Why African statistics get revised

The figure on a country profile today may not match the figure you cited last year — and that's almost always normal. This is what causes it.

One of the most common emails the editorial team receives is some version of "the population / GDP figure on this page no longer matches the value I quoted in a paper last year — which one is right?" The honest answer is usually that both are right, for different periods or under different methodologies, and the figure has been revised. This page is a reference to the revisions that affect African data most often. It is the answer to the email, and a guide to writing citations that won't trip on the next revision.

The four kinds of revision that move African headlines

1. GDP rebasing

National GDP series rest on a base year — the snapshot used to weight the contribution of each economic sector. Over time, economies change shape: services grow as a share of GDP, manufacturing shifts, the digital economy expands, informal activity gets better measured. When the gap between the base year and the current year becomes large enough, the structure of the series no longer reflects the structure of the economy, and the country rebases.

Rebasing typically reveals output that earlier base years undercounted, particularly in services and informal-sector activity. The result is usually a one-time upward step in nominal GDP. Several major African economies — Nigeria, Ghana, Kenya, Tanzania and Uganda among them — have rebased their national accounts in recent years, and most of those rebasings produced double-digit-percent upward revisions in the GDP level. The growth path before and after is recalculated on the new base, so the post-rebase series is internally consistent, but a single figure quoted from the pre-rebase series will differ from the equivalent figure on the rebased series even for the same year.

For a fuller treatment of how rebasing interacts with the nominal-vs-PPP and exchange-rate issues, see reading African economic data.

2. Census-driven population revisions

Population figures are estimates between censuses. Run a fresh census after a long gap and the count usually differs from the rolling estimate by several percent. Sometimes the census comes in higher than expected; sometimes lower. Either way, the UN Population Division and the World Bank revise their published series to align with the new benchmark, and that revision propagates back several years.

The same kind of revision also flows through every per-capita indicator. A country whose population is revised down by 5% will see its GDP-per-capita figure rise by about 5%, even if no economic change has occurred. Per-capita rankings between countries can therefore reorder when one of them runs a census.

For more on how to read the population figures themselves, see reading African demographic data.

3. Methodology updates from the source organisation

The international institutions that publish African data revise their methodologies on an ongoing basis. Some of the most consequential revisions are:

4. Series breaks

Some revisions are large enough that the post-revision series is not directly comparable to the pre-revision series. A change in classification (more or fewer items inside an indicator), a change in geographic coverage (a country that was not previously broken out separately is now), or a change in the underlying survey instrument can all create a "series break". When a series breaks, charting the indicator across the break date without flagging it can produce misleading visuals. The originating organisation usually documents these breaks; on Africa Center we follow the latest series and note the break on the country profile when relevant.

Common pitfall: citing a figure from this site without an access date and without naming the source. When the figure is revised — and it will be — your reader can no longer verify what version you used. The how to cite page has examples that handle this correctly.

How to write citations that survive a revision

Three rules cover the vast majority of revision-related citation problems:

  1. Always include an access date. "Retrieved 28 April 2026, from africacenter.net/kenya/" tells the reader which version of the page you saw, even if a later visit shows different numbers.
  2. Cite the underlying primary source for individual figures. If you cite the World Bank or IMF directly, your reader can find the same vintage of the figure on the source's archived release — the originating organisations preserve old releases more reliably than aggregators do.
  3. Name the indicator variant. "Nominal GDP, current US dollars, IMF WEO October 2025 release" is unambiguous. "GDP" alone is not.

The full set of citation examples in three formats is on how to cite.

What to do if a figure on the site is wrong

Two scenarios are easy to confuse. They are different.

Editorial checklist for revisions

When a country page is updated to reflect a new release, the editorial check covers:

Where to go next

Last reviewed: 28 April 2026.