Why Can an On-the-Run Treasury Yield Less? The Off-the-Run Spread

The on-the-run Treasury in a maturity sector is generally the most recently issued benchmark security. When a newer issue becomes the market's reference, the former benchmark becomes off-the-run. These are market-status labels: an older note does not change its promised coupons or maturity simply because a new benchmark appears, and the label itself does not change the U.S. Treasury as issuer.

Investors and dealers may prefer the current benchmark because it is quoted, traded, and used for comparisons more often. They may also value the ability to trade it quickly or use the specific issue for a hedge or a securities-financing transaction. Those services can make buyers willing to pay more for the on-the-run security. For the same promised cash flows, a higher price corresponds to a lower yield.

The resulting on/off-the-run yield spread is often positive when defined as the off-the-run yield minus the on-the-run yield. That pattern is common, not a rule that must hold every day or for every pair of securities. A useful explanation therefore starts with what the two yields actually compare, then asks whether liquidity, benchmark demand, repo, or different cash flows could account for the gap.

For a clearly stated convention, define the spread as:

off-the-run yield spread = y_off − y_on

If the two securities are sufficiently comparable and y_off is higher, the on-the-run security has the lower yield and trades richer on a yield basis. This notation does not make two bonds with different coupons or maturity dates identical. Their yields to maturity reflect different payment schedules, remaining terms, and exposure to the shape of the yield curve.

A careful comparison uses the same observation date and time, yield convention, maturity sector, and market source. A researcher can go further by comparing the market yield on the on-the-run bond with the yield of a hypothetical off-the-run bond that has identical cash flows, inferred from an off-the-run curve. The New York Fed's research uses this kind of matched-cash-flow comparison when constructing an on-the-run premium; it is a model-based reference, not necessarily a bond that can be purchased.

Do not subtract a constant-maturity Treasury series from a particular bond's yield and call the result a pure on/off-the-run premium unless the series definitions, dates, and curve construction make that comparison appropriate. A fitted curve can reduce maturity and coupon mismatches, but it introduces modeling choices and market-quote noise. For tradeable bonds, the exact CUSIP and its cash flows remain important.

Liquidity is the ability to trade in useful size without an unusually large price concession or long delay. The newest benchmark often attracts more interdealer activity and acts as a reference for other Treasury securities. Buyers may accept a lower yield in exchange for a bond that is easier to price, hedge, finance, or sell. The price premium is compensation for those market services; it is not extra coupon income.

New York Fed research using Treasury transaction data finds that trading activity and frequency decline as securities move from on-the-run to off-the-run status, while effective trading costs generally rise as securities age. The pattern is not a law for every issue: some seasoned bonds remain active because they are important for a Treasury futures delivery or another specific use. The study describes a sample and measured trading patterns, not a live quote or a guarantee about a particular security. See the New York Fed's analysis of Treasury liquidity as securities age.

Benchmark demand can reinforce that difference. A widely watched issue may be the first reference dealers show on a screen or use to describe a maturity point. Demand for a benchmark can affect its price even when the issuer, payment promise, and broad economic outlook are unchanged. The strength of that demand varies with market participants' hedging, inventory, and trading needs; the benchmark label does not guarantee a stable premium.

Liquidity also cannot be reduced to one displayed bid-ask spread. A Federal Reserve study finds that indicative spreads can misstate actual trading costs for seasoned Treasuries and that benchmark liquidity affects how those indicative spreads relate to costs. Trade size, execution venue, market depth, and the particular security matter. A wider screen quote is evidence to investigate, not a complete estimate of the cost of selling a position. See the Federal Reserve's study of off-the-run trading costs and indicative liquidity.

“Ten-year Treasury” identifies a maturity sector, not a single identical cash-flow schedule. Two notes in that sector may have different coupons, issue dates, remaining maturities, coupon dates, and prices. A small difference in remaining term can matter when the yield curve slopes or bends, and coupon size changes the timing and amount of interim cash flows. A raw yield gap between two such bonds mixes these differences with any on/off-the-run effect.

Treasury auctions also include reopenings. TreasuryDirect explains that a reopened security has the same CUSIP, maturity date, and interest rate or spread as the original security, but a different issue date and usually a different price. A reopening adds more of an existing security; it is not automatically a distinct new CUSIP. This distinction matters when identifying which bond is the current benchmark and when matching an older issue to it. See the Treasury's schedule and explanation of auction reopenings.

The auction cycle can change both supply and demand around a benchmark. When a newer issue takes over the reference role, trading attention may move with it even though the older security's contractual terms remain intact. A reopened issue can also change the amount available without changing its CUSIP. Researchers therefore may control for weeks since issuance or auction events when studying the spread; those adjustments help interpret a sample but do not turn the observed gap into a single observable cause.

Repo is a way to exchange securities for cash and later reverse the exchange. In a general-collateral, or GC, repo, the cash provider accepts securities from an eligible basket. In a specific-security repo, the cash provider wants a named issue. When that bond is difficult to obtain, the holder may be able to borrow cash against it at a repo rate below the rate on comparable GC collateral. This low issue-specific repo rate is called specialness.

Specialness matters to the cash bond because the Treasury can be valuable not only for its promised payments but also for its usefulness as collateral. A buyer may pay more for an issue that is useful to finance or deliver, which can correspond to a lower cash-market yield. For someone short the bond, however, borrowing the exact issue can be costly relative to general collateral. The repo rate on a security and its cash-market yield are different rates; the connection runs through the value and availability of the security, not a one-for-one conversion formula.

The Federal Reserve's research on repo and bond-market intermediation describes how repo frictions can affect cash-market liquidity. It also explains why a position that appears to exploit the on/off-the-run price difference can lose its apparent advantage when the cost of borrowing the on-the-run issue is included. The New York Fed's Treasury-market study separately treats the on-the-run premium as containing both liquidity and repo-specialness effects, and uses repo and auction-cycle variables in an empirical decomposition. Neither paper establishes that repo specialness explains every observed spread. See the Federal Reserve's repo-intermediation study and the New York Fed's Treasury-market functionality study.

On-the-run and off-the-run nominal notes and bonds are obligations of the same U.S. Treasury issuer. The status label does not assign them different seniority or issuer credit ratings. Their market yields can still differ because the specific cash flows, maturity, demand, trading costs, collateral value, and available supply differ. The spread alone does not isolate credit risk or prove that one security is safer.

Nor does a positive spread mean that a trader can lock in a risk-free profit. A relative-value position needs appropriate cash-flow or duration matching and may require selling or borrowing the on-the-run issue. That leg can face repo specialness, limited availability, changing financing terms, and settlement risk. Bid-ask costs, market impact, accrued interest, coupon cash flows, hedge error, margin, and the time until a spread converges can reduce or erase the apparent difference. The spread can widen, shrink, or reverse while a position is open.

These frictions are why a yield spread is better treated first as a market observation. The New York Fed's analysis controls for repo specialness and weeks since issuance when estimating a liquidity component, but the paper describes a statistical decomposition of a sample. It does not provide a guaranteed trade signal or a directly observed, perfectly separated liquidity premium. A historical relationship can help explain a mechanism without predicting what a particular pair will do next.

Suppose, purely hypothetically, that on one date a 10-year on-the-run note yields 4.06%. A fitted off-the-run curve implies that a hypothetical bond with the same cash flows would yield 4.10%. With the definition above:

4.10% − 4.06% = 0.04 percentage point = 4 basis points

The on-the-run bond is richer on this matched-yield comparison. The four-basis-point figure is an illustration, not a current Treasury quote, forecast, or estimate of a trade's profit. It also does not translate into a fixed dollar price gap without the cash-flow dates, coupon, maturity, and price-yield calculation.

If instead an actual older note has a different coupon and several months less remaining maturity, subtracting the current benchmark's yield from the older note's displayed yield gives a raw pair spread, not the same-cash-flow measure above. The two can differ because of the curve and payment schedules as well as liquidity or repo demand. If, on a different hypothetical date, the comparable off-the-run yield were 4.05% and the on-the-run yield were 4.07%, the sign would reverse by two basis points. That reversal would not itself indicate a change in the U.S. Treasury's credit quality; it would be a different market observation to investigate.

First identify both CUSIPs, whether each is an original issue or a reopening, its coupon and maturity, and when each observation was recorded. Confirm that the yields use the same convention and that the securities are in a comparable maturity sector. State whether the spread is a raw bond-to-bond difference or a comparison against a fitted, same-cash-flow off-the-run yield.

Next check what mechanism the evidence can support. Trading volume, effective transaction costs, order-book depth, and benchmark use help describe cash-market liquidity. Repo rates for the specific issue versus comparable GC collateral help show whether the bond was special on the relevant dates and tenor. A broad liquidity index or quoted bid-ask spread does not by itself identify the cause of one CUSIP's yield.

Finally, describe the measured difference before assigning a cause or discussing a trade. “The matched-cash-flow off-the-run reference yielded four basis points more on this date” is a bounded result. “The difference was entirely liquidity” or “the spread guarantees arbitrage” needs evidence and assumptions beyond two yields. For related mechanics, see the repo and reverse-repo guide and the Treasury futures price-versus-yield guide.