UK growth

UK output now, and the model's near-term path.

Where it stands

Latest readings from the stored snapshot.

Real gross domestic product
2026Q3 · vintage 2026-11-16
£m, chained volume measure
ONS · ABMI
£750.0bn 2026Q3 · ▲ +0.6pp y/y
Monthly GVA index
Sep 2026 · vintage 2026-11-12
index, chained volume measure, seasonally adjusted
ONS · ECY2
103.3 Sep 2026 · ▲ +0.3%
Real business investment
2026Q2 · vintage 2026-09-30
£m, chained volume measure, seasonally adjusted
ONS · NPEL
£83.5bn 2026Q2 · ▲ +1.8%
How these are measured

Levels exactly as the snapshot stores them — chained volume measures in £m and an index, not growth rates. The year-on-year figure quoted on this page is derived from these same stored levels rather than typed in: 749,999.0 in 2026Q3 against 735,710.0 in 2025Q3 is 1.9%.

What the models see

The model view, with its limits beside it.

BoE SVARforecasts UK GDP · round 2026-11-12 · data to 2026Q2

UK real GDP, % change on a year earlier

UK real GDP: ONS outturns, then the BoE SVAR forecast, median with 68% and 90% bands ONS outturns from the 2026-11-16 snapshot run 2024Q4 to 2026Q3 (1.9% at the end); fan chart of UK real GDP, % change on a year earlier, from the BoE SVAR round generated 2026-11-12: median 1.88% in 2026Q4 and 1.16% in 2027Q4; the quarter-by-quarter values are in the table under the chart.
  • ONS outturn, snapshot 2026-11-16
  • BoE SVAR median, round 2026-11-12
  • 68% band (raw)
  • 90% band (raw)

Source: ONS · BoE SVAR round, 12 Nov 2026

Details

Numbers by quarter

BoE SVAR year-on-year real GDP growth: median and predictive ranges, from the archived round. Bands are raw model quantiles; the fans on the track record are re-calibrated to measured coverage.
QuarterMedian68%90%
2026Q41.88%1.07% to 2.70%0.54% to 3.29%
2027Q11.57%0.42% to 2.54%−0.22% to 3.28%
2027Q21.28%0.10% to 2.53%−0.75% to 3.34%
2027Q31.16%−0.09% to 2.50%−0.74% to 3.38%
2027Q41.16%−0.09% to 2.48%−0.76% to 3.32%

Sources

Source: ONS, ABMI, growth derived from the stored level, snapshot 2026-11-16; BoE SVAR round generated 2026-11-12, current_forecast.json

Predictive validation: Weak. No demonstrated skill against a random walk with drift on any variable at any horizon. Validation →

Its limits
  • Status: partial replication of Brignone and Piffer (2025): GDP decomposition consistent with the paper; CPI global-shock share about 8.5pp short; no forecast skill against a drifting random walk.
  • It cannot answer statutory policy reform effects.
  • Uncertainty is posterior 68% and 90% intervals, which under-cover in back-testing (coverage study).
  • Estimated on 1992Q1-2025Q1 but conditioned on 2026Q2 data, so the estimation sample stops well short of the quarters being forecast.
How it movedChange since the previous reading.
  • Real gross domestic product +0.6pp on the year-on-year rate against the prior quarter (2026Q2)
  • Monthly gross value added index +0.3% against the prior month (Aug 2026)
  • Real business investment +1.8% against the prior quarter (2026Q1)
Run it yourselfThe commands and MCP tools behind this page.

The forecast above is one command. It re-estimates and re-identifies the model: minutes per estimation and identification run.

pe-macro forecast --horizons 8   # GDP and CPI, medians with 68% and 90% bands
pe-macro summary   # headline replication results, instant, from committed files

Over MCP the same forecast is the forecast_uk tool, the identified shocks behind it are latest_shocks, and get_model_status returns the limitations quoted above as structured data rather than prose. Connect a client →

Data and sourcesSeries, coverage and the append-only snapshot files.
Each value is read from the vintage file linked here.
SeriesPublisherCoverageSnapshotsThis vintageNext release
Real gross domestic productONS · ABMI1955Q1 – 2026Q342026-11-16.json22 Dec 2026
Monthly gross value added indexONS · ECY21997-01 – 2026-0952026-11-12.json11 Dec 2026
Real business investmentONS · NPEL1997Q1 – 2026Q222026-09-30.json22 Dec 2026

Quarterly national accounts are revised for years after first publication, which is why the snapshot date matters as much as the observation period.

Snapshots are append-only, so every number here is reproducible. Data store →