Corporate Default Events

Your agent translates this into the event types that signal real distress, the dollar amounts on each defaulted obligation, and the dates those disclosures reached EDGAR.

The moment a company puts a default in writing.

A default is a company failing to meet a term of its debt. That can be a missed interest or principal payment, a breach of a financial covenant such as a leverage or coverage test, or a creditor declaring the entire balance immediately due, which is called acceleration. Public companies disclose these in SEC filings, usually as prose inside an 8-K or a periodic report rather than as a tagged data point.

This dataset reads that prose and turns each disclosed default into one dated row per company. Results are filtered to default events only, so there is no non-event noise to strip out before you can use the feed.

A credit signal you can trade in the equity.

A default disclosure is one of the few equity events where the bad news keeps compounding after it is public. In the published study on this dataset, 472 default events since 2020 ran a median 34% lower three months out, with 61% of names lower a month after the event (z = 4.6). This is drift, not a one-day gap that mean-reverts by Friday.

Severity ordering is what makes the feed more than a distress list. Missed payments ran a median 37% lower at one month in the same study, while the softer events came in near flat. Filtering on event type and on whether acceleration was declared turns a broad watchlist into a graded one, and the dollar fields let you weight by how much debt is actually behind the disclosure.

• Filter a long book to drop names with a default event in the trailing quarter.
• Build graded short baskets: acceleration_declared first, softer covenant events last.
• Run drift studies from event_date, then repeat from default_date to price what the disclosure lag was worth.
• Use default events as labels or regime triggers for a credit model where CDS and bond data are out of reach.
• Weight the signal by principal_outstanding_usd, so a small facility is not treated like a billion-dollar note.

Labeled from filing text, with the sentence attached.

Filings are swept from SEC EDGAR and the default label is extracted from the filing text by a labeling model, then normalized: the event into a fixed type vocabulary, obligations and creditors into named strings, amounts into US dollars. Every row carries evidence_quote, the verbatim sentence the label came from, and confidence, the labeler's own score, so any claim can be checked against the source through filing_url in one click.

Nothing is inferred from news coverage, ratings actions, or bond prices. Where the filing text does not state a value the field is null instead of an estimate, so a blank amount means the company did not disclose the number, not that the number was zero. Rows are served newest first by event_date and are not rewritten after publication.

The fields that rank severity and size.

The headline fields below are a subset. Every field, with exact types and semantics, is documented in the API reference.

One call returns the event and the obligation behind it.

One key works across the REST API, the hosted MCP server, and this dashboard. Every response is JSON with a { count, data } envelope.

import requests

url = "https://api.alphanume.com/v1/corporate-default-events"
params = {
    "date_gte": "2026-01-01",
    "date_lte": "2026-02-28",
    "api_key": "alp_abc123"
}

r = requests.get(url, params=params)
print(r.json())

curl "https://api.alphanume.com/v1/corporate-default-events?date_gte=2026-01-01&date_lte=2026-02-28&api_key=alp_abc123"

{
  "count": 1,
  "data": [
    {
      "event_date": "2026-02-09",
      "ticker": "ESGH",
      "filing_url": "https://www.sec.gov/Archives/edgar/data/1883835/000152013826000054/0001520138-26-000054-index.htm",
      "confidence": 0.93,
      "evidence_quote": "The Company failed to make the interest payment due on the Notes...",
      "event_type": "missed_interest_payment",
      "obligation_name": "8.5% Senior Secured Notes due 2028",
      "obligation_type": "senior_secured_notes",
      "creditor_name": "Wilmington Trust, N.A. (as Trustee)",
      "principal_outstanding_usd": 145000000,
      "amount_accelerated_usd": null,
      "missed_payment_amount_usd": 6162500,
      "default_date": "2026-02-02",
      "grace_period_end_date": "2026-03-04",
      "acceleration_declared": false
    }
  ]
}

What this data does not claim.

• Only defaults a filing actually states are here. A company negotiating quietly with its lenders before disclosing anything produces no row, so the feed tracks disclosure, not distress itself.
• event_date is the date of the disclosing filing while default_date is when the filing says the default happened. The two often differ by days or weeks, so choose the one your study actually means.
• Enriched fields are null wherever the filing did not state them. A null missed_payment_amount_usd means the amount was not disclosed, not that no money was missed.
• confidence scores the extraction from the filing text. It is not a probability of bankruptcy and it is not a market forecast.
• One company can generate several rows as a situation develops, from covenant breach to missed payment to acceleration. Deduplicate by ticker before treating events as independent observations.