The Misery Index

What moved it

Each basket class and what it contributed to the essentials figure, in index points. These add up to the printed total.

• Transport+3.4
• Food and non-alcoholic drink+0.8
• Electricity, gas and other fuels−0.7

Essentials 3.5 + unemployment 4.9 = 8.4

What the City expects

The independent forecasters HM Treasury surveys, on the benchmark index — headline CPI plus unemployment — because that is the measure they actually forecast. It was 8.6 in 2025; the consensus has it at 8.5 in 2026 and 7.4 in 2027. The band is the range between the highest and the lowest forecast, and it is the point: on this release it is wider than the distance between the consensus and today.

HM Treasury, Forecasts for the UK economy, 19 August 2026— the average, median and range of independent forecasts, under OGL v3.0. CPI is Q4 on Q4 and unemployment is the Q4 rate, which is why this chart is annual: the monthly index above is a different measurement and drawing a forecast onto it would mix the two. The OBR's own forecast is shown beside the consensus, never inside it — it is the government's number, not one of the independents. We will score each consensus here once its year has closed.

In office

What the index read while each government was in office, and what it read in the month of every general election. The highest average of any premiership of a year or more is Harold Wilson's at 23.4; the lowest is Keir Starmer's at 6.7. Liz Truss's 2 months average 24.5 — higher than any of them, and too short to be a record.

At every general election

By Prime Minister

By Chancellor

Government support, by how bad it felt

Every month since January 1989 in which we have both a reading and a poll — 448 of them — grouped by the index, against the mean poll share of the party in government. Across the whole span the correlation is −0.26 — weak and negative: how the economy feels is a real pull on a government's standing and nowhere near the whole of it.

A month counts for whoever held the office for most of its days, so a handover month is never counted twice. Rows marked BENCHMARK predate the headline index (January 1989) and are read on the Okun benchmark — CPI inflation plus unemployment — which runs from February 1971. The two are never averaged together in one number. An office holder with no complete published month yet shows no average. These figures say what the index read while someone was in office; they are not a claim about what any government did to it. Government support is the mean of every GB Westminster poll published that month (Wikipedia, CC BY-SA), which is one calendar-month mean throughout rather than the live poll-of-polls: a series that changed technique half way along would show that change as a change in the country. A month with no polling is left out, never interpolated.

Sources

Every series the index is built from, and when each one next publishes. All Crown copyright, used under the Open Government Licence v3.0.