# Point-in-time *https://stockmarketstack.com/glossary/point-in-time · next to Fundamental Data & Stock Research Platforms* **Definition:** Data as it stood on a past date, rather than as it stands today. A point-in-time series returns the figure that was actually available then — the earnings first reported, the index membership that held, the identifier in use — instead of the current, restated view of that period. Most free sources give you the second and describe it as history. ## How it works Most financial data is stored as one current view of the past. Ask for a company's 2019 revenue and you are given the figure that stands today, after any restatement, reclassification or change of accounting policy. Ask for the constituents of an index "in 2019" and you are usually given today's list. Ask which sector a company was in and you get the classification it carries now. A point-in-time source stores the extra dimension: not just what the value is for a period, but when each version of it became known. A query then takes two dates — the period it is about, and the date you are pretending to stand on — and returns the figure that was available then. The distinction is easiest to see in the SEC's own data. The filing archive is inherently point-in-time: every document has a filing date, and an amended 10-K sits alongside the original rather than replacing it, so the as-known-then view can be reconstructed. The convenience layers on top are not. EDGAR's frames API returns the fact *last filed* for a period, not the fact as first reported — which is the right answer for describing a company and the wrong one for reconstructing what was known. The Financial Statement Data Sets are published per filing, which is what makes them usable for this purpose at all. Either way the reconstruction is work: you are assembling versions, not reading a field. ## Where the current view quietly substitutes for the past **Restatements.** A revised figure is the more accurate description of the period and was not available during it. Both statements are true at once, which is why the data has to carry the date. **Index membership.** The membership history is licensed data sold by the index provider rather than a public fact, which is the mechanical reason [historical constituents cost money](https://stockmarketstack.com/guides/historical-index-constituents). **Identifiers and symbols.** A ticker is reassigned; a company renames. Without a permanent identifier underneath, a series silently becomes a different company's — the problem [security identifiers exist to solve](https://stockmarketstack.com/guides/cusip-isin-figi-identifiers). **Shares outstanding and derived ratios.** Anything computed from a restated denominator inherits the restatement, usually without saying so. **Delisted securities.** Prices are mostly kept as traded, but a history that drops companies which stopped trading is not a record of the past either. That is the sibling problem, [survivorship bias](https://stockmarketstack.com/glossary/survivorship-bias). ## Why it matters here The practical value is narrow and sharp: it tells you whether a dataset can answer a question about the past, or only a question about now. A figure nobody had at the time cannot be part of a reconstruction of what was knowable, and the difference is not small — restatements cluster in exactly the periods people most want to study. So it is worth asking every fundamentals source on this site one question: is the history point-in-time or restated? Vendors built for this answer directly and price accordingly — [SimFin](https://stockmarketstack.com/tools/simfin) and [Daloopa](https://stockmarketstack.com/tools/daloopa) both address as-reported data explicitly, and [QuantRocket](https://stockmarketstack.com/tools/quantrocket) bundles point-in-time datasets because the use case demands it. [SEC EDGAR](https://stockmarketstack.com/tools/sec-edgar) is free and is the primary record, at the cost of doing the assembly yourself. A general API that returns one number per period and says nothing about versions is giving you the current view, whatever the label on the endpoint says. [What is genuinely free](https://stockmarketstack.com/guides/free-financial-data-sources) covers where the free sources stand on this, and it is the fourth of the four questions worth asking before any data invoice. ## Where you will meet this - [SEC EDGAR](https://stockmarketstack.com/tools/sec-edgar.md) - [SimFin](https://stockmarketstack.com/tools/simfin.md) - [Daloopa](https://stockmarketstack.com/tools/daloopa.md) - [QuantRocket](https://stockmarketstack.com/tools/quantrocket.md) - [Norgate Data](https://stockmarketstack.com/tools/norgate-data.md) - [AI Hedge Fund](https://stockmarketstack.com/tools/ai-hedge-fund.md) - [AlgoSeek](https://stockmarketstack.com/tools/algoseek.md) - [BLS Public Data API](https://stockmarketstack.com/tools/bls-api.md) - [ECB Data Portal API](https://stockmarketstack.com/tools/ecb-data-portal.md) - [EconDB](https://stockmarketstack.com/tools/econdb.md) - [FRED](https://stockmarketstack.com/tools/fred.md) - [QuestDB](https://stockmarketstack.com/tools/questdb.md) ## FAQ ### Is EDGAR point-in-time? The archive is, the convenience APIs are not. Every filing sits there with its own date, restatements included, so an as-known-then view can be reconstructed. But the frames API returns the fact last filed for a period rather than the fact as first reported, which is the opposite of what a point-in-time query wants. ### Does point-in-time only matter for fundamentals? No. Index membership, ticker symbols, sector classifications, shares outstanding and analyst estimates all get revised or reassigned, and each of them is usually served as the current view. Prices are the exception where most vendors do keep an as-traded record, though whether delisted securities survive is a separate question. ## Sources 1. [EDGAR Application Programming Interfaces](https://www.sec.gov/search-filings/edgar-application-programming-interfaces) — U.S. Securities and Exchange Commission, 2025-04-08 2. [Financial Statement Data Sets](https://www.sec.gov/data-research/sec-markets-data/financial-statement-data-sets) — U.S. Securities and Exchange Commission, 2026-06-30 3. [CRSP US Stock Databases](https://indexes.morningstar.com/research-data-products/crsp-us-stock-databases) — Morningstar Indexes (CRSP), read 2026-09-14 *Last updated 2026-09-21. A reference page, corrected in place — not a dated post.*