Emulator of the OBR macroeconometric model

OBR emulator

How far to trust each configuration, the check against outturns, and the limits.

How far to trust it, and where it stops

Anchored to the EFO, GDP drifts to +0.42% and consumption to +0.69% above it by 2027Q4; unemployment follows the EFO by construction. Weak when run free. Errors are MAPE against the EFO path.

What was checked, against what, and the result
CheckAgainstResultVerdict
Solver convergenceTolerance in every quarter, 15 demo shocks12 of 12 quarterspass
Anchored fit, real GDPOctober 2026 EFO, 2025Q1–2027Q4; consumption 0.40%0.24% MAPEby construction
Held add-factors, real GDPOctober 2026 EFO, 2026Q1–2027Q4; consumption 0.85%; 7 of 8 in band, two of them identities on EFO inputs0.63% MAPEweak
Free-running, real GDPOctober 2026 EFO, 2025Q1–2027Q4; the emulator’s closure, not OBR forecast error4.24% MAPEweak
ONS outturnsq/q growth, 2025Q2–2026Q1, ONS release of 12 Nov 2026; the EFO missed 2026Q1 by 0.25ppmisses up to 0.27ppweak
Income tax multiplierOBR 0.3 on impact, fading; q1, q4 and q12 in current prices; size close, profile reversed0.10 → 0.21 → 0.37weak
Spending multiplierOBR 0.6, fading; the shock lands in the GDP identity1.01 → 1.04fail
Bank Rate +1ppBank of England −0.2% to −0.35% (1999); sign right−0.06% peakfail

Evidence by dimension

The registry’s categorical judgement, and what would change it
DimensionLevelWhat would change it
Implementation fidelitymoderateEvery behavioural equation run; inactive channels scoped
Predictive validationweakRolling-origin tests vs benchmarks and outturns
Identificationnot applicableStays not applicable
Policy counterfactualsweakFiscal shocks vs official costings and multipliers
UncertaintyweakSensitivity envelopes for judgement and closure
Vintage reproducibilitymoderateThree historical EFO vintages reproduced end to end
About this emulatorIndependent; not an official OBR product

An independent implementation of the OBR's published model code and forecast data. Not produced, maintained or endorsed by the Office for Budget Responsibility; not official OBR estimates.

Three configurations, three tests

Every reform score uses the anchored configuration. The other two are checks.

How each one scoresFull figures, windows and the free-running caveat
Anchored
GDP MAPE 0.45% (0.447%) to 2031Q1, every quarter converged at 0e44ec8; CI gates tolerance, the expenditure identity and signs (step 1, unemployment).
Held add-factors
Re-run on 18 Nov 2026 (docs/forecasting_framework.md at 0e44ec8): GDP 0.63%, consumption 0.85%, scored on 2026Q1–2027Q4 from the EFO’s own 2026Q1 values (scorecard).
Free-running
Raw dynamics with GDP endogenous, not OBR forecast error. Every computed line is in step 3.
Bands and gatesWhat counts as in band, what CI enforces

in band rates within 1.0pp · net balances within 1.5% of GDP · levels within 10% MAPE.

gated CI enforces the anchored tolerance and the multiplier band; the HMRC comparison and the outturn backtest are computed for the working paper and refreshed by hand.

1

Anchored against the EFO

GDP +0.06% to +0.42%; consumption +0.10% to +0.69%

% deviation from the October 2026 EFO

OBR emulator: anchored baseline vs October 2026 EFO, quarterly deviation Line chart. Quarterly percentage deviation of the anchored emulator from the published October 2026 EFO, 2025Q1 to 2027Q4. Real GDP ranges from +0.06% to +0.42% (mean absolute percentage error 0.24%); consumption from +0.10% to +0.69% (mean absolute percentage error 0.40%). Continuous integration pins each mean absolute percentage error to within 25% of its measured value, so a drift in either direction fails; it does not test single quarters.
Current October 2026 EFO baseline. Quarterly deviation from the EFO, 2025Q1–2027Q4. CI pins each series’ MAPE to within 25% of its measured value (GDP 0.242%, consumption 0.398%), so a drift either way fails (test_anchored_reproduces_efo_published_aggregates). Computed from papers/obr-macro/figures/fig_anchored_data.csv, regenerated from the October 2026 detailed forecast tables at obr-macroeconomic-model 0e44ec8 on 19 Nov 2026. validation/check_obr_tables.py recomputes the table below and the free-running endpoints from the same CSVs.
Anchored levels against the October 2026 EFO, £bn/qtr, chained volume on the EFO’s own price base (not today’s ONS reference year, so levels differ from the current ABMI: read the deviations)
OursOfficialDeviation
Real GDP, 2025Q1704.2703.2+0.15%
Real GDP, 2027Q4727.2724.1+0.42%
Consumption, 2025Q1431.0430.0+0.24%
Consumption, 2027Q4443.0439.9+0.69%
About this stepnotes

Anchored GDP and consumption are one test, not two: the £m deviation is identical in all twelve quarters, because every other term in the demand identity is exogenous or pinned.

2

Against ONS outturns

Misses up to 0.24pp in 2025 and 0.27pp in 2026Q1

The only test against reality, and mostly a test of the inherited EFO vintage.

Real GDP growth, % q/q

OBR emulator: quarterly real GDP growth — emulator vs EFO Nov 2025 vs ONS outturn (% q/q) Bar chart with outturn markers, percentage quarter-on-quarter real GDP growth for the four quarters the November 2025 EFO study scored. 2025Q2: emulator 0.15, EFO 0.28, ONS 0.01; 2025Q3: emulator 0.14, EFO 0.20, ONS 0.18; 2025Q4: emulator 0.25, EFO 0.27, ONS 0.01; 2026Q1: emulator 0.37, EFO 0.39, ONS 0.64. Against the ONS release of 12 Nov 2026 the emulator misses the 2025 outturns by up to 0.24 points (2025Q4); it misses the 0.64 per cent 2026Q1 outturn by 0.27 points, and the November EFO it inherits misses it by 0.25. Emulator and EFO from papers/obr-macro/figures/fig_outturn_data.csv; ONS outturns from data/vintages/ons/uk_gdp_cvm/2026-11-16.json. Coordinates: value v maps to y = 258 - v * 331.4 on a 0 to 0.7 axis.
  • emulator
  • EFO Nov 2025
  • ONS outturn
% q/qemulatorEFOONS
2025Q20.150.280.01
2025Q30.140.200.18
2025Q40.250.270.01
2026Q10.370.390.64
Emulator and EFO on the November 2025 vintage, the working paper’s study; the live baseline is anchored to the October 2026 EFO. Outturns are the ONS release of 12 Nov 2026.

Source: OBR emulator · ONS, 12 Nov 2026

Source details

Source: emulator and EFO, papers/obr-macro/figures/fig_outturn_data.csv; ONS outturns, the pinned vintage file data/vintages/ons/uk_gdp_cvm/2026-11-16.json. The paper scored against the 14 May 2026 first estimate (0.1, 0.2, 0.2 and 0.6).

Four quarters of outturn

2025 outturns
misses of 0.14pp, 0.04pp and 0.24pp; on the 14 May 2026 first estimate they were within 0.06pp
2026Q1
missed by 0.27pp; the EFO missed by 0.25pp
Own contribution
0.02–0.13pp gap between model rows
Vintage
November 2025 EFO, kept on purpose
About this step3 points

Forecast versus outturn. Comparing one forecast vintage with another tests agreement, not accuracy.

  • The November EFO the emulator inherits misses 2026Q1 the same way, so this mostly tests that vintage.
  • The vintage is kept because changing it would erase the historical forecast being tested.
  • ONS quarterly estimates are themselves revised.
3

Free-running

Weak: 8 of 10 non-trivial lines in band

What the equations do with GDP made endogenous and no add-factors. The miss is the emulator’s, not OBR forecast error.

How much of the OBR emulator scorecard the model actually computes Two stacked bars. Of 21 headline variables in the OBR emulator calibration scorecard, 11 are actually computed by the model and 10 are passthrough, held at the OBR published value and therefore scoring zero error trivially. Of the 11 computed, 8 are fair or good, 1 is an identity, 2 fit badly. 9 of the 11, or 82 per cent, land within band, but one of those is an accounting identity that closes over passthrough inputs, so 8 of the 10 non-trivial computed variables are in band. The worst are company profits 12.88 per cent and RPI inflation 2.32 percentage points.
  • fair or good 8
  • identity 1
  • fit badly 2

Only 8 of the 10 non-trivial computed variables are in band.
One of the 9 in-band passes is an accounting identity over passthrough inputs.
bands: rates ±1.0pp · net balances ±1.5% of GDP · levels ≤10% MAPE

Raw calibration against the October 2026 EFO, default closure. Company profits, the worst fit, is regression-gated, not tuned.

Source: OBR emulator

Source details

Source: docs/calibration_scorecard.md, obr-macroeconomic-model at 0e44ec8.

The free-running scorecard and the October 2026 re-anchoring

What the miss measures. Not OBR forecast error.

  • In the OBR's published model GDP (GDPM) is an exogenous input, so that model cannot miss GDP.
  • The gap measures this emulator's closure swap, which makes GDP endogenous, and inputs the OBR does not publish.
  • So reform deltas are scored against the anchored baseline, never the raw one.

Company profits, the worst line.

  • The miss does not trace to households' operating-surplus constant OSHH: its ONS anchor applies in the raw solve too.
  • Under the default closure profits miss by 12.88%. Under the published closure the investment deflator drifts in the solve and the miss reaches 22.47% (nominal GDP limit).
  • The line is regression-gated, not re-tuned: tuning it would be fitting to the answer.

Household taxes, held on their data path.

  • Households’ taxes on income and wealth (TYWHH) sum ONS components that are not seasonally adjusted against a seasonally adjusted total. The gap put a first-quarter dip into raw household income (−10.6% in 2025Q1, −12.5% in 2026Q1).
  • From model pin 0fe04c9 an identity residual closes the line, as the OBR’s own listing does: GDP 4.25% → 4.05%, consumption 7.24% → 6.92%, household income 4.30% → 3.08%, profits 12.32% → 11.66%. Since #74 (pin 0e44ec8): 4.24%, 7.23%, 3.10% and 12.88%.
  • The fall is the removal of a data-basis artefact, not better fit: the household-income score no longer tests that identity.

Scope.

  • The table under this note lists all 11 computed lines, under both closures.
  • Every figure is the October 2026 vintage, after the upstream OSHH ONS anchor and the recovery of eight ONS series a failed download had dropped (355 series in all).

Real GDP, £bn per quarter

OBR emulator: real GDP level, anchored vs free-running vs the October 2026 EFO (£bn/qtr) Line chart of quarterly real GDP levels in billions of pounds, 2025Q1 to 2027Q4. The published October 2026 EFO path rises from 703.2 to 724.1. The anchored emulator tracks it closely, running from 704.2 to 727.2 (mean absolute percentage error 0.24 per cent, recomputed here from the plotted series). The free-running emulator, de-seeded and with no add-factors, contracts from 697.4 to 674.3 — a gap that widens to 50 billion pounds, 4.24 per cent mean absolute percentage error over the horizon. Free-running and EFO paths from papers/obr-macro/figures/fig_free_running_data.csv; anchored path from papers/obr-macro/figures/fig_anchored_data.csv. Coordinates: value v in billions maps to y = 292 - (v - 660) * 2.95 on a 660 to 740 axis; quarter i of 12 maps to x = 58 + i * 60.545.
  • anchored (0.24% MAPE)
  • EFO October 2026
  • free-running (4.24% MAPE)
Current October 2026 EFO baseline. The anchored path stays within 0.5% of the EFO, drifting up to +0.42% by 2027Q4; the same equations free-running (de-seeded, no add-factors) contract away from it. Computed from papers/obr-macro/figures/fig_free_running_data.csv and fig_anchored_data.csv, both regenerated at obr-macroeconomic-model 0e44ec8 on 19 Nov 2026.

Vintage. Re-anchored from the November 2025 EFO to March 2026, then to the October 2026 EFO (28 Oct 2026; obr-macroeconomic-model #69).

  • Every headline on this page is computed on the October 2026 baseline; only the outturn backtest retains November 2025.
  • Reform effects are differences between structurally identical runs, so the re-anchoring leaves the static costing untouched.
  • Anchoring add-factors are fixed in baseline and shocked runs and cancel in the difference.
  • Second-round GDP is quoted as a ratio to revenue raised, not a per-quarter percentage, so it sits beside the OBR’s published multipliers (worked example).
Free-running error against the October 2026 EFO, 2025Q1–2027Q4: the 11 of 21 headline lines the model computes (the other 10 are passthrough). In band: 8 of 10 non-trivial lines under the default closure, 9 of 11 with the identity; 8 of 11 under the published closure (scorecard at 0e44ec8)
Default closurePublished closureDefault closure band
Real GDP4.24%4.28%in band
Consumption7.23%7.30%in band
Business investment7.34%14.19%in band
Household income3.10%4.04%in band
Real household income3.30%4.25%in band
Company profits12.88%22.47%outside
Unemployment rate0.62pp0.62ppin band
RPI inflation2.32pp2.32ppoutside
Current account0.91% of GDP1.33% of GDPin band
Trade balance0.41% of GDP0.41% of GDPin band
Employment0.00%0.00%identity
The costing it is fed1p on the basic rate, against HMRC

This compares the microsimulation’s static costing with HMRC’s ready reckoner. The emulator takes that costing as given, so it tests the input, not the emulator: a benchmark, not a validation.

The figures, the interpolated 2028–29 bar and why the 2030 row has no counterpart are on the microsimulation’s validation tab.

Static revenue, £bn a year

OBR emulator: 1p on the basic rate, ours vs HMRC ready reckoner (£bn/yr) Grouped bar chart in billions of pounds per year. PolicyEngine's static costing of a 1 percentage point rise in the UK basic rate of income tax, against HMRC's Direct effects of illustrative tax changes ready reckoner, June 2025 vintage. For the basic rate +1pp, 2026–27 group, ours is 6.55 against HMRC’s 6.90, a deviation of -5.1%. For the basic rate +1pp, 2028–29 group, ours is 7.01 against HMRC’s 8.20, a deviation of -14.5%. Both microsimulation figures are scored runs on the hosted server (policyengine 6.11.1): £6.55bn in 2026 and £7.01bn in 2028.
  • ours
  • HMRC
Against HMRC's Direct effects of illustrative tax changes (June 2025 vintage). Ours are scored runs on the hosted server (policyengine 6.11.1).

Source: Hosted runs

Source details

Source: basic_rate_costing.json; the working paper’s panel B used the earlier 4.x figures (6.46 and an interpolated 6.92).

What it cannot be used for

Each limit opens to its evidence. Only income tax and NICs carry behaviour.

The spending multiplier is close to 1 by construction, against the OBR’s own published 0.6.

Why the multiplier is close to 1Spending lands straight in the GDP identity

The OBR’s 0.6 is for day-to-day public services and welfare spending. A spending shock lands straight in the GDP identity, so £1.25bn a quarter in returns about £1.27bn to £1.30bn of GDP a quarter (figures).

  • From model pin 0e44ec8 the shock is nominal-consistent by default (obr-macroeconomic-model #75): nominal spending moves by the same cash amount, so the government deflator is unchanged and borrowing rises by the cash cost, +£1,336m at q1 and +£1,419m at q12. Receipts have no equation in the listing and do not respond.
  • The real-only shock, the earlier default, spilled into the investment deflator and moved business investment by +£295m on impact; that artefact is gone (about −£2m).

Channels that do not transmit

VAT barely reaches GDP by default

A 1pp VAT rise is £2.2bn a quarter on VREC. Receipts and borrowing move one for one, but GDP barely moves, against the OBR’s published VAT multiplier of 0.35, fading.

GDP moved per pound of VAT raised, current prices
q1q4q12
Default0.0130.0260.043
Opt-in price pass-through0.1750.3450.638
  • Default. CPI is exogenous in the OBR’s listing, so VAT cannot pass into prices. The tax lands on company profits, and only the withdrawals that follow profits reach households.
  • Opt-in price pass-through. price_passthrough=True (obr_shock or the model’s run_reform) passes the tax into CPI through the listing’s own cost-push weights, with the retail markup held at baseline. CPI rises about 0.4%.
  • The opt-in is a labelled judgement overlay, off by default. Its match with 0.35 after a year is not a calibration, and the effect keeps building where the OBR’s fades.
  • In default runs only income tax and NICs transmit through household income. Their benchmark is a different lever: 0.3 on impact, fading (Overview).
  • Both rows are 12-quarter runs from 2025Q1; every quarter of each converges.

    Source: OBR emulator, demo run of 18 Nov 2026

    Source details

    Source: results/demo_shocks.json at obr-macroeconomic-model 0e44ec8 (rows vat and vat_passthrough), run on 18 Nov 2026 at commit a822045 and reproduced exactly at 0e44ec8 by Actions run 44424337333, every quarter converged. The default row is identical to the run in obr-macroeconomic-model #69.

Government investment does not transmit by default

Under the default closure CGIPS does not reach GDP, against the OBR’s published multiplier for capital spending of 1.0. An opt-in closure passes it through as an identity. Neither is a validated capital-spending multiplier.

  • Default. Total fixed investment (IF) moves exactly zero at any shock size: it has no live equation in the published listing, so the chain never reaches the GDP identity. The model warns on every such run.
  • Business investment (IBUSX) moves only as an offset. The listing makes it a residual (IBUS = IF − GGI − …, l.84), so higher CGIPS displaces it one for one in real terms.
  • £1.5bn a quarter for eight quarters from 2026Q1: the offset is −£1,408m at q4 and q8 (−£1,404m on average), linear to −£5,634m and −£5,633m for £6bn.
  • GDP falls by £3.8m at q4 and £7.3m at q8 for £1.5bn, the consumption that follows profits. Unemployment does not move.
  • The offset is accounting, not a multiplier, and must never be quoted as one.
  • Opt-in investment closure. investment_closure=True switches on the listing’s commented-out IF identity (l.86), so real government investment enters GDP one for one. A nominal £3bn a quarter for 12 quarters from 2025Q1 moves real GDP by +£3,075m at q1, +£2,901m at q4 and +£2,903m at q12.
  • That is identity passthrough at about 1.0 per real pound, which the result labels mechanical_passthrough: no taper, no crowding-out, and the nominal-GDP pin is off under this closure. The OBR’s 1.0 tapers to zero.
  • The default figures are measured at model pin 0e44ec8 under the nominal-GDP pin; every quarter of the baseline and shocked solves converges at every size.

    Source: OBR emulator, 18 Nov 2026

    Source details

    Source: default, obr-macroeconomic-model at 0e44ec8, results/site_figures.json (cgips_size_sweep), run on 18 Nov 2026 at commit b56fe3b. Opt-in, results/demo_shocks.json at 0e44ec8 (cgips_investment_closure), run on 18 Nov 2026 at commit a822045 and reproduced exactly at 0e44ec8 by Actions run 44424337333, every quarter converged.

Whole blocks do not respond to policy

This is about shock response, distinct from the baseline passthroughs.

  • In default runs the unemployment rate, employment, retail prices, exports and imports move exactly zero, not approximately zero, across every lever tested: government consumption, the household tax bridge, corporation tax, Bank Rate, exchange rate.
  • There is no Okun channel, no Phillips channel and no trade channel. A 6% sterling appreciation moves exports and imports not at all.
  • The second round is consumption and nothing else; under the investment closure it is business investment instead.
  • Pinned per block in tests/test_sign_conventions.py.
Passthrough channels

Exports, imports and CPI are exogenous here, held at the OBR value. They score 0.00% error because they are copied, not modelled. 10 of 21 scorecard lines are passthroughs, labelled as such.

Two household-income equations never fire

The profits → dividends → household-income channel is inert in every published figure.

  • The listing's only bare log() left-hand sides, log(HHTFA) and log(NDIVHH), never execute: inputs MAJGDP and CORP are absent from the databank.
  • A 5pp corporation-tax rise moves FYCPR by −£413m by quarter 12, −£191m on average over 12 quarters (results/site_figures.json), and NDIVHH by exactly zero.
  • Reviving the channel needs a CORP series.
Bank Rate: sign right, GDP response about 3–6× below the Bank of England’s range

A sustained +100bp lowers GDP (figures on the Overview); a 100bp cut raises it by a similar amount (about £134m to £519m).

  • The response runs through household spending alone. There is no monetary rule; unemployment, prices and trade do not move.
  • A 1pp rise held for four quarters lowers GDP by about 0.06% at peak, in the published closure and under the nominal-GDP pin (obr-macroeconomic-model #50).
  • The Bank of England’s published range for the same experiment is −0.2% to −0.35% (the benchmark).
  • The default holds the mortgage rate on its EFO path. The model’s opt-in monetary_conditioning=True moves it with Bank Rate at the listing’s own pass-through and gives −0.149% at peak, still below the range. It is a labelled overlay on run_reform, off by default and not exposed by obr_shock.
  • The benchmark dates from 1999. A sustained rise never turns within 12 quarters, and company profits move by an artefact of the nominal-GDP pin, so do not read profits from rate shocks.
  • Each figure is the difference between a baseline and a shocked solve under the default nominal-GDP pin; every quarter of both converges.

    Source: OBR emulator, 18 Nov 2026

    Source details

    Source: obr-macroeconomic-model at 0e44ec8, measured 18 Nov 2026 (results/site_figures.json, commit b56fe3b; four-quarter peaks from results/demo_shocks.json, rows bank_rate and bank_rate_conditioned).

Results set by construction or simplified

Income tax: similar size to the OBR, opposite time profile

The household route is the one lever with behaviour in it. Its GDP effect per pound of revenue lands near the OBR’s published figure (worked example) but rises over the horizon where the OBR’s decays. The working paper calls the profile inverted.

Corporation-tax closure converges slowly

The published TCPRO → cost-of-capital → investment chain: a controlled scenario closure, not a calibrated supply block.

  • A sustained +5pp rise heads toward a ~£0.96bn/q plateau (figures to q12), but the root is slow (~0.958/quarter): a 12-quarter run captures only ~40% of the full effect, so every result reports its plateau_fraction.
  • The deviation has reached a steady state since the anchor add-factors moved to log space (August 2026).
  • The business-investment equation is reconstructed from a line the OBR publishes commented out; MSGVA, PIF and PIRHH are frozen against uncalibrated feedback.
  • The response size rests on allowance present values estimated from statute and the OBR's own gilt assumption, not invented constants. Full expensing makes it ~0.42× its former magnitude.
Approximated add-factors

Recent corrections are averaged and held flat. The OBR's judgemental, quarter-by-quarter add-factors are not reproduced.

No behavioural micro

Aggregate equations only. Distributional questions belong to PolicyEngine.

Hosted runs, closure and vintage

Anchored unemployment is the EFO’s, by construction

Anchored, unemployment (LFSUR) follows the EFO path: 5.07% at 2027Q4 and 4.17% at 2031Q1, the same as the EFO. It is not a forecast.

  • From model pin 0e44ec8 the forecast 16+ population follows the path the EFO’s own unemployment, employment and participation imply (obr-macroeconomic-model #74). Before, it was held flat at 56.4m, so unemployment drained to 1.0% by 2031Q1.
  • Run free, the unemployment rate misses the EFO by 0.62pp on average (step 3).
  • Fiscal and rate shocks still leave it exactly zero (channels that do not transmit).
Hosted results depend on the model pin and window

This repository pins the hosted service to 0e44ec8, on the October 2026 EFO. The server takes a new pin when it redeploys; every result records the one it ran at in provenance.source_revision.

  • At 0e44ec8 the nominal-GDP pin is the default (obr-macroeconomic-model #54–#58), the ended EU budget payments are held at zero (#65), the household-tax ratio is reported on both bases (#66), household taxes are held on their data path (#71), the two opt-in routes are labelled (#70), the forecast population follows the EFO (#74) and the spending shock is nominal-consistent (#75).
  • So obr_shock and score_reform(model="obr") results at this pin converge in every quarter and are scored around the anchored baseline, like the figures on these pages.
  • The default window differs from the worked example: score_reform scores five years, from 2026Q2 to 2031Q1; the example is a 12-quarter run from 2025Q1.
  • A hosted score_reform takes over ten minutes on the default window, mostly the five yearly microsimulation costings. Run it through start_job, or pass years=1 or 2 (macro #243).
Nominal GDP needs the pin to converge

The listing writes one identity twice (lines 337 and 688), to set the investment deflator (PIF) and nominal GDP (GDPMPS), so the published closure pins neither down.

  • PIF drifts, nominal GDP follows it, and company profits (FYCPR), a residual, follow nominal GDP (obr-macroeconomic-model #42). Under that closure no quarter converges.
  • The nominal-GDP pin restores the OBR’s own line 690. It is the default at this pin, the Overview figures use it, and with it every quarter converges.
  • The free-running scorecard on this page uses the pin: company profits miss by 12.88%, with 8 of 10 non-trivial lines in band, 9 of 11 counting one identity (scorecard at 0e44ec8, October 2026 EFO). The published closure gives 22.47% with 8 of 11 (obr-macroeconomic-model #52).
Vintage

October 2025 equation listing, October 2026 EFO, current-vintage ONS series. Where the ONS has revised history the identities do not close exactly; the slack lands in the add-factors.