How the American Distress Index Works

The Five Domains

The ADI draws on 10 federal data series organized into five domains. Every domain carries the same weight, and within a domain every member series carries the same weight. There are no fitted weights anywhere in the chain — the weight rule is structural, so no era of the data gets to vote on which dimension matters most.

How the Score Is Calculated

Each input is measured against its own full quarterly history. A composite of 50 means the inputs, on average, sit higher than in half of their own recorded quarters since 2005.

1. Aggregate to quarters. A series-quarter exists only if at least one raw observation is dated within it; the quarterly value is the mean of the observations present. No interpolation, no carry-forward.
2. Rank each series against its own history. Every quarterly value is converted to a Hazen percentile within that series’ full record: (average_rank − 0.5) / n × 100, ties averaged. One yardstick — no baseline window, no regime split, no winsorization, no caps. Inputs where lower means more distress are inverted before ranking.
3. Average within domains. A domain’s score is the mean of its member percentiles present that quarter.
4. Average the domains. The composite is the mean of the five domain scores. A quarter publishes only when all five domains have at least one member present — the engine refuses to publish a partial quarter.

The composite is a mean of input percentiles, not itself a percentile of quarters. The composite’s own rank within the published record is computed and published separately, so neither number gets quoted as the other.

The Five Bands

The 0–100 scale splits into five equal bands. The thresholds are uniform, not calibrated to any narrative period, and the labels apply to the national time axis only — states and counties get ranks, never these labels. A band label always publishes alongside the literal reading sentence.

Technical Details

The Record

The published series runs 2005-Q1 through 2025-Q4 — 84 quarters. Its peak is 85.3 in 2009-Q3, the depth of the financial crisis. Its low is 15.7 in 2021-Q4, when pandemic-era support programs left household buffers unusually full. Nothing in the method points at those quarters; they emerge from ranking each series against its own record.

Leading Indicator Research

Beyond tracking current conditions, the American Default research pipeline systematically tests all pairwise indicator combinations for statistically validated leading relationships — cases where one indicator consistently precedes another by multiple quarters. These are documented as structural projections: historical patterns observed across multiple economic crises, not forecasts. The pipeline operates on individual indicators, separately from the composite index.

Five-Stage Validation Pipeline

Every candidate pair must survive all five filters before being classified as "validated." This pipeline tested 64,897 raw pair-lag combinations and produced 7 fully validated relationships.

1. Cross-correlation on raw levels — Identify pairs where one indicator's past values correlate with another's future values, corrected for multiple testing using the Benjamini-Hochberg false discovery rate (FDR α = 0.05).
2. Cross-correlation on first-differenced series — Repeat on quarter-over-quarter changes. This eliminates spurious correlations driven by shared trends (e.g., two series that both trend upward over time).
3. Multi-crisis validation — The relationship must hold during at least two of three economic crises: the 2001 recession, the 2008 financial crisis (GFC), and COVID-19. A relationship that only works in calm periods is not useful.
4. Granger causality — The leader must statistically Granger-cause the follower (p < 0.05), meaning past values of the leader improve the follower model beyond what the follower's own history provides.
5. Out-of-sample validation — Calibrate on 2000–2012 data, validate on 2013–2025. The relationship must hold in data the model never saw during calibration (OOS r > 0.3).

Validated Leading Relationships

7 indicator pairs passed all five filters:

Active structural projections based on these relationships are tracked on the Structural Outlook page.

Data Sources and Update Schedule

All ADI data comes from public federal sources. No proprietary data, no paywalled sources, no models. Everything is independently verifiable.

Most inputs update quarterly after source data lands; claims and unemployment series arrive faster and are aggregated to quarters. The composite recomputes whenever its inputs refresh, and a quarter publishes only once all five domains have data. Analysis of each release lands under Analysis.

Attribution requirement: FRED data use is subject to the Federal Reserve Bank of St. Louis Terms of Use. Full source attribution appears on every indicator detail page.

Variable Definitions

The American Distress Index tracks 92 economic indicators across nine categories. Each indicator is linked to its detail page with full time-series data, methodology notes, and downloadable datasets.

The Published History (2005–2025)

The table below shows the last published quarter of each year. The full quarterly series, domain scores, and member percentiles ship in the ADI JSON API and the CSV download.

How to Cite This Methodology

If you use the American Distress Index or its methodology in academic work, policy analysis, or journalism, please cite as follows.

APA Format

BibTeX

@techreport{adi_methodology_2026,
  title     = {American Distress Index: Methodology and Composite Scoring},
  author    = {Kilburn, Ross},
  year      = {2026},
  institution = {American Default Research},
  url       = {https://americandefault.org/methodology/},
  note      = {Five-domain equal-weighted composite; each input ranked
               against its own quarterly history (Hazen percentile);
               published 2005--present}
}

For individual indicator citations, use the citation tools on the press page or download .bib / .ris files from any indicator detail page. For a print-friendly summary of current ADI data, see the ADI one-pager.

Interactive Tools

For national methodology, use the ADI methodology page. For state-level analysis, the Zip Code Stress Test shows your state's distress score and foreclosure protections.