Replicate RecessionAlert indicator: Recession Forecasting with the Anxiety Index [PREMIUM] | RecessionAlert

Objective

Follow-up to #911 for https://recessionalert.com/recession-forecasting-with-the-anxious-index/ pinned at sha256:91b3792b9d00ecf469c3331c00ed642ab51bb5588deacb3eaec868ffb1cba053. Review only the durable offline snapshot; do not live-crawl or invent thresholds. Publish a fail-closed indicator/strategy specification with page and asset SHA256 provenance, disclosed semantics, native and effective spans, timing/cadence, revision limitations, and an explicit no-go disposition where evidence is insufficient.

Approach

1. Verified offline page cache and SPF fixtures (page 91b3792b…, spf_prob.xlsx 71b1e506…, spf_recess.csv f5ada538…) without live crawl or DB mutation.
2. Wrote fail-closed spec research/findings/specs/recessionalert-anxiety-index-1555.json defining recessionalert-anxiety-index as an Other-indicators sell-side recession warning (anxious >= 28.2), with explicit no-go for portfolio allocation rules and AnxiousIndex.gif redistribution (page_url asset rows=0).
3. Locked provenance/spans: native SPF RECESS2 1968Q4→2026Q2 (231 quarters); effective monthly 1968-11→2026-06 (692 months) via release months 2/5/8/11; never-revised; ~6-week publication lag.
4. Validated causally via research.recessionalert.anxiety_index: AUC 0.927 full / 0.944 pre-2012 (page claim 0.94), accuracy 0.877 at 28.2; values change only on release months; extra +1m lag AUC drag ≈0.010–0.012. Labeled adjacent diagnostics at thresholds 25/30 retained as inferred sensitivity only.
5. Decision=promote for Other indicators (already registered from qm-09tp.24); strategy overlay remains no-go.
6. Confirmed canonical producers: registry research/fixtures/buy-signals-v1.json, history docs/site-data/signals/history.json, page docs/other-indicators.html; shipped this changelog via python3 -m research.reports.results.

Files / code / data changed

• research/findings/specs/recessionalert-anxiety-index-1555.json — fail-closed indicator specification for issue 1555
• research/extensions/1555-*.json — pipeline node evidence (intake through publish)
• research/results/1555.md — this durable task note
• docs/site-data/results.json, docs/results.html, docs/results/1555.html — rebuilt changelog
• Reused read-only: research/recessionalert/anxiety_index.py, research/fixtures/buy-signals-v1.json#recessionalert-anxiety-index, docs/site-data/signals/history.json, docs/other-indicators.html, data/recessionalert.sqlite, SPF fixtures

Validation

• python3 -m research.recessionalert.anxiety_index — exit 0, output digest sha256:2a890062d547fe45d89fd3df259467aeba1ee6c51ff649631ac556af6a531a74
• pipeline temporal.causality check-1 — passed (release months {2,5,8,11}, extra-lag delta stable)
• pipeline core.behavioral_proof check-2 — passed
• Adjacent diagnostics: threshold 25/28.2/30 and +1m lag on full monthly history (documented vs inferred labeled)
• No live crawl; no secrets; no hand-edit of generated canary HTML

Results

Fail-closed specification published for the pinned Anxiety Index page. Promoted Other-indicators signal recessionalert-anxiety-index retains disclosed threshold 28.2 with SHA256 provenance for page and SPF assets. Portfolio buy/sell/allocation rules and uncrawled chart GIF bytes are explicit no-go. Distinct from issue 1355 (3Q2012 update narrative without threshold).

Links

• Spec: research/findings/specs/recessionalert-anxiety-index-1555.json
• Page: https://recessionalert.com/recession-forecasting-with-the-anxious-index/ (cached sha256:91b3792b9d00ecf469c3331c00ed642ab51bb5588deacb3eaec868ffb1cba053)
• Runner: research/recessionalert/anxiety_index.py
• Registry: research/fixtures/buy-signals-v1.json
• Other indicators: docs/other-indicators.html
• Prior replication: research/results/qm-09tp.24.md