An argument, its evidence, and what could change it.
What we’re watching
31 Aug 2025 to 31 Aug 2026
At publication ·
Change over this period
The memo · WEDNESDAY, 23 SEPTEMBER 2026
From August last year to 2026-08, the Treasury's average rate paid rose from 3.372% to 3.49% while its average bill rate fell from 4.283% to 3.788%, raising the question of how far financing relief extends. These are average interest rates on existing debt. The bill result therefore cannot stand for the whole debt stock.
Franc treats the average bill rate's move from 4.283% in August last year to 3.788% in 2026-08 as the start of general financing relief. His strongest case is a staggered adjustment: cheaper replacement financing could reach other debt categories over time. That requires the effect to outweigh any offset from debt composition, a mechanism these averages do not establish.
If a security's price assumes that transition, its failure could produce a loss. A cash obligation could force a sale during a temporary price decline, making the capital loss permanent. Dismissing the bill evidence could also miss a beneficial adjustment if relief spreads before prices anticipate it. These averages do not establish whether prices offer that opportunity.
The next test is the Treasury's month-end average rates release, comparing the next matching annual changes for the overall average and bills. Opposite signs would preserve the split, requiring Franc to concede that no common direction is established. If both are negative, his case gains support, but he must concede that convergence does not establish his mechanism; if both are positive, he must abandon the relief claim. If either is zero, the split ends without a shared direction, requiring him to leave the broader turn unresolved.
The Treasury's bill average has not earned promotion into a verdict on the whole debt stock.
The judgment
This edition has no separate structured assessment on record. These questions offer a way to examine the memo.
Which facts support the argument?
What interpretation is being placed on those facts?
What is the strongest competing explanation?
Where could a loss become difficult to recover from?
What could excessive defensiveness miss?
Which observable development would weaken the argument?
Experimental · snapshot 21 Sept 2026
Exploratory context · predictive value under evaluation.
People · employment level · Aug 2026
Systematic COVOL · PC1 index · 31 Aug 2026
Sept 2025Aug 2026
COVOL measures volatility shared across markets. Higher values indicate more synchronized market-wide volatility; they do not indicate the direction of prices or jobs.
We are testing whether shared market volatility adds information about the size of future changes in temporary-help employment. The historical lead is exploratory.
This research snapshot stays dated. It is separate from the latest instrument data and does not update automatically.
Source: V-Lab, the Volatility Laboratory at NYU Stern Volatility Institute. For reference.
Exploratory context · predictive value under evaluation.
Yield · percent · 31 Aug 2026
Annualized return volatility · 31 Aug 2026
Sept 2025Aug 2026
This model describes variability in Treasury returns. Its percentage is separate from the yield level and is not a range around that yield.
We are testing whether Treasury return volatility helps estimate unusually large yield variability over the next 20 trading days. No prospective forecast probability is shown.
This research snapshot stays dated. It is separate from the latest instrument data and does not update automatically.
Source: V-Lab, the Volatility Laboratory at NYU Stern Volatility Institute. For reference.
+1.56% YoY
Exploratory context · predictive value under evaluation.
CPI index · 1982–84 = 100 · Aug 2026
Annualized volatility · 31 Aug 2026
Sept 2025Aug 2026
Modeled annualized variability of S&P 500 returns. The value does not predict their direction.
Five market measures are being evaluated together for changes in medical-price variability. The selected input is one component, not a standalone medical-price signal.
This research snapshot stays dated. It is separate from the latest instrument data and does not update automatically.
Source: V-Lab, the Volatility Laboratory at NYU Stern Volatility Institute. For reference.
Compared with the collection on 2026-09-27. 63 instruments are unchanged.
No instrument values or periods changed between these collections.
No new memo for 28 Sept. The observations did not support a new argument.
Collection overview
Latest edition · 23 Sept 2026
63 instruments · at collection: 54 current · 7 provisional · 2 unavailable
None of 63 moved since 27 Sept
Last collection · 28 Sep 2026, 12:00 UTC
These are dated observations. This page does not stream live market data.
Find a measure. Inspect its history. Check its source.
No instruments match. Try another term or reset the filters.
The cost of money. Rates describe financing conditions; they do not establish what an asset is worth.
The pace of price changes. An easing inflation rate still allows the price level to rise.
Employment, pay, and hours. Look for agreement across measures before generalizing from one release.
Production, spending, and trade. A reported change can be revised as more evidence arrives.
Physical supply and use. Quantities alone do not establish demand, margins, or investment value.
Public borrowing, cash, and spending. Stocks and flows answer different questions.
Lending conditions and realized bank losses. Positive survey standards mean net tightening; positive demand means net strengthening. Neither is a corporate-bond default rate.
A three-company reporting pilot. Cash is not all liquidity; current liabilities are not all debt. Fiscal-year-to-date cash flows can cover different lengths of time.
Pilot
Watch the latest episode →