PolicyEngine Macro
compare models
Which model answers your question?
Start with the question and country, then check the comparison. The suite covers both the UK and US; each model's country scope is stated explicitly. Compare method and run surface here, then use the Papers and Validation tabs for the supporting evidence.
start here
Choose by question and country.
Household taxes, benefits, or distributional effectsUK + US · PolicyEngine
US monetary or fiscal shocksUS · frb-us
Selected near-term UK fiscal or macro scenariosUK · obr-macro
Long-run work, saving, and capital effectsUK · psl-og
Drivers of UK GDP and inflationUK · boe-svar
Check evidence and limitations before using a result. Review validation.
compare
Five model classes, side by side.
| OBR macroeconometric obr-macro | Bank of England structural VAR boe-svar | Federal Reserve model frb-us | Overlapping generations psl-og | PolicyEngine microsimulation pe-microsim | |
|---|---|---|---|---|---|
| Question class | Selected near-term UK fiscal and macro scenarios | What drove the UK economy at the latest model-data vintage; short-run forecasts and revision narratives | US macro dynamics: monetary and fiscal shocks, quarter by quarter | Long-run incentives: labour supply, saving, the capital stock over decades | What a specific family pays and receives; distributional effects |
| Method | Structural macroeconometric emulator — Gauss–Seidel over 372 equations per quarter | Bayesian structural VAR — posterior sampling with zero + sign restrictions naming the shocks | Large-scale macroeconometric model — 284 equations solved under VAR expectations | Dynamic general equilibrium (OLG) — steady-state root-find and transition-path iteration | Static tax-benefit microsimulation — direct rule evaluation, no behavioural response |
| Country | UK | UK | US | UK (OG-UK) | UK & US |
| Data vintage | OBR Economic and Fiscal Outlook, March 2026 | ONS, BoE and FRED series — coefficients estimated through 2025Q1; conditioned through the 2026Q1 data edge | The Fed's model.xml and LONGBASE database, April 2026 | ONS, OBR and BoE national accounts as calibration targets; OG-UK 0.3.2 (from GitHub) | UK & US statute; enhanced FRS 2023–24 microdata for UK population runs |
| Evidence class (numbers →) | Validated for selected scenarios; published anchor is not out-of-sample validation | Validated replication for selected outputs | Validated software replication; substantive comparisons are approximate | Research prototype; calibrated counterfactual with no ground truth | Deterministic rule calculation; population estimates add survey uncertainty |
| Reform scoring | Yes — shocks to exogenous model variables; static-costing bridge from PolicyEngine reforms | No — the baseline/conditioning member: it reads the economy reforms are scored against | No — shock experiments only (funds-rate or fiscal shocks under VAR expectations); score_reform refuses model="frbus", as no PolicyEngine reform bridge exists | Yes — a PolicyEngine policy translated into estimated tax functions | Yes — a PolicyEngine parameter + a new value, applied directly to households |
| Cannot establish | Arbitrary reform incidence or borrowing through the current adapter | The causal effect of a specific statutory reform | PolicyEngine reform effects, anticipated-policy paths, or MCE scenarios | A short-run forecast or independently validated reform effect | GDP, inflation, interest rates, or macro feedback |
| Run surface | Hosted (CLI, MCP, Python API) | Hosted (CLI, MCP, Python API) | Hosted (CLI, MCP, Python API) — raw shocks only | Local only | Hosted |
| Typical runtime | Seconds to minutes per scenario | Minutes per full estimation + identification run | Seconds to minutes per simulation | ~17+ min per steady-state solve (two per score); hours for transition paths | Sub-second per household; minutes for population runs |
Comparison rule. These models use different horizons,
baselines, mechanisms, and output definitions. Related results should
be interpreted together, not added, averaged, or ranked. The CLI and
JSON result schema retain units, time basis, provenance, and an explicit
comparability label for this reason.