What you need, as it was known at the time
Two databases built from one source, with a link to the original SEC filing on every row. Parquet and CSV, updated as EDGAR is.
Point-in-time fundamentals for every US company that files with the SEC, including the thousands that have since delisted.
- 27 fields, quarterly and annual, 2009 to today
- First-reported value and every restatement, side by side
- Filing date and accession number on every value
Every EDGAR event the day it happens, with acceptance times to the second.
- Earnings releases with market session
- Insider trades, incl. the 10b5-1 flag
- 13D activist and 13G stakes, 13F holdings
- Late filings, 8-K red flags, IPOs, proxy fights
Built for research that has to hold up
Point-in-time backtesting
Test strategies on the numbers investors could actually see, for the companies that actually existed, with no look-ahead and no survivorship.
Event-driven research
Earnings release times to the second, insider buying, auditor changes and restatement warnings, timestamped by EDGAR.
Ownership and flows
Who holds what: 13F institutional holdings each quarter and 13D/13G stakes as they're filed.
An audit trail
Every value names its source filing, so any number can be checked against the original document in one click.
Latest filings, as they land
Earnings releases
All earnings releases →Largest insider buys (30 days)
All insider activity →8-K red flags
All red flags →What we found building the data
About 1 in 10 revenue figures is later restated. Most datasets only keep the new number.
Why a backtest on restated data quietly uses information that didn't exist yet, and how first-reported values fix it.
SEC's bulk filing history has timestamps off by four to five hours for a third of companies
How we found it, how we corrected it against SEC's own XBRL data, and why unverifiable times are shown as dates only.
Three years after the rule changed, some fund managers still report 13F values in thousands
The unit errors that put quarters at $15,000 trillion, and the checks that bring the total back to $82 trillion.
Frequently asked
Where does the data come from?
Only from SEC EDGAR: the SEC's financial statement data sets, its insider, Form D and 13F data sets, and the daily EDGAR filing feed. Every row carries the accession number of the filing it came from.
What does "point-in-time" mean here?
Each value is dated by the filing that made it public, and the first-reported number is kept even when a later filing restates it. A backtest therefore sees only what investors could have known on the day.
Are delisted companies included?
Yes. The data is built from every filer in every period, so bankrupt, acquired and delisted companies are present, with their historical tickers where they can be recovered.
How often is it updated?
Varve Fundamentals updates monthly as the SEC publishes new financial-statement data. Varve Events updates every business day from EDGAR.
What formats do you deliver?
Parquet and CSV for every table, with a manifest of SHA-256 checksums, a methodology guide and a full data dictionary.
Start with a free sample
Full history for 25 companies, the last week of events, and the methodology guide.