Key figures (as of October 2026)
| World real GDP growth, 2025 / 2026 / 2027 (projection) | 3.5% / 3.1% / 3.4% |
| G20 national members with a fall in real GDP | 0 of 19 in 2025; 0 projected for 2027 |
| Lowest projected 2027 growth among G20 members | Italy 0.6%, France 0.7%, Japan 0.7%, Russia 0.9% |
| Years since 1993 with no G20 member in decline | 9 of 33, on the Centre's own calculation |
| Years since 1993 with five or more G20 members in decline | 3 of 33 (1993, 2009 and 2020), on the Centre's own calculation |
| United States, third quarter of 2026 | real GDP up 3.4% at an annual rate on the second quarter |
| Euro area, third quarter of 2026 (flash) | real GDP up 0.4% on the second quarter and 1.3% on a year earlier |
| China, third quarter of 2026 | real GDP up 4.5% on a year earlier |
| US yield curve, 29 October 2026 | 10-year yield 4.98%, 3-month yield 4.18%: a positive gap of 0.80 percentage points |
| Policy rates | United States held at 3.75% to 4.00% (28 October); euro area deposit rate 2.50%, held on 29 October; Japan held at 1.25% (30 October) |
What a recession count measures
There is no single official definition of a world recession. The US business-cycle dating committee dates turning points for the United States alone; its last recession ran from a peak in February 2020 to a trough in April 2020. Many countries use two quarters of falling output as a working rule. Both are hard to compare across 19 economies with different release calendars.
This study uses a simpler, comparable measure: whether a country's real GDP for the full calendar year is lower than in the year before. An annual fall is a narrower test than a recession, because a short downturn inside a year can be hidden by growth in the other quarters, but it is published for every G20 member in one official database, on one date.
The Centre's study "Growth in 2027: The Gulf and the World Economy Outlook" carries a separate forecast on UAE growth in 2027. This study looks at the G20 as a whole.
The baseline: slower, but positive everywhere
The main international financial institution's October 2026 outlook projects world growth of 3.1% in 2026 and 3.4% in 2027, against 3.5% in 2025. Advanced economies are projected to grow 2.1% in 2026 and 1.8% in 2027. Two forces pull in opposite directions: higher energy and commodity prices after temporary regional supply and shipping disruptions, and strong investment in artificial intelligence. The central assumption is that energy markets and shipping normalise gradually through 2027.
In the October 2026 database, every G20 national member is projected to grow in 2027. The projections run from 0.6% for Italy to 7.9% for Saudi Arabia, the fastest projected rate in the group on the Centre's reading of the database. The United States is projected to grow 2.3%, China 4.0% and India 7.0% on its fiscal-year basis.
In one of its adverse scenarios, a longer conflict keeps oil at $130 a barrel from the fourth quarter of 2026 to the first quarter of 2027 and raises long-term borrowing costs; together these lower world growth by 0.2 percentage points in 2026 and 0.5 percentage points in 2027 compared with the baseline. In another, a sharp fall in investment in artificial intelligence and in asset prices lowers world output by 1.1% by 2028 compared with the baseline, including a fall of 1.8% for the United States. It puts the probability of world growth falling below 2% in 2027 at about 10%.
The latest data
United States. Real GDP rose at an annual rate of 3.4% in the third quarter of 2026, after 2.2% in the second, led by investment, consumer spending and government spending. The central bank held its target range at 3.75% to 4.00% on 28 October, with two members preferring a rise of 0.25 percentage points, and said that economic activity was expanding at a strong pace while inflation remained elevated.
Euro area. The flash estimate shows output 0.4% higher in the third quarter of 2026 than in the second, and 1.3% higher than a year earlier. Germany grew 0.2% on the quarter and Spain 0.6%, while Italy was unchanged and France was 0.1% lower. France's output has now been lower than in the previous quarter in two of the three quarters of 2026 so far, and was 0.1% below its level of a year earlier in the third quarter. The euro area deposit rate stands at 2.50% after a rise with effect from 16 September, and was held on 29 October.
Asia. China's real GDP was 4.5% higher in the third quarter of 2026 than a year earlier, and 4.6% higher over the first three quarters. Japan's output rose 0.4% in the second quarter on the first, an annual rate of 1.4%. The Japanese central bank held its policy rate at 1.25% on 30 October.
Business surveys. Flash purchasing managers' indices for October 2026 stayed above the 50 mark that separates growth from contraction in the four largest advanced economies: 57.6 in the United States, 53.1 in the United Kingdom, 52.3 in Japan and 51.9 in the euro area. The euro area figure was down from 53.1 in September.
Bond markets. On 29 October 2026, the yield on 10-year US government bonds was 0.80 percentage points above the 3-month yield, and 0.37 percentage points above the 2-year yield. An inverted curve, with short-term yields above long-term yields, has often come before US recessions; the curve is not inverted today.
How often G20 economies shrink
On the Centre's own calculation from the October 2026 database, the number of G20 national members whose annual real GDP fell has been:
- 0 in 9 of the 33 years from 1993 to 2025, most recently 2021 and 2025.
- 1 in 9 years, most recently 2022 and 2024.
- 2 in 6 years, most recently 2023.
- 3 or 4 in 6 years, most recently 2019, when three fell.
- 5 or more in 3 years: 1993 (five), 2009 (14) and 2020 (17).
The series starts in 1993 because that is the first year with figures for all 19 members. These counts use the figures as they stand in the October 2026 database, after later revisions, not as first published. Revisions matter at the edge: Germany's 2024 growth now stands at 0.0% in the database, so a small revision can move a country across the line.
Since 2010, leaving aside 2020, Argentina accounts for 7 of the 17 annual falls among G20 members, on the Centre's own calculation. Argentina's real GDP fell in 20 of the 46 years from 1980 to 2025, including 2023 and 2024. It is projected to grow 3.5% in 2027.
The Gulf: a strong 2027
The Gulf economies enter 2027 with some of the strongest growth projections in the world. International projections put UAE growth at 7.7% in 2027 and GCC growth at 9.1%, as energy output and trade normalise and non-oil activity keeps expanding. The G20's Gulf member is projected to grow 7.9% in 2027, the fastest rate in the group.
What it means
The G20 enters 2027 with positive growth projected for all of its national members, and the central case is that none, or at most one, shows a fall in output. On the Centre's estimates, the combined weight of those two bands is 68%. The risks are concentrated in a few slow-growing economies and in two shared shocks, energy prices and a reassessment of investment in artificial intelligence. A count of five or more would need a global shock of a kind seen only twice in the past two decades.
Scenarios: the number of G20 economies with a fall in real GDP in 2027
These probabilities are the Centre's analytical estimates, reasoned from the figures above. Because the bands are ordered, the forecast is scored with both the multi-category Brier score and the ranked probability score.
The variable is the number of the 19 national members of the G20 (Argentina, Australia, Brazil, Canada, China, France, Germany, India, Indonesia, Italy, Japan, Korea, Mexico, Russia, Saudi Arabia, South Africa, Türkiye, the United Kingdom and the United States) whose real GDP growth for 2027 is below zero in the outlook database released with the main international financial institution's April 2028 world outlook. The two regional bloc members, for Europe and for Africa, are excluded, so the count runs from 0 to 19 and no economy is counted twice. Each member's annual figure is used on the basis the database uses, which for India is the fiscal year. Each growth figure is rounded half up to one decimal place, once, before it is judged, so -0.05% counts as 0.0%, which is not a fall, and -0.06% counts as -0.1%, which is a fall. A member counts only if its rounded growth figure is below 0.0%; exactly 0.0% does not count. A figure reported as a range counts at its midpoint. If the database has no 2027 figure for a member, that member's first annual 2027 figure from its national statistics office is used; if neither exists by the cut-off date, the member counts as not falling. Later revisions are ignored. The bands cover every possible count, and each edge falls in the higher band: exactly 1 falls in band 2, exactly 2 in band 3, exactly 3 in band 4 and exactly 5 in band 5. If the April 2028 database is not published by 31 July 2028, the first full edition of the database published between 1 March and 31 July 2028 is used; if none is published by 31 July 2028, the forecast is withdrawn.
No G20 economy shrinks: 0
This is the latest actual value: no member's GDP fell in 2025, and none is projected to fall in 2027. It occurred in 9 of the 33 years since 1993, about 27% of them on the Centre's own calculation. The weight of 38% sits above the historical rate of about 27% because the latest projection shows no member falling in 2027. It is the most likely band and gets the highest weight.
One economy shrinks: 1
One fall occurred in 9 of the 33 years, as often as none. Four members are projected to grow by less than 1% in 2027, France's output fell in two of the first three quarters of 2026, and Argentina's GDP has fallen in about two years in five since 1980. Any one of these could turn negative without a wider downturn, so this band gets the second-highest weight.
Two economies shrink: 2
Two falls occurred in 6 of the 33 years, most recently 2023. This would need two of the weaker economies to slip at the same time, for example after a slower return of energy markets to normal, and it gets a moderate weight.
A broader slowdown: 3 or 4
Three or four falls occurred in 6 of the 33 years, often in years of regional strain, such as 1998 and 1999. It would need a shock that hits several energy-importing economies together, such as the adverse oil scenario in the October outlook, which on that outlook's own estimate still leaves world growth positive. The weight is below the historical share of about 18%, because no member is projected to shrink in 2027.
A synchronised world downturn: 5 or more
Only three years since 1993 reached five or more, and two of them, 2009 and 2020, were global shocks that hit 14 and 17 members. Its adverse scenarios include a sharp fall in investment in artificial intelligence, and it puts the chance of world growth below 2% in 2027 at about 10%, so the weight is low, but the history shows how quickly the count can jump.
What to watch
- Quarterly GDP: the euro area's second flash estimate for the third quarter of 2026, due on 13 November 2026, with France's figure; Japan's first estimate for the third quarter, due on 16 November 2026; and the second US estimate, due on 25 November 2026.
- Business surveys: whether the euro area and Japanese indices stay above 50 as energy costs stay high.
- Energy and shipping: how quickly energy markets and shipping normalise in early 2027, the main assumption behind the baseline.
- Interest rates: the next US and euro area rate decisions in December 2026, and whether the US yield curve stays positive.
- Forecast updates: the main international financial institution's January 2027 update, and its April 2027 outlook with the first full figures for 2026.
Figures as of late October 2026. Probabilities are analytical estimates by the Ghaf Centre for Studies & Research.