Almost every discussion about which Treasury bond to buy stalls at the same point: where interest rates are headed. If they’re rising, they say, prefer the coupon bond. If they’re falling, the long fixed-rate one. The whole conversation orbits a rate forecast that nobody gets right with any consistency — and that, deep down, is the least important part of the decision.
There’s a variable that decides more and that you control completely, unlike the Selic (Brazil’s benchmark interest rate). It isn’t on the Focus Report (the central bank’s weekly market-expectations survey) or on the Tesouro Direto (Brazil’s government bond platform) screen. It’s in your own financial life, and the question is embarrassingly simple: when this bond pays you, are you going to reinvest the money or spend it? The answer doesn’t tweak the decision at the margins. It inverts it.
Are you in the accumulation phase or the spending phase?
Every financial-independence plan has two halves, and they obey opposite logics. In the first, you turn earned income into wealth: every real that comes in is reinvested, and the goal is to maximise the balance down the line. In the second, the wealth is what sustains your life: the portfolio has to spit out cash, month after month, to pay the bills without you having to sell an asset in a panic.
A coupon bond serves the first half badly and the second half well. And it’s easy to get confused, because most people aren’t cleanly in either — they’re in a middle ground, saving part and spending part, without ever having asked the question explicitly.
The diagnostic error is expensive. If you’re accumulating and you pick the coupon for the sake of “passive income,” you’re paying tax early and taking on reinvestment risk to receive a cash flow you don’t even use — it goes back into the portfolio anyway, just already bitten by the taxman. It’s a cost you get nothing in return for. We covered that trap in The High-Coupon Myth; here the point is the flip side of the coin.
The question isn’t about your risk profile. It’s about the job that money does in your life right now.
What does the coupon do to someone who reinvests?
It gets in the way, and the math is merciless about it. The contracted rate that appeared on the screen — say, 13% a year — only materialises if you reinvest every coupon at exactly 13% until maturity. It’s an assumption baked into the bond, not a promise. If the Selic softens and you can only reinvest at 9%, the realized return collapses.
| Reinvesting coupons at | Realized return (coupon bond) | Zero-coupon |
|---|---|---|
| 13% (contracted rate) | 13.00% p.a. | 13.00% p.a. |
| 9% | 11.58% p.a. | 13.00% p.a. |
| 17% | 14.56% p.a. | 13.00% p.a. |
Look at the right-hand column: it doesn’t move. The zero-coupon delivers 13% no matter what the Selic does, because there’s nothing to reinvest — the return is locked in at purchase. The coupon bond swings almost three percentage points around that, and that swing is pure risk you carry for free.
For someone accumulating, the coupon is, at best, a directional bet on rising rates dressed up as income. Add the early tax, which quietly erodes compounding, and the verdict is clear: in the accumulation phase, money that doesn’t go back to earning intact is money working at half power.
And if the money is for spending — does anything change?
Everything changes. The coupon’s two big costs — reinvestment risk and early tax — only exist because they assume you were going to reinvest. Take reinvestment out of the equation and both evaporate.
Reinvestment risk disappears because there’s no reinvestment: that 1.42-point loss in the falling-Selic scenario was, entirely, the difference between reinvesting at 13% and at 9%. If the coupon becomes dinner, you never suffer that loss — it only existed relative to a reinvestment that, in your case, isn’t going to happen. And there’s a bonus the long fixed-rate bond doesn’t give you: the NTN-F (Brazil’s fixed-rate treasury bond with semiannual coupons) pays a coupon fixed in reais, so your nominal income doesn’t shrink when the Selic falls. A rate cut doesn’t cut your dinner.
The early tax stops being a disadvantage for the simplest reason in the world: you can’t defer money you need to spend. To get cash out of a zero-coupon, you have to sell part of the position — and the sale also triggers tax on the gain, only with two aggravating factors. You’re exposed to mark-to-market, so you can be forced to sell at the worst moment, and you liquidate principal unevenly, never quite sure how much is left working. The coupon hands you the cash with no sale at all, on a set date, in a predictable amount.
For someone who spends, the coupon’s tax cost stops being a toll and becomes the price of never having to sell an asset in a bad market.
That’s genuine protection against sequence-of-returns risk — and it’s worth putting a number on it, because “selling in a bad market” sounds abstract until you see the size of the hole.
How much does it cost to be forced to sell in the wrong year?
Picture two retirees, each holding a ten-year Treasury bond bought at 13%. The first has the coupon-paying NTN-F; the second, the Tesouro Prefixado (Brazil’s fixed-rate government bond) with no coupon. In the second year of retirement, rates jump from 13% to 16% — nothing exotic by Brazilian standards. Each needs to pull out R$ 60,000 to live on that year.
The retiree with the coupon does nothing. The NTN-F coupon is fixed in reais — R$ 48.81 per bond every semester — and the rate spike doesn’t touch it. He collects, spends, and moves on.
The one holding the Prefixado has to sell, and sell at the worst moment. With rates at 16%, the bond that was worth R$ 376 “on track” is marked at R$ 305 — a mark-to-market drop of nearly 19%. To pull the same R$ 60,000, he liquidates 23% more units than he would at fair value. And each unit sold at a depressed price leaves the portfolio for good — it doesn’t come back when rates fall again.
| Rate jump (8 years to maturity) | Mark-to-market loss if you sell |
|---|---|
| 13% → 15% | 13.1% |
| 13% → 16% | 18.9% |
| 13% → 18% | 29.3% |
| 13% → 20% | 38.2% |
This is sequence-of-returns risk in its rawest form: it isn’t the average of the returns that breaks the retiree, it’s the order they arrive in. A bad year early on, forcing the sale of principal, does damage that a decade of good returns afterward can’t undo. The coupon doesn’t eliminate that risk — it swaps “selling in a panic” for “collecting on schedule,” which is already quite a swap for anyone who depends on the cash flow.
(Yes, the same Prefixado marked at 16% will yield 16% to the end if you hold it. The problem was never the yield at maturity; it was needing the money before then.)
To assess yourself: what fraction of your retirement income comes from receiving — coupons, rent, dividends — and what fraction depends on selling a piece of your wealth? The more the scale tips toward “selling,” the more exposed you are to the wrong year.
The border is blurrier than the two words suggest
Here it’s worth distrusting your own ruler. “Accumulation” and “spending” sound like two territories separated by a fence, and almost nobody lives that way. The transition from career to retirement usually takes years, and is sometimes partial — Barista FIRE, where a small work income covers part of the expenses, blurs the line of purpose. Some people reinvest coupons for a decade and suddenly start spending them at 55, without having changed a single bond in the portfolio.
And there’s a behavioural detail the spreadsheet doesn’t catch. Even people who intend to reinvest often don’t — the coupon lands in the checking account, blends in with the salary, and vanishes. The person thinks they’re in the accumulation phase, but their behaviour is that of someone spending. In that case the honest diagnosis isn’t what they declare; it’s what they do.
I’ll leave that tension unresolved on purpose, because it doesn’t resolve on paper. It only resolves by looking at your last two years of statements.
How to use what you found
Put the two ends together — the phase and the actual behaviour — and you land in one of three places.
If you’re accumulating and actually reinvesting, the zero-coupon is the clean choice: Tesouro Prefixado or Tesouro IPCA+ Principal (its inflation-linked, no-coupon cousin, IPCA being Brazil’s official inflation index). You lock in the rate, defer the tax to a single event at the end, and carry no reinvestment risk at all. It’s not that the coupon is forbidden — it’s that you pay for three things (early tax, reinvestment risk, the temptation to spend) and get none of them in return.
If you live off your portfolio, or are close to it, the logic flips and the coupon bond starts to make sense: predictable cash flow, no forced sale, with nominal income that doesn’t sink in a rate cut. Here the tax on each coupon isn’t waste — it’s the price of protection worth having. Just one caveat about which coupon: the NTN-F’s is fixed in reais, and inflation erodes what it buys over a thirty-year retirement. For anyone who’ll live off income for decades, the Tesouro IPCA+ with Semiannual Interest is usually the better pick — it pays a coupon like the NTN-F, but adjusted for inflation, preserving the cash flow’s purchasing power.
And if you think you’re accumulating but spend the coupons without noticing? Then you’ve got the worst of both worlds: you pay the coupon’s tax cost and sabotage your own compounding on top of it. The fix might be changing the bond, or it might be changing the behaviour — automating the reinvestment until it becomes invisible. Which of the two is easier for you, only you know.
It’s worth less energy guessing the Selic and more modelling both trajectories against your real horizon. Run both paths through the Retirement Calculator and see in which year, exactly, you stop reinvesting and start needing the cash flow. That date — and not the next Copom (Brazil’s monetary policy committee) meeting — is what decides which bond belongs in your portfolio.
This article is for general educational purposes and does not constitute investment advice. The return figures are illustrative and assume a hypothetical contracted rate of 13% a year, with full reinvestment of coupons at the stated rate — assumptions that rarely hold exactly. The tax treatment of Tesouro Direto can vary with the bond, the purchase date, and changes in legislation. Confirm the current rules with Tesouro Direto and consult a qualified financial adviser before deciding.