Measures of the market itself rather than of its prices: what is outstanding, what is held in what form, and how that has changed. Each is read from a public Treasury file and published with its statement date.
Debt outstanding
Every Treasury security in existence, by class, from Treasury's own Monthly Statement of the Public Debt. Marketable debt is what trades and what the curves and indices on this site are fitted from; non-marketable debt is the larger half by count of classes and never trades at all.
Statement of August 31, 2026
Total public debt
$40.18tn
Held by the public
$32.41tn
Intragovernmental
$7.76tn
Marketable share
79.2%
Marketable
Marketable by class, with each class's share of the group and the split between debt held by the public and intragovernmental holdings.
Class
Outstanding
Share
Held by public
Bills
$7.25tn
22.8%
$7.25tn
Bonds
$5.53tn
17.4%
$5.51tn
Federal Financing BankHeld only intragovernmentally
$3.6bn
0.0%
$0.0m
Floating Rate Notes
$680.0bn
2.1%
$679.9bn
Notes
$16.22tn
51.0%
$16.21tn
Treasury Inflation-Protected Securities
$2.15tn
6.8%
$2.15tn
Total MarketableTreasury's own total, not a sum of the rows above
$31.83tn
100.0%
$31.81tn
Non-marketable
Non-marketable by class, with each class's share of the group and the split between debt held by the public and intragovernmental holdings.
Class
Outstanding
Share
Held by public
Domestic SeriesHeld by the Resolution Funding Corporation
$10.7bn
0.1%
$10.7bn
Government Account SeriesFederal trust funds, mostly Social Security
$8.11tn
97.1%
$366.4bn
Other
$4.9bn
0.1%
$4.9bn
State and Local Government SeriesMunicipal issuers parking bond proceeds
$78.9bn
0.9%
$78.9bn
United States Savings SecuritiesSeries EE and I savings bonds
$146.9bn
1.8%
$146.9bn
Total NonmarketableTreasury's own total, not a sum of the rows above
$8.35tn
100.0%
$607.8bn
The classes add up to Treasury's totals.
Marketable
$0.0m apart (1.2e-15 of the total)
Non-marketable
$0.0m apart (1.2e-16 of the total)
Both figures come from the same statement, so this is not two independent sources agreeing. What it tests is that Treasury computed the total and we computed the sum, so a class dropped or double-counted on our side shows up, while an error in Treasury's own total would not. Across all 308 statements the two differ on 25 marketable months, never by more than 3.4e-7 of the figure, and the gaps repeat as round numbers across consecutive months, which is a reporting artefact in the source rather than a missing row here.
STRIPS float
How much of the Treasury market is held as separate principal and interest components rather than as whole bonds. This is the float figure the Safe Rate indices deliberately do not deduct, published so that decision can be checked rather than taken on trust.
Statement of August 31, 2026
Held stripped
$626.5bn
Share of strippable debt
2.62%
Reassembled this month
$32.2bn
Securities
406
Why the indices do not deduct it
A stripped bond has not left the market. It has been decomposed into components any dealer can reconstitute, and $32.2bn was reassembled in this month alone. Over the last 12 statements, $364.7bn was reassembled against a stripped stock of $626.5bn, a turnover of 58% a year. The stripped pool is a turnstile, not a warehouse.
That is what separates it from the deductions the indices do make. A buyback is irreversible, because the debt is destroyed. Federal Reserve holdings are reversible only by a policy decision, on the Federal Reserve's timetable. Stripping is reversible by any dealer, on the day, and more than half of it reverses every year.
Deducting it anyway costs 3.8 basis points a year. And because the components are excluded too, it would leave an index holding neither side of a fifth of the long sector.
There is a coherent index that would deduct it, and this is not it. An index measuring what can be bought as a coupon bond today should subtract the stripped portion, because that par is not currently purchasable in that form. That is a purchasable-float index, and a different product. This family is a market representation: the debt is counted once, in whatever form it is being held.
Twenty-five years of it
Stripping is a demand signal rather than a supply one: it is what happens when somebody wants duration without a coupon, which in practice means pension and insurance liability matching. The share fell for fifteen years as the debt grew faster than the demand for zeros, bottomed under two per cent, and has climbed since.
308 monthly statements. The y axis is a percentage of strippable debt, not a rate.
It is almost entirely a long-bond phenomenon
Which follows from what stripping is for. A buyer wanting a thirty-year zero strips a thirty-year bond; nobody strips a two-year note to get a two-year zero they could buy as a bill.
Stripped par and its share of outstanding, by security class, on the August 31, 2026 statement.
Security class
Held stripped
Share of class
Treasury Bonds
$605.3bn
10.95%
Treasury Notes
$21.2bn
0.13%
Treasury Inflation-Protected Securities
$12.5m
<0.01%
The issues it concentrates in
The ten most heavily stripped securities, as a share of their own size. An index deducting the float would be removing this much of these particular bonds.
The ten securities with the largest stripped share of their own size on the August 31, 2026 statement.
Treasury publishes this as a month-end statement, so every STRIPS figure here is as of August 31, 2026, a month-end position beside the daily closes elsewhere on this site, not a stale one. The 3.8 basis point cost and the four-in-five-years comparison are results of rebuilding the index history with the deduction applied, and are stated rather than derived on this page.
Retail demand
What households bought directly from Treasury, and what they still hold. Every other measure on this site is the wholesale market; this is the only public series that shows a household reacting to a rate.
Net savings bond sales through TreasuryDirect, monthly
Gross sales less returns, so a household that buys and then cancels inside the month is not counted. 299 months from October 31, 2001.
Series ISeries EE
Millions of dollars. The spike is October 2022, the last month Series I paid 9.62%: households put $6,781.9m into it against $42.5m in August 2026, a factor of a hundred and sixty. Series EE, whose rate is fixed for the life of the bond rather than reset against inflation, averages $5.6m a month across 280 months and never exceeds $90.4m, which is what makes the Series I line a statement about the rate rather than about TreasuryDirect. A savings bond has no secondary market, so every dollar here is a household's own decision rather than a dealer position.
Savings bonds outstanding
Report of July 31, 2026
The latest report only, by series. Each series' outstanding amount is what has been issued less what has been redeemed; matured and unredeemed is the part that has stopped earning interest and has not been cashed.
Savings bonds outstanding by series at the July 31, 2026 report, with issued, redeemed and matured-unredeemed amounts.
Series
Outstanding
Issued
Redeemed
Matured, unredeemed
EESeries EE
226,434,073
1,976,146,478
1,749,712,405
82,236,813
ISeries I
24,205,123
179,339,685
155,134,562
0
ESeries E
19,832,306
4,599,418,701
4,579,586,395
19,832,306
HHSeries HH
465,729
14,941,799
14,476,070
465,729
SNSavings Notes
134,583
30,850,113
30,715,530
134,583
HSeries H
26,818
16,386,525
16,359,707
26,818
FSeries F
9,092
5,624,339
5,615,247
9,092
A-DSeries A-D
7,855
18,559,885
18,552,030
7,855
RSeries R
7,063
562,439
555,376
0
GSeries G
5,519
17,260,002
17,254,483
5,519
PSeries P
2,139
194,335
192,196
0
JSeries J
1,767
793,354
791,587
1,767
KSeries K
96
936,780
936,684
96
All seriesTreasury's own total, not a sum of the rows above
271,132,163
6,861,014,435
6,589,882,272
102,720,578
Counts of bonds, not dollars. A savings bond's face value depends on its denomination, so the number outstanding and the amount outstanding are different figures. The dollar amount for savings securities as a whole is in the non-marketable table above, under United States Savings Securities. Current Series I and EE rates are here.
Index levels before September 30, 2026 are back-tested. They were computed after the fact by applying the rules to historical data, which benefits from hindsight in the choice of rules, and an index cannot be invested in directly. Methodology v1.0 takes effect at that rebalance, when levels begin to be struck on the day; the rulebook is identical either way, and the version is published on every row.
A fitted curve is a fit, not a quote. Daily error averages 3.8 basis points across the history and reaches about 20 on the worst days, in December 2008, when the market was genuinely hard to fit one smooth curve to. Every curve page publishes its own fit error rather than burying it.
Free to benchmark against. Paid only to track. Measuring anything against these curves and indices is free. No license, no fee, no registration, and no permission needed to say that you did, including in a prospectus. A fee applies to one thing: launching a product that tracks an index.
Safe Rateâ„¢, Safe Rate Indicesâ„¢ and the Safe Rate US Treasury Indexâ„¢ are trademarks of Safe Rate, claimed through use in commerce and not registered. Third-party marks are the property of their owners, which are not affiliated with and do not endorse this data.
Curves are fitted from public Treasury data and carry fit error; figures are not a record of trading, and an index cannot be invested in directly. No claim of compliance with the IOSCO Principles for Financial Benchmarks is made or implied. Not investment advice, not an offer, and not a recommendation to buy or sell any security.