Connect every model page to the microsimulation scoring pipeline - #142
Merged
Conversation
- Pipeline strip (reform → pe-microsim → obr-macro → score) on /score and /models - /pe declares itself the hub and links to /score - /olg body now describes the two-way micro coupling incl. og+microsim - /svar, /frb-us, /us-hank callouts route to /pe and /score with the reason no bridge exists - Homepage gains a Score-a-reform strategy card; /connect gains the score CLI command and a /score link; /economy ties the baseline to scoring and the forecast record Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
- score.py picks the newest boe-svar round instead of hardcoding 2026-07-21; §03 names the displayed round and explains the surge-sample mean reversion behind the sticky CPI path; §02 states the error-sign convention - us-hank: 'It cannot' callout becomes a bullet list - connect: centered client tabs, full-width hero lede, compact Python picker - forecasts charts widen to 1080px Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Integration: - EconomicAssumptions.from_frbus_result / from_hank_result: annual-mean wage(+hours/labor) deviations become a pre-tax earnings factor with the OG overlay's double-counting invariant and plausibility gates - frbus_shock_incidence, hank_shock_incidence, svar_inflation_incidence: reform-free automatic-stabilizer incidence via _pe_population_incidence; score_reform refusals unchanged, now pointing at the incidence tools - hank_shock surfaces the pre-tax real wage w and labor N IRFs - svar_inflation_incidence prices the SVAR-vs-EFO CPI gap through a curated slice of CPI-uprated benefits (validated run: +0.94pp gap -> GBP 0.09bn in 2028, gains in deciles 2-5) - frbus_payload/hank_payload two-process escape hatches (8GB hosts jetsam-kill the combined solve+microsim process); CLI flags to match - 30 unit tests; full suite 245 passed Website: - DEFINE-UK added as model 07 (experimental, pre-validation; unlicensed upstream run unmodified at a pinned commit via PolicyEngine/define-uk-model) - /svar, /frb-us, /us-hank gain incidence examples; /pe and /score describe the macro-to-micro return direction; counts six -> seven Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…ero steady state raises instead of omitting; stale connect comment Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…; experimental label) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
/score was only reachable through in-page links; it is now a top-level nav destination (and breadcrumb section) generated by site_nav.py. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…odel pages Per review: no separate score page. The /models scoring section (now anchored #score, open by default) absorbs the run-it command and the uncertainty caveat; the worked mechanics were already on /pe and /obr. All internal links repoint to /models#score, /score 301-redirects there, nav item removed, sitemap updated. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
/papers and /validation index pages were retired when /models became the hub (main); the disclosures the tests guard moved to obr/code, obr/validation, svar/validation, and the paper landings. Same guarantees, new locations. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Score-a-reform is the first strategy card and the chooser band comes before the outturns table, matching the site's center of gravity: PolicyEngine microsimulation and policy/economic analysis first, economy monitoring second. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
… and model page Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…parisons Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…ility note beside the machine-precision claim The April 2026 LONGBASE misses realized 2026Q1 core PCE by -1.7pp while tracking GDP, unemployment, and the funds rate within ~0.3pp — now stated next to the replication-fidelity metrics, sourced to the model repo's weekly outturn backtest. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What this PR does
This branch makes the PolicyEngine microsimulation the hub of the suite, in both directions, and adds DEFINE-UK as the seventh model.
Integration — macro→micro incidence bridges. A shared incidence-bridge pattern in
integration/src/policyengine_macro/core.pyand the extendedassumptions.py:frbus_shock_incidenceandhank_shock_incidenceturn a macro shock's pre-tax wage/hours (FRB-USpl/lhp) or wage/labor (HANKw/N, now surfaced from the GE Jacobian) deviations into an employment-income overlay on the US microsim — automatic-stabilizer incidence, explicitly experimental and NOT reform scoring;score_reformcontinues to hard-refusefrbus/hank.svar_inflation_incidenceprices the SVAR-vs-EFO CPI gap through a curated slice of CPI-uprated UK benefits (validated run: +0.94pp 2027 gap → £0.09bn 2028, largest relative gains in deciles 2–5). CLI + MCP surfaces added (20→23 tools), withfrbus_payload/hank_payloadescape hatches for splitting the macro solve and microsim across processes on memory-constrained machines. 563 lines of new incidence tests; suite: 245 passed / 70 skipped.Website. Pipeline strip (reform → pe-microsim → obr-macro → score) on /score and /models; every model page now states its microsim connection (or exactly why none exists); homepage gains a Score-a-reform card; /define added as model 07 (experimental, pre-validation — upstream DEFINE-UK code is unlicensed, so PolicyEngine/define-uk-model fetches it at a pinned commit and runs it unmodified; no hosted surface, no quotable numbers until replication passes); counts updated six → seven with site_contract enforcement.
Forecasts. /forecasts model-vs-official now dynamically shows the latest archived SVAR round (was hardcoded to the superseded 2026-07-21 round), names the round displayed, explains the surge-sample mean-reversion behind the sticky CPI path, and states the error-sign convention. Scoring/archive/ingest checks pass; nothing under
forecasts/rounds/touched.Review process. Three rounds of independent review agents (code + website + final verification): all findings fixed in-branch; final verdict merge-ready. The only failing tests (6 in
test_site_contract.py, missingpapers/index.html) pre-exist on main.🤖 Generated with Claude Code