Bank reserves vs GDP
Reserves fell 0.4 pp on the week to 9.0 %, 1.0 pp above the cited 8.0 % breach line.
The overview
Follow the forces shaping US debt and liquidity. Every reading sourced. Every threshold explained.
Nothing scheduled
No scheduled events
No scheduled events
Reserves fell 0.4 pp on the week to 9.0 %, 1.0 pp above the cited 8.0 % breach line.
Overnight RRP take-up fell $838 M over the week to $1 B, $24 B below the project's $25 B stress line.
Net interest $1 054 B (trailing twelve months) vs national defense $952 B: interest is 111 % of defense.
10-year Treasury yield fell 0.01 pp over the week to 4.96 %, 0.04 pp below the project's 5.00 % breach line.
30-year Treasury yield fell 0.050 pp over the week to 5.290 %, 0.020 pp above the project's 5.270 % stress line.
Door 4 open: fed_assets_13w_change +$10 B (calm); buyback_quarter_capacity +$67 B (tracking); doors austerity, taxes locked.
Reserves fell $115 B on the week to $2.92 T, $78 B below the cited $3.00 T watch line.
Standing Repo Facility take-up was $0.0 B on 22 Sep 2026; 4 of the last 5 sessions saw take-up.
Fed funds − IORB is unchanged at −2.0 bp (20-day mean; −2 bp on 21 Sep 2026), 2 bp below Lavish's 0.0 bp stress line. His reading: abundant near −7 bp, ample near −1 bp, positive means scarce.
19-Year 11-Month auction on 15 Sep 2026: bid-to-cover 2.57× vs 2.64× trailing-12 mean; dealers 16.9 %, indirects 52.5 %; high yield 5.420 %; tail not computed.
7Y to 10Y buyback on 17 Sep 2026: $9.74 B offered against a $4.0 B cap (2.4× oversubscribed); Treasury accepted $2.39 B (−$1.62 B vs the cap).
Bills as share of marketable debt rose 0.6 pp over the month to 22.8 %, 2.2 pp below the project's 25.0 % stress line.
Net interest as share of receipts fell 0.2 pp over the month to 19.6 %, 0.6 pp above the cited 19.0 % watch line.
Deficit as share of GDP, trailing 12 months fell 0.5 pp over the month to 5.4 %, 0.6 pp below Lavish's 6.0 % stress line.
The deferred asset rose $147 M on the week to −$232.8 B, $17 B above the cited −$250.0 B stress line; 47 months since the zero line broke.
Wage growth (4.1 %) minus CPI (3.7 %) is +0.4 pp, 0.1 pp below the project's 0.5 pp watch line.
Headroom under the $41.104 T limit is $1 192 B, $308 B below the project's $1 500 B watch line.
QT-to-QE trigger: not triggered (rrp_balance stress; reserves_level watch).
The TGA rose $155 B over the week to $973 B.
The TGA changed +$186 B over four weeks, $262 B below the project's $300 B watch line.
SOFR − EFFR fell 2 bp over the week to −3 bp, 8 bp below the project's 5 bp watch line.
Reserve-management purchases: the Desk's schedule is $0 B for the period from 15 Sep 2026; none since 14 Aug 2026 (last schedule $10 B per period). Outright Treasury purchases settled: $28.7 B in Aug 2026 (all purposes, including reinvestment of agency principal).
The H.4.1 line for 16 Sep 2026 is zero-equivalent at $0.00 B (below the $50 M rounding band).
The H.4.1 line for 16 Sep 2026 shows $94 M, $406 M below Lavish's $500 M watch line.
18 operations scheduled 1 Jul–29 Sep 2026 with caps totalling $67.0 B ($42 B liquidity support, $25 B cash management).
Foreign official holdings fell $5.0 B on the month to $3 773.1 B (no consecutive monthly rise).
Total debt rose $89 B over the month to $40.102 T; the last trillion took 113 days.
Net interest over the trailing twelve months fell $7 B to $1 054 B, $196 B below the project's $1 250 B watch line.
Gross interest over the trailing twelve months fell $14 B to $1 359 B; net interest $1 054 B, so $305 B was interest received by trust funds.
Average rate on interest-bearing debt rose 0.118 pp over the year to 3.490 %.
$10.8 T of marketable debt matures within 12 months (34 % of the stock) at a weighted-average coupon of 3.37 %; recent coupon auctions cleared 1.27 pp higher.
Deficit, trailing 12 months fell $178 B over the month to $1 768 B.
Receipts as share of GDP is 16.6 %, inside the 16 %–19 % band (inside = Hauser's Law holding (Door 2 locked)).
Nominal GDP growth 6.6 % minus the average rate on interest-bearing debt 3.49 %: a gap of +3.1 pp (2.1 pp of the growth was real output).
Primary deficit as share of GDP, trailing 12 months fell 0.5 pp over the month to 2.2 %.
Growth 6.6 % vs rate 3.49 %: a gap of 3.1 pp. Primary deficit 2.2 % of GDP (trailing twelve months). Net: gaining. On his fiscal-year-to-date basis (11 months annualised): primary deficit 3.2 % of GDP, net −0.1 pp: treading water.
10-year term premium (ACM) fell 0.13 pp over the month to 0.64 %, 0.11 pp below the project's 0.75 % watch line.
10-year real yield 2.62 % (+0.02 % over the week); breakeven 2.34 % (−0.03 pp).
2s30s curve fell 16 bp over the week to 53 bp, with no threshold set.
The Fed raised the target range on 16 Sep 2026: +25 bp to 3.75 %–4.00 %, unanimous (FOMC statement).
30-year since the last decision: −6 bp since the 25 bp hike on 16 Sep 2026 (5.35 % → 5.29 %).
Fed balance sheet since the last decision: $0 B since the 25 bp hike on 16 Sep 2026 ($6.75 T → $6.75 T).
Securities held outright are +2.6 % over four quarters (+1.3 % over the quarter); his Response line is 6 %.
2-year 4.76 %, 76 bp above the target ceiling of 4.00 %: the market is priced for tightening beyond the current range (above the ceiling for 153 days).
Door pulled: hike. The balance sheet is $0 B and the 30-year is −6 bp since 16 Sep 2026: the floor is not showing. A hike with a flat or shrinking book is the one branch of the framework where the floor claim is not yet supported. Four-quarter holdings growth 2.6 % (his Response line: 6 %).
Fed balance sheet, 13-week change fell $5 B over the week to +$10 B, $40 B below the project's $50 B watch line.
Outlays growth, trailing 12 months fell 2.3 pp over the month to +0.4 %, with no threshold set.
Open items nearest their resolution date
The FIMA line stayed at zero through August, which points to Tokyo selling rather than borrowing dollars. The August TIC data (published around 16 Oct) is where that shows up in Japan's holdings.
Resolves on 16 Oct 2026 via data release · from The Monday Brief: Four Days and $119 Billion (8 Sept 2026)
His falsification test: two consecutive monthly rises would cut against the thesis. At send the July print was already down ($3,773.1 B vs $3,778.1 B in June and $3,845.9 B in May), so the October print can only fail to break the thesis. Resolves from the TIC 'Of Which: Foreign Official' row: fewer than two consecutive rises = as expected (thesis survives); two or more = contrary. The Japan-specific twin is the 16 Oct TIC item.
Resolves on 16 Oct 2026 via data release · from The Informationist Institutional, No. 1 — The Global Bond Bar Tab (20 Sept 2026)
At send (20 Sep): the first enlarged operation on 10 Sep had a $6 B cap, $10.5 B offered and $5.19 B taken. Resolves from the long-end operations through 4 Nov and the Q4 tentative buyback schedule published at the refunding: 'kept' if long-end caps of at least $6 B per operation persist (as expected), 'cut' if they are lower (mixed), 'lapsed' if no long-end operations are scheduled (contrary). The $2 B caps on the pre-August schedule rows for 27 Oct and 4 Nov are not read as a cut until the Q4 schedule replaces them.
Resolves on 4 Nov 2026 via the post-refunding buyback schedule · from The Informationist Institutional, No. 1 — The Global Bond Bar Tab (20 Sept 2026)
A CNBC report said so; Treasury has not confirmed. Evidence to weigh: the TGA path around operation days against bill issuance.
Resolves on 5 Nov 2026 via editor · from The Monday Brief: Last Words Before Fed Blackout Period (31 Aug 2026)
The one test currently running against the thesis. At send (16 Sep): fed funds printed 2 bp under IORB, the standing repo facility had gone days without a taker, and reserves were 9.3 % of GDP (week to 16 Sep) against the Fed staff's 10 % line. Closed by the editor on 31 Dec (his next quarterly) from the states of the EFFR − IORB, standing-repo and reserves-to-GDP gauges on that date: plumbing still loose = contrary to the thesis; tightening = as expected.
Resolves on 31 Dec 2026 via editor · from The Informationist Institutional, No. 1 — The Global Bond Bar Tab (20 Sept 2026)
How earlier watch items turned out
Pre-registered test: weakness at the 30-year auction (a tail, a low bid-to-cover, a heavy dealer take) would say the term premium is structural rather than a passing scare.
Resolved 22 Sept 2026 via TreasuryDirect auction result · from The Monday Brief: The Long End Didn't Take It (10 Aug 2026)
The belly is outside the expanded buyback window, so these auctions meet the market on their own.
Resolved 22 Sept 2026 via TreasuryDirect auction result · from The Monday Brief: Friday at Ten (24 Aug 2026)
Resolved 22 Sept 2026 via TreasuryDirect auction result · from The Monday Brief: Friday at Ten (24 Aug 2026)