UK fiscal reform, quarter by quarter.

The OBR's model is a computational representation of the UK National Accounts — first built by HM Treasury in 1970 and jointly maintained by the OBR and the Treasury since 2010. The OBR publishes its full equation listing, but as EViews source that few can run. This project transpiles that listing to Python and solves it: 372 published equations in 17 thematic groups over 636 variables — 383 solved by the model (roughly 29 econometrically estimated, 216 identities, and 138 calibrated or technical relationships) and 253 exogenous inputs, which are the judgements and assumptions you get to change.

It is a structural macroeconometric model in the HM Treasury / Cowles-Commission tradition. Where the OLG model derives behaviour from optimising households and firms, this one reproduces the estimated relationships the official forecaster actually relies on.

For forecasting rather than one-off shocks, the emulator fits add-factors over a base window of 2024Q1–2025Q4, holds them constant, and projects to 2027Q4 with exogenous inputs extrapolated (seasonal cash series at their trailing annual mean). Scenario solves run twelve quarters, 2025Q1–2027Q4. The anchored horizon extends to 2031Q1 (GDP reproduced to 0.29% at 2031Q1); anchored unemployment (LFSUR) is unreliable beyond 2027Q4, drifting to 0.9% against the EFO's 4.1% by 2031Q1.

The package installs straight from GitHub — the equation listing, the March 2026 EFO forecast tables, and the ONS snapshot ship inside it, so there is no data step. A reform is one call: pick an exogenous variable, apply a shock, read the deviation from baseline.

pip install git+https://github.com/PolicyEngine/obr-macroeconomic-model

from obr_macro import run_reform

# £5bn/yr government spending shock (£1.25bn per quarter)
results = run_reform(
    name="Fiscal Stimulus",
    var="CGG",        # government consumption (exogenous)
    shock=1250,        # £1.25bn per quarter, in £m
    periods=12,
)
print(results[["period", "delta_gdp_bn", "pct_gdp"]])
# ΔGDP ≈ £1.25bn in the impact quarter, near-flat thereafter —
# under the demand closure the impact multiplier is ~1 by construction

two solves, so about a minute. run the five bundled scenarios with charts via python -m obr_macro.reform_analysis — a £5bn spending rise, a 5pp corporation-tax cut and rise, £10bn public real investment (~£3bn nominal CGIPS per quarter), and a £10bn austerity cut. Charts land in outputs/.

Reforms enter through the 253 exogenous inputs; results come out of the 383 the model solves. The test is simply whether a variable has its own equation. The most useful levers:

the full searchable glossary — all 636 codes with descriptions, ONS series and equations — is the model package's variable and equation references.

The emulator's designed job is the second half of the official workflow: microsimulation supplies the direct Exchequer effect, the macro model supplies the economy-wide feedback. A static-costing bridge wires PolicyEngine's microsimulation into the emulator so a statutory reform can be scored end to end.

Worked example — 1p on the basic rate of income tax from April 2026, scored 2026–2030. PolicyEngine puts the direct yield at £6.46bn in 2026, rising to £7.38bn by 2030; HMRC's June 2025 ready reckoner puts 1p at £6.9bn in 2026–27 rising to about £8.2bn by 2028–29, so the static costing sits inside the official range, toward its lower end. Converted to a positive quarterly revenue-costing path of £1.62bn to £1.85bn, the HHDI_ADDFACTOR adapter applies the corresponding negative held add-factor to household disposable income. Propagated through the consumption function, GDP falls 0.020% (−£0.14bn) on impact in 2026Q1, deepening to 0.058% (−£0.40bn) by 2027Q4 on the March 2026 baseline — the sign and order of magnitude the OBR's own indirect-effect conventions imply.