Mortgage Points: When Buying Down the Rate Actually Pays

A point is 1 percent of the loan, not of the price

The vocabulary trips people before the arithmetic does. A discount point is one percent of the amount borrowed, paid once, at closing. On a $375,000 house bought with 20 percent down, the loan is $300,000, so one point is $3,000 — not $3,750. Get that base wrong and every downstream number is wrong by the size of your deposit. Lenders may also quote fractions: half a point, a quarter point, or an odd figure like 0.875 points that falls out of their pricing engine rather than out of a menu.

A point is also not the same thing as an origination fee, even when both are quoted in the same units. An origination fee pays the lender for arranging the loan and buys you nothing; a discount point buys a lower contractual rate that follows the loan for its whole life. Both belong in the APR, which is precisely why the APR on a loan with points sits closer to its nominal rate than you would expect, and why comparing two offers on nominal rate alone is how you end up paying for a rate you did not need.

The mirror image exists too and is worth knowing: a lender credit, sometimes called a negative point. You accept a higher rate and the lender pays part of your closing costs. It is the same trade run backwards — cash now against a worse rate later — and it is the right trade for exactly the borrower for whom points are wrong: someone who expects to move or refinance soon. The break-even arithmetic below works identically, with the signs flipped.

The break-even everyone computes, and the one they should

Take the worked case: $300,000 over 30 years, 6.00 percent without points, 5.75 percent with one. The instalment comes from the standard annuity formula, derived in full in our article on the EMI formula; here we simply use it. At 6.00 percent it returns $1,798.65, at 5.75 percent $1,750.72. The saving is $47.93 a month, the cost is $3,000, and the sum everyone writes down is $3,000 ÷ $47.93 = 62.6, rounded up to 63 months — five years and three months.

That number is wrong, and it is wrong in an instructive way. It adds a payment saved in month 60 to a payment saved in month 1 as if the two were interchangeable, while the $3,000 was handed over in month 0 in full. Money has a price. If the alternative use of $3,000 was earning 4 percent, then the honest question is when the present value of the stream of $47.93 savings first reaches $3,000. Discounting each monthly saving at 4 percent nominal — that is, dividing the saving in month k by 1.003333 to the power k — the running total crosses $3,000 in month 71. Not 63: 71, five years and eleven months. Eight months of difference, entirely from a step the shortcut skips.

There is a second correction, and it pushes the other way. A lower rate does not only shrink the payment; it changes how the payment is split. At 5.75 percent a larger slice of each instalment goes to principal, so the balance falls faster. After 63 months the no-points loan owes $920 more than the points loan. If you sell or refinance, you settle that balance, so those $920 are real money in your pocket on that date. Counting both effects — payment savings and the balance gap, each discounted at 4 percent — the point has repaid itself by month 54, four years and six months.

Three numbers, one loan: 63, 71 and 54 months. They are not competing estimates of the same quantity. They answer three different questions — how many payments does the raw saving take to add up, how long before the discounted saving repays the outlay, and how long before I am ahead if I sell on that date. The last one is the question a buyer actually faces, which is why the naive figure is not so much imprecise as beside the point.

Three things that reset the clock: selling, refinancing, overpaying

Points are bought on an assumption about the future that is almost never stated out loud: that the loan will still be there in five or six years. Every quoted break-even is really a bet on your own tenure. Sell the house at month 36 and the point has returned $1,725 of payment savings plus $527 of extra equity — $2,252 against $3,000 spent. You are $748 down, and no future saving will ever arrive, because the loan is gone.

Refinancing does the same thing faster. Rates fall a percentage point, you switch lender at month 30, and the point has returned $1,438 in payments and $439 in equity, $1,877 in all — leaving $1,123 unrecovered. The cruel part is the correlation: the environment that makes refinancing attractive is a falling-rate environment, and the borrower most tempted to buy down a rate is the one who took out the loan when rates were high, which is exactly the borrower most likely to refinance within three years.

Overpaying is the subtlest of the three, and here the popular claim needs correcting. Overpaying does not push the break-even date out — we ran it: with an extra $200 a month on top of the required payment, the sale-date break-even moves from month 54 to month 53, which is noise. What overpaying destroys is the tail. Without overpayment, one point saves $17,256 of interest over the full 30 years. Add $200 a month to each loan and the base loan finishes in 279 months with $256,341 of interest while the points loan finishes in 280 months with $244,892 — the point is now worth $11,449, a third less. The lower rate is applied to fewer months, so the long tail of savings that justified the fee simply never happens.

The rate cut per point is a price list, not a law of nature

The rule of thumb says a point buys a quarter of a percentage point. It is a rule of thumb about one lender's pricing at one moment, and it changes with the loan type, the term, your credit profile, the loan-to-value ratio and where the secondary market happens to be that week. The rate sheet in the table above is a realistic one, and it is not linear: the first point buys 0.25, the second buys 0.20, the third buys 0.15. Diminishing returns are the norm, because the lender is selling you progressively deeper discounts out of a shrinking margin.

The consequence is that the break-even lengthens as you buy more. In the table, half a point and one point both break even at 63 months naive and 71 discounted, because their cost and their benefit scale together. Two points push those to 70 and 80; three points to 79 and 92. Deep buydowns are therefore a bet on a much longer tenure than shallow ones — the opposite of the intuition that if a point is good, three are better. Always ask for the whole sheet, not the one line the loan officer picked, and compute each row.

Where points do not exist, and what replaces them

Discount points as a named, itemised product are a United States and Canadian convention. They exist because the American mortgage is sold into a secondary market that prices rate and fee as two dials on the same instrument, and because the thirty-year fixed loan is long enough for the trade to make sense. In most of continental Europe you will not find a line called points on a mortgage offer, and asking for one will get you a blank look. That does not mean the trade is unavailable — it means it is packaged differently, and the packaging changes the arithmetic.

Germany has the closest structural equivalent in the Disagio, also called Damnum: the contract is written for a nominal amount, the bank pays out less than that — say 95 percent — and in exchange the nominal rate is lower. The difference is a prepaid interest charge, exactly like a point, and it is amortised over the fixed-rate period rather than the whole loan, which shortens the horizon the trade has to survive. In France the analogous levers are the negotiated rate itself, the frais de dossier and, above all, the insurance premium, which since the Lemoine reform can be moved to another insurer at any time and often saves more than any rate haggling. In Spain and Portugal the standard mechanism is the bundled discount: the spread falls if you take the payroll account, the home insurance and the life cover with the same bank, which is a rate buydown paid monthly in kind rather than once in cash.

The second half of the divergence matters more than the first: what it costs to leave. In Italy the surroga created by the 2007 Bersani law lets a borrower move a mortgage to another bank at no cost, which makes any large upfront payment for a lower rate structurally unattractive — you are prepaying for a rate you can replace for free. France caps the early-repayment indemnity at the lesser of six months of interest on the capital repaid and 3 percent of the outstanding balance. Spain caps it too under the 2019 mortgage law, at levels that fall to zero after the early years. Germany allows a borrower to terminate a fixed-rate loan after ten years with six months notice and no penalty, but charges a Vorfälligkeitsentschädigung before then. In the United States, penalties on owner-occupied loans have become rare. Every one of those rules changes the probability that the loan survives to the break-even date, and therefore changes whether points are worth buying at all.

A short checklist before you pay a point

First, write down the horizon before you look at the break-even, not after. If your honest answer is under five years, the arithmetic is already decided and no rate sheet will change it. Second, check that the money has no better job: paying a point is a guaranteed return equal to the loan rate, which is excellent if the alternative is a savings account and poor if the alternative is clearing a credit-card balance or funding the deposit that would take you below the threshold for mortgage insurance. Third, ask for the whole price grid and compute the discounted break-even on every line, because it lengthens as you go down. Fourth, read the prepayment clause, since it decides how expensive it will be to end the loan before the point has paid for itself.

Frequently asked questions

Are mortgage points tax-deductible?

In the United States, points paid on a loan to buy or build a main home can often be deducted in full in the year they are paid, provided a list of conditions is met; points on a refinance normally have to be spread over the life of the loan instead. IRS Publication 936 sets out the tests. Outside the United States the question usually does not arise in the same form, because the product does not exist and because interest on a home you live in is generally not deductible at all in France, Germany, Spain, Portugal or Italy — deductibility there is mostly confined to rental property, where financing costs are a business expense. Never treat a tax break as part of the break-even until you have confirmed it applies to your own situation.

Is it better to pay a point or to put the same money into the deposit?

Compare like with like: both are cash spent at closing to reduce future cost. A point buys a rate cut on the whole balance; extra deposit reduces the balance itself and, if it crosses a loan-to-value threshold, can also cut the rate or remove mortgage insurance entirely. That threshold effect usually wins, because it is discontinuous — a few thousand that takes you from 81 percent to 79 percent loan-to-value can be worth several times what the same money buys as a point. Away from a threshold, the two are close, and the tiebreaker is again tenure: extra deposit keeps its value if you sell next year, a point does not.

What discount rate should I use for the discounted break-even?

Use the return you would genuinely have earned on the money instead, after tax. For most buyers that is the yield on a safe deposit or short-dated government paper, which is why 4 percent is a reasonable working assumption in a normal rate environment. Do not use the mortgage rate itself: that would assume the money's alternative use was repaying this very loan, which is a different comparison. The choice is not very sensitive in any case — on our example the break-even is 69 months at a 3 percent discount rate, 71 at 4 percent, 73 at 5 percent and 76 at 6 percent. Any of them is closer to the truth than the undiscounted 63.

Can I negotiate points, or are they fixed?

The grid comes from the lender's pricing engine and is not usually negotiable line by line, but which line you sit on very much is, and so is the choice of lender. Two banks pricing the same borrower on the same day will offer materially different amounts of rate per point, because they fund and hedge differently. The practical move is to ask each lender for the full sheet — rate at zero points, at half, at one, at two — and compare the discounted break-even at the same horizon rather than comparing headline rates. That comparison also exposes the offer that has quietly folded an origination fee into what it is calling a point.

Do points make sense on a variable-rate loan?

Much less, and often not at all. On a variable or periodically reset loan, what you are buying down is usually the initial rate or the margin, and the initial rate lasts only until the first reset. If the fixed period is five years and the discounted break-even is 71 months, the point expires before it has paid for itself. Where the buydown reduces the margin permanently the trade can still work, but you have to read the contract to know which of the two you are buying — the offer document will not volunteer the distinction, and the difference between them is the whole decision.