Flow Of Funds Mortgage Debt Data: Dataset Structure and Field Coverage

The current release covers 2026:Q1 and was issued June 11, 2026, then corrected and reposted June 12, 2026 after a technical error in table M3.s - a reminder that these are living statistical series rather than frozen archives. Mortgage REITs matter to this framework because they are levered holders of mortgage assets, and the Z.1 is where the whole-market stock those books draw on gets counted. For the accounting vocabulary, start with the flow of funds definition.

The mortgage detail lives in specific table families and in the series mnemonics themselves. The release's companion file z1_table_mapping.csv (9.3 KB, verified August 21, 2026) maps new table IDs to retired series, including the F4.5, F4.5a and F4.5b tables for total mortgages, one-to-four-family residential mortgages and multifamily residential mortgages by holder group and property type - the direct successors of the former Mortgage Debt Outstanding releases.

Each CSV column header encodes its measure. FA893065005.Q parses as FA (annual transactions flow), 89 (all sectors), 3065 (mortgages) and an asset-side position; FL prefixes level series and FG growth rates, with the .Q suffix marking quarterly frequency. Missing history appears as ND (No Data), never as zero - a detail that silently corrupts averages if ignored.

The Z.1 grid answers the question few lender surveys can: which sector owns the loans. Coverage splits by sector and instrument, and the 2026:Q1 print for total mortgages reads 593,832 at seasonally adjusted annual rates, of which one-to-four-family residential mortgages account for 303,356 and multifamily for 96,400. The holder map below shows how the six sector groups divide that ledger.

Two practical notes before charting shares. Some sector series begin later than 1945:Q4, so early-decade holder comparisons can be thinner than the headline start date implies. And because each release refreshes the full history, share calculations should be re-run on the latest vintage rather than cached.

Six steps take the raw Z.1 bundle to a defensible mortgage REIT research file:

Observed mortgage tables begin at 1945:Q4, and the printed sample rows show the arc: total mortgages read 6,179 at 1946:Q4 (one-to-four-family 4,394, multifamily 348) against 593,832 at 2026:Q1 - roughly 96 times higher across eight decades. Few series in the 1,744-dataset Datadory catalog reach that depth, and none in this industry's slice matches it.

The sibling benchmarks are shorter but complementary. The H.15 runs daily from July 1, 1954 - its first federal funds observation prints 1.13 - while Nareit's monthly workbook starts at December 1971 and now holds 665 observations. Splicing all three lets one study cover the pre-securitization era, the savings-and-loan unwind and the post-2008 agency boom on consistent public data.

Depth comes with a revision obligation. Every Z.1 release restates the full history, and the June 12, 2026 correction of table M3.s shows even intra-month fixes happen. Archive vintages if your work needs point-in-time values, and read the head-to-head in Fed Z.1 Financial Accounts vs H.15 Rates before choosing which release anchors a given question.

The Z.1 counts the whole mortgage market; two Nareit-rooted records show the listed sliver of it. The Nareit Mortgage REIT Sector Overview tracks roughly 30 vehicles, with FTSE Russell aggregate metrics as of 7/31/2026 showing a 13.14% dividend yield, and a sortable roster of about 22 companies per pull carrying tickers, prices as of the 08/21/2026 close and one-year total returns that range widely - AGNC Investment Corp at +28.88% against ACRES Commercial Realty at -27.98%.

For packaged exposure, Yahoo Finance's REM quote and holdings feed covers the iShares ETF tracking the FTSE Nareit All Mortgage Capped Index since May 1, 2007: net assets of $532.81 million, a 0.48% expense ratio and top holdings led by Annaly Capital Management at 25.74% of assets, AGNC at 16.67% and Starwood Property Trust at 7.63%. That top-ten array is also the quickest way to see the agency home-financing versus commercial-credit split that sector aggregates blur. Quants weighing these sources side by side will find the ranking logic on the investors quants use cases page.

Start with the mortgage-reits data guide for the full seven-record landscape, then read how the same Z.1 series feed models in the data scientists guide to mortgage REITs data. For the raw series, go straight to the Federal Reserve Financial Accounts of the United States - Z.1 Statistical Release or the free mortgage-reits datasets list, and browse the wider pool on the mortgage-reits data hub.