Methodology — how the metro stress reading is built

By Ranjan Gupta · YATU framework · Last updated May 25, 2026

Every metro page on this site is anchored in source-verified data, structured by a single diagnostic triad, and labeled with a stress tier that means something specific. This page documents what each piece is and how it's calculated, so you can read the pages with full transparency about their underlying logic.

Stress tiers — what each one means

The four tiers describe the YATU framework's reading of the metro's current institutional condition. They are not credit ratings, not insolvency forecasts, not collapse predictions. They are a structural read of where on the Earth-to-Air trigon transition arc the metro currently sits.

Important: None of the 20 metros in the 2026 dataset is at "collapse tier." The YATU framework reads what's happening as correction — institutional forms returning to a scale the new era can sustain — not as failure. The full framework reading is at The Compelled Correction · Institutional Form.

How tier moves work — escalation discipline

Tier moves are not single-cycle reactions to one news event. They are framework-level claims about a metro's cumulative institutional-form-stress signature, and they require the full claim-validation pipeline before the framework will move a metro from one tier to another.

What a single signal-scan cycle does — and does not

The autonomous signal-scan cron operates every other day. Each cycle surfaces new news-items to the per-metro pages — HIGH / MEDIUM / COUNTER-SIGNAL entries with framework reading per the per-metro empirical signature. The cron has no authority over tier assignments. A metro does not move from Tier 2 to Tier 3 because one bad cycle. It moves because across four to six consecutive cycles the signature pattern across multiple dashboard dimensions consistently reads as the next tier's signature, not the prior tier's.

Criteria for tier escalation

• Tier 1 → Tier 2. Stress signatures appearing across two or more of the seven dashboard dimensions, sustained over four or more consecutive cycles.
• Tier 2 → Tier 3. Signatures across four or more dimensions, sustained four or more cycles, with at least one of them structural (a K-12 endgame vote, a fiscal-cliff with no available bridge, an anchor-employer cluster, a federal-credit contraction landing) rather than cyclical.
• Tier 3 → Tier 4. Signatures across five or more dimensions sustained six or more cycles, with at least two structural, and counter-signals failing to offset. Chicago's pattern — pension obligation + Boeing departure + CTA reductions + CPS structural deficit + governance volatility, with no offsetting expansion of comparable magnitude — is the framework's reference signature for what Tier 4 looks like.
• Tier 2.5 (special — Nashville). When the framework reads conflicting category signatures simultaneously — major contraction in one institutional category alongside clear expansion in another, of comparable magnitude — neither Tier 2 nor Tier 3 captures the reading honestly. The half-tier flags split-form metros. Nashville's K-12 (MNPS deficit) against its private-school surge + healthcare-corporate expansion is the reference.

What holds a metro in place — the counter-signal discipline

Counter-signals are reasons not to escalate a metro even when contraction signals are accumulating. Atlanta's Authority Brands +390-job Sandy Springs headquarters announcement, DFW's Moody's outlook upgrade to STABLE, Florida Citizens' 2026 standard-homeowner rate reductions, Miami's Citizens reform context, the City of Austin's Fitch AAA upgrade — these are not noise. They are the form-rotation evidence that some institutional forms within a metro thrive under the same era pressure that compresses others. Surfacing them honestly is what allows the framework to read where the form-stress concentrates and where it does not.

Why Ranjan makes tier calls, not the cron

Tier assignments are framework-level interpretations of cumulative signatures, not numerical rollups. They must pass the claim-validation pipeline: framework-strengthening test, source verification, counter-signal honoring, the three-cornerstone triplet check (yatubook tone + YATU framework concept + Yogananda Gita lineage), and substrate-as-authority discipline. This is a deliberate decision by the framework author at the planning-session level, not an automated output. The cron supplies evidence; the framework reads. That separation is what keeps the tier readings credible when the empirical evidence accumulates faster than the framework discipline can integrate it.

The three-component diagnostic triad

Every metro's K-12-layer stress signal is read through the same three-component diagnostic. Each component alone is a stressor most districts could absorb. The three together produce the fiscal squeeze districts structurally cannot.

Component 1 — Enrollment decline

Below-replacement birth rates. Family formation delayed by housing costs. Migration to lower-cost regions slowing because lower-cost regions are no longer low-cost. None of these is reversible by school policy. Districts are responding to a demographic and economic reality they did not create.

Component 2 — Fixed bonded debt

School districts that grew during 2010-2022 issued bonds at low interest rates with debt-service obligations running 15-40 years forward. The interest payment is the same whether campuses are full or half-empty. As enrollment declines, per-student debt service rises mechanically. This is not a flaw in how any specific district managed itself — it's the structural mismatch between bonds-designed-for-one-demographic-trajectory and the demographic-trajectory now actually in motion.

Component 3 — School-choice expansion channel

Voucher programs (Texas TEFA, Arizona ESA, Georgia Promise, Florida FES, Tennessee ESA, NC Opportunity Scholarship) redirect education funding to follow the student rather than the building. They do not move bonded debt off the district's balance sheet. The combined math — fewer students per district, same bond service, less revenue per remaining student — produces the fiscal squeeze.

The framework reads the school-choice channel as one of the operational mechanisms through which the broader institutional-form contraction is moving faster, not as cause of the contraction. The public-district math would shift even without expanded school choice. Parents who chose alternatives were responding to real and reasonable concerns about educational fit for their specific children. The framework holds both honestly together — the structural mathematics is real AND the parent choice is lawful and often well-considered. This is the framework's Claim 32 anti-exculpation discipline applied at the parent-choice layer.

Source-verification pipeline (three gates)

Every numerical claim that appears on a metro page passes three verification gates before being published. This is the operational discipline that lets you trust the citations.

Gate 1 — URL liveness, source allowlist, date

URL must return 200 (HEAD request). Source domain must be on the allowlist (Tier A = auto-eligible; Tier B = always human-reviewed; Tier C = rejected outright). Article publish-date metadata extracted from <meta property="article:published_time"> must be within the relevant time window.

Gate 2 — Headline-source match + numerical-claim grounding

Article full text is fetched. Haiku model verifies: does the article content directly support the proposed headline? Every numerical claim in the headline must appear in the source article text (verbatim or as recognizable paraphrase). If either check fails, the item is rejected, not published.

Gate 3 — Cross-corroboration for significant items

Items meeting significance triggers — layoffs ≥100 jobs, new closure announcements, credit-rating changes, housing-data revisions >3% MoM, voucher-policy changes, federal-funding actions affecting research universities or healthcare systems — require an independent second source before publishing. If only one source is available, the item goes to human review even when the source is Tier A.

Source tiers — allowlist

Tier A — auto-eligible after probation period

• Government: .gov domains, district transparency portals, state comptroller releases, US Census, BLS, NIH grant databases
• Major national: NYT, WaPo, WSJ, Reuters, AP, Bloomberg, Axios, Financial Times
• Established secondary local journalism (with editorial review): Community Impact, Chalkbeat, Houston Chronicle, Atlanta Journal-Constitution, Boston Globe, San Antonio Report, Charlotte Observer, Tampa Bay Times, KERA, KUT, WBEZ, WBUR, KQED, GBH, MPR
• Specialty trade publications: Bond Buyer, Higher Ed Dive, K-12 Dive, Manufacturing Dive, Healthcare Dive, Banking Dive, EdSource, EdNC

Tier B — always human-reviewed

• Local newsletters and blogs with editorial standards (Hoodline, Patch, Local Profile, Citizen Portal, BizWest, Dallas Express)
• Aggregators with original attribution preserved
• Industry-specific outlets (Inside Higher Ed, The Real Deal, GeekWire)
• Government-adjacent research organizations (Civic Federation, Empire Center, GBPI, JLBC)

Tier C — rejected outright (never published)

• Reddit posts (used for signal/discovery only, never as source citation)
• Personal blogs without editorial review
• AI-generated content farms
• Substack posts (unless from a known established voice — case-by-case exception list)
• Paywalled-headline-only sites where the article body isn't accessible
• Press-release-only sites without independent journalism

Data sources used

• Housing: Zillow Research Data (free monthly CSVs), Redfin Data Center (free monthly), St. Louis Fed FRED (free, API), regional MLS where publicly accessible (NTREIS, HAR, ARMLS, SABOR, NWMLS, GLVAR, MIAMI Realtors, REBNY, etc.), M&D Real Estate, local journalism analysis
• Schools: NCES (National Center for Education Statistics), state DOE portals (TX TEA, CA CDE, NY SED, FL FLDOE, etc.), Texas Bond Review Board (TX gold-standard), district transparency portals, Chalkbeat / EdSource / Education Week, local journalism
• Employment: WARN notices (state labor departments — most authoritative US layoff source), BLS State and MSA Employment, layoffs.fyi (tech-specific), Challenger Gray & Christmas monthly reports, TrueUp Layoffs Tracker, company press releases, 10-Q/10-K filings
• Higher education: US Dept of Ed Financial Responsibility Composite Scores, IPEDS (Integrated Postsecondary Education Data System), Higher Ed Dive closure tracker, Moody's higher-ed ratings, SHEEO, Hechinger Report
• Local government fiscal: Moody's / S&P / Fitch / KBRA credit rating actions, city budget documents, comptroller transparency portals
• Voucher / school choice: State comptroller / treasurer offices, NCSL, EdChoice, state-specific portals (Texas Comptroller TEFA, AZ DOE ESA, etc.)

Every per-metro research file at /data/metroplex/ cites the specific source URL behind each numerical claim. Where data is unverifiable, the file explicitly marks "DATA GAP" rather than guessing.

Update cadence

• Canon pages (the /metro/[name] AEO pages and the stress dashboard tiles): quarterly updates. Major events between updates flagged via the every-other-day signal scan.
• Every-other-day signal scan (the "News this week" section on each per-metro page): launches June 8, 2026. Scheduled task runs every other day at 6am, surfaces verified news items per metro from the prior 48 hours, with one-sentence framework annotation per item.
• Raw research files (the /data/metroplex/[name] source-of-truth files): quarterly bulk refresh. Snapshot date visible on every file.
• Framework canon (claims #119-122 at /canon Section 17): forever-stable claim numbers; corrections recorded as version increments in master-index.json revision history.

Audit + accuracy guardrail

Monthly: random sample 20 published items. Manually verify (1) headline accurately reflects source content, (2) all numerical claims are sourced, (3) source is on Tier A or B, (4) framework annotation matches canon. Calculate accuracy rate.

Threshold: if accuracy < 95% in any monthly audit, the daily news task reverts to human-review-for-all until the failure mode is diagnosed and corrected.

Per-item audit trail (archived, never deleted): unique ID, run timestamp, source URL, headline as published, framework annotation as published, Haiku confidence score, gates passed, second-source corroboration result, approval timestamp, any subsequent corrections. This means retrospective review is always possible.

Corrections policy

Found an error? Reach Ranjan directly via either channel:

• Email: ranjan.gupta@jyoling.com — include the specific page, the specific claim, and the source you believe is correct.
• X / Twitter: @jyolingapp — public or DM, whichever you prefer.

All corrections are logged + archived for retrospective audit. Correction acknowledgments aimed for within 24 hours during business days; corrections published within 72 hours if substantiated. Persistent errors trigger a public correction note added to the relevant page's footer.

If you're disputing the framework reading rather than the underlying data — that's a different conversation, and the same channels work. The framework operates under nistrai-guṇya middle-position discipline; substantive disagreement is welcome.

What this methodology is not

It is not a forecast that any specific metro will collapse, get worse, or recover. The framework reads structural mechanism, not future timing of specific events.

It is not a credit rating. Moody's, S&P, Fitch, and KBRA assess creditworthiness. The YATU framework reads institutional-form structural alignment with the era-shift. These can correlate but they measure different things.

It is not partisan. The structural condition the framework names operates regardless of which political coalition is in power in any specific metro or state. Texas TEFA, Arizona ESA, Georgia Promise are mechanisms the framework reads structurally, not endorsements or attacks.

It is not advice. The pages document what's happening and what the framework reads as the structural mechanism. They do not tell you whether to move, where to send your kid to school, whether to sell your house, or which job to take. Those are your decisions.

Found an error or have a correction? Reach Ranjan at ranjan.gupta@jyoling.com or @jyolingapp on X · all corrections logged + archived for retrospective audit