Form 13F-HR Files Dataset

Form 13F-HR is the quarterly holdings report that institutional investment managers with at least $100 million in Section 13(f) securities must file with the SEC. The Form 13F-HR Files Dataset packages every 13F-HR and 13F-HR/A submission on EDGAR as a structured folder of five files — metadata JSON, two XML source documents (cover page and information table), and two HTML renditions — within monthly ZIP archives. Each record maps to one accession number and represents one manager's complete disclosure of long positions in Section 13(f) securities as of a calendar quarter-end date. The dataset covers filings from February 1998 to the present and includes both original quarterly reports and amendments.

Dataset APIs

Dataset Files

What This Dataset Contains

The Form 13F-HR Files Dataset is built from Form 13F-HR and Form 13F-HR/A filings submitted through the SEC's EDGAR system. Form 13F-HR is the quarterly holdings report required by Section 13(f) of the Securities Exchange Act of 1934. Institutional investment managers exercising investment discretion over $100 million or more in Section 13(f) securities — a defined universe of exchange-traded equities, certain equity options, and convertible debt instruments enumerated on a list the SEC publishes quarterly — must file within 45 days of each calendar quarter-end. The filing discloses the manager's long positions in every reportable security held as of the quarter's last trading day, together with the nature of investment discretion and voting authority over each position. Form 13F-HR/A is an amendment to a previously filed quarterly report; amendments may correct errors, add omitted holdings, or replace the entire holdings table. The dataset includes both initial filings and amendments from February 1998 to the present.

Unlike most major SEC forms, the 13F-HR is not a narrative disclosure document. It consists almost entirely of structured data: a cover block with filer identification and summary statistics, a signature, and a detailed tabular schedule of individual security positions. There is no MD&A, no risk-factor discussion, and no financial statements. The filing's value is in the granular position-level data and the ability to track portfolio composition across quarters.

The dataset is distributed as monthly ZIP archives. Within each archive, folders are named by accession number with dashes removed, yielding an 18-digit zero-padded string (e.g., 000082043425000005). For XML-era filings (Q4 2013 onward), every folder contains exactly five files: a dataset-supplied metadata.json, two raw XML source documents from the EDGAR submission, and two HTML renditions in an xslForm13F_X02/ subfolder. For pre-XML filings (1998 through 2013), root-level documents may be .txt or .html files rather than .xml, and the XSL rendition subfolder may be absent or structured differently. The metadata.json wrapper is uniformly present across all eras.

Content Structure of a Single 13F-HR Filing Record

Each record folder contains five files organized in the following layout:

          1
          000082043425000005/
        

          2
            metadata.json
        

          3
            primary_doc.xml
        

          4
            <informationtable>.xml
        

          5
            xslForm13F_X02/
        

          6
              primary_doc.xml
        

          7
              <informationtable>.xml
        

The three root-level files are the dataset metadata wrapper and the two raw XML source documents from the EDGAR submission. The xslForm13F_X02/ subfolder contains HTML renditions of those same XML documents. Image files from the original EDGAR submission are excluded. No SGML <DOCUMENT> wrapper envelopes are present; all files are standalone XML or HTML.

metadata.json — Filing-Level Metadata

This JSON file is generated by the dataset provider (not by the SEC) and consolidates EDGAR index and header information into a single structured object. It is the primary machine-readable entry point for identifying, filtering, and linking the filing.

Core identification fields:

• id (string): Unique hex identifier assigned by the dataset provider, e.g., "f1bbfdf3ebd40910e8a856c6320f750e".
• accessionNo (string): EDGAR accession number in dashed format, e.g., "0000820434-25-000005".
• formType (string): "13F-HR" or "13F-HR/A".
• filedAt (string): ISO 8601 timestamp with timezone offset indicating the EDGAR acceptance datetime.
• effectivenessDate (string): Date in YYYY-MM-DD format.
• periodOfReport (string): Calendar quarter-end date in YYYY-MM-DD format, e.g., "2025-06-30".
• description (string): Human-readable label, e.g., "Form 13F-HR - Quarterly report filed by institutional managers, Holdings". Amendments include "[Amend]".

Entity information:

The entities array typically contains one object identifying the filing manager:

• companyName (string): Filer name with role suffix, e.g., "PVG ASSET MANAGEMENT CORP (Filer)".
• cik (string): Central Index Key, the SEC's unique numeric filer identifier.
• fileNo (string): SEC file number, typically in the 028-XXXXX series for 13F filers.
• irsNo (string): IRS Employer Identification Number.
• type (string): Mirrors the top-level formType.
• fiscalYearEnd (string): Four-digit MMDD string, e.g., "1231".
• stateOfIncorporation (string): Two-letter state code; may be absent.
• sic (string): Standard Industrial Classification code with text description; present mainly when the filer is itself a publicly traded company.
• tickers (array of strings): Ticker symbols; present when the filing manager is a public company.
• act (string): Governing securities act, typically "34".
• filmNo (string): EDGAR film number.

Document references:

The documentFormatFiles array lists each document in the submission with its sequence number, size in bytes, documentUrl (SEC.gov Archives path), type (e.g., "13F-HR" for the primary document, "INFORMATION TABLE" for the holdings table), and optional description.

Link fields:

• linkToFilingDetails: URL to the XSL-transformed primary document on SEC.gov.
• linkToTxt: URL to the complete EDGAR submission text file.
• linkToHtml: URL to the EDGAR filing index page.
• linkToXbrl: Always empty for 13F filings (XBRL is not used).
• dataFiles: Always an empty array.
• seriesAndClassesContractsInformation: Always an empty array.

primary_doc.xml — Cover Page, Summary, Amendment Info, and Signature

This XML document uses the http://www.sec.gov/edgar/thirteenffiler namespace (schema version X0202 in current filings). It functions as the formal wrapper for the holdings report: identifying the filer, declaring the report type, providing aggregate summary statistics, and carrying the signature. It does not contain individual holdings.

Header data block:

Contains submissionType (e.g., 13F-HR), the filer's cik, a masked ccc field (always XXXXXXXX), and periodOfReport in MM-DD-YYYY format. Note this date format differs from the YYYY-MM-DD format used in metadata.json.

Cover page block:

• reportCalendarOrQuarter: Quarter-end date, matching periodOfReport.
• filingManager: Manager's legal name and mailing address, structured with name, street1, street2, city, stateOrCountry, and zipCode elements (address elements use the http://www.sec.gov/edgar/common namespace).
• reportType: Declares the filing category. Common values: "13F HOLDINGS REPORT" (manager reports all holdings in a single filing) and "13F COMBINATION REPORT" (manager files jointly with other reporting managers).
• form13FFileNumber: SEC-assigned 13F file number for the manager.
• provideInfoForInstruction5: "Y" or "N" flag indicating whether the filer provides additional information under Instruction 5 of Form 13F (used when another manager reports holdings on the filer's behalf).

Amendment information block (13F-HR/A only):

Present only in amendments:

• isAmendment: Boolean-like flag.
• amendmentNo: Sequential amendment number.
• amendmentType: Nature of the amendment — "RESTATEMENT" (replaces the entire prior information table) or "NEW HOLDINGS" (adds previously omitted positions).
• otherManagersInfo: Optional sub-structure listing other managers involved, each identified by CIK, name, and form13FFileNumber.
• additionalInformation: Free-text field for explanatory notes.

Summary page:

Provides pre-computed aggregates that are analytically useful for validation and screening:

• tableEntryTotal: Total number of individual holdings entries in the information table (integer count of position lines).
• tableValueTotal: Aggregate market value of all reported holdings, in thousands of USD.
• otherIncludedManagersCount: Number of other managers whose holdings are included (relevant for combination reports; zero for standalone filings).
• isConfidentialOmitted: Optional flag indicating whether a confidential treatment request has caused omission of certain holdings from the public information table.

Signature block:

Contains name of the signatory, title, phone number, signatureDate, city and stateOrCountry of signing, and the signature element carrying the typed name (the electronic equivalent of a manual signature).

Information Table XML — Position-Level Holdings Data

The information table is the analytical core of the 13F-HR filing. It is a separate XML document in the http://www.sec.gov/edgar/document/thirteenf/informationtable namespace, containing a flat list of <infoTable> elements, one per reported security position.

File naming: The filename is chosen by the filer and varies widely. Common names include infotable.xml (roughly 36% of recent filings), Form13F_InfoTable.xml, informationtable.xml, and numerous custom names (e.g., f1306302025.xml, HPM13FQ424.xml). Despite inconsistent naming, the XML schema is uniform across all post-2013 filings.

Fields per holding entry:

• nameOfIssuer (string): Name of the issuing company or fund, e.g., "APPLE INC". Free text; not standardized across filers. Varying capitalization, abbreviations, and parenthetical ticker annotations are common.
• titleOfClass (string): Security class or type, e.g., "COM" (common stock), "CL A" (Class A), "ETF", "SH BEN INT" (shares of beneficial interest), "CALL", "PUT", "NOTE", "CONV DEB".
• cusip (string): 9-character CUSIP identifier. This is the sole standardized security identifier in 13F filings.
• value (integer): Market value of the position in thousands of USD, rounded to the nearest thousand. A value of 1260458 represents approximately $1.26 billion.
• shrsOrPrnAmt: Nested element containing:
  • sshPrnamt (integer): Number of shares (for equity) or principal amount (for debt).
  • sshPrnamtType (string): "SH" (shares) or "PRN" (principal amount).
• investmentDiscretion (string): "SOLE" (manager has sole discretion), "SHARED" (shared with another manager), or "NONE" (no discretion; position reported because manager has voting authority).
• votingAuthority: Nested element with three integer sub-elements:
  • Sole: Shares over which the manager has sole voting authority.
  • Shared: Shares with shared voting authority.
  • None: Shares with no voting authority.
• otherManager (optional, string): Sequence number referencing another manager listed in the cover page's otherManagersInfo; used in combination reports.
• putCall (optional, string): Present only for options positions; "PUT" or "CALL".

For equity positions, the sum of Sole + Shared + None voting authority shares should equal sshPrnamt, though filers occasionally report inconsistent totals.

xslForm13F_X02/ Subfolder — HTML Renditions

This subfolder contains HTML-transformed versions of both primary_doc.xml and the information table XML. These are complete HTML documents with embedded CSS classes (FormData, FormTitle, SectionTitle, etc.) and tabular layouts that reproduce what a viewer sees on SEC.gov. The HTML renditions are presentational duplicates of the XML source data with no additional information content. They are useful for visual inspection but redundant for data extraction.

The stylesheet name xslForm13F_X02 corresponds to the current XSL transformation version used by EDGAR. Earlier XML filings may reference xslForm13F_X01 or other versions, but recent dataset containers consistently apply the X02 rendition.

Included Content

Each record folder contains the complete machine-readable content of the EDGAR submission: the cover page with filer identification and summary statistics, the full position-level holdings table, and the styled HTML renditions. The dataset-supplied metadata.json adds consolidated EDGAR index metadata, unique identifiers, and direct URLs to the SEC.gov source. Together, these five files provide everything needed to identify the filer, determine the reporting period, extract every individual security position, and link back to the original submission.

Excluded or Separate Content

• Image files that may have been part of the original EDGAR submission are excluded.
• The raw SGML submission text file (the monolithic .txt concatenation of all documents with SGML headers) is not included but is accessible via the linkToTxt URL in metadata.json.
• XBRL data is not applicable to Form 13F; linkToXbrl is always empty and dataFiles is always an empty array.
• The SEC's Official List of Section 13(f) Securities (the quarterly CUSIP list defining which securities are reportable) is external to the filing and not part of the dataset.
• Confidentially omitted holdings that the SEC has approved for non-disclosure are absent from the public information table. Their existence may be signaled by the isConfidentialOmitted flag on the summary page; the omitted positions are typically disclosed in a subsequent amendment after a delay of approximately one year.

Historical Format Variation

1998 through 2013 — text and HTML era. The earliest filings in the dataset (from February 1998) were submitted as plain text or HTML. The information table appeared as a formatted text table with columns for issuer name, title of class, CUSIP, market value, shares, and discretion/voting fields. Column widths, header labels, and spacing varied by filer and filing software. Some filers used HTML <table> elements; others used monospaced text with column alignment. The cover page was similarly unstructured. Automated extraction of pre-XML filings requires format-specific parsing that accounts for substantial filer-to-filer variation.

2013 XML mandate. Effective for the quarter ending December 31, 2013, the SEC required all 13F filings to use structured XML for both the primary document and the information table. This introduced the http://www.sec.gov/edgar/thirteenffiler and http://www.sec.gov/edgar/document/thirteenf/informationtable namespaces. The mandate dramatically improved data consistency and machine readability.

Schema evolution. The schema version progressed from X0101 at the initial XML mandate to X0202 in current filings. Changes between versions are minor — refinements to element optionality, namespace declarations, and validation rules rather than fundamental structural changes.

Pre-XML folder structure in the dataset. Folders for pre-XML filings may contain .txt or .html files instead of .xml, and the xslForm13F_X02/ subfolder may be absent or structured differently. The metadata.json wrapper is present uniformly regardless of era.

Interpretation Notes

Value units are thousands, not dollars. All value fields in the information table and tableValueTotal in the summary page are in thousands of USD. A reported tableValueTotal of 29381810 represents approximately $29.38 billion.

CUSIP is the sole standardized identifier. Ticker symbols do not appear as structured fields in the holdings table (though filers sometimes embed them in the nameOfIssuer string). Mapping holdings to tickers, ISINs, or other identifiers requires an external CUSIP crosswalk.

Issuer name is free text. The same company may appear as "APPLE INC", "Apple Inc", "APPLE INC.", or "APPLE INC (AAPL)" across different filings. CUSIP is the reliable join key.

Combination reports and double-counting. When multiple managers file a combination report, the same positions may appear across multiple 13F-HR submissions. The otherManager field and otherManagersInfo structure provide the information needed to identify and de-duplicate overlapping holdings, but doing so requires cross-filing analysis.

Amendments may replace or supplement. A "RESTATEMENT" amendment supersedes the entire prior information table. A "NEW HOLDINGS" amendment adds previously omitted positions and should be interpreted as additive to the original. The amendmentType field in the primary document distinguishes these cases. When processing a manager's holdings for a given quarter, the latest restatement amendment, if any, replaces all prior filings for that period.

Late filings and container-period mismatches. A monthly ZIP archive may include filings for quarter-end dates other than the most recent one. Late filings, amendments to prior quarters, and filings with unusual reporting periods appear in the container corresponding to their EDGAR acceptance date, not their periodOfReport. Always use periodOfReport as the authoritative indicator of the quarter covered.

Put/Call positions. Options holdings include the putCall element to distinguish puts from calls. The sshPrnamtType for options is typically "SH" representing shares underlying the contracts, and titleOfClass carries values like "CALL" or "PUT". Reporting consistency for options positions varies across filers.

Date format inconsistency. The periodOfReport in primary_doc.xml uses MM-DD-YYYY format, while metadata.json uses YYYY-MM-DD. Parsers must account for this difference when cross-referencing the two sources.

Who Files or Publishes This Dataset, and When

Who files

The filer is the institutional investment manager — the entity or natural person exercising investment discretion over one or more accounts. The filer is not the fund, trust, or account vehicle, and not the beneficial owner of the securities. A single advisory firm managing multiple funds files one consolidated 13F-HR covering all accounts over which it exercises discretion.

Under Section 13(f) of the Securities Exchange Act of 1934 and Rule 13f-1 (17 CFR 240.13f-1), any institutional investment manager whose accounts hold, in aggregate, at least $100 million in fair market value of Section 13(f) securities as of the last trading day of any month during the calendar year must file. Once the threshold is crossed in any month, the obligation applies for every quarter of that calendar year, even if holdings later fall below $100 million. The manager may stop filing at the end of the calendar year if it no longer meets the threshold.

The filing population includes:

• Registered investment advisers — the largest category, covering firms managing mutual funds, hedge funds, separately managed accounts, and other pooled vehicles.
• Banks and bank holding companies — including trust departments exercising discretion over customer and proprietary accounts.
• Broker-dealers — only to the extent they exercise investment discretion, not merely execute orders.
• Insurance companies — managing general-account and separate-account portfolios.
• Pension fund managers — internal investment teams or external managers exercising discretion over corporate or public pension assets.
• Endowments, foundations, and sovereign wealth fund managers — any entity exercising discretion over accounts meeting the threshold.
• Natural persons — rarely, an individual manager exercising discretion over accounts aggregating to $100 million or more in 13(f) securities.

Where two or more managers share discretion over the same account, each must report those holdings. They may file jointly or arrange for one manager to report on behalf of the others using the "Other Managers" column in the information table.

When the record is created

Form 13F-HR is periodic, not event-driven. It is filed quarterly, keyed to calendar quarter-ends: March 31, June 30, September 30, and December 31. The deadline is 45 calendar days after the quarter-end date. When the 45th day falls on a weekend or federal holiday, the deadline moves to the next business day.

Typical deadlines:

The filing is a single point-in-time snapshot of all long positions in Section 13(f) securities held across the manager's accounts as of the last calendar day of the quarter. Short positions and synthetic short exposures are excluded.

A manager that first crosses the $100 million threshold mid-year files its first 13F-HR for the quarter in which the threshold was met.

Amendments (Form 13F-HR/A) are filed when the manager discovers errors or omissions in a prior quarterly report. There is no fixed amendment deadline; amendments are filed as needed. Most amendments restate the full information table for the quarter. The dataset includes both original filings and amendments; multiple amendments for the same filer and quarter appear as separate records.

Important distinctions

The $100 million threshold, set in 1978, has never been adjusted for inflation. A 2020 SEC proposal to raise it to $3.5 billion was withdrawn. The filing population has grown substantially as asset values have risen.

EDGAR coverage of Form 13F begins in approximately February 1998. Earlier filings (1978 through mid-1990s) exist only in SEC paper archives.

How This Dataset Differs From Similar Datasets or Filings

Form 13F-HR occupies a specific niche among SEC disclosure regimes that touch on securities holdings. The comparisons below cover the forms most likely to be confused with it or used alongside it.

Closest Related Datasets or Filing Types

Form 13F-NT and Form 13F-CTR are siblings within the 13F family. 13F-NT is a procedural notice of late filing that contains no holdings data; it signals a delayed 13F-HR, nothing more. 13F-CTR requests confidential treatment for specific positions, meaning some holdings are temporarily redacted from the public 13F-HR. Users building complete position histories should watch for amendments (13F-HR/A) that restore previously withheld positions once confidential treatment expires.

Schedule 13D / Schedule 13G are the most commonly confused alternatives. The core differences:

Many 13F filers never file a 13D/13G (diversified holdings stay below 5% per issuer). Many 13D/13G filers are not institutional managers and never file a 13F. The two regimes overlap only where a large manager also holds a concentrated position.

Form N-PORT is the closest functional analog in content type — both disclose security-level portfolio holdings periodically. The differences are substantial:

• Filer population: N-PORT is filed by registered investment companies (mutual funds, ETFs, closed-end funds) under the 1940 Act. 13F-HR is filed by investment managers under the 1934 Act. A fund complex typically files both: the fund files N-PORT, the adviser files 13F-HR.
• Asset-class coverage: N-PORT covers the complete portfolio including fixed income, derivatives, repos, and other instruments well beyond the Section 13(f) security list. 13F-HR is limited to 13(f) securities (principally exchange-traded equities, certain options, and convertible debt).
• Entity scope: N-PORT reports one fund's portfolio. 13F-HR aggregates holdings across all accounts the manager controls.
• Public availability: N-PORT is filed monthly but only the quarter-end filing is public, with a 60-day delay. 13F-HR is public immediately upon EDGAR acceptance.
• Data structure: N-PORT uses a detailed XML schema with dozens of fields per position (delta, payoff profile, liquidity classification). 13F-HR uses a simpler tabular format with fewer fields.

Form 4 reports insider transactions (officers, directors, 10% beneficial owners) in their own company's securities. It overlaps with 13F-HR only when an institutional manager is also a 10% owner. Form 4 is transaction-based (individual buys/sells within two business days) and single-issuer, while 13F-HR is snapshot-based (quarter-end aggregate positions) and portfolio-wide. Form 4 captures transaction price and date; 13F-HR does not. The two complement each other for studying large institutional holders who are also insiders.

Key Differences

• 13F-HR vs. 13D/13G: Portfolio-wide quarterly snapshot vs. single-issuer event-driven disclosure at the 5% threshold. Different filer populations with limited overlap.
• 13F-HR vs. N-PORT: Same concept (periodic holdings), different filer regimes, different asset-class scope, different availability timing. N-PORT is deeper per fund; 13F-HR is broader per manager.
• 13F-HR vs. Form 4: Quarter-end position snapshot vs. real-time transaction report. Different triggers, different granularity, minimal filer overlap.
• 13F-HR vs. 13F-NT/CTR: Data-bearing filing vs. procedural companions (delay notice, confidentiality request).

Boundary Summary

Form 13F-HR is the only SEC dataset that provides quarterly, security-level, portfolio-wide holdings snapshots for the full population of large institutional investment managers, publicly available upon filing. No other form replicates this combination. Schedule 13D/13G captures only concentrated single-issuer positions on an event-driven basis. N-PORT covers richer multi-asset detail but only for registered funds and with delayed public release. Form 4 captures insider transactions, not quarter-end portfolios. The 13F family siblings (13F-NT, 13F-CTR) carry no holdings data of their own. For institutional equity ownership analysis at the manager level, 13F-HR remains the primary and often irreplaceable source.

Who Uses This Dataset

Quantitative researchers and systematic portfolio managers

Quant teams parse holdings schedules across thousands of filings per quarter, extracting CUSIP, share count, and market value to reconstruct position-level snapshots. Quarter-over-quarter differencing produces entry, exit, and sizing signals for crowding, momentum, and consensus-deviation strategies. Amendment filings matter because uncorrected stale positions introduce bias into backtests. Manager identity and report-period-end date are essential for aligning holdings to a consistent calendar across filers.

Hedge fund analysts and event-driven investors

Analysts at multi-strategy and event-driven funds track peer positioning around mergers, activist campaigns, and restructurings. The investmentDiscretion field distinguishes sole-discretion positions from shared or delegated authority, signaling conviction level. Monitoring share-count changes for a target CUSIP across many filers reveals whether institutional capital is accumulating or exiting ahead of known catalysts.

Fundamental equity analysts and long-only portfolio managers

Fundamental analysts aggregate all filings reporting a given CUSIP each quarter to build institutional ownership profiles by manager type and size. Quarter-over-quarter share-count changes flag accumulation or distribution that complements bottom-up thesis work. Long-only portfolio managers use the same data for peer benchmarking — comparing their holdings against managers with similar mandates to identify differentiated bets and crowded positions.

Activist monitoring and corporate advisory teams

Proxy solicitation firms and corporate advisors scan 13F filings to detect early-stage activist accumulation. Critical fields are manager identity (matched against known activist filers), CUSIP, share count, and voting authority (sole, shared, or none). A new or growing position by a known activist, visible in the quarterly delta, often triggers defensive preparation before any Schedule 13D appears. Advisory teams also aggregate ownership across all filers to model likely vote outcomes in contested proxy situations, weighting shares by voting-authority classification.

Compliance officers

Compliance teams at broker-dealers and asset managers verify that their own firm's published holdings schedule matches internal portfolio-accounting snapshots as of the report-period-end date. They also monitor counterparty and sub-advisor filings for undisclosed concentration risk. The investmentDiscretion field confirms that shared-discretion positions are properly attributed. Amendment filings are tracked as potential indicators of internal control weaknesses at the filing institution.

Investor relations teams

IR professionals at public companies extract manager identity, share count, and market value for every filer reporting their CUSIP each quarter. Tracking changes over time identifies new entrants, departing holders, and shifts in concentration — informing roadshow targeting, engagement priorities before annual meetings, and board-level governance planning. The voting-authority breakdown helps estimate how much of the institutional base will vote independently versus defer to a proxy advisor.

Financial data engineers

Data engineering teams at financial data vendors and internal analytics groups ingest the full dataset to build normalized quarterly holdings panels. They parse XML and legacy plain-text schedules, standardize CUSIPs, resolve manager-entity mappings across accession numbers, and reconcile originals against amendments. Metadata files (filer CIK, report period, filing date, document index) drive automated parsing pipelines. Format variation across the dataset's historical range — plain-text tables in the late 1990s, structured XML from 2013 onward — is the main engineering challenge.

Academic researchers

Researchers studying institutional herding, ownership concentration, information asymmetry, and market microstructure use 13F data as the primary source of position-level institutional holdings in U.S. equities. The full cross-section of filers and securities per quarter supports measures such as breadth-of-ownership changes and portfolio overlap among manager groups. The time series back to 1998 enables event studies around regulatory changes and market crises. Amendments matter for research accuracy: studies using only original filings risk measurement error for managers who routinely file corrections.

Financial journalists

Reporters covering asset management and corporate control use 13F data to surface newsworthy position changes. They focus on large quarter-over-quarter share-count swings in high-profile securities, new positions in companies facing public controversy, and exits from formerly large holdings. The gap between report-period-end date and filing date reveals disclosure lag, and amendment filings can shift the narrative around a manager's involvement in contested situations.

LLM and retrieval-augmented-generation teams

Teams building financial retrieval systems use the 13F dataset as a structured-yet-textual corpus combining metadata files, tabular holdings schedules, and cover pages in mixed formats (XML, HTML, plain text). Security-level fields — name, CUSIP, shares, value, discretion — provide ground-truth labels for entity extraction and normalization tasks. The amendment relationship between 13F-HR and 13F-HR/A filings tests a system's ability to resolve document versioning in retrieval.

Specific Use Cases

Reconstructing manager-level portfolio snapshots for crowding analysis

Quantitative researchers extract CUSIP, share count (sshPrnamt), and market value (value, in thousands of USD) from the information table XML across all filers for a given periodOfReport. Differencing these snapshots quarter-over-quarter produces entry, exit, and sizing changes per manager. Aggregating across managers reveals crowded positions — securities where many large institutions hold overlapping exposure — which feed signals for crowding risk models and consensus-deviation strategies. Amendment filings (13F-HR/A with amendmentType of RESTATEMENT or NEW HOLDINGS) must be resolved to avoid stale-position bias in backtests.

Detecting early-stage activist accumulation before Schedule 13D filings

Corporate advisory and proxy solicitation teams monitor 13F filings by matching filer CIK against a watchlist of known activist managers. A new or materially increased share count for a target CUSIP, visible in the quarterly delta, often surfaces weeks before a Schedule 13D is filed (which is only triggered at the 5% beneficial-ownership threshold). The investmentDiscretion field set to SOLE and voting authority concentrated in the Sole sub-element strengthen the signal that the activist has direct control over the position. This early detection supports defensive preparation and shareholder outreach planning.

Building institutional ownership profiles for investor relations targeting

Investor relations teams at public companies filter all filings in a quarter by their company's CUSIP to extract every reporting manager's identity (companyName, cik), share count, and market value. Tracking these across quarters identifies new institutional entrants, departing holders, and shifts in concentration. The voting-authority breakdown (Sole, Shared, None) helps estimate how much of the institutional base votes independently versus defers to proxy advisors, directly informing engagement strategy ahead of annual meetings and contested votes.

Reconciling amendments and confidential treatment to maintain accurate holdings panels

Financial data engineering teams ingest both original 13F-HR filings and amendments (13F-HR/A), using the amendmentType field in primary_doc.xml to determine whether an amendment replaces the entire holdings table (RESTATEMENT) or adds omitted positions (NEW HOLDINGS). The isConfidentialOmitted flag on the summary page identifies quarters where positions were withheld under confidential treatment requests, signaling that a later amendment will restore them. This reconciliation pipeline, keyed on accessionNo and periodOfReport, ensures downstream analytics reflect the most complete and current version of each manager's reported portfolio.

Tracking counterparty and sub-advisor concentration risk for compliance monitoring

Compliance teams at broker-dealers and asset managers pull 13F filings for counterparties and sub-advisors, comparing reported holdings against internal records as of the same periodOfReport. Position-level CUSIP and share-count data reveal undisclosed concentration. The investmentDiscretion field confirms whether shared-discretion positions are properly attributed between co-managers. Amendment frequency for a given filer, tracked via the amendmentNo field, can also serve as an indicator of internal control quality at the filing institution.

Measuring quarter-over-quarter institutional demand shifts for individual securities

Event-driven and fundamental analysts aggregate share-count changes across all filers reporting a specific CUSIP between consecutive quarter-end periods. Net buying or selling pressure from the institutional base — derived from the difference in sshPrnamt values — complements bottom-up thesis work on catalysts such as mergers, restructurings, or earnings inflections. The putCall field isolates options positions (puts vs. calls) from equity holdings, preventing options exposure from distorting the directional ownership signal.

Dataset Access

Dataset Index JSON API: https://api.sec-api.io/datasets/form-13fhr-files.json

This endpoint returns metadata about the Form 13F-HR Files Dataset, including the dataset name, description, last updated timestamp, earliest sample date, total records and total size, covered form types, container format, and content file types. It also returns the download URL for the entire dataset and a list of all individual container files with per-container metadata such as file size, record count, last updated timestamp, and download URL. No API key is required to access this endpoint.

Use this endpoint to monitor which containers have been updated in the most recent refresh run, so you can decide which containers to download on a day-by-day basis rather than re-downloading the entire dataset.

          1
          {
        

          2
            "datasetId": "1f13365b-9ade-61d8-9517-1d1b1e45f257",
        

          3
            "datasetDownloadUrl": "https://api.sec-api.io/datasets/form-13fhr-files.zip",
        

          4
            "name": "Form 13F-HR Files Dataset",
        

          5
            "updatedAt": "2026-04-21T02:53:45.361Z",
        

          6
            "earliestSampleDate": "1998-02-01",
        

          7
            "totalRecords": 1362205,
        

          8
            "totalSize": 10639010779,
        

          9
            "formTypes": ["13F-HR", "13F-HR/A"],
        

          10
            "containerFormat": "ZIP",
        

          11
            "fileTypes": ["TXT", "JSON", "PDF", "HTML", "XML"],
        

          12
            "containers": [
        

          13
              {
        

          14
                "downloadUrl": "https://api.sec-api.io/datasets/form-13fhr-files/2026/2026-04.zip",
        

          15
                "key": "2026/2026-04.zip",
        

          16
                "size": 8491523,
        

          17
                "records": 312,
        

          18
                "updatedAt": "2026-04-21T02:53:45.361Z"
        

          19
              }
        

          20
            ]
        

          21
          }
        

Download Entire Dataset: https://api.sec-api.io/datasets/form-13fhr-files.zip?token=YOUR_API_KEY

Downloads the full dataset as a single ZIP archive containing all containers. This endpoint requires an API key passed via the token query parameter.

Download Single Container: https://api.sec-api.io/datasets/form-13fhr-files/2026/2026-04.zip?token=YOUR_API_KEY

Downloads one individual container file, such as a monthly archive, instead of the full dataset. Container paths and their download URLs are listed in the dataset index JSON API response. This endpoint requires an API key passed via the token query parameter.

Frequently Asked Questions

What form does this dataset cover?

The Form 13F-HR Files Dataset covers Form 13F-HR (quarterly institutional holdings reports) and Form 13F-HR/A (amendments to those reports), both filed on the SEC's EDGAR system.

What does one record in this dataset represent?

One record is a single EDGAR submission of Form 13F-HR or 13F-HR/A, represented as a folder of five files — metadata.json, two XML source documents (cover page and information table), and two HTML renditions — within a monthly ZIP archive. Each folder maps to exactly one accession number.

Who is required to file Form 13F-HR?

Any institutional investment manager exercising investment discretion over $100 million or more in Section 13(f) securities as of the last trading day of any month during the calendar year must file. The filing population includes registered investment advisers, banks, broker-dealers, insurance companies, pension fund managers, and endowment managers.

How often are records added to the dataset?

Form 13F-HR is filed quarterly, with deadlines 45 calendar days after each calendar quarter-end (March 31, June 30, September 30, December 31). Amendments (13F-HR/A) may be filed at any time and appear as separate records.

What time period does the dataset cover?

The dataset includes filings from approximately February 1998 to the present. Pre-2013 filings use plain-text or HTML formats; filings from Q4 2013 onward use structured XML following the SEC's XML mandate.

Are market values reported in dollars or thousands of dollars?

All value fields in the information table and tableValueTotal in the summary page are reported in thousands of USD. A reported value of 1260458 represents approximately $1.26 billion.

How does this dataset differ from Form N-PORT?

Both disclose security-level portfolio holdings periodically, but Form 13F-HR is filed by investment managers under the 1934 Act and covers only Section 13(f) securities (principally equities), while N-PORT is filed by registered investment companies (mutual funds, ETFs) under the 1940 Act and covers all asset classes. N-PORT filings are public with a 60-day delay; 13F-HR filings are public immediately upon EDGAR acceptance.