Benchmark vs. Off-the-Run Treasuries

The U.S. Treasury market holds roughly $27.8 trillion in marketable debt spread across about 400 distinct CUSIPs (Q4 2025), yet at each key maturity point — 2-, 5-, 10-, and 30-year — a single security captures 60–70% of all trading volume for that tenor. That single security is the on-the-run benchmark: the most recently auctioned issue. Everything else is off-the-run.

Consider two 10-year Treasury notes sitting side by side in February 2026. One yields 4.460%, the other 4.490% — separated by just 3 basis points and a single auction cycle. Yet the benchmark trades roughly $18 billion per day while its predecessor manages about $3 billion. In normal markets, that 2–6 bps spread is a quiet, reliable feature of Treasury pricing. During the COVID-19 dislocation in March 2020, it spiked to 25 bps in under two weeks, revealing how violently liquidity pricing can shift when markets seize.

Understanding this distinction matters whether you're building a bond portfolio, evaluating fixed-income fund holdings, or simply trying to read the yield curve correctly. What follows covers what benchmark and off-the-run actually mean, how the auction cycle creates the divide, a worked example of the yield pickup, historical stress episodes (LTCM 1998, COVID 2020, October 2023), the real risks of chasing the spread, and a practical checklist for navigating both sides of the market.

TL;DR: On-the-run Treasuries are the most recently auctioned issues at each maturity — they dominate trading volume and serve as pricing benchmarks. Off-the-run issues are everything else, typically yielding 2–6 bps more to compensate for lower liquidity. That small spread can be free money for buy-and-hold investors or a dangerous trap during market stress.

What "On-the-Run" and "Off-the-Run" Actually Mean

On-the-run (benchmark) Treasury is the most recently auctioned security at a given maturity point — the latest 10-year note, for example — and it serves as the primary pricing reference for that tenor. When a financial news headline quotes "the 10-year yield," it means the on-the-run 10-year. Dealers use it to quote yield levels, build the yield curve, and price other fixed-income instruments from corporate bonds to mortgage-backed securities.

Off-the-run Treasury is any previously issued security that has been superseded by a newer auction in the same maturity bucket. The moment a new 10-year note is auctioned, every prior 10-year issue becomes off-the-run. These securities are perfectly valid U.S. government obligations — same credit quality, same tax treatment — but they typically trade at a slight yield premium to the on-the-run issue because fewer market participants actively trade them.

The distinction starts with a CUSIP number. Each Treasury auction creates a new CUSIP (Committee on Uniform Securities Identification Procedures number), and tracking CUSIPs is how the market distinguishes the benchmark from everything else. The Treasury issues 4 new 10-year CUSIPs per year, plus 8 reopenings (12 total auctions), so at any moment roughly 40+ distinct 10-year CUSIPs are outstanding.

Reopenings vs. new issues matter here. A reopening adds to an existing CUSIP — same coupon, same maturity date — increasing the outstanding amount. A new 10-year auction is typically $42 billion, while a reopening runs about $35 billion (U.S. Treasury quarterly refunding data). Reopenings don't change which issue is on-the-run; they simply make the existing benchmark larger.

The benchmark roll is the transition point when a new auction replaces the prior issue as on-the-run. Trading activity and pricing references shift to the new CUSIP — often overnight. Volume in the former benchmark typically drops 40–60% within two weeks of the roll, and bid-ask spreads widen. If you're holding the old benchmark, you wake up one morning owning an off-the-run security with meaningfully less liquidity.

When-issued (WI) trading — secondary-market activity between the auction announcement date and settlement — is where the new benchmark begins establishing its pricing even before it formally exists. And the 25 primary dealers (financial institutions authorized to participate directly in Treasury auctions and required to make markets) are the main conduit for on-the-run liquidity, concentrating their quoting and inventory in the current benchmark.

The liquidity premium — the additional yield investors require for holding a less liquid security — is what the on-the-run/off-the-run spread directly measures. It's not a credit difference (both are backed by the full faith of the U.S. government). It's purely about how easily and cheaply you can trade.

How the Benchmark Cycle Works in Practice

The 10-year auction cycle runs on a predictable calendar: 4 original issues plus 8 reopenings equals 12 auctions per year. Each original issue creates a new CUSIP that becomes on-the-run; the prior CUSIP shifts to "first off-the-run" status. The cycle repeats quarterly, so the benchmark 10-year note changes identity roughly every three months.

Current market snapshot (February 2026): The 10-year benchmark carries a coupon of approximately 4.375% and yields about 4.50%. The first off-the-run 10-year yields roughly 4.53%, implying a 3 bp on-the-run/off-the-run spread — squarely within the normal 2–6 bps range.

Auction mechanics are straightforward but worth understanding. A typical 10-year auction offers $42 billion in securities. The bid-to-cover ratio (total bids divided by amount offered) measures demand — a higher ratio signals stronger appetite. Results including high yield, allotment percentages, and the new CUSIP are published on TreasuryDirect within minutes of the auction close.

The liquidity concentration is remarkable. Average daily Treasury trading volume runs about $900 billion (2025, SIFMA data). On-the-run 10-year notes account for 60–70% of all 10-year trading volume despite being just one CUSIP among dozens. This isn't irrational — it's self-reinforcing. Dealers quote the benchmark because clients want to trade it. Clients want to trade it because dealers quote it tightly. The result is a bid-ask spread of roughly 1/64 on the on-the-run 10-year versus approximately 2/64 on off-the-run issues — double the transaction cost.

Other maturity points follow the same pattern at different scales: 2-year auctions run about $69 billion, 5-year about $70 billion, and 30-year about $25 billion. The February 2025 quarterly refunding totaled $125 billion across coupon-bearing maturities (U.S. Treasury Quarterly Refunding Statement).

There's also a repo market dimension that many investors overlook. On-the-run issues often trade "special" in the repo market — meaning they can be borrowed at rates below general-collateral (GC) levels because of high demand to short or deliver them. Off-the-run issues typically trade at GC rates. This financing cost difference can offset or even exceed the yield pickup from buying off-the-run, particularly for leveraged investors who fund positions in the repo market. The point is: the yield spread alone doesn't tell the full cost story.

Worked Example — Calculating the On-the-Run / Off-the-Run Yield Pickup

Your situation: You're evaluating two 10-year Treasury notes in February 2026, deciding which belongs in your portfolio.

The on-the-run issue: CUSIP 91282CKX (illustrative), coupon 4.375%, maturing February 15, 2036. Price: 99-08 (99.25), yield-to-maturity: 4.460%. Outstanding amount: $42 billion. Daily trading volume: approximately $18 billion.

The first off-the-run issue: CUSIP 91282CKA (illustrative), coupon 4.250%, maturing November 15, 2035. Price: 98-16 (98.50), yield-to-maturity: 4.490%. Outstanding amount: $112 billion. Daily trading volume: approximately $3 billion.

The spread calculation: The off-the-run yields 3 bps more than the benchmark despite having a slightly shorter remaining maturity (9.75 years vs. 10 years). That premium compensates entirely for lower liquidity — the off-the-run trades roughly one-sixth the daily volume of the benchmark.

What 3 bps is worth: On a $10 million position, 3 basis points equals approximately $3,000 per year in additional income. That may sound modest, but it compounds, it's essentially risk-free from a credit perspective, and across a large institutional portfolio it adds up meaningfully.

If you're a buy-and-hold investor (insurance company, pension fund, individual building a bond ladder), the off-the-run note is likely the better choice. You capture the 3 bp yield pickup, and since you don't plan to trade frequently, the wider bid-ask spread is irrelevant. The larger outstanding amount ($112 billion vs. $42 billion) doesn't matter if the goal is hold-to-maturity.

If you're an active trader or dealer, the on-the-run note wins. The tighter bid-ask spread (roughly 1/64 vs. 2/64) and deeper order book mean your transaction costs on frequent trading would quickly exceed the 3 bp yield advantage. You need to be able to get in and out cleanly, and the benchmark lets you do that.

One additional angle worth knowing: off-the-run issues are often used to create STRIPS (Separate Trading of Registered Interest and Principal of Securities) — zero-coupon instruments stripped from coupon and principal payments. This adds another source of demand for certain off-the-run CUSIPs, which can occasionally compress their spreads relative to other off-the-run issues.

The durable lesson: the "right" choice depends entirely on your time horizon and trading frequency. The spread exists because different investors have different needs — and both sides of the trade are rational.

When the Spread Blows Up (Historical Stress Episodes)

The on-the-run/off-the-run spread is stable enough in normal markets that it can lull you into complacency. Then a crisis arrives, and what looked like a riskless 3 bps of yield pickup becomes a liquidity trap. Three episodes illustrate the pattern.

LTCM, 1998 — The Original Convergence Trade Gone Wrong

Long-Term Capital Management built enormous convergence trades: long off-the-run, short on-the-run Treasuries, betting that the spread would narrow as the off-the-run issues aged. The trade was directionally correct — spreads do tend to converge over time. But in August–September 1998, after Russia's debt default triggered a global flight to liquidity, on-the-run/off-the-run spreads for the 30-year bond widened from roughly 6 bps to over 15 bps. The 29.5-year off-the-run yielded approximately 15 bps more than the 30-year benchmark (Krishnamurthy, 2002, Journal of Financial Economics).

LTCM couldn't survive the margin calls as spreads moved against them before eventually (and inevitably) converging. The trade was right. The leverage killed them. This remains the canonical warning for anyone running leveraged convergence strategies in the Treasury market.

COVID-19 Market Dislocation, March 2020 — Liquidity Evaporates in Days

Between March 9 and March 18, 2020, the 10-year on-the-run/off-the-run spread widened from approximately 3 bps to 25 bps — an eightfold increase in under two weeks (Federal Reserve Bank of New York; BIS Quarterly Review, June 2020). Bid-ask spreads on off-the-run issues widened 5–10x their normal levels. Some off-the-run issues saw no executable bids for hours.

The Federal Reserve intervened with $75 billion per day in Treasury purchases starting March 13 — an unprecedented pace. The intervention worked, but the episode exposed a structural vulnerability: the Treasury market had grown far faster than dealer balance-sheet capacity to intermediate it. Off-the-run holders who needed to sell during those days faced execution costs that dwarfed years of yield pickup.

Why this matters: 3 bps of annual yield pickup is worthless if you lose 50+ bps in execution costs during the one week you need to sell.

October 2023 Term Premium Repricing — A Milder but Instructive Episode

Between September and October 2023, the 10-year benchmark yield rose from 4.35% to 5.00% as term premium repriced sharply. Off-the-run spreads widened modestly to 5–8 bps as dealers reduced inventory. Average daily volume in off-the-run 10-year notes dropped roughly 20% versus the prior quarter (SIFMA Trading Volume Data; Federal Reserve Financial Stability Report, November 2023).

This wasn't a crisis — but it showed how quickly off-the-run liquidity can deteriorate even during an orderly selloff. Dealers protecting their balance sheets pulled back from market-making in less liquid issues first, widening the gap.

The pattern across all three episodes is consistent: stress flows toward the benchmark, and away from everything else. On-the-run issues become the safe harbor within the safe-haven asset class.

The Real Risks of Chasing the Off-the-Run Spread

The yield pickup is real, but so are the risks. Here's what can go wrong — and why the spread exists in the first place.

Liquidity evaporation in stress. Off-the-run issues can become very difficult to sell in a crisis. During March 2020, bid-ask spreads on off-the-run 10-year notes widened from 1/32 to 4/32 or more, and some issues saw no executable bids for hours. If you might need to liquidate quickly, this is the risk that matters most.

Convergence trade risk. The on-the-run/off-the-run spread can widen before it narrows. LTCM's convergence trades were directionally correct, but the fund couldn't survive the margin calls as spreads widened in 1998. Leverage amplifies this risk exponentially — being right eventually doesn't help if you're bankrupt today.

Roll risk and benchmark transitions. When a new 10-year note is auctioned, the prior benchmark loses on-the-run status overnight. Trading volume in the former benchmark typically drops 40–60% within two weeks of the roll, and bid-ask spreads widen. If you're holding a security that's about to go off-the-run, the liquidity profile of your position changes without any action on your part.

Repo and financing costs. Off-the-run issues often trade at higher repo rates (general collateral) compared to on-the-run issues, which can trade "special" at below-GC rates. The financing cost difference can offset or exceed the yield pickup, particularly for leveraged positions. Always check the specific CUSIP's repo rate before assuming the yield spread is pure profit.

Mark-to-market volatility. Off-the-run issues may have stale or wider pricing, especially in less-liquid tenors. Portfolio managers relying on end-of-day pricing from dealers may see more mark-to-market noise relative to the on-the-run benchmark — a nuisance for daily NAV calculations and a real problem for margin accounts.

The test: Can you hold this position through a March 2020–style event without being forced to sell? If the answer is yes, the off-the-run spread is likely worth capturing. If there's any doubt, the liquidity of the benchmark is worth paying for.

Practical Checklist (Before You Trade Either Side)

Identify the current on-the-run CUSIP for each benchmark maturity (2-, 5-, 10-, 30-year) using TreasuryDirect auction results
Calculate the on-the-run/off-the-run spread for your target maturity; the typical range is 2–6 bps in normal markets — anything outside that range warrants investigation
Check the bid-ask spread on any off-the-run issue before buying; compare it to the on-the-run to estimate true transaction costs
Evaluate your holding period: the yield pickup from off-the-run issues mainly benefits investors who plan to hold to maturity or for extended periods
Review repo rates for the specific CUSIP: if the off-the-run trades at general collateral and the on-the-run is trading special, account for that financing difference
Monitor the Treasury auction calendar; know when benchmark rolls occur so you're not surprised by a sudden drop in liquidity on your position
Assess portfolio-level liquidity needs: if you may need to liquidate quickly, overweighting off-the-run issues increases execution risk during stress
For convergence trades (long off-the-run, short on-the-run), stress-test the position for a 20+ bps spread widening and ensure you can meet margin requirements — LTCM is the permanent cautionary tale