Z.1 flow of funds tables: what the sector matrices hold

The Z.1 - formally the Financial Accounts of the United States - is where the American credit system gets counted: transactions and levels of financial assets and liabilities, cut by sector and financial instrument at once, plus full balance sheets with net worth for households and nonprofits, nonfinancial corporate and noncorporate business, the Integrated Macroeconomic Accounts and supplemental detail. Federal Reserve Z.1 Financial Accounts of the United States (Flow of Funds) is the Datadory product built on that release, one of five primary records among the eleven pooled datasets on the asset-management-custody-banks hub.

It is also the strongest record Datadory catalogs for this industry - a perfect 10 out of 10 quality score, shared only with SEC Form N-PORT, against a catalog-wide mean of 7.81; just 145 of 1,744 records reach ten.

Every table in the bundle falls into one of five families, and picking the right one saves hours of column-hunting:

The matrices cut the economy by sector and instrument simultaneously: households, nonfinancial business, federal and state-local government, the rest of the world, commercial banks, credit unions, money market funds, mutual funds, ETFs, finance companies, insurers, private and government pensions, agencies and GSEs, securitizers. Fund sectors are named directly - S123.s covers money market funds, S124.1.s covers mutual funds - with related instruments such as L.206 money market fund shares, L.226 mutual fund shares, L.228 life insurance reserves and L.229 pension entitlements.

A concrete read shows how deep the levels run. In the mutual-fund table, the first column LM654090000.Q - "Mutual funds; total financial assets," millions of dollars at market value - opens at 1,245 for 1945:Q4 and 1,298 for 1946:Q4: roughly $1.2 billion held by an industry whose scale the same column then tracks quarter by quarter into six figures.

Mnemonic anatomy makes the rest navigable. Codes combine an instrument prefix with a sector suffix - LM654090000.Q, FA384104005.Q - and every table ships a matching dictionary entry that resolves each code to its description, line number, table and units: "Mutual funds; total financial assets | Line 1 | S124.1.s Mutual funds | Millions of dollars; amounts outstanding end of period, market value, not seasonally adjusted." Six fields describe every series: period, mnemonic, description, line, table, units.

That structure is why quants modeling custody-bank custody flows start at the sector matrices rather than at filings: one quarterly refresh restates the whole who-holds-what matrix, so joining fund-sector levels to the banking system needs no entity resolution. The trade-off runs the other way too - security-level positions live in Form N-PORT's holdings corpus instead.

Temporal - quarterly flows and levels open at 1945:Q4 and annual series reach back to 1919, which makes multi-cycle structural work possible without stitching vintages together. Rate regimes, deleveraging episodes, household net worth swings: all of them fit inside one consistent definition.

Each quarterly release refreshes that full history rather than appending a single quarter - the June 11, 2026 release covered 2026:Q1, and corrections are reposted with notices, as with the June 12, 2026 correction to matrix table M3.s. Two structural caveats belong in any model built on top. First, publication lands roughly 10-11 weeks after quarter end, so align any trigger to the actual schedule rather than the quarter boundary. Second, table numbering changed between vintages - S124.1.s maps to legacy L.122 Mutual funds, F521.s to legacy L.206 Money market fund shares - so cross-vintage joins should ride the published mapping rather than hard-coded numbers.

Household leverage and net-worth cycles. The B.x balance-sheet family turns household net worth, liabilities and debt-service-relevant stocks into a quarterly panel spanning eight decades - the natural base for deleveraging studies and housing-cycle work.

Fund-industry growth accounting. S123.s and S124.1.s give money market and mutual fund assets quarter by quarter from 1945, so market-share and flow-of-capital narratives get measured instead of anecdotal.

Who-lends-to-whom structure. The M3.s flow matrices expose instrument-by-instrument borrowing and lending between sectors each quarter - credit transmission traced end to end rather than inferred.

Benchmarking micro positions. Aggregate N-PORT holdings up and the result must reconcile to these sector levels; the two records are designed to be read side by side, and both carry the pool's perfect quality score.

Citable macro numbers. Every figure traces to the Board's own published accounts with line numbers and units attached, which is what publication teams and academics need before a chart goes in a report.

Within Asset Management & Custody Banks, four records answer adjacent questions, and the useful move is knowing which question you are asking:

The pairing that earns its keep is the Z.1 against N-PORT: one aggregates fund and bank sectors back to 1945:Q4 across 286 tables per release, the other descends to named securities held by named funds from October 2019. Both score 10/10, and the head-to-head at Z.1 vs N-PORT works the trade-off attribute by attribute. For banking features at daily frequency beside the quarterly matrices, FRED Category 13 - Banking Data Sub-tree carries roughly 5,200-6,700 addressable banking series - though it does not restate the sector-by-instrument matrix itself.

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

Because the hard part was never obtaining a release - it is keeping eight decades of restated history coherent as one time series. As published, each quarterly drop replaces its predecessor wholesale, numbering shifted between vintages, and every mnemonic needs a dictionary lookup before a value means anything.

Datadory handles that upstream so the rows just arrive: mnemonics resolved against their descriptions, lines, tables and units; current and legacy table codes carried side by side; corrections folded forward while revision vintages stay distinguishable; and each new quarter appended beside prior decades under one schema instead of replacing them. A sector panel accumulates history rather than resetting at every posting.