The EDGAR system
What is XBRL?
Updated
XBRL, short for eXtensible Business Reporting Language, is the tagging standard that makes SEC financial statements machine-readable. Every number in a modern 10-K or 10-Q carries an invisible label identifying exactly what it is: which concept (say, revenue from contracts with customers), which period, which unit, which segment. It is the reason software can compare revenue across a thousand companies without a human re-keying a single figure.
How it works
Tags come from a standard dictionary, the US GAAP taxonomy, maintained by the FASB and updated annually, which defines thousands of concepts with precise accounting meanings. A company maps each reported fact to a taxonomy element, adds context (period, entity, currency), and files the result alongside the human-readable document. When a company reports something the taxonomy lacks, it can create a custom extension tag. That flexibility is also the format’s biggest analytical headache, since extensions defeat cross-company comparison.
From bolt-on to Inline XBRL
The SEC phased in XBRL for financial statements between 2009 and 2011, initially as a separate data file bolted onto the filing, which meant the tagged data and the visible document could quietly disagree. Inline XBRL (iXBRL), phased in from 2019 to 2021, fixed this by embedding the tags directly in the HTML document itself: one file is simultaneously human-readable and machine-readable, and opening a recent filing in the SEC’s viewer lets you click any number to reveal its tag, period and unit. Cover-page facts such as ticker, filer category and share counts are tagged too, and tagging requirements have steadily expanded deeper into the footnotes.
What it enables, and where it misleads
Tagged data powers everything downstream: the SEC’s own company-facts APIs, academic research, screeners, and products like Scrutar. The standardized statements on Microsoft’s financials page or NVIDIA’s exist because the underlying filings are tagged. Regulators use it to screen for anomalies at scale.
The caveats are real, though. Tagging errors happen: wrong signs, wrong periods, misapplied concepts. Custom extensions also vary widely in quality. Two companies may tag economically similar items differently, and restated figures can linger in derived datasets. Serious work treats XBRL as the index and the filing text as the authority: when a number matters, click through and read the statement it came from.
Adoption also stretches well beyond the SEC. European issuers file annual reports in the ESEF inline-XBRL format, banking regulators collect call-report data in XBRL, and dozens of other jurisdictions have mandated variants, so a skill built reading tagged U.S. filings transfers almost everywhere structured financial disclosure exists.
The bigger picture
XBRL quietly converted EDGAR from a document warehouse into a structured database, arguably the largest free financial dataset in existence. Every company in Scrutar’s directory sits on top of it.