Public Macro-Liquidity & Plumbing Monitor · Updated 2026-10-17
Most market commentary confuses central-bank balance sheets, government cash drains, and commercial-bank credit creation. This page separates the five distinct pipes of dollar liquidity using official Federal Reserve, US Daily Treasury Statement, and Bank for International Settlements data—and tests which signals have genuinely mattered for markets.
2026-10-14 it holds just $0.20bn—meaning Treasury cash rebuilds now drain commercial-bank reserves dollar-for-dollar.
2026-10-13 shows the Treasury General Account rose to $920.2bn (up from $885.8bn on Wednesday 2026-10-14), pulling cash out of private bank accounts.
SOFR 3.89%) sits 1 basis point below the interest rate the Fed pays on reserves (IORB 3.90%), showing that $3,022bn of bank reserves has not yet triggered acute funding scarcity.
What it is: Fed Total Assets ($6,756bn) minus Treasury Cash ($886bn) minus All Reverse Repos ($327bn) as of 2026-10-14. Bank reserves alone are $3,031bn (1y low: $2,828bn).
What it is: Exact closing cash balance from the Daily Treasury Statement on 2026-10-13 (6 days newer than the weekly Fed H.4.1 release). Over the past 20 business days, net non-debt fiscal spending was $+51.2bn.
What it is: Commercial-bank loans and securities as of 2026-10-07. Unlike Fed reserves, commercial bank lending directly creates new private bank deposits when banks extend loans.
What it is: Cost of borrowing cash overnight against US Treasuries (SOFR) minus the rate the Fed pays banks on reserves (IORB). A sustained positive spread warns that bank reserves are becoming scarce.
What it is: Combined balance sheets of the Fed, ECB, Bank of Japan, and People's Bank of China as of 2026-09-30. Fixed-FX removes exchange-rate swings to isolate actual balance-sheet changes.
What it is: BIS total private non-financial credit to GDP and debt-service burden through 2026-03-31, paired with daily high-yield corporate bond spreads (2026-10-12) and curve slope.
Pipe 1 · Weekly H.4.1 Balance Sheet (to 2026-10-14)
When the Federal Reserve holds assets, those dollars sit on the liability side as currency, Treasury cash (TGA), reverse repos, or commercial-bank reserves. Over the past 260 weeks (2021–2026), weekly changes in the Treasury's cash account and bank reserves have a correlation of -0.68: when the Treasury builds cash, reserves typically fall unless offset by money-market funds leaving the reverse-repo facility.
Pipe 2 · Daily Treasury Statement (5,282 business days to 2026-10-13)
Unlike the Fed's weekly Wednesday snapshot, the US Treasury publishes exact cash deposits and withdrawals every business day. Separating Net Non-Debt Fiscal Spending (operating payments to the economy minus taxes) from Net Cash Borrowing (cash raised from issuing bills, notes, and bonds minus redemptions) shows why the TGA is rising or falling in real time.
| Cash Flow Component | 20-Day (4w) | 60-Day (13w) |
|---|---|---|
| Net Non-Debt Spending Operating outlays minus non-debt receipts |
$+51.2bn | $+490.6bn |
| Net Cash Borrowing Debt cash issues minus cash redemptions |
$+100.2bn | $+595.5bn |
| Tax Cash Credited to TGA Withholding, corporate & other taxes |
$484.2bn | $1,219.1bn |
| Net Cash Released from TGA Negative = TGA buildup (cash drain) |
$-49.0bn | $-104.8bn |
How to read this: Over the last 20 business days (through 2026-10-13), the Treasury injected $+51.2bn via net non-debt spending, but raised $+100.2bn in net debt cash—resulting in a net TGA build of $+49.0bn.
Pipe 3 · Commercial Bank Credit Creation (H.8 to 2026-10-07)
Central banks create reserves, but commercial banks create the deposit money used by businesses and households when they make loans. Tracking loan categories reveals a dramatic divergence: traditional real-estate and consumer loan growth is modest (+2.6% to +5.0% YoY), while bank lending to non-bank financial intermediaries (private credit funds, BDCs, mortgage REITs, and securitization vehicles) has surged +21.6% YoY.
| H.8 Credit Category | Level | 1Y YoY | 13W Ann. | Pace | Rank |
|---|---|---|---|---|---|
| Total bank credit TOTBKCR · 2026-10-07 | $19,949bn | +6.2% | +5.6% | decelerating | 52th |
| Total loans and leases TOTLL · 2026-10-07 | $14,183bn | +7.9% | +9.2% | accelerating | 71th |
| Business (C&I) loans BUSLOANS · 2026-09-30 | $2,959bn | +10.0% | +8.7% | decelerating | 66th |
| Real estate loans REALLN · 2026-09-30 | $5,842bn | +2.7% | +2.9% | steady | 32th |
| Consumer loans CONSUMER · 2026-09-30 | $1,927bn | +5.0% | +5.2% | steady | 54th |
| Loans to non-bank financial firms LNFDCBM027SBOG · 2026-09-30 | $1,535bn | +21.6% | +15.7% | decelerating | 78th |
Note on limitations: H.8 measures outstanding loan stocks (affected by repayments, charge-offs, and loan sales) classified by borrower, not the economic purpose of the spending. Rank is the percentile of 1Y growth since 2000 (since 2015 for non-bank financials).
Pipe 4 · International Central-Bank Balance Sheets (Converted via ECB FX)
When non-US central banks (ECB, Bank of Japan, People's Bank of China) are summed in US dollars, exchange-rate moves can create the illusion of money printing or tightening. Comparing the Actual USD Total ($24,981bn) against the Fixed-Exchange-Rate Total ($24,888bn) separates genuine balance-sheet expansion from currency translation.
| Central Bank | Own Currency | In USD | 1Y Local | 1Y USD | Status |
|---|---|---|---|---|---|
| Federal Reserve 2026-10-14 | 6,756 USD bn | $6,756bn | +2.4% | +2.4% | current |
| European Central Bank 2026-10-09 | 5,944 EUR bn | $6,661bn | -4.1% | -7.1% | current |
| Bank of Japan 2026-09-30 | 625,025 JPY bn | $3,981bn | -10.2% | -15.3% | current |
| People's Bank of China 2026-09-30 | 50,500 CNY bn | $7,532bn | +7.1% | +13.8% | current |
| Swiss National Bank 2026-08-31 | 951 CHF bn | $1,176bn | +11.5% | +10.8% | current |
| Bank of England 2025-06-30 | 855 GBP bn | $1,172bn | -8.3% | -0.7% | stale (>100d) |
| Bank of Canada 2026-10-14 | 240 CAD bn | $169bn | +4.7% | +3.4% | current |
| Reserve Bank of Australia 2026-10-14 | 347 AUD bn | $243bn | -13.8% | -7.2% | current |
Pipe 5 · Falsifiable Empirical Scorecard (No Cherry-Picked Charts)
Financial media frequently overlays Net Liquidity or TGA onto the S&P 500 over a 6-month window and claims a mechanical trading edge. We tested these rules across their full available histories using non-overlapping return windows, publication-lag allowances, 10,000 random shuffles, and circular time-shifts.
| Liquidity Signal & Index | Periods (Up/Down) | Mean 13W Return | Up − Down | Shuffle p | Shift p |
|---|---|---|---|---|---|
| US net liquidity NASDAQ Composite (2003-03-19 to 2026-08-21) | 55 / 39 | +3.7% vs +3.3% | +0.4 pp | 0.825 | 0.819 |
| US net liquidity S&P 500 (2003-03-19 to 2026-08-21) | 55 / 39 | +2.7% vs +2.3% | +0.4 pp | 0.801 | 0.894 |
| Bank credit pace NASDAQ Composite (1974-01-02 to 2026-08-10) | 95 / 116 | +2.7% vs +3.9% | -1.2 pp | 0.447 | 0.502 |
| Bank credit pace S&P 500 (1974-01-02 to 2026-08-10) | 95 / 116 | +1.9% vs +2.8% | -0.9 pp | 0.408 | 0.370 |
| Big-four assets in USD NASDAQ Composite (2003-04-01 to 2026-10-01) | 67 / 27 | +3.1% vs +4.8% | -1.7 pp | 0.439 | 0.553 |
| Big-four assets in USD S&P 500 (2003-04-01 to 2026-10-01) | 67 / 27 | +2.1% vs +3.8% | -1.7 pp | 0.301 | 0.426 |
How to interpret p-values: None of the 6 simple 13-week directional rules has a p-value below 0.05 (they range from 0.270 to 0.894). A rising 13-week Net Liquidity reading alone is not a validated standalone stock-timing rule.
79 non-overlapping quarterly windows (2006–2026), commercial-bank credit acceleration correlates negatively with cap-weight minus equal-weight performance (r = -0.35, p = 0.008): expanding bank credit broadens participation to the average stock, while tighter credit concentrates gains in mega-caps.
r = -0.68 over 260 weeks) and repo funding pressure once ON RRP is depleted.
2026-03-31, -1.90 pp YoY) with a 13.8% debt-service ratio—acting as a slow 1-to-2-year macro constraint rather than a short-term trading trigger.