Federal Reserve Bank of New York Markets Datasets

Federal Reserve Bank of New York Markets Datasets is the Investment Banking & Brokerage catalog's monetary-operations record: the New York Fed's entire markets hub, covering everything its Open Market Trading Desk touches. Nine product families sit under one roof - Reference Rates (EFFR, OBFR, SOFR, SOFR Averages & Index, TGCR, BGCR); repo and reverse repo operations with announcements, results and propositions; securities lending out of the System Open Market Account portfolio; Treasury securities operations with outright purchases and sales; agency MBS operations across outrights, dollar rolls and coupon swaps; agency CMBS purchases; SOMA holdings summaries by security type; primary dealer statistics; and primary dealer market share, quarterly and year-to-date.

The scale is the story: hundreds of primary-dealer timeseries of positions and financing on their own, weekly portfolio snapshots stretching back to 2003, and rate prints that arrive with the volume and percentile distribution attached rather than a bare number. Get a sample of this dataset cut to the series your desk actually quotes.

Three verified rows ship with the record - two consecutive rate prints and one primary-dealer statistics cell:

reference rates
effectiveDate : 2026-08-20    type             : SOFR
percentRate   : 3.63          volumeInBillions : 2,922
distribution  : 1st 3.58 | 25th 3.60 | 75th 3.68 | 99th 3.71

effectiveDate : 2026-08-19    type             : SOFR
percentRate   : 3.62          volumeInBillions : 2,923

primary dealer statistics
asOfDate   : 2024-07-03
timeSeries : PDPOSMBS-TOT
value      : 116,392  (millions of dollars)

Read what the rows prove rather than what they quote. Behind a single 3.63 SOFR print sits $2.92 trillion of overnight secured volume, spread across thirteen basis points from the 1st to the 99th percentile - dispersion you can model, not just a level to chart. The third row is the other half of the hub: dealer mortgage positioning reduced to a keyed timeseries, PDPOSMBS-TOT, reading $116.4 billion as of July 2024. Every row arrives with its join key already in place, so panels assemble without reshaping anything.

Nine verified fields carry every shape the hub publishes. Three belong to the rate prints - effectiveDate, percentRate and volumeInBillions - joined by the four-volume percentile set that exposes the distribution behind each headline. Operation days add operationId / operationType / operationMethod, securityType and the submitted-versus-accepted amounts. The statistics side keys on As Of Date / Time Series / Value (millions) for dealer surveys and asOfDate for weekly SOMA holding snapshots. Six further field families - securities lending, MBS operations, Treasury operation prices, CMBS purchases, dealer market share and the Dodd-Frank transaction elements - fold under additional fields on request: defined and verified, but not itemized in the core dictionary. Name the ones your models need when you request a sample.

Geography - United States throughout. The Desk's operations, the SOMA portfolio and the primary dealer survey respondents are all U.S. institutions, which makes the record a clean domestic funding-and-operations panel rather than a comparative one.

Temporal - unusually deep for operational data. SOFR runs from its April 2018 inception and EFFR reaches considerably further back; primary dealer surveys and weekly SOMA holdings extend to 2003; transaction-level disclosure files documenting individual Desk trades run back to July 2010.

Granularity - daily rate prints and operation records, weekly portfolio holdings snapshots, weekly to monthly dealer surveys, quarterly and year-to-date market-share cuts, and quarterly transaction-level disclosures. One dataset therefore serves a same-day funding monitor and a twenty-year structural study without a second source.

API, files, or your warehouse. Daily, weekly, or hourly.

Your cadence is your call regardless of the publishing rhythm underneath - a morning warehouse load keeps the funding dashboard current, and nobody on your team babysits a pipeline. Every delivery ships the complete field dictionary above, the sample rows and the coverage profile mapped to the rate series and dealer statistics you named.

1. Funding and treasury desks calibrate short-term funding cost against SOFR, TGCR and BGCR levels with volume and percentile spread attached - basis models get dispersion, not just a level.
2. Rates strategists read repo and reverse repo operation results to see where the Desk supplied or absorbed liquidity, split by collateral class and tenor.
3. Quants and flow researchers mine hundreds of primary-dealer timeseries - positions and financing by instrument class - as features ahead of auctions and issuance calendars.
4. Portfolio and balance-sheet analysts track SOMA holdings by security type week over week to watch the composition of the Fed's book, not just its size.
5. Journalists and academics cite an operational record of policy implementation - operation identifiers, amounts accepted, counterpart detail - instead of paraphrase.

Investors and quant researchers get the funding-rate and dealer-positioning layer beneath every rates strategy; see investors quants use cases. Data scientists and ML engineers get uniformly keyed timeseries ready for feature pipelines; see data scientists use cases. Developers and data-product builders get benchmark rate content for dashboards and products without assembling it from fragments; see developers builders use cases. Market researchers and consultants get policy-implementation context behind capital-market narratives; see market researchers use cases. Journalists, academics and students get citable, identifier-stamped operational records.

Within investment banking and brokerage data, this record owns the central-bank operations and benchmark-rate category. U.S. Treasury Interest Rate Data (Daily Yield Curve) prices time across maturities while this set prices today's overnight funding - we score the pairing directly in our vs U.S. Treasury Interest Rate Data comparison. SIFMA Capital Markets Fact Book reports annual volumes and issuance; WhaleWisdom covers institutional ownership; FINRA BrokerCheck profiles individual firms. None of them document what the Desk actually did and what dealers actually held - that is this record's lane.

Three things worth having in hand.

First, scope is domestic by construction. Desk operations, SOMA holdings and dealer survey respondents are United States institutions; pair with an international source for cross-country funding comparisons.

Second, operation-level fields populate on operation days. The rate prints are the always-on core - operation identifiers, security types and submission amounts appear when the Desk acts, so time-series work should treat them as event records rather than a continuous calendar.

Third, the fold-under list is real inventory. Securities lending, MBS operations, Treasury operation prices, CMBS purchases, dealer market share and the Dodd-Frank transaction elements are verified parts of the hub outside the core nine-field dictionary - say which you need and the sample confirms them end to end.