Seven estimation exhibits and ten funding-market charts, drawn live from the committed data snapshot. Drag across a chart to zoom, double-click to reset, hover for a readout, and save any of them as SVG, PNG or CSV. The y axis clips to the 1st–99th percentile by default, which is what makes a series like SOFR−IORB legible next to its September-2019 tail.
One chart per estimation, in the order of the results note. Coefficients are in basis points; standard errors are Newey–West with 10 daily lags. The COVID dislocation (March–May 2020) is excluded from the rate-spread fits; September 2019 is retained, because it is the observation that identifies the steep part of the reserve-demand curve — and it is why the y axis clips outliers by default.
SOFR−IORB against reserves as a share of GDP, with the published four-parameter logistic fit. Midpoint 9.42% of GDP, abundance elbow 10.97%; reserves sit near 9.5% — on the steep part. Points are coloured by year, cool to warm — the legend below the chart is clickable, so a single year can be isolated. Drag to zoom; the eight September-2019 points reaching 315bp sit outside the default view, which is the point of clipping it.
SRF take-up against how far SOFR prints through the SRF rate. On days the facility is tapped, SOFR is about 16bp above it (t=12.7). Full allotment cut the maximum leak from 22bp to 12bp but raised the frequency. Outliers are shown in full here — the spikes are the subject, not noise.
SOFR's 99th percentile minus its median, bucketed by net new supply and split at the fitted elbow. Chunky settlement days lift the tail — but only when reserves are scarce, which is the interaction the regression finds at t=2.9.
The quarter-end profile of SOFR−IORB by business days to the reporting date, and the year-end mean by year. The turn flipped from −10bp in the abundant years to +10bp in 2025 — the balance-sheet tax is priced again.
Weekly SOFR−IORB against leveraged-fund short Treasury futures. The positioning coefficient is insignificant (t=−1.2) and reserves dominate: position size is a vulnerability gauge, not a daily price driver. Points are coloured by year, cool to warm; click a year in the legend to isolate it.
Three-month bills against IORB, with the ON RRP balance and stablecoin AUM alongside. RRP cash on the sidelines cheapens bills; stablecoin growth richens them.
Weekly changes: the Treasury's cash balance against reserves. The accounting identity puts the slope at −1 before other H.4.1 items move; the fitted partial is −0.40, attenuated because they move together.
The chart set the 2019 notes ran on, rebuilt from the panel. Drag across any chart to zoom, double-click to reset, hover for a readout, click a legend entry to hide a series, and save any of them as SVG, PNG or CSV. The Bloomberg-dependent exhibits — cross-currency basis, FX-hedged yields, OIS curves — are not here because they cannot be rebuilt from public data; the exact tickers are on the sources page.
EFFR and SOFR inside the target range against IORB — Pozsar's anchors of US rates. The corridor is the band; the question is always whether the secured rate stays inside it.
SOFR, TGCR and BGCR against IORB. In calm markets the segments sit a few basis points apart; the spread between them is the dealer balance-sheet rent.
Reserves against the Treasury's cash balance, the ON RRP and the foreign repo pool — the four claims on the Fed's balance sheet that decide how much cash the system has.
Take-up from the $2.55trn peak in December 2022 to roughly zero by mid-2025. This is the buffer that made the second round of QT safe, and it has been spent.
Total assets and bills held outright, through QE, taper and QT. Reserve management purchases restart the bill line from December 2025.
Tri-party, DVP and GCF volumes. The growth is in cleared and sponsored repo, which is where the mandatory-clearing rule will land.
Four-week and three-month bills against IORB. When bills cheapen relative to the administered rate, cash is abundant relative to paper — and the reverse.
Money-market fund AUM against the repo footprint — the shadow system's depositors, and the marginal cash lender in every funding market on this page.
SOFR minus IORB across the whole sample. September 2019 is the outlier that defines the series — clipped by default so the rest is legible, one click from being shown.
EFFR against IORB — bank-side scarcity, which lags repo. The 2019 lesson was that the unsecured rate is a slow gauge of a fast problem.
