Which corporate events get disclosed on Fridays?

Study #3 · window 2025-09-01 → 2026-08-31 · every number below regenerates from one script on a free key

An 8-K is the SEC's current report: when something material happens — earnings, an executive departure, a restatement, a delisting notice — the company generally has four business days to disclose it, and the filing itself declares what happened via the SEC's official item codes (2.02 = results of operations, 5.02 = officer and director changes, 4.02 = non-reliance on previously issued financials, …). That declaration makes disclosure timing measurable by topic: this page takes a year of 8-Ks and asks how each item code's filings distribute across the trading week — the measurable core of the pattern popularly called the “Friday news dump”. What it finds is a descriptive record of when disclosures reached EDGAR, by their own labels and dates; it attributes no intent and characterizes no filing as good or bad news.

One EDGAR mechanic shapes how to read “Friday” here, so it is worth being precise. Under Regulation S-T Rule 13 (17 CFR 232.13), a filing accepted after 5:30 p.m. Eastern is deemed filed the next business day — and the 10 p.m. same-day window covers Forms 3/4/5 and Schedules 13D/G, not 8-Ks. So a Friday filed_date means EDGAR accepted the filing by Friday 5:30 p.m., and a submission pushed out Friday evening carries Monday's date. Both edges of the week are shown below.

What gets filed: a year of 8-Ks by item code

Filings carrying each item code (a filing with several items counts once under each). The shares tell you what the 8-K stream actually is before any weekday question: mostly exhibits riding along with substantive items (9.01), earnings (2.02), other events (8.01), Reg FD (7.01), and personnel changes (5.02).

Items per filing: 13,060 with 1 item · 32,112 with 2 items · 13,683 with 3 items · 3,567 with 4 items · 2,039 with 5+ items. 51,401 filings (79.7%) carry more than one item, which is why the shares above sum past 100%.

The combinations that travel together most:

The weekday baseline

All original 8-Ks by the day of week of their filed date. The per-day column normalizes for how often each weekday even occurred as a filing day in the window (holidays fall unevenly across weekdays), and reads as filings per active day.

Weekend filed dates in the window: 0 (EDGAR accepts filings on business days, so a weekend filed date is an exceptional event, reported here rather than folded into a weekday).

Friday, by item code

Each item code's filings across the week, for every code with at least 500 filings in the window — below that, weekday percentages are noise. Sorted by Friday share; the Δ column is percentage points versus the all-8-K Friday baseline of 16.8%. Monday is shown alongside because of the 5:30 p.m. rule above: a filing transmitted Friday evening lands there.

Read the table as composition, not intent: an item code's weekday profile also reflects when its underlying events are scheduled (earnings calls cluster mid-week; annual meetings — item 5.07 — happen when they happen). The measurement is what reached EDGAR when; why is outside what a filing date can say.

Is the population complete?

A weekday study over an incomplete population is quietly wrong in the worst way — ingestion gaps are not uniform across weekdays — so the script cross-checks every day's filing count against SEC EDGAR's own daily master indexes: distinct accessions, form types 8-K and 8-K/A, bucketed by each row's own Date Filed column across every index in the sweep. The sweep matters: EDGAR disseminates some filings only in the next day's index under the prior day's date, and re-disseminates corrections under their original date, so no single day's index is the whole population for that date. Result: 240 of 248 index days match exactly; 117 days had no published index (weekends/holidays) and zero filings in the API — consistent.

Mismatched days are reported, not hidden — see the results file for the full per-day table.

Worth knowing: every remaining mismatch above was chased to a named accession rather than shrugged at, and they decompose into two classes, neither a coverage hole. Most are date-attribution disagreements inside EDGAR's own records — the daily index's Date Filed column names one day while the filing's own header (FILED AS OF DATE) names an earlier one; the API serves dates exactly as filed, so those filings are in the population under their header date. The rest are filings the index lists but EDGAR's archive no longer serves — the document folders return 404 (verified per accession), so they can enter no one's dataset. The first run of this check also surfaced real ingestion gaps — 36 filings across 13 days that our live feed or backfill had dropped — which were re-ingested from the daily indexes before publishing. The numbers on this page are from the repaired population.

Doing this with the API

Everything above is the item= filter on GET /v1/events plus a fixed date window; webhook subscriptions take the same items filter, so an integration can watch one item code in real time instead of polling. The 8-K API page walks the filters; query patterns covers windows and paging.

Reproduce this

The whole study is one standard-library Python script that runs on the free tier. The full 12-month window is ~66,000 filings ≈ 700 requests (free tier: 10 req/min, so expect about seventy minutes; a single month runs in minutes on any tier). The --verify-coverage flag hits EDGAR directly and needs a contact identity in the User-Agent, per SEC fair-access policy — set your own.

# free key, no card: https://filingpulse.io/signup.html
export FILINGPULSE_API_KEY=fp_...
export EDGAR_CONTACT=you@example.com

curl -O https://filingpulse.io/examples/eightk_weekday_study.py
python eightk_weekday_study.py \
    --since 2025-09-01 --until 2026-08-31 \
    --out results.json --verify-coverage

The committed results this page renders from: eightk_weekday_2025-09_2026-08.json. A fixed since=/until= window returns the same population on any tier, any day — that is what makes the study reproducible rather than a screenshot.

Limitations

• Filed date is not acceptance time. EDGAR assigns an 8-K accepted after 5:30 p.m. Eastern the next business day's date (Regulation S-T Rule 13), and this dataset carries dates, not timestamps — so “Friday” here means accepted by Friday 5:30 p.m., and true Friday-evening submissions surface inside Monday's numbers. The two columns bound the effect; they cannot separate it hour by hour.
• Item codes are the filer's own declaration in the EDGAR submission header. They are standardized and almost universally present, but this study classifies by that label, not by reading the filing body.
• One year is one year. Weekday mix varies with earnings seasons and holiday placement; treat these numbers as a measurement of this window, not a law of nature. This is an observed disclosure-timing record, not a compliance determination and not a list of violations — 8-K deadlines are measured from the triggering event, which a filing date alone cannot see.
• A filing counts once per distinct item code it declares, so per-item populations overlap (2.02 and 9.01 usually ride together).
• Multi-filer 8-Ks are attributed to the first filer, and amendments (8-K/A) are excluded as re-filings — both stated in the method box, with counts.

Sources

• 17 CFR § 232.13 (Regulation S-T Rule 13) — the 5:30 p.m. / next-business-day filing-date rule
• SEC Form 8-K — the current report and its official item codes
• EDGAR daily index — the per-day ground truth the coverage check reads
• Study #1: How late is Form 4 data? · Study #2: When insider filings get corrected
• FilingPulse API docs · 8-K event API · query patterns