — is on the reference page.

The rule

A number appears on Afronomics only when a publisher has printed it. If the publisher has no value for a country, the cell is blank and says so. We never estimate, interpolate or let a language model write a figure.

Sources

LayerPublisherRefresh
Macro, debt, trade, technology and climate seriesWorld Bank Open Data (World Development Indicators, International Debt Statistics)Daily check; the publisher updates quarterly
Development-finance projectsWorld Bank Projects & OperationsEvery 6 hours
Currency referenceExchangeRate-API mid-market ratesHourly
Headlines (the Wire)12 African publishers’ public RSS feedsEvery 15 minutes

The 22 tracked series

Rankings and comparisons

Rankings use each country’s latest print from the last three years, so a country is not dropped because its statistics office is a year behind. The year of every value is shown; an older print carries a small ’YY mark. Continental totals are summed only for the year in which at least 90% of the best-covered year’s countries reported, and the number of reporting countries is printed next to the total. “Movers” compare a country’s latest annual value with the year immediately before it.

Currencies

The live reference is a mid-market rate for comparison, not a dealing rate or a central-bank fixing. The annual figure is the official period average reported to the World Bank. Where a parallel market exists, it can differ materially from both.

Development finance

Commitments are IBRD and IDA amounts as the World Bank publishes them; co-financing is not included. Pipeline projects carry the Bank’s expected Board date, which moves. Themes are assigned from each project’s title and abstract and are a guide, not the Bank’s own sector coding.

Headlines

The Wire shows a publisher’s headline, a short excerpt from its feed, and a link to the original article. We tag countries and desks automatically from the text; tags can be wrong and are corrected on request. Publishers can ask to join or leave the Wire at desk@afronomicsfeed.com.

Kenya Poverty Premium

A monthly file of what the same everyday need costs a Kenyan on the options open to small amounts, against the option a better-off saver or borrower can reach: a short overdraft, a mobile loan, a year’s saving and money sent home. Every figure is worked out by code from data files we hold, each with its publisher and date; lines with no data are shown as gaps, never filled, and the lines are never added into one total. The full method, the gaps and every month’s file are on the Poverty Premium page.

Analysis

Briefs are labelled Facts, Analysis or Opinion, cite their primary documents with dates, and carry an as-of stamp. We do not give buy, sell or hold recommendations.

Shillings found

Five tools end with one line worked out by code: what the reader’s option costs or earns, the cheapest or best-earning option we hold for the same use, and the difference over a year, with the source and date of both sides. Is my rate fair? (saving, in Kenya, Ghana and Nigeria) compares with the highest-yielding money market fund or Treasury bill; mobile loans and Fuliza with the cheapest mobile loan whose charge is published as a fixed figure, for the same amount and days; money market funds with the highest-yielding fund after 15% withholding tax; sending money home with the cheapest service the World Bank surveyed on the same corridor. Only options a reader can take are compared: a gap to an average, such as the Central Bank’s bank lending average, is shown on the page but never counted. If we hold no comparable option, the tool says nothing new. If the reader’s option is already the best we hold, it says so and nothing is counted.

When a line shows a positive difference, it is added to a public total: “shillings found” on Kenya pages, “money found” in other currencies, and currencies are never added together. Our server works the difference out again from the reader’s figures and our own data before counting it, then drops the figures; only the tool, the market, the currency and the amount are kept, as a daily total. The count started on the day the feature was released, with nothing added for earlier days.

Its limits, plainly:

  • A difference shown is not money moved. We do not know what any reader did next.
  • One reader can be counted more than once (once per tool each time they open a page).
  • Amounts are what readers typed. One result counts at most KES 500,000 a year on savings and loans (the same converted into cedis or naira at the day’s reference rate, or not counted when we show none) and USD 2,000 a year on sending money home; a larger difference is not counted at all.
  • The comparable is the published price on its date: a fund’s yield can change, and a provider’s “from” rate is never used as the cheapest option.
  • Our own visits are not counted. Information, not advice.

Corrections

If a figure, date or tag is wrong, tell us and we fix it and log it publicly on the corrections page.

Reusing our data

Every series can be downloaded as CSV from its data page. Cite it as “Source: Afronomics, compiled from [publisher]”. The underlying World Bank data is licensed CC BY 4.0.

Source registry and reuse terms

Every publisher our paid products draw on, what we take and on what terms. A source that needs written permission stays out of paid products until it is granted; facts from all of them are reported with attribution. The same list is at /api/v1/sources, and the fingerprint, row count and newest date of every dataset deployed is at /api/v1/provenance. Every row is checked for a source link, a date and a rate in range before it is published.

  • Central Bank of Kenya, treasury bill and bond auction results (weekly (bills), monthly (bonds)). No open licence; facts reported with attribution. Written permission for bulk redistribution through the paid API not yet requested (founder action, strategy pack section N).
  • Central Bank of Kenya, bank deposit and lending averages, policy rate (monthly). As above.
  • Capital Markets Authority, collective investment schemes quarterly report (quarterly). Public report; cited as context with attribution. No ranking of named managers (D-028).
  • Fund managers, published money market and fixed-income fund yields (daily to monthly). Published facts with source and date; listed A to Z, never ranked.
  • SACCO Societies Regulatory Authority, licence list and annual supervision report (yearly). Public; attributed. Industry figures only in published products.
  • Retirement Benefits Authority, industry briefs on pension assets (half-yearly). Public; attributed.
  • Nairobi Securities Exchange, equity prices (daily). Redistribution in any medium needs an NSE distribution licence. Not shown until licensed.
  • Bank of Uganda, treasury bill and bond auction results (fortnightly). Reproduction needs written permission. Not in the paid regional pack until granted.
  • Central Bank of Nigeria, treasury bill results (fortnightly). Copying allowed with CBN named as the source and the content unaltered.
  • World Bank, consumer price inflation, annual (yearly). CC BY 4.0.