Home › Guides › How Treasury Auctions Work
How the U.S. government actually borrows money - it auctions new bonds to buyers and lets them set the rate.
Every week and every month, the U.S. government needs to borrow money. The way it does this is genuinely interesting: it holds auctions. The Treasury Department announces well in advance that, say, $50 billion of new 10-year notes will be auctioned next Wednesday. On Wednesday at 1:00 PM Eastern, pre-qualified buyers around the world submit secret bids saying what yield they want to be paid. Treasury then sells the bonds at whatever yield clears the demand. That yield becomes the rate the U.S. government pays for that batch of debt.
How the U.S. government actually borrows money - it auctions new bonds to buyers and lets them set the rate.
Every week and every month, the U.S. government needs to borrow money. The way it does this is genuinely interesting: it holds auctions. The Treasury Department announces well in advance that, say, $50 billion of new 10-year notes will be auctioned next Wednesday. On Wednesday at 1:00 PM Eastern, pre-qualified buyers around the world submit secret bids saying what yield they want to be paid. Treasury then sells the bonds at whatever yield clears the demand. That yield becomes the rate the U.S. government pays for that batch of debt.
This happens constantly. Treasury bills (very short-term, from four weeks to a year) auction every week. Treasury notes (2 to 10 years) and Treasury bonds (20 and 30 years) auction every month. TIPS - inflation-protected versions - auction a few times a year. Together, these auctions raise the trillions of dollars the U.S. government needs to fund itself.
Why should you care? Because the auction results are a real-time signal of how willing investors are to lend to the United States. If demand is strong, the bonds sell at low yields (cheap borrowing for the government). If demand is weak, the bonds sell at higher yields. The most-watched metric is the "bid-to-cover ratio" - total bids submitted divided by bonds sold. A ratio above 2.5 means a lot more was bid than was available, signaling healthy demand. Below 2.0 signals weakness. A weak auction can move bond yields meaningfully on the day.
The auctions also reveal who's buying. Primary dealers (about two dozen large banks designated by the New York Fed) are required to bid at every auction. "Direct bidders" are large institutions buying for their own account. "Indirect bidders" are mostly foreign central banks and other foreign accounts bidding through intermediaries. The mix of buyers shifts over time and tells you something about the global demand for U.S. debt.
For an individual, you can actually participate in Treasury auctions directly through TreasuryDirect.gov - buying bills and bonds at the same auction price that the big institutional bidders get. It's one of the few financial markets where retail participants get the same pricing as institutions, with no commission.
Single-price (Dutch) auctions; primary dealers, direct bidders, indirect bidders; bid-to-cover and tail matter.
Since 1998, all Treasury auctions use a single-price (also called "Dutch" or "uniform-price") format. Bidders submit either competitive bids (specifying the yield they want) or non-competitive bids (accepting whatever the clearing yield turns out to be). Treasury accepts the lowest-yield bids first, working up until it has sold the announced amount. The highest accepted yield is the "stop-out" yield, and every winning bidder - competitive and non-competitive alike - pays that single price. The format reduces the winner's curse and tends to produce more aggressive bidding than the pre-1992 multiple-price format.
Three bidder categories report results separately:
Primary dealers - currently 25 banks and brokerage firms officially designated by the Federal Reserve Bank of New York as Treasury market-makers (the list changes occasionally). They are required to bid at every Treasury auction and to make markets in Treasury securities. Awards in the auction itself go strictly to the most competitive bids regardless of bidder category, but because dealers are obligated to bid meaningfully across a range of yields, in practice they absorb whatever the other bidder classes don't take. Primary dealer share above ~30% is sometimes read as a signal of weak demand from other bidder types.
Direct bidders - institutions submitting competitive bids for their own account directly to Treasury via TAAPS (the Treasury Automated Auction Processing System) rather than through a primary dealer. Includes major U.S. asset managers, sovereign wealth funds, large pension funds, and some bank treasuries. Direct bidder share has grown gradually over the past decade as more institutions have invested in direct-access infrastructure.
Indirect bidders - bidders accessing the auction through a primary dealer or direct bidder. The category includes foreign central banks (a major source of demand historically), foreign private accounts, and some U.S. institutional bidders that prefer not to set up direct accounts. Indirect bidder share is the commonly-cited proxy for foreign demand, though the proxy is imperfect.
Auction schedule:
Bills (very short-term, no coupon): 4-week, 6-week, 8-week, 13-week, 17-week, 26-week, 52-week. The 4/6/8/13/17/26-week bills auction weekly; the 52-week every four weeks. CMBs (Cash Management Bills) auction irregularly for short-term cash-management needs.
Notes (2 to 10 years): 2-year, 3-year, 5-year, 7-year auction monthly. The 10-year is auctioned monthly but with original issues in February, May, August, November and re-openings in other months. New issues set a new coupon; re-openings sell more of an existing security at the prevailing market yield.
Bonds (long-term): 20-year and 30-year auction monthly with similar new-issue/re-open pattern. The 30-year bond is the longest-dated regular Treasury issue.
TIPS (inflation-protected): 5-year (new in April and October, re-opened in June and December), 10-year (new in January and July, re-opened four times a year), 30-year (new in February, re-opened in August). TIPS issuance is smaller than nominal Treasury issuance.
FRNs (Floating-Rate Notes, 2-year): monthly. FRNs pay a floating rate based on the 13-week T-bill auction rate plus a fixed spread.
The Quarterly Refunding Announcement (QRA) - released at 8:30 AM ET on the Wednesday of the mid-quarter refunding week, typically at the start of February, May, August, and November (sometimes the last Wednesday of the preceding month) - announces the next quarter's note and bond auction sizes. QRA shifts can meaningfully move long-end Treasury yields (the August 2023 QRA upsizing of long-end issuance was a major contributor - alongside the Fitch downgrade, strong economic data, and the Fed's higher-for-longer messaging - to the August–October 2023 back-up that took the 10Y to 5% in late October).
A weak auction (low bid-to-cover, positive tail vs WI, low indirect-bidder share) can push yields up several basis points that day. A strong auction can pull them down. The market watches every major auction result.
Stop-out − WI = 'tail'; healthy bid-to-cover 2.0-3.5x; indirect % ~ foreign demand proxy.
Pre-auction reference: the When-Issued (WI) market trades the not-yet-auctioned security in the days preceding the auction. WI provides forward price discovery. The WI yield at the moment of auction close (1:00 PM ET) is the benchmark for measuring auction performance. Auction tail = stop-out yield − WI yield at auction close. A positive tail (auction yielded higher than WI predicted) signals weak demand; a negative tail (auction yielded lower than WI) signals strong demand.
Published metrics per auction:
- High-rate (stop-out yield): the highest yield accepted; this is the yield every winning bidder pays
- Median-rate: the yield at which half the awarded amount has been allocated
- Low-rate: the lowest yield bid (the best bid)
- Bid-to-cover ratio: total tendered amount divided by accepted amount (the higher the better)
- Percent allotted at the high yield: how much of the issuance was sold at the marginal stop-out rate (high percent allotted at high yield can signal a marginal demand environment)
- Allotment percentages by bidder category: primary dealer share, direct bidder share, indirect bidder share
- Allotment in dollar amounts by tenor and issuance type
Bid-to-cover benchmarks: healthy historical range is 2.0–3.5x. Above 3.5x suggests very strong demand. Below 2.0x suggests genuinely weak demand. The auction's bid-to-cover should be interpreted in context - different tenors have different historical bid-to-cover averages; bills typically run higher than long bonds.
Indirect-bidder share interpretation: historically a proxy for foreign demand. Indirect share above 60% at a 10-year or 30-year auction signals strong foreign demand; below 40% signals weak. The proxy is imperfect - some indirect bidders are domestic, and some foreign demand comes through direct or even primary-dealer channels. Indirect share has trended up secularly - recent 10-year auctions often run ~65–70% indirect, with primary-dealer takedown trending down - though levels vary by tenor.
Auction tail interpretation: a positive 1–2 bp tail is unremarkable. Positive 3+ bp tail signals meaningful demand weakness and typically catalyzes secondary-market yield moves. Negative tails are uncommon; large negative tails signal strong demand. The tail metric is more informative than the absolute yield level, which is set by market conditions independent of the auction.
Auction schedule with frequency detail:
Bills - 4W, 6W, 8W, 13W, 17W, 26W auction weekly: 13W/26W announce Thursday, auction Monday, settle Thursday; 4W/8W announce Tuesday, auction Thursday, settle the following Tuesday; the 17W auctions Wednesday and the 6W Tuesday. The 52W auctions every four weeks. CMBs (Cash Management Bills) supplement when Treasury needs additional short-term cash.
Notes - 2Y, 3Y, 5Y, 7Y auction monthly. 10Y new issues in February, May, August, November; re-opens in other months. New issues set a fresh coupon at the auction; re-opens sell more of an existing security at the prevailing market yield.
Bonds - 20Y monthly new/re-open mix. 30Y new issues in February, May, August, November; re-opens in other months.
TIPS - 5Y new in April and October, with re-opens in June and December. 10Y new in January and July, with re-opens in March, May, September, and November. 30Y in February (new), August (re-open). TIPS auction sizes are typically smaller than nominal Treasuries.
FRNs - 2Y monthly. New issue at quarter starts; re-opens in other months.
Dealer-research auction calendars: Goldman, JPMorgan, BofA, Citi rate-research teams publish auction previews and post-auction reviews with analytics on bid-to-cover trends, tail trends, indirect-share trends. SIFMA publishes monthly Treasury market statistics with bidder-category breakouts and total issuance summaries.
Foreign demand secular trends: foreign official holdings of Treasuries have declined as a share of outstanding debt, even though indirect-bidder takedown at coupon auctions has trended up over the same period. Drivers: foreign central bank reserve diversification (China, Saudi Arabia, Russia, others over various periods), shifts in foreign-private demand for hedged-USD assets, growth of U.S. domestic Treasury demand (money funds, retail, banks). The secular decline in the foreign share shifted some Treasury demand burden onto domestic institutions.
Monetary policy interaction: when the Fed is doing QE (buying Treasuries via SOMA), it absorbs duration from the private market, supporting auction demand structurally. When the Fed is doing QT (allowing SOMA holdings to roll off without reinvestment), private-market duration absorption requirements rise, putting more pressure on auction demand. The 2022–24 QT period coincided with several auction-demand weakness episodes.
TGA dynamics: the Treasury General Account at the Fed is where Treasury holds its cash balance. Auction settlements move cash from primary dealers and direct/indirect bidders into the TGA. Large auction settlements drain bank reserves; large benefit payments and contractor payments rebuild them. The TGA-vs-reserves dynamic is one of the more important short-term drivers of money-market and SOFR conditions.
A practical example: in November 2023, the Treasury 30-year bond auction tailed by ~5 bps. Indirect bidders took ~60% versus a recent average around 65–67%, leaving primary dealers with roughly 25% of the issue - about twice the norm (a ransomware attack on ICBC's U.S. broker-dealer that day also complicated dealer bidding). The result helped catalyze a several-bp back-up in long-end yields over the following day and contributed to broader market questions about long-end demand absorption capacity.
Pre-1992 multi-price English format on long end; post-2008 supply dynamics, foreign-share secular decline.
Pre-1992 long-end auctions used multiple-price (English) format - each winning bidder paid the yield they bid (rather than a single uniform clearing price). The format created winner's-curse incentives: bidders systematically shaded their bids below their true valuation to avoid overpaying, which Treasury staff worried was lowering auction proceeds. The shift to single-price format was supported by auction-theory work including Bikhchandani-Huang (1989, Review of Financial Studies) and subsequent research showing that uniform-price auctions reduce winner's curse and lift expected revenue in common-value-asset settings.
Treasury's transition was gradual: experimented with single-price on 2-year and 5-year notes in 1992; expanded to all coupon issuance by 1998. The 1998 transition completed the shift; bills, which had been auctioned multiple-price since their inception in 1929, were among the last to convert. The empirical analysis after the change (Malvey-Archibald 1998, Garbade-Ingber 2005) suggests modest revenue improvements consistent with the theoretical prediction.
Foreign-share secular decline detail: the decline shows up in ownership data - foreign holdings have fallen substantially as a share of outstanding marketable Treasury debt since the mid-2010s - not in auction takedown, where indirect-bidder share has trended up (recent 10Y auctions often print ~65–70% indirect). The ownership decline reflects: (1) foreign-central-bank reserve diversification - China's foreign reserves peaked in 2014 at ~$4T and have since declined; (2) foreign-private demand shifts - hedged USD-funded carry trades face cross-currency basis costs that can compress relative returns vs unhedged JGBs or Bunds; (3) growth of U.S. domestic Treasury demand - MMF AUM grew $2T during 2022–24, absorbing supply that foreign accounts previously bought.
The secular decline shifts the marginal Treasury demand burden onto domestic institutions and households. Money-market funds, bank treasuries, pension funds (LDI strategies), insurance company treasuries, and retail (via brokerage holdings and Treasury ETFs) absorb the marginal supply. The depth of domestic demand became a focus during 2023–24 as auction demand metrics softened.
Quarterly Refunding Announcement (QRA) mechanics: published at 8:30 AM ET on the Wednesday of the mid-quarter refunding week, typically early February, May, August, and November (sometimes the last Wednesday of the prior month). The refunding process has three pieces: (1) marketable borrowing estimates - Treasury's financing needs for the coming quarters, released separately on the Monday two days before the refunding statement; (2) coupon issuance sizes by tenor, set in the Wednesday statement itself - how much will be auctioned at each tenor across the upcoming quarter; (3) TBAC (Treasury Borrowing Advisory Committee) recommendations and meeting minutes - institutional perspectives on Treasury market functioning.
The August 2023 QRA was the consequential recent event. Treasury upsized long-end (10s, 30s) issuance by more than the market expected, partly in response to expanded financing needs from Fed QT (which removed Fed as a marginal duration buyer) and Treasury's pivot from bills toward coupon issuance. 10-year yields backed up from ~3.95% pre-QRA to ~5.0% by late October, with the supply surprise compounding the Fitch downgrade and the Fed's higher-for-longer repricing. The 30-year backed up correspondingly. The episode established that QRA decisions can be primary drivers of long-end yield moves.
Primary-dealer awards above ~30% are sometimes interpreted as weak indirect demand (the dealers are stuck with paper that other bidders didn't want). This interpretation requires care - dealers are required to bid and may be intentionally taking down inventory for client distribution. The signal is most informative when combined with low indirect share, positive tail, and low bid-to-cover.
Auction analytics by tenor have meaningfully different baselines: 2Y auctions typically run bid-to-cover ~2.5–3.0x; 5Y ~2.4–2.8x; 7Y ~2.5–2.8x; 10Y ~2.4–2.7x; 30Y ~2.3–2.5x; bills ~3.0–4.0x; TIPS ~2.4–2.6x. Period averages drift; recent tenor-specific benchmarks should be referenced.
SIFMA's monthly Treasury statistics (https://www.sifma.org/resources/research/us-treasury-securities-statistics/) provide the best free industry dataset for analysis of auction trends, supply patterns, and bidder-category dynamics. The data goes back to 2000.
Treasury's auction-result data feed at TreasuryDirect (https://www.treasurydirect.gov/auctions/auction-query/) provides full historical data on every Treasury auction since the late 1990s. Variables include announced size, accepted size, bid-to-cover, high/median/low rate, allotment by bidder category, WI yield where available.
For institutional research, the NY Fed's primary-dealer page (https://www.newyorkfed.org/markets/primarydealers) lists current primary dealers. The NY Fed's Treasury Market Practices Group (TMPG) publishes white papers on Treasury market structure and best practices, including topics like fail charges, settlement conventions, and trading conventions.
Academic literature on Treasury auctions: Bikhchandani-Huang (1989) on auction format theory; Malvey-Archibald (1998) on the U.S. Treasury single-price transition; Garbade-Ingber (2005) on Treasury auction format experiments; Hortacsu-Kastl-Zhang (2018, American Economic Review) on bid shading and bidder surplus in the U.S. Treasury auction system (building on Hortacsu-Kastl 2012, Econometrica, which developed the methodology on Canadian auctions); Allen-Kastl-Wittwer on primary dealers and the demand for government debt. The Federal Reserve Bank of New York's Liberty Street Economics blog frequently posts on Treasury market dynamics.
Market-microstructure relevance: on-the-run (newly auctioned) vs off-the-run (previously auctioned) Treasury bonds trade with different liquidity characteristics. OTR bonds command a small specialness premium in repo markets. The OTR-vs-off-OTR rotation at quarterly refunding is closely watched. The specialness compresses OTR yields by 1–10 bps; when an OTR loses specialness on quarterly rotation, the bond re-rates wider.
Research caveats for empirical work: (1) auction-day yield moves can confound auction-specific effects with broader market drivers; isolating auction effects requires careful event-window construction; (2) tail metrics depend on WI yield at exact 1:00 PM ET, which is a snapshot; intraday WI volatility can produce noise; (3) bidder-category definitions and reporting have evolved over time (post-2000 categorization is the current standard); (4) supply effects from QRA changes propagate over weeks to months, complicating contemporaneous identification.