Research & Data Access
The American Distress Index is a free, open dataset tracking U.S. household financial distress across 88 economic indicators. This page provides everything researchers need to access, reproduce, and cite ADI data. No API keys, no registration, no restrictions on academic use.
Dataset Overview
The ADI composite is a 0–100 score built from 5 domains, each capturing a distinct dimension of household financial distress. Every member series is converted to a Hazen percentile within its own full quarterly history since 2005. Each domain score is the mean of its members' percentiles. The composite is the mean of the 5 domain scores.
| Domain | Weight | Member series | Measures |
|---|---|---|---|
| Delinquency | 20% | Delinquency Rate on Single-Family Residential Mortgages (90+ days); Delinquency Rate on Credit Card Loans; Delinquency Rate on Consumer Loans (ex credit card); Auto Loan Serious Delinquency Rate (90+ days) | Share of borrowers behind on mortgage, credit card, consumer, and auto loans. |
| Default & Legal | 20% | Charge-Off Rate on Credit Card Loans; Charge-Off Rate on Single-Family Residential Mortgages | Charge-offs and the foreclosure-stage outcomes that follow late payments. |
| Debt Burden | 20% | Household Debt Service Ratio | Required debt payments as a share of household income. |
| Labor | 20% | Unemployment Rate; Initial Unemployment Claims (SA) | Unemployment and new jobless claims. |
| Safety Net & Buffer | 20% | Personal Saving Rate | The savings cushion households hold against a bad quarter. |
Each domain carries the same weight: 20% of the composite. A composite of 62 means the index's inputs, on average, sit higher than in 62 percent of their own quarterly histories. The composite is a mean of percentiles, rather than a percentile of quarters. The composite's own rank among published quarters is published separately as rank_in_history.
Current reading: ADI 44.6 Band 3 of 5 · Typical for 2025-Q4. On average, its inputs sit higher than in 45% of their own quarterly histories since 2005. Full methodology at /methodology/.
Data Access
All data is served as static JSON from Cloudflare's edge network. No authentication, no rate limits, no usage restrictions for academic purposes.
JSON API Endpoints
Latest quarter with the composite, band, band label, the literal reading, and the composite's own rank in history, plus the full quarterly series with five domain scores and member percentiles (84 observations, 2005–2025).
curl https://americandefault.org/api/adi.json All 88 indicators with current value, trend, YoY change, units, source, and metadata.
curl https://americandefault.org/api/indicators.json Full time series for a single indicator. Replace {slug} with any indicator slug from the index endpoint.
curl https://americandefault.org/api/indicators/the-buffer.json CSV Downloads
- All indicators — latest values (one row per indicator)
- State debt & delinquency (51 rows, 5 metrics per state)
- Individual indicator time series:
/api/downloads/{slug}.csv(available from every indicator page)
Embeddable Charts
Every indicator has a self-contained iframe-embeddable chart. Pattern: /embed/{slug}/. See the embed documentation for examples.
Code Examples
Python (requests + pandas)
import requests
import pandas as pd
# Fetch all indicators
resp = requests.get("https://americandefault.org/api/indicators.json")
indicators = resp.json()["indicators"]
# Load into DataFrame
df = pd.DataFrame([{
"name": i["display_name"],
"value": i["latest_value"],
"trend": i["trend"],
"yoy_change": i.get("yoy_change"),
"source": i["source"]
} for i in indicators])
# Fetch ADI composite with full history
adi = requests.get("https://americandefault.org/api/adi.json").json()
adi_df = pd.DataFrame(adi["history"]) R (httr2 + jsonlite)
library(httr2)
library(jsonlite)
# Fetch ADI composite
adi <- request("https://americandefault.org/api/adi.json") |>
req_perform() |>
resp_body_json()
# Convert history to data frame
history <- do.call(rbind, lapply(adi$history, as.data.frame))
history$date <- as.Date(history$date)
# Fetch single indicator time series
savings <- request("https://americandefault.org/api/indicators/the-buffer.json") |>
req_perform() |>
resp_body_json()
ts_data <- do.call(rbind, lapply(savings$data, as.data.frame)) The API returns standard JSON. Any language with HTTP client and JSON parsing support can access the data. No authentication headers or API keys needed.
Variable Definitions
88 indicators organized into 9 categories. Each indicator includes full time series data, source attribution, and trend analysis.
For detailed metadata on each variable (units, source agency, series ID, update frequency), fetch the individual indicator JSON endpoint: /api/indicators/{slug}.json.
Reproducibility
The ADI is designed for full reproducibility. Every input, transformation, and output is documented and accessible.
last_updated timestamps.Historical Series
The full quarterly ADI history is available via the ADI API endpoint. Each observation includes the quarter, composite, band, band label, all five domain scores, and the member percentiles behind them. This data can be used to reproduce the published series end to end, decompose any quarter by domain, or extend the analysis.
Leading Indicator Research
Beyond tracking current conditions, the ADI research pipeline systematically tests all pairwise indicator combinations for statistically validated leading relationships. A five-filter pipeline identifies cases where one indicator consistently precedes another by multiple quarters across multiple economic crises.
Validated Leading Relationships
Scanner has not yet been run. See the Structural Outlook page for current status.
Additionally, the foundational Buffer Depletion → Debt Stress relationship (9 quarters with r = 0.69) was validated manually. Full analysis: What the Savings Rate Told Us Nine Quarters Before the Last Crisis.
API Access
All validated leading relationships with correlation coefficients, lag periods, Granger causality statistics, and out-of-sample validation metrics.
curl https://americandefault.org/api/research/leading-indicators.json Python Example
import requests
# Fetch leading indicator research
resp = requests.get("https://americandefault.org/api/research/leading-indicators.json")
data = resp.json()
# List validated pairs
for pair in data["validated_pairs"]:
leader = pair["leader"]["name"]
follower = pair["follower"]["name"]
lag = pair["lag_quarters"]
r = pair["correlation"]
print(f"{leader} → {follower} ({lag}q lag, r={r:.2f})") Active structural projections based on these relationships are tracked on the Structural Outlook page. The Leading Indicator Discovery article provides the full cross-correlation methodology.
Cite This Data
Suggested Citations
APA (ADI composite):
American Default Research. (2026). American Distress Index [Data set]. https://americandefault.org/adi/
APA (individual indicator):
American Default Research. (2026). [Indicator name] [Data set]. https://americandefault.org/indicators/[slug]/
BibTeX:
@misc{adi2026, title={American Distress Index}, author={{American Default Research}}, year={2026}, url={https://americandefault.org/adi/}, note={Composite household financial distress index}}
Citation Tools
- Per-indicator citations: Every indicator page has a "Cite" button generating APA, MLA, Chicago, and BibTeX formats
- Bulk citation widget: The press page has a dropdown selector for all 88 indicators with formatted citations
- BibTeX files:
/citations/{slug}.bibfor any indicator — import directly into Zotero, Mendeley, or EndNote - RIS files:
/citations/{slug}.ris— compatible with all major reference managers - Printable summary: The ADI one-pager provides a print-friendly reference with the current score and citation block
Attribution for Underlying Data
The ADI composite score, component analysis, and editorial content are original work by American Default Research. When citing specific government statistics (e.g., the personal savings rate), also cite the original agency:
U.S. Bureau of Economic Analysis, Personal Saving Rate [PSAVERT], retrieved from FRED, Federal Reserve Bank of St. Louis; https://fred.stlouisfed.org/series/PSAVERT, March 2026.
Research Contact
Available for research collaboration, custom data requests, methodology questions, and speaking engagements. We respond to all academic inquiries within 48 hours.