The Best Time to Buy a Car (What the Data Actually Says)

The holiday sale event is the worst time to buy

This is the strongest and least intuitive finding in the data, so it goes first.

iSeeCars analysed more than 40 million used-car sales from 2024 through October 2025, counting how often a listing was a genuine deal — priced at least 10% below its estimated market value. On an average day, 13.7% of listings qualify. They then measured how each holiday and month compared to that baseline.

The heavily advertised summer sale events all came in below average:

• July 4th: 22.4% fewer deals than an average day
• Father's Day: 19.7% fewer
• Juneteenth: 19.7% fewer
• Memorial Day: 18.0% fewer
• Mother's Day: 9.2% fewer
• Labor Day: 5.0% fewer

The holidays that actually deliver are the cold ones nobody markets: Martin Luther King Jr. Day (65.5% more deals), New Year's Eve and Day (58.6% more), Presidents' Day (47.0% more), and Veterans Day (23.1% more).

This is not a one-off result. iSeeCars ran a comparable study a decade earlier on 2013–2016 data and reached the same conclusion about summer holidays. Two independent samples, ten years apart, agreeing on a counterintuitive result is about as good as consumer evidence gets.

The explanation is simple: a "sale event" is a marketing calendar, not a pricing decision. Demand is high in warm months, so dealers do not need to discount. The advertising exists because people shop then, not because prices fall.

The best and worst months

From the same study, by month, against that 13.7% baseline:

• January: 55.6% more deals — the clear best
• February: 36.2% more
• December: 15.2% more
• March: 6.3% more
• November: 5.8% more
• October: 2.1% more

And the months to avoid:

• June: 22.8% fewer deals — the worst month of the year
• May: 22.2% fewer
• July: 22.1% fewer
• April: 15.0% fewer
• August: 7.8% fewer

An iSeeCars analyst summarised the pattern neatly: used-car prices follow the average temperature, with both falling between November and March.

Note that these figures describe how often good deals appear, not how much you save. A month with 55% more deals is a month where searching is more likely to turn one up — it does not mean the same car is 55% cheaper.

One caveat on December, since it is the month folklore fixates on. For new cars it is genuinely the strongest month, but the effect is modest rather than dramatic: J.D. Power and GlobalData put late-December 2025 incentives at $3,433 per vehicle, about 6.5% of MSRP. Real, worth having, not transformative. For used cars, January and February now beat it clearly.

End of month, quarter, and year

Here the mechanism is real but the numbers you will read are not.

What is documented: manufacturers run stair-step volume bonus programmes. These pay a dealer a per-unit bonus that increases at volume thresholds and often applies retroactively to every unit sold that period — hit 20 cars instead of 19 and the bonus can jump on all twenty. That structure genuinely can make a dealer willing to sell the last car of a period at or below cost. Quarter-end and year-end bonuses are generally larger than month-end ones.

What is not documented: how much you actually save. There is no credible public measurement.

Two figures circulate constantly and both appear to be fabricated. One claims prices drop 5–10% in the final days of the month "per Edmunds and TrueCar" — neither company publishes anything of the sort. Another claims end-of-month transactions run 4–8% lower out-the-door "controlling for trim, region and incentives," describing a controlled study that does not exist. The false precision is the tell.

It is also worth hearing the skeptical case. Lyle Romer, a two-decade industry veteran, has argued that dealers want to move stale inventory on any day of any month, and that we measure in months simply because months are a convenient unit — not because a genuine end-of-month crisis occurs. That is opinion rather than data, but it comes from inside the business.

Practical read: shopping at period-end costs you nothing and occasionally helps. Planning your whole purchase around it, or accepting a worse car because it is the 30th, is not supported by anything measurable.

Model-year changeover

New model years generally arrive between July and November, concentrated in the August to October window, staggered model by model. We could not find an authoritative source for this timing — it is industry common knowledge rather than published fact, so treat it as a rough guide.

What is measured: iSeeCars examined 2.6 million listings in late November 2025 and found leftover prior-model-year inventory averaging 21.3% for the outgoing year. Some models were dramatically higher — the BMW i4 at 89.2%, Lexus GX 550 at 87.8%, Subaru BRZ at 87.1%.

The important caveat: that study measures inventory share, not discounts. iSeeCars' own analyst framed high leftover share as a sign dealers "may be more willing to negotiate" — explicitly an inference. Claims that leftover models save a specific $2,000 to $8,000 are unsourced.

The sound version of this advice: an outgoing model year that carries into the new year substantially unchanged is the best value on the lot, because you take the model-year depreciation hit on a car that is mechanically current. If the model was redesigned, the calculation is different — the old one will feel dated for its whole life.

Where the market actually is, mid-2026

Timing advice is worthless without knowing what market you are timing. As of the June 2026 data published by Cox Automotive:

• Average new-vehicle transaction price: $49,758, up just 0.6% year over year — essentially flat.
• Average MSRP: $51,654, actually down 0.9% year over year.
• Incentives: 7.0% of transaction price, where they have sat for 13 consecutive months.
• New inventory: 2.82 million units, 80.3 days' supply — above the roughly 60-day balanced norm.
• Used inventory: tighter, at 47 days' supply, with wholesale values still rising.

The single most useful comparison: incentives at 7% today versus 10.9% in December 2019. Automakers are holding pricing discipline. Waiting for a return to pre-pandemic discounting has been a losing strategy for several years running.

Our read — and this is interpretation rather than a sourced verdict — is that the new market is balanced to slightly buyer-favourable, while the used market is tightening. The real leverage is not when but what: supply is wildly uneven by brand. Toyota, Lexus, Honda and Subaru inventory is tight, so there is little room to negotiate. Stellantis brands are heavily oversupplied, with Jeep running about twice the industry average days' supply. That gap is worth far more than any calendar trick.

Tariffs, and one widely misreported ruling

Section 232 tariffs of 25% on imported vehicles and certain parts remain in effect as of this writing.

In February 2026 the Supreme Court struck down a separate set of tariffs issued under the IEEPA. This was widely reported as relief for car buyers. It was not — the automotive tariffs rest on Section 232 and were untouched by that decision.

Cox Automotive's assessment after the first year put the industry cost at roughly $30 billion, with average MSRP up 10.4%. The detail that matters most to a buyer is the split: consumers paid about 5.9% more on average, while dealers and automakers absorbed roughly 4.5 points of the increase. That absorption is why sticker prices rose while average transaction prices stayed nearly flat.

This is the fastest-moving section of this guide. Trade policy can change between our writing and your reading — verify current status before making a decision that depends on it.

Day of the week

Skip it. The figures in circulation trace to analyses from roughly a decade ago that have never been refreshed, and two commonly-cited versions contradict each other outright — one naming Monday as best, another naming Sunday. Even taken at face value, the spread was under one percentage point, roughly $300 on a $50,000 car, and is confounded by who happens to shop when.

Edmunds' own position is the honest one: early in the week is better because you get more of the salesperson's attention, not because the price is lower. Attention is worth something. A measurable discount is not on offer.

What actually to do

1. Buy in winter if you can wait. January and February are meaningfully better for used cars, and the effect is large enough to matter.
2. Ignore the summer sale events. They are demonstrably worse than an ordinary day.
3. Pick an over-supplied brand over a clever date. Days' supply varies more between brands than the calendar varies across the year.
4. Shop at period-end if it is convenient. Free option, unquantified benefit, no reason to organise your life around it.
5. Do not wait for a discount cycle that is not coming. Incentives have been flat near 7% for over a year, well below pre-pandemic norms.

And the timing factor that outweighs all of the above: your own financial readiness. Edmunds put the average new-car payment at $777 a month in Q2 2026, with 23.9% of loans now running 84 months or longer — both records. Buying in the best month on a loan you cannot comfortably carry is worse than buying in the worst month on one you can.