Form 13F Holdings Dataset — Quarterly Portfolio Holdings of Institutional Investment Managers

The Form 13F Holdings dataset provides structured data from the information tables filed in SEC Form 13F-HR and 13F-HR/A filings. Each record represents one complete information table — the entire portfolio holdings of a single reporting entity for a calendar quarter-end. A single record may contain 10 securities or over 20,000 securities, depending on the manager's portfolio size. The dataset covers all 13F filings published since May 2013, encompasses approximately 3,000 to 4,000 filing managers per quarter, and is completely survivorship-bias-free, including all portfolios of all reporting entities irrespective of their current status. Data is distributed as compressed JSONL files (.jsonl.gz) organized by calendar month.

This is the canonical source for analyzing what institutional investors own across US-listed equities and equity-linked securities at the CUSIP level.

Update Frequency
Daily
Updated at
2026-04-04
Earliest Sample Date
2013-05-01
Total Size
3.9 GB
Container Format
.jsonl.gz
Content Types
JSONL
Form Types
13F-HR, 13F-HR/A

Dataset APIs

Programmatically retrieve the full list of dataset archive files, download URLs and dataset metadata.

Dataset Index JSON API

Download the entire dataset as a single archive file.

Download Entire Dataset:

Download a single container file (e.g. monthly archive) from the dataset.

Download Single Container:

Dataset Files

160 files · 3.9 GB
Download All
2026-04.jsonl.gz801.8 KB
2026-03.jsonl.gz7.1 MB
2026-02.jsonl.gz83.6 MB
2026-01.jsonl.gz34.2 MB
2025-12.jsonl.gz2.9 MB
2025-11.jsonl.gz79.8 MB
2025-10.jsonl.gz32.8 MB
2025-09.jsonl.gz1.7 MB
2025-08.jsonl.gz78.3 MB
2025-07.jsonl.gz36.0 MB
2025-06.jsonl.gz3.1 MB
2025-05.jsonl.gz84.2 MB
2025-04.jsonl.gz33.0 MB
2025-03.jsonl.gz1.2 MB
2025-02.jsonl.gz78.5 MB
2025-01.jsonl.gz30.6 MB
2024-12.jsonl.gz4.8 MB
2024-11.jsonl.gz71.9 MB
2024-10.jsonl.gz34.8 MB
2024-09.jsonl.gz1.1 MB
2024-08.jsonl.gz78.1 MB
2024-07.jsonl.gz33.3 MB
2024-06.jsonl.gz2.3 MB
2024-05.jsonl.gz75.7 MB
2024-04.jsonl.gz27.3 MB
2024-03.jsonl.gz2.4 MB
2024-02.jsonl.gz72.5 MB
2024-01.jsonl.gz25.4 MB
2023-12.jsonl.gz2.9 MB
2023-11.jsonl.gz71.4 MB
2023-10.jsonl.gz23.8 MB
2023-09.jsonl.gz1.8 MB
2023-08.jsonl.gz68.2 MB
2023-07.jsonl.gz24.0 MB
2023-06.jsonl.gz3.0 MB
2023-05.jsonl.gz70.2 MB
2023-04.jsonl.gz19.0 MB
2023-03.jsonl.gz6.6 MB
2023-02.jsonl.gz71.1 MB
2023-01.jsonl.gz18.8 MB
2022-12.jsonl.gz1.1 MB
2022-11.jsonl.gz65.1 MB
2022-10.jsonl.gz20.1 MB
2022-09.jsonl.gz1.3 MB
2022-08.jsonl.gz66.2 MB
2022-07.jsonl.gz19.0 MB
2022-06.jsonl.gz1.4 MB
2022-05.jsonl.gz67.1 MB
2022-04.jsonl.gz17.9 MB
2022-03.jsonl.gz1.3 MB
2022-02.jsonl.gz69.2 MB
2022-01.jsonl.gz17.4 MB
2021-12.jsonl.gz1.1 MB
2021-11.jsonl.gz64.9 MB
2021-10.jsonl.gz16.8 MB
2021-09.jsonl.gz4.9 MB
2021-08.jsonl.gz63.4 MB
2021-07.jsonl.gz17.7 MB
2021-06.jsonl.gz886.4 KB
2021-05.jsonl.gz59.6 MB
2021-04.jsonl.gz16.5 MB
2021-03.jsonl.gz4.6 MB
2021-02.jsonl.gz60.3 MB
2021-01.jsonl.gz12.7 MB
2020-12.jsonl.gz947.8 KB
2020-11.jsonl.gz52.3 MB
2020-10.jsonl.gz14.8 MB
2020-09.jsonl.gz624.2 KB
2020-08.jsonl.gz51.0 MB
2020-07.jsonl.gz15.3 MB
2020-06.jsonl.gz1.5 MB
2020-05.jsonl.gz51.7 MB
2020-04.jsonl.gz14.2 MB
2020-03.jsonl.gz1.1 MB
2020-02.jsonl.gz52.8 MB
2020-01.jsonl.gz14.8 MB
2019-12.jsonl.gz356.7 KB
2019-11.jsonl.gz49.3 MB
2019-10.jsonl.gz17.5 MB
2019-09.jsonl.gz978.8 KB
2019-08.jsonl.gz51.7 MB
2019-07.jsonl.gz16.0 MB
2019-06.jsonl.gz2.2 MB
2019-05.jsonl.gz52.2 MB
2019-04.jsonl.gz18.6 MB
2019-03.jsonl.gz1.3 MB
2019-02.jsonl.gz50.2 MB
2019-01.jsonl.gz12.6 MB
2018-12.jsonl.gz763.2 KB
2018-11.jsonl.gz50.6 MB
2018-10.jsonl.gz14.5 MB
2018-09.jsonl.gz884.7 KB
2018-08.jsonl.gz49.7 MB
2018-07.jsonl.gz13.0 MB
2018-06.jsonl.gz584.2 KB
2018-05.jsonl.gz51.0 MB
2018-04.jsonl.gz11.8 MB
2018-03.jsonl.gz1.3 MB
2018-02.jsonl.gz49.6 MB
2018-01.jsonl.gz10.9 MB
2017-12.jsonl.gz638.4 KB
2017-11.jsonl.gz46.5 MB
2017-10.jsonl.gz11.8 MB
2017-09.jsonl.gz813.1 KB
2017-08.jsonl.gz45.3 MB
2017-07.jsonl.gz12.0 MB
2017-06.jsonl.gz764.8 KB
2017-05.jsonl.gz47.2 MB
2017-04.jsonl.gz10.9 MB
2017-03.jsonl.gz1.6 MB
2017-02.jsonl.gz45.7 MB
2017-01.jsonl.gz12.0 MB
2016-12.jsonl.gz651.6 KB
2016-11.jsonl.gz43.9 MB
2016-10.jsonl.gz12.3 MB
2016-09.jsonl.gz1.1 MB
2016-08.jsonl.gz45.1 MB
2016-07.jsonl.gz10.9 MB
2016-06.jsonl.gz2.0 MB
2016-05.jsonl.gz44.1 MB
2016-04.jsonl.gz10.9 MB
2016-03.jsonl.gz902.2 KB
2016-02.jsonl.gz46.2 MB
2016-01.jsonl.gz9.6 MB
2015-12.jsonl.gz1.5 MB
2015-11.jsonl.gz43.4 MB
2015-10.jsonl.gz11.1 MB
2015-09.jsonl.gz456.5 KB
2015-08.jsonl.gz45.9 MB
2015-07.jsonl.gz11.2 MB
2015-06.jsonl.gz1.2 MB
2015-05.jsonl.gz44.4 MB
2015-04.jsonl.gz10.3 MB
2015-03.jsonl.gz1.7 MB
2015-02.jsonl.gz47.8 MB
2015-01.jsonl.gz8.0 MB
2014-12.jsonl.gz1.7 MB
2014-11.jsonl.gz39.5 MB
2014-10.jsonl.gz11.6 MB
2014-09.jsonl.gz1.6 MB
2014-08.jsonl.gz42.6 MB
2014-07.jsonl.gz9.4 MB
2014-06.jsonl.gz306.8 KB
2014-05.jsonl.gz43.0 MB
2014-04.jsonl.gz8.6 MB
2014-03.jsonl.gz766.9 KB
2014-02.jsonl.gz41.5 MB
2014-01.jsonl.gz9.2 MB
2013-12.jsonl.gz809.0 KB
2013-11.jsonl.gz37.5 MB
2013-10.jsonl.gz10.4 MB
2013-09.jsonl.gz650.7 KB
2013-08.jsonl.gz41.0 MB
2013-07.jsonl.gz7.4 MB
2013-06.jsonl.gz767.0 KB
2013-05.jsonl.gz706.4 KB
2013-04.jsonl.gz20 B
2013-03.jsonl.gz20 B
2013-02.jsonl.gz20 B
2013-01.jsonl.gz20 B

What This Dataset Contains

Each record is one complete 13F information table — the entire portfolio holdings of a single reporting entity for a given calendar quarter. A single JSONL object contains default filing attributes (id, CIK, ticker, accession number, etc.) alongside a holdings array, where each item in the array is one security position. A manager holding 500 securities produces one record with 500 entries in the holdings array. The unit of observation is the manager-quarter pair.

Cover pages are excluded from this dataset and available in the dedicated 13F Cover Pages dataset.

Example Record Structure

Example
1 {
2 "default filing attributes: id, cik, ticker, etc.": "...",
3 "holdings": [
4 {
5 "nameOfIssuer": "ADVANCED MICRO DEVICES INC",
6 "titleOfClass": "COM",
7 "cusip": "007903107",
8 "ticker": "AMD",
9 "cik": "2488",
10 "value": 1565000,
11 "shrsOrPrnAmt": {
12 "sshPrnamt": 18527,
13 "sshPrnamtType": "SH"
14 },
15 "investmentDiscretion": "SOLE",
16 "votingAuthority": {
17 "Sole": 0,
18 "Shared": 0,
19 "None": 18527
20 }
21 }
22 ]
23 }

Holdings Array Fields

Each object in the holdings array contains:

FieldDescriptionExample
nameOfIssuerIssuer name"ADVANCED MICRO DEVICES INC"
titleOfClassSecurity class or title"COM", "SHS CL A", "PUT"
cusip9-character CUSIP identifier"007903107"
tickerTicker symbol (enriched, not in raw EDGAR data)"AMD"
cikIssuer CIK (enriched, not in raw EDGAR data)"2488"
valueAbsolute market value in USD (standardized)1565000
shrsOrPrnAmt.sshPrnamtShare count or principal amount18527
shrsOrPrnAmt.sshPrnamtTypeShares ("SH") or principal ("PRN")"SH"
investmentDiscretionSole, defined, or other discretion"SOLE", "DEFINED", "OTHER"
votingAuthority.SoleShares with sole voting authority0
votingAuthority.SharedShares with shared voting authority0
votingAuthority.NoneShares with no voting authority18527
putCallPut/call flag for option positions"Put", "Call", or absent
otherManagerManager ID sharing discretionOptional

Filing-Level Metadata Fields

Each record also includes top-level filing attributes alongside the holdings array. These follow the standard Filing Query API response structure:

FieldDescriptionExample
formTypeOriginal or amendment"13F-HR" or "13F-HR/A"
cikReporting manager's SEC Central Index Key"1067983"
periodOfReportQuarter-end date (as-of date for positions)"2024-12-31"
filedAtTimestamp submitted to the SEC"2025-02-14T..."
accessionNoUnique SEC filing identifier"0001067983-25-000123"

What Is Included

  • All Section 13(f) securities positions from every 13F-HR and 13F-HR/A filed since May 2013.
  • Long positions in equities, equity options (puts and calls), convertible debt, warrants, and other securities on the SEC's official 13(f) securities list.
  • Both original filings and amendments. Both versions appear in the dataset; consumers must de-duplicate by keeping only the most recent filing per manager-quarter.
  • Each holding is enriched with the issuer's ticker symbol and CIK, mapped from the CUSIP. While CUSIP is the only identifier in the raw EDGAR data, this dataset adds ticker and CIK for easier cross-referencing.

What Is Excluded

  • Sub-threshold managers: Holdings of managers below $100 million in 13(f) securities.
  • Non-13(f) securities: Corporate bonds, municipal bonds, money market instruments, private placements, and foreign securities not listed on US exchanges.
  • Short positions: Form 13F requires only long position disclosure.
  • Confidential treatment positions: Withheld from the initial filing; may appear in later amendments.
  • Option contract details: Strike price, expiration, and contract terms are not reported in 13F filings.

The Underlying Filing — Form 13F-HR

Form 13F-HR is the quarterly report filed by institutional investment managers who exercise investment discretion over $100 million or more in Section 13(f) securities. Each filing has three components:

  1. Cover Page — Manager identity, report date, amendment status, confidentiality requests. Available separately in the 13F Cover Pages dataset.
  2. Summary Page — Aggregate holding count and total market value.
  3. Information Table — Structured XML listing every 13(f) security held at calendar quarter-end, with value, shares, CUSIP, voting authority, and discretion type per position.

This dataset extracts the information table entries and combines them with filing-level metadata into structured, analysis-ready records.

Who Files and When

Filing Population

Form 13F is filed by institutional investment managers exercising investment discretion over $100 million or more in Section 13(f) securities. The SEC defines this term broadly under Section 13(f)(6)(A) of the Securities Exchange Act of 1934, covering:

  • Registered investment advisers —The largest filer category. Includes BlackRock, Vanguard, Fidelity, and thousands of smaller advisory firms.
  • Hedge fund management companies —Bridgewater Associates, Citadel Advisors, Renaissance Technologies, and others. Only long positions are disclosed.
  • Mutual fund and investment company sponsors —File at the manager level, aggregating holdings across all sub-advised funds.
  • Pension funds and endowments —CalPERS, New York State Common Retirement Fund, university endowments with internally managed portfolios.
  • Insurance companies —Berkshire Hathaway, MetLife, and other insurers with large equity portfolios.
  • Bank trust and asset management divisions —JPMorgan Chase, Goldman Sachs, Bank of America.
  • Sovereign wealth funds —Norges Bank Investment Management and other foreign institutions with US equity holdings above the threshold.
  • Family offices and broker-dealers —Those with 13(f) holdings exceeding $100 million.

Managers whose 13(f) holdings never reach $100 million, those holding only non-13(f) securities, and foreign managers with no US-listed holdings do not file.

Filing Schedule

Filings are due quarterly, within 45 calendar days of each calendar quarter-end:

QuarterHoldings As-OfDeadline
Q1March 31May 15
Q2June 30August 14
Q3September 30November 14
Q4December 31February 14

Amendment Filings (13F-HR/A)

Amendments replace the entire information table when the original contained errors, when previously omitted positions are discovered, or when confidential treatment expires. They are full restatements, not incremental corrections.

Regulatory Framework

The filing obligation is established by Section 13(f) of the Securities Exchange Act of 1934 and implemented by Rule 13f-1. The related Rule 13f-2 governs short position reporting for institutional investment managers. The SEC publishes an official list of Section 13(f) securities quarterly, defining the reportable universe: exchange-traded equities, equity options, convertible debt, and closed-end fund shares.

vs. Schedules 13D/13G (Beneficial Ownership)

13F reports a manager's full portfolio quarterly. Schedules 13D/13G are event-driven filings triggered when a single position crosses 5% beneficial ownership. 13D adds the investor's intentions, source of funds, and plans — information absent from 13F. 13D is filed within 10 days of crossing 5%; 13F lags up to 45 days. Use 13F for portfolio-level analysis; use 13D/13G for concentrated ownership and activist tracking.

vs. Form N-PORT (Fund-Level Holdings)

13F reports at the manager level, covering only 13(f) securities. Form N-PORT reports at the individual fund level and covers all asset classes (equities, bonds, derivatives, repos). N-PORT includes richer per-position data (coupon, maturity, delta, liquidity classification) but is filed by registered investment companies only, not by hedge funds or pensions.

vs. Forms 3, 4, 5 (Insider Transactions)

Insider forms report individual buy/sell transactions within 2 business days and cover only the insider's holdings in their own company. 13F provides quarterly snapshots with no transaction detail but covers the manager's entire 13(f) portfolio across all issuers.

vs. Raw 13F Filings on EDGAR

This dataset contains the same underlying information as raw EDGAR XML filings, but in a structured, analysis-ready JSONL format. Raw filings require XML parsing, schema handling, and multi-document joins. This dataset delivers each information table as a single JSON record with all holdings in a holdings array alongside filing-level metadata, enriched with ticker and CIK per security.

Who Uses This Dataset

Quantitative investment professionals build crowding signals, institutional flow estimates, and "smart money" replication strategies from quarter-over-quarter changes in 13F holdings. Crowding metrics (how many managers hold the same security, and how concentrated that overlap is) feed directly into risk models.

Fundamental equity analysts and portfolio managers monitor peer positioning, detect activist fund accumulation, screen for under-owned stocks, and assess consensus vs. variant positioning.

Risk managers at multi-manager platforms and prime brokers measure cross-manager portfolio overlap, detect concentration risk, and cross-check 13F disclosures against internally reported positions.

Institutional allocators (pension investment staff, endowment CIOs, fund-of-funds analysts) use 13F filing histories during manager due diligence to verify strategy consistency, detect style drift, and compare portfolio composition against peers.

Investor relations officers at public companies identify their institutional shareholder base each quarter, target outreach to new holders, monitor activist positioning, and support proxy solicitation with ownership data.

Sell-side equity research and sales teams map shareholder bases for covered companies, detect ownership shifts, and support investment banking advisory mandates with institutional holder analysis.

Academic researchers rely on 13F data for empirical studies of institutional herding, the predictive content of holding changes, mutual fund skill measurement, common ownership effects on competition, and the relationship between institutional ownership and corporate governance.

Financial journalists report on 13F filings as quarterly events, covering celebrity investor portfolio moves and broader sector rotation trends.

Regulatory and legal professionals reference 13F data in market surveillance, enforcement investigations, securities litigation, and antitrust analysis of common ownership.

Specific Use Cases

1. Institutional Crowding and Liquidation Risk Scoring

Quantitative risk managers build crowding scores for each security by counting distinct 13F holders, computing ownership concentration (HHI across managers), and tracking quarter-over-quarter holder count changes. Securities where a small number of large managers control outsized institutional ownership relative to float are flagged as high-crowding-risk names. This analysis is a standard input to portfolio risk models because crowded positions suffer amplified drawdowns during forced-selling episodes.

2. Smart Money Portfolio Replication

Quantitative researchers and retail platforms track quarterly holding changes of curated high-conviction managers. New position entries, full exits, and significant size changes (>25% share-count change) are the highest-signal observations. Academic research demonstrates that replicating skilled managers' high-conviction positions generates risk-adjusted alpha even with the 45-day reporting lag. Multiple ETFs and fintech platforms operationalize this methodology.

3. Quarterly Shareholder Base Identification

Investor relations teams compile a complete institutional holder list for their company each quarter by filtering all 13F records matching the company's CUSIPs. Quarter-over-quarter comparison reveals new holders, departing holders, and position size changes. Managers are classified by type (index, active, hedge fund, pension) for targeted engagement. This is the foundational IR workflow after each 13F deadline.

4. Style Drift Detection in Manager Due Diligence

Institutional allocators map a manager's 13F CUSIPs to security attributes (market cap, sector, style factor) and compute portfolio-level statistics each quarter: weighted-average market cap, sector concentration, top-10 position weight, number of holdings. Significant deviations from the manager's stated mandate or historical norms trigger review. A "small-cap value" manager shifting into large-cap growth positions is a concrete red flag that 13F data reveals independently of the manager's own reporting.

5. Sector Rotation and Thematic Flow Analysis

Macro strategists and sector analysts map CUSIPs to GICS sectors or custom themes (AI/semiconductors, clean energy, defense) and compute total institutional value, holder count, and quarter-over-quarter changes by sector. Normalizing for market-cap changes isolates flow-driven vs. price-driven movement. The resulting sector rotation signals are published by sell-side strategists and consumed by portfolio allocators.

6. Common Ownership and Antitrust Cross-Holding Measurement

Researchers compute the degree to which competing firms in the same industry are held by overlapping institutional managers. For each company pair, 13F data reveals shared holders, their ownership stakes, and concentration of overlapping ownership. Modified HHI metrics quantify the potential competitive effects. This analysis is the empirical foundation for the "common ownership" hypothesis debated by the DOJ, FTC, and competition scholars.

7. Activist Campaign Early Warning

Corporate governance teams and event-driven funds maintain watchlists of known activist manager CIKs. Each quarter, they extract new positions and significant increases for these managers, flagging when an activist's stake exceeds meaningful thresholds (>1% of outstanding shares or a top-20 position in the activist's portfolio). Early detection provides lead time for proactive engagement or defense preparation before a public campaign is announced.

Important Data Nuances

Values are standardized absolute USD amounts. Raw EDGAR 13F filings inconsistently mix absolute values and values in thousands of USD. This dataset normalizes all values to absolute dollar amounts (e.g., a value of 1565000 means $1,565,000). No manual conversion is needed.

CUSIP, ticker, and CIK are all available per holding. While CUSIP is the only security identifier in the raw EDGAR data, this dataset enriches each holding with the issuer's ticker symbol and CIK, pre-mapped from the CUSIP. CUSIPs change after corporate actions; building a CUSIP history table or using a commercial security master is necessary for accurate time-series analysis.

Amendments create apparent duplicates. Both the original 13F-HR and any 13F-HR/A amendment appear in the dataset. De-duplicate by keeping only the most recent filing date per manager-quarter combination.

Data is inherently stale. Positions are as of quarter-end but filings appear up to 45 days later. Active managers may hold materially different portfolios by publication date.

Aggregate vs. disaggregate reporting varies. Some managers consolidate all accounts; others report separately by sub-adviser. This affects position summation.

No cost basis, returns, or weights. 13F filings contain no purchase price, gain/loss, or performance data. Portfolio weights must be computed from position values.

Option positions lack contract details. Put/call positions report the underlying CUSIP, value, and share-equivalent quantity, but not strike price, expiration, or contract terms.

Dataset Access

The full dataset is available for download as a ZIP archive containing monthly .jsonl.gz files. Each line in a JSONL file is one complete 13F information table record — the entire portfolio holdings of a single reporting entity for a given quarter. Monthly files are organized by calendar year (e.g., 2026/2026-03.jsonl.gz). Coverage begins May 2013 and updates as new filings are processed from EDGAR.

Download URL: https://api.sec-api.io/datasets/form-13f-holdings.zip?token=YOUR_API_KEY

Form types included: 13F-HR, 13F-HR/A

Total archive size: ~3.9 GB compressed

Frequently Asked Questions

What is the difference between 13F-HR and 13F-HR/A? 13F-HR is the original quarterly filing. 13F-HR/A is an amendment that replaces the entire information table. If a manager files both for the same quarter, use the amendment (most recent filing date) and discard the original to avoid double-counting.

Why do some managers appear to hold very few positions? Some managers request confidential treatment for portions of their portfolio, withholding those positions from the initial filing. The withheld positions may appear in a later 13F-HR/A amendment. Additionally, managers that primarily hold non-13(f) securities (e.g., fixed income) will show only their 13(f) holdings, which may be a small fraction of their total AUM.

Can I compute a manager's total AUM from 13F data? Only partially. Summing the value field across all positions gives the total 13(f) portfolio value, but this excludes non-13(f) assets (bonds, private investments, foreign securities, cash). For total AUM, cross-reference with the manager's Form ADV filing.

How do I handle CUSIPs that change due to corporate actions? CUSIPs change when companies merge, spin off divisions, or undergo other corporate actions. Each holding in this dataset includes the ticker and CIK alongside the CUSIP, which helps with identification. For accurate time-series analysis, building a CUSIP history table from corporate action data or using a commercial security master is still recommended.

Is the value field in thousands or absolute dollars? Raw EDGAR 13F filings inconsistently mix both conventions — some filers report in thousands, others in absolute dollars. This dataset standardizes all values to absolute USD amounts. A value of 1565000 means $1,565,000. No manual conversion is needed.

Are ETF and mutual fund holdings included? Shares of ETFs and closed-end funds that appear on the SEC's 13(f) securities list are included as holdings. The 13F shows which managers hold these funds. For the internal holdings of those funds, use Form N-PORT.