OG-UK overlapping generations

OG-UK

Shows where the UK economy settles under a tax reform, and the year-by-year path there.

  • UK
  • Calibrated
  • Runs locally
  • Long-run steady state
Targets met by construction; no outcome benchmark. Evidence ONS and OBR aggregates are calibration targets. The nearest external benchmark, OBR Working Paper 22 Table 5.1, has not yet been run.

Good for

  • Long-run effects of a tax reform on GDP, work, saving and capital
  • Year-by-year transition paths, and eight industry sectors
  • Wages back to the microsimulation for a dynamic score

Not for

What it returns

A stylised reform, not announced policy, against baseline on the committed transition path.

1pp on the basic rate from 2027–28

Reform against baseline in 2030, the last committed transition year. Source: olg/figures/tpi_data.json, the OG-UK run of 19 March 2026 (OG-UK a314659), regenerated and CI-checked by olg/figures/make_showcase.py.
AggregateBaseline 2030Reform 2030Change
GDP (£bn, current prices)3,627.93,622.8−5.0 (−0.14%)
Tax revenue (% of GDP)32.06%32.39%+0.34pp
Consumption (% of GDP)61.33%61.19%−0.15pp
Investment (% of GDP)18.50%18.30%−0.20pp
Government consumption (% of GDP)20.95%21.16%+0.21pp
Debt (% of GDP)95.32%95.38%+0.05pp

The showcase reform

Reform
+1pp basic rate (20% → 21%) from 2027–28
Baseline
OBR November 2025 EFO
Window shown
2026–2030 of a 60-period solve
OBR outturn / forecast Reform +1pp basic rate Outturn / forecast boundary (2023) Reform start (2027)
ONS outturn / forecast Reform +1pp basic rate Outturn / forecast boundary (2023) Reform start (2027)

Government consumption rises because the extra revenue finances it in this run. Every share divides by the reform run’s own GDP (the debt measure).

Model-consistent magnitudes, not forecasts. No ground truth exists to score them against, and the committed run predates OG-UK’s July 2026 labour-tax-rate fix (limits).

With the microsimulation

Joined both ways: statute in, long-run wages back.

Reform OG-UKlong-run wages PolicyEngine microsimulationsecond run Dynamic scorelocal only

The dynamic score comes from a second microsimulation run (every join in the suite).

Model cardinputs, outputs, horizon, data, access
Answers
Where the UK economy settles under a tax reform, and the year-by-year path there
Inputs
A PolicyEngine parameter reform, the same dict the microsimulation takes; calibration parameters
Outputs
GDP, work, saving, capital, wages, interest rates, debt
Horizon
Long-run steady state; the package also solves transition paths
Data
OG-UK packaged calibration inputs, with tax functions from PolicyEngine microdata
Uncertainty
None quantified; sensitivity analysis not yet comprehensive
Access
Local CLI steady state and Python package; not hosted (Code)
Runtime
A steady-state solve takes many minutes and transition paths far longer (not timed)
Joins the microsimulation
Both ways: tax functions in, long-run wages and labour supply out (the join above)
Status
Calibrated; no outcome benchmark; research prototype
Cannot answer
Short-run forecasts; fast hosted custom scenarios; price-level effects (the model is real)