Lookback, Lockout and Payment Delay

You set the lookback, lockout, and payment delay terms when you calculate interest on loans that use rates compounded in arrears, when you calculate accrued interest on floating rate notes, and when you align terms with a derivatives hedge.

Knowing the interest amount in advance

The rate for compounding in arrears is not set until the overnight rate for the last business day of the interest period is published. The Bank of Japan publishes the final value of TONA (Tokyo Overnight Average Rate) on the next business day¹, so without some adjustment, the rate becomes known only on the end date of the period. For loans and bonds that pay interest on the end date of the period, this leaves no time to confirm the amount and prepare the payment. The formula and day counting for compounding in arrears are covered in Compounding in Arrears.

The Alternative Reference Rates Committee (ARRC) in the United States lists three conventions that give advance notice of payments on floating rate bonds (floating rate notes)². These notes use SOFR. The three are the lookback, the lockout, and the payment delay, and they are sometimes used in combination.

Lookback: using rates from earlier business days

A lookback is a convention that uses, for each day in the interest period, the rate from a number of business days before that day. With a five-business-day lookback, even the last business day of the period uses the rate from five business days earlier, so the rate is known five business days before the end of the period. In an example in the ARRC's loan paper, a 30-day period starts on June 1. June 1 uses the rate for May 25, five business days earlier, and June 2 uses the rate for May 26. The agent (the financial institution that handles the administration of the loan) knows the rates through the end of the period by June 23, and the borrower can pay interest by June 30³.

A lookback comes in two methods, depending on which days are used to weight the rates. Without an observation shift, only the rates are taken from a number of business days earlier, and the day weights stay those of the interest period. The loan paper of the Working Group on Sterling Risk-Free Reference Rates in the United Kingdom also calls this method "Lag"⁴. With an observation shift, the period in which rates are observed (the observation period) is itself moved earlier by a number of business days. Each rate is weighted by the number of days it applies to within that period⁵. With a lookback of $L$ business days, the rate without an observation shift is:

$$R = \left( \prod_{i=1}^{d_b} \left( 1 + \frac{r_{i-L} \times n_i}{365} \right) - 1 \right) \times \frac{365}{D}$$

Here, $d_b$ is the number of business days in the interest period. The term $r_{i-L}$ is the rate for the day $L$ business days before the $i$th business day of the period. The term $n_i$ is the number of calendar days from the $i$th business day to the next business day, and $D$ is the number of calendar days in the interest period. The rate with an observation shift uses, for each business day in the observation period, that day's rate and the number of days until the next business day, and is annualized over the days in the observation period.

$$R = \left( \prod_{j=1}^{d_o} \left( 1 + \frac{r_j \times m_j}{365} \right) - 1 \right) \times \frac{365}{D_o}$$

Here, $d_o$ is the number of business days in the observation period. The term $r_j$ is the rate for the $j$th business day of the observation period. The term $m_j$ is the number of calendar days in the observation period that the rate applies to, and $D_o$ is the number of calendar days in the observation period. Under either method, the interest is the principal × $R$ × $D \div 365$, using the days in the interest period. The worked example in the Working Group's paper also annualizes the rate with an observation shift over the days in the observation period, and multiplies by the days in the interest period to get the interest⁶.

The two methods differ in how weekends and holidays are weighted. Without an observation shift, the three days of interest for a Friday are multiplied by the rate for a day before that Friday. As a result, the rate applied to weekends and holidays may not match the rate that reflects transactions over those days. A five-business-day lookback uses the rate for the same day of the week, so weekends are weighted consistently, but holidays are not⁷. With an observation shift, the number of days in the observation period may differ from the number of days in the interest period (for example, when the observation period starts just before a holiday)⁸.

Lockout: fixing the rate for the last few days

A lockout repeats the rate for the cut-off date, which falls a set number of business days before the end of the period, through the last business day. The period in which the cut-off rate is used may also be called a "suspension period"². Here, a lockout of $K$ business days sets the cut-off date $K$ business days before the end date of the period. The rate for the cut-off date is repeated for the $K - 1$ business days after it. For a rate published on the next business day, such as TONA, the rate is known on the business day after the cut-off date, that is, $K - 1$ business days before the end of the period.

If the rate moves during the lockout, the move is not reflected in the rate. Consider a hypothetical example in which the 28-day period runs from Monday, February 23, 2026, to Monday, March 23, 2026, using the sample rates in the simulator below. The sample treats national holidays (February 23 and March 20) as business days and uses rates for them, so the result differs from a calculation based on actual Tokyo business days⁹.

In the sample, the rate rises on March 19 from around 0.477% to around 0.727%, so only the last 2 business days of the period use the higher rate. With a three-business-day lockout, the cut-off date is Wednesday, March 18, and the 2 business days of the 19th and 20th also use the rate for the 18th. The table below shows the last 5 business days of the period.

The period rate is 0.5128% without a lookback or lockout, but 0.4772% with a three-business-day lockout. The interest on a principal of 100 million yen falls from 39,336 yen to 36,604 yen, or 2,732 yen less¹⁰. In this example, the borrower pays less interest, and the lender receives less by the same amount. If the rate falls at the end of the period instead, the lockout raises the rate and the borrower pays more interest. Rate increases and decreases during the lockout are not reflected in interest payments, which can work for or against either the lender or the borrower¹¹.

Payment delay: moving the payment date later

A payment delay is a convention that leaves the rate calculation unchanged and sets the payment date a number of business days after the end of the period. Each day in the period uses that day's own rate. The rate is set on the end date of the period, and the parties confirm and pay the amount during the few business days before payment. In a Japanese yen OIS example, the floating amount is calculated on the end date of the period after TONA is published, and the net amount is settled 2 business days later¹². This is also a form of payment delay.

In the ARRC paper's example of a bond with a payment delay, a two-business-day lockout is added only for the final period, so that its interest can be paid on the maturity date. Because the payment for the first period arrives two business days after the period ends, investors may lose reinvestment income for those days¹³.

Rates and payment dates for the same period

Consider a hypothetical example in which the 88-day period runs from Monday, February 2, 2026, to Friday, May 1, 2026, with rates of around 0.477% through March 18 and around 0.727% from March 19. The table compares the rate and payment date under each convention¹⁰. The sample treats national holidays (February 11, February 23, March 20, and April 29) as business days and uses rates for them, so the result differs from a calculation based on actual Tokyo business days. The payment date of May 5 in the payment delay row is also a national holiday in practice⁹. "Rate known on" assumes a rate published on the next business day.

A lookback lowers the rate here because it delays the day on which the higher rate from mid-period starts to be used. Without a lookback or lockout, the higher rate is used for 43 of the 88 days. With a two-business-day lookback without an observation shift, it is first used from March 23, which makes 39 days. With a two-business-day lookback with an observation shift, the observation period runs 90 days, from Thursday, January 29, to Wednesday, April 29, and the higher rate is used for 41 of them. That is why the rate with a shift is slightly higher than the rate without one.

With a five-business-day lookback, every day uses the rate for the same day of the week, and the observation period stays at 88 days, so the two methods give the same rate. The rates at the end of the period are almost flat, so the rate with a two-business-day lockout is almost the same as without a lookback or lockout. In a period when rates fall, a lookback raises the rate.

Figure description

This figure puts the rate-known dates and payment dates from the 88-day comparison table in date order. The more business days in the lookback, the earlier the rate is known; with a lockout it is known later than that, and with no shift it is known at the end of the period. Only with a payment delay does the payment date fall after the end of the period. The filled dot marks the end of the period. The dates come from a count that treats holidays as business days, so they differ from actual Tokyo business days. The spacing does not reflect the length of time between dates.

Use in loans, bonds, and derivatives

For Japanese yen loans, the Cross-Industry Committee on Japanese Yen Interest Rate Benchmarks (the Committee) published its conventions for calculating interest on loans that use TONA compounded in arrears in December 2020¹⁴. In an earlier consultation, an overwhelming majority of participants in the Committee's loan subgroup chose the lookback without an observation shift ("Lookback without Observation Shift"). The loan subgroup continued its work with this method as the core¹⁵.

In the next consultation, a large majority supported presenting five business days as an example of the lookback length, citing consistency with work in the United Kingdom and the United States and the practice of interest payments¹⁵. The UK body recommended five business days and the US body presented five business days as an example¹⁶. Some respondents, however, said a shorter lookback is better in economic terms. An overwhelming majority also agreed not to preclude the use of the method with an observation shift ("Observation Period Shift")¹⁵.

For US syndicated loans, the ARRC also recommends a business day lookback without an observation shift. With an observation shift, if a loan is prepaid in a period when the number of days in the observation period and the interest period differ, the borrower may pay too much or too little interest. The same mismatch can arise between the seller and buyer when a loan is traded¹⁷.

For bonds, as of the ARRC's August 2019 paper, market conventions for floating rate notes using SOFR had not settled on one approach. Among bond issues using the sterling overnight rate SONIA (Sterling Overnight Index Average)¹⁸, those recent as of the paper had adopted a five-business-day lookback. At the time, the ARRC's floating rate notes working group preferred, among the structures that use a lookback, one with an observation shift. This is because it weights each rate correctly and allows the use of a published compounded SOFR index¹⁹.

In the working group's view, a longer lookback or lockout gives more time to confirm the amount for tax calculation and clearing systems, while a shorter one keeps the rate closer to market rates¹⁹.

In the ARRC working group's assessment, the payment delay convention is common in the derivatives market¹³. Using an observation shift or a payment delay in floating rate notes makes it easier to align the terms with a derivatives hedge. Floating rate notes whose terms do not match the hedge can add cost and risk¹⁹.

Notes

1. Bank of Japan Financial Markets Department, "「コール市場関連統計」の解説" (Explanation of call market statistics, in Japanese), the table in "３．公表方法" (Publication method), page 2. The final value is published at around 10:00 a.m. on the next business day. ↩

2. Alternative Reference Rates Committee, "SOFR Floating Rate Notes Conventions Matrix" (August 2019), page 6, the descriptions of Lockout, Lookback and Payment Delay, and note 11. ↩ ↩²

3. Alternative Reference Rates Committee, "SOFR 'In Arrears' Conventions for Syndicated Business Loans," page 2, "Lookback/Lockout/Payment Delay," and note 10 (the example assumes that May 25 is not a holiday). The paper's term for the agent is "administrative agent." ↩

4. Working Group on Sterling Risk-Free Reference Rates, "Detailed Loans Conventions" (September 2020, updated March 2021), note 1 on page 2 (Observation Lag convention) and note 1 on page 4 (Lag). ↩

5. Alternative Reference Rates Committee, "SOFR Floating Rate Notes Conventions Matrix," page 8, the description of the third structure (FRNs with Two-Day Backward-Shifted Observation Period and No Lockouts). ↩ ↩²

6. Working Group on Sterling Risk-Free Reference Rates, "Detailed Loans Conventions," page 4 (under either method, interest is calculated using the calendar days in the interest period), and the worked example on page 10 (the rate is annualized over the 28 days of the observation period, and interest is calculated over the 31 days of the interest period). ↩

7. Alternative Reference Rates Committee, "SOFR Floating Rate Notes Conventions Matrix," page 8, the description of the second structure (FRNs with Five-Day Lookback and No Lockouts). ↩

8. The same document, page 8, note 17. ↩

9. Cabinet Office, "国民の祝日について" (About national holidays, in Japanese), the section "令和８年（２０２６年）及び令和９年（２０２７年）の「国民の祝日」" (National holidays in 2026 and 2027). ↩ ↩²

10. The worked examples and the simulator calculate interest without rounding the rate along the way, and show interest rounded to the nearest yen. ↩ ↩²

11. Alternative Reference Rates Committee, "SOFR Floating Rate Notes Conventions Matrix," page 7 (the description of the first structure), and the second point of the working group's views on page 10. ↩ ↩²

12. Study Group on Risk-Free Reference Rates, "日本円ＯＩＳ（Overnight Index Swap）─取引の概要と活用事例─" (Yen OIS: overview of transactions and examples of use, April 2018, in Japanese), page 6, "例２：１年未満のOIS 清算取引で、満期時のみに利払いを行うケース" (Example 2: a cleared OIS of less than one year with interest paid only at maturity). ↩

13. Alternative Reference Rates Committee, "SOFR Floating Rate Notes Conventions Matrix," page 9, the description of the fourth structure (FRNs with Two-Day Payment Delay, No Lookback, and Two-Day Lockout prior to Maturity). ↩ ↩²

14. Bank of Japan Financial Markets Department and Financial System and Bank Examination Department, and Financial Services Agency Strategy Development and Management Bureau and Supervision Bureau, "円LIBOR移行対応の振り返りと今後の取り組み" (Bank of Japan Review 2022-J-8, May 2022, in Japanese), page 3, Chart 3. An English version is available: "Review of JPY LIBOR Transition and Future Initiatives" (Bank of Japan Review 2022-E-4, May 2022). ↩

15. Cross-Industry Committee on Japanese Yen Interest Rate Benchmarks, "貸出における TONA（後決め）のコンベンション（利息計算方式）について" (Conventions for TONA compounded in arrears in loans (interest calculation methods), in Japanese), the opening of the main text and "Lookback 期間（調整日数）" (Lookback period (adjustment days)) on page 1, and "①以外の方式" (Methods other than 1) on page 2. ↩ ↩² ↩³

16. The same document, main text, page 1, note 3. ↩

17. Alternative Reference Rates Committee, "SOFR 'In Arrears' Conventions for Syndicated Business Loans," pages 2–3, "Lookback/Lockout/Payment Delay." ↩

18. Bank of England, "SONIA interest rate benchmark," the opening description. ↩

19. Alternative Reference Rates Committee, "SOFR Floating Rate Notes Conventions Matrix," page 6 (conventions not settled), page 7 (the second structure), the third and fourth points of the working group's views on page 10, and the days entry on pages 10–11. ↩ ↩² ↩³

20. Alternative Reference Rates Committee, "SOFR 'In Arrears' Conventions for Syndicated Business Loans," page 4, "SOFR Index." ↩

21. Alternative Reference Rates Committee, "SOFR Floating Rate Notes Conventions Matrix," pages 7–10, "Secondary trading consideration" for each structure, and note 14. ↩