Type “used car loan rates” into anything and you get percentages. Sometimes one, sometimes a grid of them sorted by credit band, always presented as though the percentage were the thing being bought. It is not. What you are buying is a schedule — a sum of money, repaid over a stated number of months, with interest charged along the way on whatever is still outstanding. The rate is one input into that schedule. It is not even reliably the most important one.

This page does the arithmetic that rate comparisons stop short of. It takes the two used-car rates that are actually published in the United States, runs them against the Federal Reserve’s own figure for what an average used-car buyer borrows, and shows what comes out the other end: the monthly payment, the total interest, where each payment goes, and the month at which the debt is finally half repaid. Those are the quantities that decide what a loan costs, and no document in a car transaction puts any of them in front of you.

The finding, stated up front so you can stop reading if it is all you needed: on a realistic used-car loan, moving the term from four years to six costs almost twice what the entire published gap between the cheapest and dearest lender channel costs. The rate is worth negotiating. The term is worth more, and it is the one that gets moved for you.

A rate is one of three numbers

Three inputs produce a car loan, and every other quantity in the transaction falls out of them.

The amount financed. Not the price of the car. The amount financed is whatever the lender advances, which routinely includes sales tax, title and registration, a documentation charge, any add-on products bought in the finance office, and any shortfall carried across from a loan on the car you traded. Every one of those raises the sum that interest is charged on.

The rate. The annual charge applied to whatever is outstanding, month by month. It is the input the whole industry advertises on and the one this page will spend the least time on, because it is the one already best covered.

The term. The number of months over which the schedule runs. It is presented as a convenience — how much do you want the payment to be? — and it is priced, negotiated and sold like one, which disguises the fact that it changes the size of the debt rather than merely its shape.

The three do not have equal weight, and the ordering surprises people. A percentage point sounds large because it is a percentage; a couple of extra years sound small because they are just time. In money, on an ordinary used-car balance, it runs the other way round.

Exactly one federal publication prints a used-car rate at all, and it prints two of them side by side. The National Credit Union Administration compares credit unions with banks each quarter — on figures it extracts from a commercial survey rather than collects itself, which is worth knowing before leaning on them — and its most recent table — for the last Friday of the quarter, 26 December 2025, retrieved 7 September 2026 — puts the average 48-month used-car rate at 5.53% at credit unions and 7.73% at banks. That is a spread of 2.20 points, and it is the full observed distance between the two channels the table covers. Where those figures come from and what they can and cannot be held against is the subject of our page on how a used-car finance rate is arrived at; here they are simply the two ends of a published range, used as raw material.

Now put both of them against the Federal Reserve’s figure for what a used-car borrower actually borrows. The G.19 consumer credit release records the average amount financed on used-car loans at $24,897.75 in the March 2026 reading. Over 48 months at 5.53% that is a payment of $579 and total interest of $2,912. At 7.73% it is $605 and $4,127. The entire published channel spread, all 2.20 points of it, comes to $25 a month and $1,214 across the whole loan.

Hold that $1,214 in mind. Now change nothing except the term. Keep the 7.73% rate exactly where it is and run the same $24,897.75 over 72 months instead of 48. The payment falls from $605 to $433 — a saving of $171 a month, which is the number the finance office will show you. Total interest rises from $4,127 to $6,297. The extra cost is $2,170, or nearly twice what the whole rate spread was worth.

That is the page in one comparison. A buyer who fights for two full percentage points and then accepts two extra years has lost more than they won, and both moves felt like victories at the time.

The term is the lever nobody prices

It is worth setting the whole ladder out, because the pattern is more instructive than any single comparison. The table below runs the same $24,897.75 at the same 7.73% across four terms. Only the number of months changes.

One thing needs saying before you read it. The NCUA publishes a used-car rate at 36 and at 48 months and at no term beyond that, so the 60- and 72-month rows here hold the 48-month rate still rather than pretending to know what a longer used-car loan is priced at. That is deliberate. Holding the rate constant is the only way to see what the term alone does, and in the real market the two move together in ways that are covered further down.

Read the second column downwards and the case for the long term is obvious: the payment falls by $344 between the shortest and longest schedules, which on most household budgets is the difference between comfortable and not. Read the third column downwards and total interest goes from $3,078 to $6,297. The same loan, at the same rate, for the same car, costs $3,219 more because of a choice that was presented as a matter of preference.

Nothing improper is happening. A lender charging an annual rate on an outstanding balance, over a schedule where the balance falls more slowly, earns more money. That is not a penalty and it is not a trap; it is what the words mean. What is worth naming is the asymmetry in what the buyer can see. The payment is printed in large type on the advertisement, on the desk pad and on the contract. The total interest exists in none of those places unless somebody works it out.

There is a related trap in how a longer term gets sold. Because the monthly figure falls, a longer term makes a more expensive car affordable at the same payment, which is very often how it is used — not to make the same car cheaper but to make a dearer car reachable. Our page on what a car should be allowed to cost you works through why a payment is the wrong unit for that decision. The point that belongs here is narrower: if the term stretches and the amount financed rises at the same time, both of the two big levers have moved in the expensive direction at once, and the payment on the contract will not show it.

All of that is one borrower, one balance, one rate. Yours differs in every one of the three, and the trade above can come out differently once real numbers are in it — so the version that decides anything is the one run on the two offers actually in front of you.

The two totals are the comparison; the payments are not. Where the offer with the lighter payment also carries the larger total, that is the trade this page is about, priced in your money rather than an average buyer’s. Where the totals land close together despite a visible gap in the percentages, the schedules have quietly cancelled most of that gap out, and the rate argument was worth less than it felt.

Why the balance falls more slowly than the calendar

The totals are only half of what the schedule does. The other half is the order the money is charged in, and it is the part that produces most of the unpleasant surprises in used-car ownership.

Interest accrues on what is outstanding. At the start of a loan almost the whole sum is outstanding, so almost the whole month’s interest is charged on almost the whole sum; the part of your payment left over to reduce the debt is whatever survives that. As the balance falls the monthly interest falls with it, and the principal share of a fixed payment grows. The payment is level. What it buys is not.

On the $24,897.75 loan at 7.73% over four years, the first payment is $605 and $160 of it is interest — 26.52% of the money you sent. Stretch the same loan to six years and the payment drops to $433, but the interest in the first payment is still $160, because the balance interest is charged on has not changed. That single fact is the whole mechanism: 37.02% of the smaller payment is interest, so the amount left to reduce the debt is much smaller, and the debt therefore takes much longer to move.

Follow it to its conclusion and the numbers stop being abstract. On the four-year schedule the balance falls below half the sum borrowed in month 26. On the six-year schedule it does not happen until month 41 — you spend forty months of a seventy-two-month loan still owing more than half of what you started with. And after exactly two years of payments, twenty-four of them, the four-year borrower owes $13,406 while the six-year borrower owes $17,840 against the same original $24,897.75.

Now put a used car next to that. A vehicle loses value fastest early and the loss does not pause while your balance is barely moving, which is why so many loans spend a stretch in the middle underwater. The mechanism is not mysterious and it is not the borrower’s fault — it is these two curves, one falling steeply and one falling slowly, crossing later than anybody expects. Our page on negative equity on a car loan covers what to do when you are in that stretch; what this section adds is where the stretch comes from and how the term decides its length.

Two practical consequences follow. A payoff figure obtained partway through a loan is almost always higher than people guess, because they mentally divide the balance by the calendar and the schedule does not work that way. And any decision that lengthens the term — a longer loan at the outset, a shortfall rolled into a new one, a refinance that adds years to buy a smaller payment — pushes the crossing point further out. It does not shorten the difficult period. It extends it.

A published rate has a shelf life

Every figure quoted so far carries a date, and the dates are not decoration. Rates move, and they move by more than the amount most people treat as normal variation.

The NCUA has published a used-car line in its quarterly comparison for forty-four quarters, from 27 March 2015 through 26 December 2025. Across that run the 48-month credit-union average has been as low as 2.76%, on 25 September 2015, and as high as 6.46%, on 29 March 2024. The bank average has run from 5.03%, on 25 March 2022, to 7.84%, on 27 December 2024. Both readings are for the same product from the same publication; the only thing that changed was when you asked.

Set that range against the arithmetic above and the practical implication is uncomfortable. The whole channel gap on the most recent reading is 2.20 points. The credit-union average has moved by more than three and a half points within one decade. Which is to say: the difference between two lenders on one afternoon is smaller than the difference the same lender makes between one year and another, and no amount of shopping will get you last year’s price.

One structural feature of the published data is durable enough to be worth carrying, though. In every one of the forty-four quarters, the longer of the two published terms carried the higher average rate — used at 48 months above used at 36, new at 60 above new at 48, in the credit-union column and the bank column alike. That is the ordinary shape: more time, more risk, a higher price. It is worth knowing because the Federal Reserve’s own bank series appear to run the other way, with shorter terms quoted above longer ones, and a reader who meets that ordering first can conclude that a long loan is being offered as a favour. It is not. In the one publication that holds the product constant, longer is dearer every single time.

Two loans at the same rate can still accrue differently

The schedule assumed above is the ordinary one: interest calculated on the outstanding balance, month by month, so the balance and the interest fall together. Most modern car loans work that way. Not all of them do, and the difference only becomes visible if something changes.

The reason this matters on a page about rates is that two contracts can carry an identical percentage and behave differently the moment your life does something unscheduled — a bonus you want to put against the car, a month you pay late, a decision to settle early because you found a buyer. On the balance-interest version, every one of those changes what you pay. On a precomputed contract, the finance charge was settled before you drove away and your options are whatever the document says they are.

Two smaller mechanics belong in the same paragraph, because both are ways a stated rate turns into more money than expected. The first is the deferred first payment: an offer to skip a month at the start is not free time, it is a longer first accrual period on the full balance, and the interest is charged. The second is the payoff quote. On a balance-interest loan interest keeps accruing between statements, so the figure to settle the account today is the balance plus whatever has accrued since the last payment landed — which is why a payoff quote has an expiry on it and why the number is not the one printed on your last statement.

The rate and the APR are not the same number

A quote contains more than one percentage and they are not interchangeable, which matters here because everything above has been arithmetic on a single rate.

The interest rate on the note is the charge applied to the balance. The annual percentage rate is broader: under Regulation Z it is built from the finance charge, which is the dollar cost of the credit itself. A fee you incur only because you are borrowing belongs inside that figure and pushes the APR above the note rate. Anything you would have paid in a cash purchase — the price of the car, most obviously — sits outside it.

The practical rule is short. When you set two offers beside each other, compare APR against APR, because two loans at the same note rate with different fee structures are not the same loan and only the APR shows it. When you feed a calculator, feed it the APR and the amount financed rather than the note rate and the sticker price, or the total it returns will be smaller than the one you actually sign for.

What moves the number, in order of size

Pulling the mechanics together, here is what is actually worth your attention, roughly ordered by how much money each is worth on a typical used-car balance.

Two of those are worth acting on before you are anywhere near a finance office. Getting an underwritten offer of your own first converts the rate from something you are told into something you hold, and our page on car loan pre-approval covers what to apply for and what the offer is conditional on. And deciding the term in advance — writing down the number of months you are prepared to accept before anybody asks about the monthly figure — removes the single largest lever from a conversation in which you are the only person without a calculator.

Working it in order

The sequence that makes the arithmetic work for you rather than around you is not complicated, and it is mostly about doing things in the right order.

Settle the price of the car first, in full, including everything required to take it away. Until the amount financed is fixed, no rate comparison means anything, because the same percentage on two different balances is two different loans.

Decide the term yourself, on paper, before the subject comes up. This is the step people skip, and skipping it is expensive in a way nothing later can recover.

Get one quote from a lender with no interest in the sale. That is the only number in the process that measures you rather than measuring you plus somebody’s commercial decision, and the published spread between channels is large enough to make the phone call obviously worth making.

Then compare on totals rather than payments. A payment can be produced to order by moving the term; a total cannot be moved without moving the money. Run both offers to their end and read the two totals side by side, which is what the calculator above is for.

And when the loan is running, revisit it rather than treating it as fixed. A rate is a moment in a market, the market moves, and the same borrower at the same institution in a different year is a different price. Whether that is worth acting on depends on where you are in the schedule and what the car is worth, both of which the sections above should now let you work out.

Common questions

What is the average used car loan rate?

There are two published averages and they are far apart. The NCUA’s quarterly comparison, for the last Friday of the quarter on 26 December 2025, puts the 48-month used-car average at 5.53% at credit unions and 7.73% at banks. The Federal Reserve’s own auto series carry no used-car line at all — every rate in the G.19 release is a new-car rate at commercial banks — so any page attributing a used-car average to the Federal Reserve has relabelled something. Treat both published figures as scale rather than as a quote: they are national averages of posted rates for a standard product, and your offer is priced against your file, the car, the term and how much is being advanced.

Does a longer car loan really cost more if the payment is lower?

Yes, and by more than most people expect. On the $24,897.75 the Federal Reserve records as the average amount financed on a used car in its March 2026 reading, held at 7.73%, four years costs $4,127 in interest and six years costs $6,297. The payment falls from $605 to $433 and the total rises by $2,170. Both of the numbers a shopper can see — the monthly payment, and often the advertised rate — improve as the term stretches. The one that gets worse is printed on nothing.

How much is the difference between rates actually worth?

It depends entirely on the balance and the term, which is why no page can answer it for you in the abstract. On the worked example — $24,897.75 over four years — the full 2.20-point distance between the two published channel averages comes to $25 a month and $1,214 over the life of the loan. On a smaller balance it is less; on a longer schedule it is more. It is worth having, and it is smaller than what the term is worth, which is the thing most rate comparisons never get round to saying.

Why is so much of my early payment going to interest?

Because interest is charged on what is still outstanding, and at the start almost everything is. On the four-year schedule above, $160 of the first $605 payment is interest, or 26.52%. On the six-year version the payment falls to $433 but the first month’s interest is still $160, so 37.02% of it is interest and much less is left to reduce the debt. Nothing is being taken from you unfairly; it is the definition of charging a rate on a balance. It does mean the balance moves slowly early on, which is why the six-year loan does not fall below half of what was borrowed until month 41.

Why do I still owe so much two years in?

Because two years of a six-year schedule is not a third of the debt. On the worked example, after twenty-four payments the six-year borrower still owes $17,840 of the original $24,897.75, while the four-year borrower is down to $13,406. Add the fact that a used car keeps losing value throughout, and you have the ordinary mechanism behind being upside down — two curves crossing later than intuition suggests, rather than anything having gone wrong.

Should I take the lower rate or the shorter term?

That is a question about your budget rather than about the arithmetic, and the arithmetic will not answer it. What the arithmetic will do is price the choice: run both offers to their totals and the trade stops being abstract. What is worth avoiding is the version where the decision is never made explicitly — where the term drifts longer during a conversation about the monthly figure, and the total is discovered years later.

Do these published rates apply to a private-party purchase?

Not directly. The published averages describe a standard used-car loan product; a purchase from a private seller is a different transaction for the lender, with the vehicle inspection, lien payoff and title work all needing to be arranged rather than assumed, and it is offered by a narrower set of institutions. The mechanics on this page — how the term moves the total, how interest front-loads — are the same whatever the channel, because they are properties of amortisation rather than of the seller.

How stale are the figures on this page?

The rates are for 26 December 2025 and were retrieved on 7 September 2026; the average amount financed is the March 2026 reading of the Federal Reserve’s series. Rates are republished quarterly and the last decade of readings has spanned 2.76% to 6.46% at credit unions and 5.03% to 7.84% at banks, so a figure here can be materially out of date within a year. The arithmetic does not go stale. Substitute today’s numbers into the same schedule and every conclusion above still follows.