Why do official figures for the same thing differ?
The same country can have two different official unemployment rates, and the UK publishes two inflation rates. Here are the figures side by side and the definitions that explain the gap.
Unemployment rate, 2025: national statistics vs ILO estimate
| Country | National statistics (12-month average) | ILO modelled estimate | Gap |
|---|---|---|---|
| U.S. (2024) | 4% | 4.02% | 0 pp |
| UK | 4.9% | 4.75% | +0.2 pp |
| Canada | 6.9% | 6.91% | 0 pp |
| Australia | 4.2% | 4.09% | +0.1 pp |
| South Korea | 2.8% | 2.68% | +0.1 pp |
Why they differ
- National figures come from each country’s own monthly labour force survey. Vernumo averages the 12 months of the year.
- The ILO figure is a modelled estimate built to make countries comparable. It adjusts for differences in age limits, survey design and definitions, and it is revised when new data arrive.
- Some national series are seasonally adjusted and some are not; for a full-year average this matters little, but monthly figures can differ more.
- Neither figure is wrong. Use the national figure to follow one country month by month, and the ILO figure to compare countries.
UK: CPI vs CPIH inflation (both from the ONS)
| Month | CPI | CPIH | Gap (CPIH − CPI) |
|---|---|---|---|
| September 2026 | 3.5% | 3.6% | +0.1 pp |
| August 2026 | 3.1% | 3.3% | +0.2 pp |
| July 2026 | 2.9% | 3.1% | +0.2 pp |
| June 2026 | 2.6% | 2.8% | +0.2 pp |
| May 2026 | 2.8% | 3% | +0.2 pp |
| April 2026 | 2.8% | 3% | +0.2 pp |
| March 2026 | 3.3% | 3.4% | +0.1 pp |
| February 2026 | 3% | 3.2% | +0.2 pp |
| January 2026 | 3% | 3.2% | +0.2 pp |
| December 2025 | 3.4% | 3.6% | +0.2 pp |
| November 2025 | 3.2% | 3.5% | +0.3 pp |
| October 2025 | 3.6% | 3.8% | +0.2 pp |
CPIH adds the costs of living in your own home (owner occupiers’ housing costs) and council tax. When those housing costs rise more slowly or faster than other prices, the two rates move apart.
Sources
- ILO modelled estimates via World Bank World Development Indicators (CC BY 4.0)
- BLS, ONS, Statistics Canada, ABS, Statistics Korea (KOSIS), Eurostat