Update: Household Saving Rate Falls to 2.8% — Debt Prints $18.8T as Student 90+ Hits 10.3%

What changed since the research vintage

In mid-August we rolled up the household balance sheet in consumer finance markets research: ~$18.42 trillion in NY Fed–style debt, a personal saving rate near 3.9%, credit-card stress elevated against calm mortgages, card APRs still ~17 pp over funds, and liquid cash split between ~$15T deposits and $7.85T money market funds. That post answered the structure question — how households save, borrow, and park retail money after the pandemic cash flood and hiking cycle.

This update answers the vintage question: what moved in the newest official prints, and does the three-ledger story still hold? Three information events force a refresh. The New York Fed’s 2026Q1 Household Debt and Credit report (May 2026) prints total balances at $18.794 trillion — up $18 billion QoQ and about $374 billion above our research stock. BEA’s personal saving rate for 2026Q2 prints 2.8%, a −1.1 percentage-point collapse from the research ~3.9% anchor (Q1 was still 3.9%; June’s monthly rate sat near 2.7%). ICI’s weekly MMF tally for the week ended August 19, 2026 puts fund assets at $7.93 trillion, +$80 billion versus the research $7.85T.

The dashboard above is built as a vintage delta: signed meter bars, the saving-rate path into Q2, product prior→new balances, a stress multi-line into 2026Q1, liquid-cash sleeve deltas, and the sticky APR−funds gap.

The headline table: research print vs newest official vintage

Treat the debt +$0.37T carefully. Part of it is simply that the research print used a late-2025 estimated stock; NY Fed’s closed 2026Q1 table is the disclosed vintage. The QoQ move itself was tiny (+$18B / +0.1%). The saving-rate move is the opposite: a true period delta that rearranges the flow ledger without rewriting the stock wealth story.

The saving rate is the loudest vintage delta

BEA’s personal saving rate is a flow: saving as a share of disposable personal income. The research post sat near 3.9% — already a thin post-stimulus norm versus the 26% spike of 2021 Q2. 2026Q1 held at 3.9%. 2026Q2 printed 2.8%. That is a −1.1 pp quarter-to-quarter drop into territory that looks less like “pre-pandemic normal” and more like an active drawdown of the marginal dollar.

A low saving rate still does not mean households are broke in the Z.1 sense. Aggregate net worth remains enormous; equities and housing still dominate assets. It does mean buffers rebuild slowly while revolving borrowers who already show stress in credit-bureau data have less new income left after consumption. Pair this flow meter with our dedicated personal saving rate deep dive — the update’s job is narrower: the research ~4% floor just cracked lower in the newest quarterly vintage.

Monthly color reinforces the quarterly print. June 2026’s personal saving rate near 2.7% (BEA Personal Income and Outlays) sits under the Q2 average, so the soft patch is not a single noisy month averaged away. For consumer-finance desks, the practical read is simple: income is being spent, not warehoused into deposits or MMFs at the margin — even while the stock of MMFs keeps grinding higher on yield.

Debt: flat QoQ, higher vs the research stock

NY Fed’s May release is almost boring on the quarter: total household debt +$18 billion to $18.8 trillion. Mortgages rose $21 billion to $13.19 trillion. HELOCs extended a sixteen-quarter climb (+$12B to $446B). Non-housing balances actually fell $15 billion, led by a seasonal $25 billion drop in credit cards to $1.25 trillion. Auto loans rose $18 billion to $1.69 trillion; student loans were roughly flat at $1.66 trillion.

Versus the research dashboard’s $18.42T / $1.21T cards, the disclosed Q1 stock is higher on both totals and revolving balances even after the seasonal card dip. Mortgages still dominate — about 70% of balances. The shareable framing is not “households levered up another half-trillion in one quarter.” It is “the research estimate was light versus the closed NY Fed table, and the product mix still says housing stock / revolving stress.”

Origination color stayed steady: about $530 billion in new mortgages and $182 billion in new auto loans appearing on credit reports in 2026Q1. Card limits rose another ~$60 billion. Capacity to borrow is still expanding even as early card delinquency transitions tick down a tenth of a point.

Stress: student 90+ worsened; cards paused; mortgages nudged

Aggregate delinquency was flat: 4.8% of outstanding debt in some stage of delinquency, matching 2025Q4 and tied for the highest share since 2017 on the report’s framing. Inside the product cut, the research post’s 7.2% vs 1.1% social card (card vs mortgage) remains directionally right but needs measurement hygiene on this update:

• Student loans: the share of balances 90+ days delinquent rose to 10.3% from 9.6% in 2025Q4the cleanest stress delta in the new PDF.
• Mortgages: transition into serious (90+) delinquency edged up 1.4% → 1.5% annuallyized. Absolute levels remain low; the direction is no longer pure calm.
• Credit cards: transition into early (30+) delinquency ticked down 8.7% → 8.6%. Serious transitions were “mostly unchanged.” Do not read a tenth of a point as a cycle turnbut the research narrative of monotone card deterioration paused for a quarter.
• Auto: early transitions held steady on the Fed’s summary.

For product-level color on the revolving-vs-housing split, keep the household debt delinquency post beside this update. Bank income-statement twins still live in bank loan charge-offs. The vintage message is narrower: student reporting/repayment stress accelerated; mortgage serious flow nudged; card early flow cooled one tick.

Liquid cash: MMFs still absorbing the yield sleeve

ICI’s August 19 week put money market fund assets at $7.93 trillion, up $80 billion from the research $7.85T anchor. Retail MMFs alone sit near $3.11 trillion. Indicative taxable MMF yields still clear money-market deposit accounts by roughly three percentage points, so the dual-pile map from the research post — sticky bank deposits beside yield-sensitive fund shares — has not inverted.

Deposits ex-large time remain enormous (we hold a ~$15T estimate pending a cleaner H.8 restatement in this update). The point for allocation desks is unchanged in geometry and slightly richer in dollars: cash did not leave the household sector; it keeps preferring the wrapper that pays. See money market funds vs deposits for the deep dive; this dashboard’s cash panel is the vintage check that the sleeve grew, not shrunk, into mid-August.

The APR gap barely moved — and that is the point

Commercial-bank card APRs still hover in the low 21% area while effective fed funds eased toward roughly 4.1% on our mid-2026 estimated path. The APR − funds gap remains near 17.2 pp — essentially unchanged from the research ~17.15 pp. Funds moved more than lending rates. Sticky card pricing plus lagged deposit betas remain the same industrial-organization story from opposite sides of the bank book.

That wedge matters more when the saving rate is 2.8%. Households that revolvingly finance consumption at ~21% while earning sub-1% on many deposit products — and only mid-3s in MMFs — are paying for rate stickiness with both cash-flow and delinquency risk. The dedicated credit-card APR vs fed funds chart remains the rate-path companion; this update only needs to say the gap did not close when the saving flow cracked lower.

What would rewrite this update

1. BEA 2026Q3 saving snapping back above 3.5% would mark Q2 as a soft patch rather than a new floor.
2. NY Fed 2026Q2 showing card serious transitions rising againor student 90+ stabilizing — would re-rank the stress panel.
3. A deposit-rate war that closes the MMDA–MMF gap would pull cash back onto bank books without fixing card APRs.
4. Faster funds easing that finally compresses card APRs would shrink the borrowing wedge; if deposit betas still lag, MMFs may keep the cash anyway.
5. Unemployment jump would hit auto and card delinquency harder than mortgages firstthe same sequencing the research post flagged.

Until those print, the live frame versus the research post is narrow: saving flow broke lower (2.8%); debt stock printed higher ($18.8T); student 90+ worsened to 10.3%; MMFs edged up to $7.93T; the APR gap stayed wide.

Caveats and methodology

1. Research print ≠ a single NY Fed PDF cell. The prior post mixed disclosed anchors with estimated late-2025 points. Debt +$0.37T is versus that print, not only versus 2025Q4’s QoQ +$18B.
2. Delinquency concepts differ. This update prefers disclosed 2026Q1 language (early vs serious transitions; student stock 90+ shares). Do not splice the research 7.2% card figure into the new early-transition series without a footnote.
3. Saving rate uses BEA quarterly SAAR (FRED A072RC1Q156SBEA). Monthly June 2.7% is color, not a substitute for the quarterly print.
4. MMF AUM is ICI weekly (week ended Aug 19, 2026). Weekly prints revise; treat $7.93T as the latest disclosed weekly, not a quarter-end Financial Accounts total.
5. Deposit levels in the cash panel remain an estimated hold from the research framing where a fresher H.8 cut is not restated here.
6. Card APR / funds mid-2026 anchors are estimated from Board G.19 / effective funds pathsthe gap is arithmetic, not a risk-adjusted credit spread.
7. Student-loan series still carry repayment-reporting normalization noise; cross-product comparisons remain imperfect.

Primary sources: NY Fed Quarterly Report on Household Debt and Credit, 2026Q1 (May 2026); BEA / FRED personal saving rate through 2026Q2; ICI Money Market Fund Assets (Aug 20, 2026 release); Board G.19 consumer credit / APR context; prior theme baseline in consumer finance markets research.

The shareable takeaway

Versus the research print, the household saving rate fell to 2.8% in 2026Q2 (−1.1 pp), while NY Fed debt printed $18.8T and student-loan 90+ delinquency rose to 10.3%. Money market funds edged up to $7.93T; the card APR−funds gap stayed near 17 pp. The three-ledger household — thin saving flow, mortgage-heavy debt stock, yield-sensitive cash — still disagrees until you read the ledgers together. For the structural essay see the research roll-up; for revolving stress mechanics see the delinquency split.