Document RF nowcasting methodology (#1400 #1179)

changelog.d/1400.md

@@ -0,0 +1 @@
+- Add a standalone documentation page summarising the Resolution Foundation's nowcasting methodology (LSO 2025 appendix: wage-floor and spillover modelling, NLW age extensions, PIPR-based rent uprating, UC-statistics caseload anchoring, four- to five-year horizon) and pointing back at PolicyEngine's own assumptions for comparison.

docs/book/assumptions/rf-nowcasting-methodology.md

@@ -0,0 +1,114 @@
+# Resolution Foundation nowcasting methodology
+
+This page summarises the Resolution Foundation (RF) microsimulation
+nowcasting methodology, drawing on the appendix to RF's [Living Standards
+Outlook 2025][lso] and the methodological notes in [Black holes and
+consolidations][bhac]. It is a paired reference for PolicyEngine UK's
+own assumptions — see [`growthfactors.md`](./growthfactors.md) for the
+PolicyEngine growth rates and [`nowcasting-comparison.md`](./nowcasting-comparison.md)
+for the high-level differences.
+
+## Data foundation
+
+RF's nowcast starts from the same UK household microdata that
+PolicyEngine uses — the **Family Resources Survey (FRS)** and its
+**Households Below Average Income (HBAI)** companion — combined with a
+range of administrative aggregates used to anchor outputs to published
+totals.
+
+## Earnings growth
+
+RF projects earnings forward year-by-year, with explicit minimum-wage
+handling rather than uniform indexation:
+
+1. **Wage floor**. In every projected year RF lifts anyone whose hourly
+   wage falls below their age-appropriate **National Minimum Wage /
+   National Living Wage** up to that floor. Floors come from the
+   announced NMW/NLW schedule and (beyond 2025-26) are assumed to grow in
+   line with average earnings.
+2. **Spillover effect**. For workers whose wage sits just above the
+   floor, RF applies a partial uplift to capture wage compression at the
+   bottom of the distribution. The magnitude is calibrated to historical
+   evidence on NMW spillovers (RF cites the Low Pay Commission research).
+3. **NLW age extensions**. RF models the **2024-25 extension** of the
+   NLW to 21- and 22-year-olds and, **provisionally**, the **2029-30
+   extension** to 18- to 20-year-olds. These show up as discontinuities
+   in the modelled wage floor for the affected cohorts.
+4. **Uniform uprating beyond the floor**. Earnings above the floor are
+   uprated equally for employees and the self-employed using OBR/BoE
+   wage projections (constrained to ONS outturn data once published).
+
+## Housing costs
+
+RF separates rent and mortgage interest:
+
+- **Private rents** are uprated using the ONS **Price Index of Private
+  Rents (PIPR)** rather than a generic CPI-housing measure.
+- **Social rents** are uprated according to the published social-rent
+  policy in force in each year (typically CPI + 1 percentage point, with
+  exceptions and caps where the government has announced freezes or
+  variations).
+- **Mortgage interest** flows are projected using OBR base-rate paths
+  and the FRS-recorded interest portion of mortgage payments.
+
+## Benefits
+
+RF takes the legislated/announced statutory uprating of working-age and
+pension-age benefits as given, layering on:
+
+- **Caseload anchoring** to current **Universal Credit statistics** for
+  in-payment caseloads, so simulated UC numbers track DWP outturn rather
+  than drift with FRS take-up.
+- **Take-up modelling** based on historical receipt rates by household
+  type, used to scale the eligible population down to expected caseload.
+- **Discretionary CoL payments** modelled as announced (and reset to
+  zero once the policy ends).
+
+## Inflation indices
+
+RF uses the OBR EFO inflation paths (CPI, CPIH, RPI) as headline
+indices, supplemented by ONS-published outturns once available. Beyond
+the EFO forecast horizon RF holds rates flat at their final forecast
+value rather than constructing a separate long-run rule.
+
+## Calibration
+
+RF aligns outputs to a mix of HMRC, DWP and ONS aggregates — most
+prominently income tax receipts, NI receipts, UC caseload and
+expenditure, and Pension Credit caseload. The exact target list and
+weights vary across reports.
+
+## Horizon
+
+The Living Standards Outlook publishes detailed nowcasts about **four
+to five fiscal years ahead** of the current outturn year. Longer-run
+sensitivity analyses are run separately and are not part of the headline
+nowcast.
+
+## Where this differs from PolicyEngine
+
+For a side-by-side comparison see
+[`nowcasting-comparison.md`](./nowcasting-comparison.md). The three
+biggest divergences are:
+
+1. **Wage-floor / NLW handling** — RF models the floor and spillovers
+   explicitly; PolicyEngine uprates earnings by the OBR average-earnings
+   index and only invokes the NLW when a reform query asks for it.
+2. **Housing-cost projection** — RF uses PIPR for private rents and
+   social-rent policy for social rents; PolicyEngine uprates rents using
+   `policyengine-uk-data` local-authority targets and ONS rent indices.
+3. **Calibration targets** — PolicyEngine carries finer-grained
+   local-authority targets for council tax and rents that the public RF
+   nowcast does not surface.
+
+## References
+
+- Resolution Foundation, [Living Standards Outlook 2025][lso] — methodology
+  appendix.
+- Resolution Foundation, [Black holes and consolidations][bhac] — fiscal
+  context and short methodology summary.
+- Low Pay Commission, [National Minimum Wage research reports](https://www.gov.uk/government/publications/national-minimum-wage-low-pay-commission-research-reports) — empirical basis for the spillover assumptions.
+- ONS, [Price Index of Private Rents (PIPR)](https://www.ons.gov.uk/economy/inflationandpriceindices/bulletins/priceindexofprivaterentsuk/latest).
+
+[lso]: https://www.resolutionfoundation.org/publications/living-standards-outlook-2025/
+[bhac]: https://www.resolutionfoundation.org/publications/black-holes-and-consolidations/