Key Takeaways
- The IMF's July 2026 WEO Update projects 3.0% global growth in 2026 and 3.4% in 2027. The World Bank's June 2026 Global Economic Prospects headlines 2.5% for 2026 and 2.8% for 2027.
- The World Bank's own Table 1.1 carries a second world line, weighted at purchasing power parity: 2.9% for 2026 and 3.4% for 2027.
- Like for like, the two institutions differ by 0.1pp in 2026 and by nothing at all in 2027. Roughly 0.4pp of the apparent gap is a weighting convention.
- The World Bank's 2026 forecast assumes Brent averages $94/bbl. Brent closed at $69.56 on 6 July 2026 and $91.82 on 27 July, averaging $90.25 across the year to that date.
The gap is 0.5 points, and most of it is not a disagreement
On 9 July 2026 the IMF put global growth at 3.0% for 2026. Four weeks earlier the World Bank had put it at 2.5%. Two institutions across the street from each other, modelling the same planet, half a percentage point apart — which on a roughly $126trn world economy is several hundred billion dollars of output.
Except that they are not measuring the same quantity. The World Bank's headline aggregate is built with GDP weights at average 2010-19 prices and market exchange rates. The IMF's world aggregate is built with purchasing-power-parity weights. PPP weights, in the World Bank's own words, "attribute a greater share of global GDP to emerging market and developing economies (EMDEs) than market exchange rates." Emerging economies grow faster. Give them a bigger share of the index and the index grows faster.
The World Bank prints the reconciliation itself
This does not have to be inferred. Table 1.1 of the June 2026 Global Economic Prospects carries a memorandum line, "World (PPP weights)", directly beneath the headline. It reads 2.9% for 2026 and 3.4% for 2027.
Set that against the IMF's 3.0% and 3.4%. The two institutions are 0.1pp apart on 2026 and identical on 2027. The chart above puts the three figures side by side. About 0.4pp of the 0.5pp headline gap is arithmetic — a units problem, not an argument about the world.
The residual 0.1pp is inside either institution's own revision cycle. The World Bank marked its headline down 0.1pp from January, and its PPP line by 0.3pp. The IMF called its July numbers "broadly unchanged on a cumulative basis" against April. Two forecasters agreeing to within a tenth of a point after a war shock are not disagreeing about the state of the world.
What the two institutions actually disagree about
The interesting divergence is in the narrative, not the number. The World Bank's June report calls 2026 "the lowest rate since the COVID-19 pandemic" and, in its foreword, "its weakest pace outside of outright recession in close to 20 years." Its framing is a growth shock driven by energy: Brent at $94/bbl for 2026, 36% above 2025 and more than 50% above the January projection.
The IMF, publishing four weeks later, is calmer, and specifically says the oil shock did not arrive. Petya Koeva Brooks of the IMF's Research Department: "A large spike in oil prices was avoided thanks to inventory drawdowns, expanded production outside the Gulf …" The IMF's worry moved elsewhere — to prices. Global headline inflation was revised up to 4.7% for 2026, with the observation that "the disinflation trend in place since the beginning of 2024 has stalled."
So the growth numbers converge and the risk assessments diverge. The World Bank is worried about output; the IMF is worried about inflation. That is a more useful thing for an allocator to know than a spurious half-point gap.
The oil assumption is already under strain
Both institutions hang their 2026 numbers on an energy path, and that path is testable in real time. The World Bank assumes Brent averages $94/bbl across 2026, a figure built on the premise that disruptions abate in July. Brent averaged $90.79 in 2026 through 6 July, close to the assumption — but the monthly path is doing something the annual average conceals. It peaked at $138.21 on 7 April and closed at $69.56 on 6 July.
This piece's own conditional read: if the second half simply held near $70, the full-year average would land near $81 rather than $94, making the World Bank's energy-driven downgrade too pessimistic by construction. It did not hold. Brent closed at $105.32 on 23 July 2026 and $91.82 on 27 July, putting the 2026 average to date at $90.25 — within $4 of the assumption, not $13 below it. The Brent strip is still the number that reconciles these forecasts, which is exactly why it should not be read off one close.
What this framework cannot settle
PPP weights are not a neutral fix. They are fixed at average 2010-19 prices, so they embed a snapshot of relative price levels that is now years stale, and they systematically raise the weight of economies whose statistical systems are weakest. Neither aggregate is "correct": market-exchange-rate weights are the basis that translates growth into dollar corporate revenue, which is the 2.5%, and PPP weights are the basis for measuring real activity, which is the 3.0%.
The two reports also have different cut-off dates: the GEP's data cutoff was 2 June 2026, and the IMF priced its commodity assumptions off futures as of 10 June. That is also how the IMF's "spike avoided" language and this piece's own $138.21 print both stand — 7 April is a daily high in a spot series, and the IMF is describing an annual average priced two months later. The residual 0.1pp is small enough to be noise. And a 0.1pp agreement between two forecasters is not evidence that either is right; both were revising in the same direction, from the same data, with overlapping staff traditions. Correlated errors look like consensus.
The practical consequence for an allocation
The gap has a portfolio-sizing analogue. An investor who benchmarks emerging-market equity weight to "global growth" is implicitly choosing an aggregation method. Using the 2.5% market-FX figure understates EM's contribution to real world activity; using the 3.0% PPP figure overstates EM's contribution to the dollar earnings that actually accrue to shareholders. The wedge between the two — 0.4pp of world growth in 2026 — is the size of the error, and it runs in opposite directions depending on which number is picked. The honest resolution is that the growth forecast does not settle the EM weight at all, and any process in which a tenth of a point of global GDP moves an allocation is a process being driven by rounding.