FI 10 Interest Rate Risk and Return
Earlier yield work showed that an investor earns exactly the yield-to-maturity (YTM) quoted at purchase only when a specific set of conditions holds. Here we open up the return and look at what actually drives it. A holder of a fixed-rate bond collects return from three distinct places: (1) the promised coupon and principal payments received on schedule, (2) the reinvestment of each coupon as it arrives, and (3) any capital gain or loss if the bond is sold before it matures. The first source is contractual and, ignoring default, certain. The other two move with market interest rates, and that is where the risk lives.
The realized return over a holding period is called the horizon yield: the single annualized rate that links what an investor pays today to everything received later. To compute it, add the future value of the reinvested coupons to the sale or redemption amount, divide by the purchase price, and annualize over the holding period T.
That reinvested-coupon total is itself an annuity accumulation: each coupon PMT is rolled forward at the prevailing reinvestment rate r across the N periods that remain until the horizon date.
When the horizon yield equals the original YTM
To see the benchmark case, follow two investors who each buy the same instrument: BRWA’s new 10-year, 6.2 percent annual coupon eurobond, priced at par. The first, Viswan Family Office (VFO), holds to maturity. The second, Baywhite Financial, plans to sell after four years. We start with interest rates unchanged, then shock them.
VFO buys BRWA’s 10-year, 6.2 percent annual coupon eurobond at par (settlement 15 October 2025, maturity 15 October 2035) and holds it for the full ten years. Coupons are reinvested at 6.2 percent.
Baywhite buys the identical BRWA bond at par but sells immediately after collecting the fourth coupon. Rates stay at 6.2 percent throughout.
Interest income is the return earned simply from the passage of time: it covers the coupons, the reinvestment of those coupons, and the gradual amortization of any purchase discount or premium back toward par. A capital gain or loss is separate; it comes from a change in the YTM (the implied market discount rate). The constant-yield price trajectory traces the bond’s carrying value over time at the original yield. For a par bond, that trajectory is a flat line at 100, so selling on it produces no gain or loss. A sale above the trajectory is a capital gain; a sale below it is a capital loss.
Shocking the interest rate
Now raise rates by 100 basis points (bps) the instant after purchase, lifting the yield and the reinvestment rate from 6.2 percent to 7.2 percent, then lower them by 100 bps to 5.2 percent. The two investors respond very differently because one sells and one does not.
Rates jump to 7.2 percent immediately after purchase.
Instead, rates fall to 5.2 percent immediately after purchase.
The two forces at work are the two faces of interest rate risk. The future value of reinvested coupons rises when rates rise and falls when rates fall; that is reinvestment risk. The sale price of a bond that still has life left after the horizon falls when rates rise and rises when rates fall; that is price risk. They pull in opposite directions. Someone who exits even before collecting a single coupon carries price risk alone; a buy-and-hold investor carries reinvestment risk alone. Two investors holding the very same bond can therefore carry entirely different interest rate risk, decided by their investment horizon.
If reinvestment risk and price risk always offset, is there a horizon at which they cancel exactly? To find it, add a third investor to the same BRWA bond: Hightest Capital, with an eight-year horizon. This is long enough that reinvestment matters, but short enough that a sale still happens with two years of bond life left, so both effects are present.
Hightest buys BRWA’s 10-year, 6.2 percent bond at par and sells after eight years, when two years remain to maturity.
Collecting the three investors together makes the pattern sharp. VFO and Baywhite each swing several basis points as rates move, but in opposite directions; Hightest barely moves at all.
| Investor (horizon) | Rates stable 6.20% | Rates rise to 7.20% | Rates fall to 5.20% |
|---|---|---|---|
| VFO (10 years, buy and hold) | 6.20% | 6.43% | 5.98% |
| Baywhite (sells after 4 years) | 6.20% | 5.28% | 7.16% |
| Hightest (sells after 8 years) | 6.20% | 6.24% | 6.17% |
The long-horizon investor gains from rising rates through reinvestment; the short-horizon investor gains from falling rates through price. The eight-year investor is close to indifferent.
Macaulay duration as the balancing horizon
The eight-year horizon was not arbitrary. It is close to the Macaulay duration of this bond: the holding period at which the reinvestment gain (loss) from a one-time parallel shift in the yield curve is exactly cancelled by the price loss (gain) on the sale. It takes its name from Frederick Macaulay, a Canadian economist who set out the idea back in 1938. As time passes after a rate rise, the bond price is pulled back to par while the extra reinvestment income builds; at the Macaulay duration the two are equal in size and opposite in sign, and the same balancing holds when rates fall. For this BRWA bond the figure works out to 7.7429 years (computed in the next section), which is exactly why Hightest, at eight years, sits so close to fully hedged.
This gives clean general statements about the relationship among horizon, duration, and interest rate risk.
| Relationship | Dominant risk | Source of interest rate risk |
|---|---|---|
| Investment horizon > Macaulay duration | Reinvestment risk | Falling interest rates |
| Investment horizon = Macaulay duration | Price risk = reinvestment risk | None |
| Investment horizon < Macaulay duration | Price risk | Rising interest rates |
The gap between the two quantities has a name. The duration gap is the Macaulay duration minus the investment horizon.
VFO, at ten years, has a negative duration gap, so its main threat is falling rates. Baywhite, at four years, has a positive gap and fears rising rates. Hightest, at eight years, has a gap near zero and is close to hedged. One caution: both quantities shrink as time elapses, since the remaining horizon and the bond’s duration both fall, so the match does not stay perfect without rebalancing.
The same logic on a longer bond
The offset is a general property, not a quirk of one bond. Take three investors who each buy, at par, a 5.61 percent Romanian eurobond maturing in 30 years and carrying a 15.16-year Macaulay duration. Rook Point holds for five years (well below the duration), Fyleton fifteen years (almost equal to it), and Amy thirty years (well above it). Watch the middle row: Fyleton’s horizon yield stays near 5.61 percent in every scenario, while the mismatched investors swing.
| Investor (horizon) | FV of coupons, stable | Sale price, stable | Yield, stable 5.61% | Yield, rise to 6.11% | Yield, fall to 5.11% |
|---|---|---|---|---|---|
| Rook Point (5 years) | 31.379 | 100.000 | 5.610% | 4.626% | 6.659% |
| Fyleton (15 years) | 126.765 | 100.000 | 5.610% | 5.613% | 5.624% |
| Amy (30 years) | 414.223 | 100.000 | 5.610% | 5.861% | 5.367% |
Under the rate-rise scenario the sale prices are 93.675, 95.179, and 100.000; under the rate-fall scenario they are 106.970, 105.152, and 100.000. When the horizon nearly equals the Macaulay duration (Fyleton), the coupon reinvestment gain or loss offsets the price loss or gain, so the return holds steady.
The balancing horizon has a precise definition. It measures, as a present-value-weighted average, when the payments on a bond land: each receipt date is weighted by the slice of the full price (the present value) that the flow arriving then contributes. Cash flows further out carry more time but, once discounted, a smaller weight, so the duration lands below the time-to-maturity for any coupon-paying bond.
Here i indexes the periods, PVi is the present value of the cash flow in period i, and PVFull sums those present values into the full price. The term t counts the days elapsed since the last coupon, while T gives the length of the whole coupon period in days. On an issuance or coupon date, t is zero, so every flow arrives exactly at its period number.
Compute the Macaulay duration of BRWA’s 10-year, 6.2 percent annual eurobond at issuance, priced at par with a 6.2 percent YTM, on a 30/360 basis (settlement 15 October 2025, maturity 15 October 2035).
| Period | Time to receipt | Cash flow | Present value | Weight | Time × weight |
|---|---|---|---|---|---|
| 1 | 1 | 6.2 | 5.8380 | 0.0584 | 0.0584 |
| 2 | 2 | 6.2 | 5.4972 | 0.0550 | 0.1099 |
| 3 | 3 | 6.2 | 5.1763 | 0.0518 | 0.1553 |
| 4 | 4 | 6.2 | 4.8741 | 0.0487 | 0.1950 |
| 5 | 5 | 6.2 | 4.5895 | 0.0459 | 0.2295 |
| 6 | 6 | 6.2 | 4.3216 | 0.0432 | 0.2593 |
| 7 | 7 | 6.2 | 4.0693 | 0.0407 | 0.2849 |
| 8 | 8 | 6.2 | 3.8317 | 0.0383 | 0.3065 |
| 9 | 9 | 6.2 | 3.6080 | 0.0361 | 0.3247 |
| 10 | 10 | 106.2 | 58.1942 | 0.5819 | 5.8194 |
| Total | 100.0000 | 1.0000 | 7.7429 |
Beyond building the table, Macaulay duration can be read straight from the DURATION function in Excel or Google Sheets, using the settlement date, maturity date, coupon, yield, coupon frequency, and a day-count basis (0 or omitted is 30/360). It can also be computed in one closed-form expression derived with algebra and calculus, which is handy between coupon dates:
Here r is the periodic yield, N the number of periods to maturity from the start of the current period, c the periodic coupon rate, and t/T the fraction of the coupon period elapsed. Applied to the BRWA bond at issuance (r = c = 0.062, N = 10, t/T = 0), this returns 7.7429 years, matching the table.
Duration falls as time passes
Macaulay duration is not fixed. Each day that elapses shortens every cash flow’s time to receipt, so the duration drifts lower. Recomputing the BRWA bond just 57 days after issuance makes the drift visible.
Recompute the BRWA bond’s Macaulay duration on 11 December 2025, which is 57 days after issuance, on a 30/360 basis, with the YTM still 6.2 percent. Now t/T equals 57/360, so every payment sits 57 days closer than at issuance.
| Period | Time to receipt | Cash flow | Present value | Weight | Time × weight |
|---|---|---|---|---|---|
| 1 | 0.8417 | 6.2 | 5.8939 | 0.0584 | 0.0491 |
| 2 | 1.8417 | 6.2 | 5.5498 | 0.0550 | 0.1012 |
| 3 | 2.8417 | 6.2 | 5.2258 | 0.0518 | 0.1471 |
| 4 | 3.8417 | 6.2 | 4.9207 | 0.0487 | 0.1872 |
| 5 | 4.8417 | 6.2 | 4.6335 | 0.0459 | 0.2222 |
| 6 | 5.8417 | 6.2 | 4.3630 | 0.0432 | 0.2525 |
| 7 | 6.8417 | 6.2 | 4.1082 | 0.0407 | 0.2784 |
| 8 | 7.8417 | 6.2 | 3.8684 | 0.0383 | 0.3005 |
| 9 | 8.8417 | 6.2 | 3.6426 | 0.0361 | 0.3190 |
| 10 | 9.8417 | 106.2 | 58.7511 | 0.5819 | 5.7273 |
| Total | 100.9570 | 1.0000 | 7.5845 |
Annualizing a semiannual bond
When a bond pays more than once a year, the table is built in periods and then annualized by dividing by the number of periods per year. The next example shows the two-period-per-year case.
A bond has three years to maturity, a 4 percent coupon paid semiannually, and is priced at 100 (so the periodic YTM is 2 percent).
| Period | Time to receipt | Cash flow | Present value | Weight | Time × weight |
|---|---|---|---|---|---|
| 1 | 1 | 2 | 1.961 | 0.0196 | 0.01961 |
| 2 | 2 | 2 | 1.922 | 0.0192 | 0.03845 |
| 3 | 3 | 2 | 1.885 | 0.0189 | 0.05654 |
| 4 | 4 | 2 | 1.848 | 0.0185 | 0.07391 |
| 5 | 5 | 2 | 1.811 | 0.0181 | 0.09057 |
| 6 | 6 | 102 | 90.573 | 0.9057 | 5.43438 |
| Total | 100.00000 | 1.00000 | 5.71346 |