August's Sales Comp Is Broken. Here Are the Numbers Worth Reading Instead.

The one month when your year-over-year number means almost nothing

Every store in the country is closing August this week, and a lot of those recaps will open with the same slide: this month versus last August. That comparison is going to look ugly. It is also, this particular month, close to meaningless - and unusually, both of the industry's major forecasters said so out loud before the month even ended.

J.D. Power projects August 2026 new-vehicle retail sales at 1,142,700 units, down 6.9% year over year on a selling-day-adjusted basis and down 10.3% without that adjustment. Total sales land at roughly 1.35 million, down 4.8% adjusted. And yet the same forecast puts the total seasonally adjusted annual rate at 16.4 million units, just 200,000 below last year, with retail SAAR at 13.0 million. Cox Automotive, forecasting separately, landed within a rounding error: a 16.3 million SAAR holding steady for a second straight month, on total volume near 1.35 million.

Two independent forecasters, two different models, same conclusion. The rate of sale is flat. The comp is broken.

The myth: A double-digit year-over-year drop in August means demand is falling apart.

The data: J.D. Power calls this month's year-over-year volume comparisons "not especially useful" for judging the underlying health of new-vehicle demand - while the retail SAAR holds at 13.0 million and total SAAR at 16.4 million. Cox's Charlie Chesbrough told dealers to "stay focused during these volatile times." Two forecasters, one message: read the rate, not the comp.

Sources: J.D. Power / GlobalData, Aug. 21, 2026; Cox Automotive, via CBT News, Aug. 26, 2026.

Why the comp broke, in two mechanical pieces

Neither reason has anything to do with how many customers want a car this month.

The first is a pull-ahead. August 2025 was inflated by the announced expiration of the $7,500 federal EV tax credit on September 30, 2025, which dragged EV purchases forward into the month that is now the comparison base. You are measuring a normal month against a deadline-driven one.

The second is the calendar. Labor Day weekend - traditionally one of the largest new-vehicle selling weekends of the year - fell inside the August reporting period in 2025 and falls in September in 2026. An entire high-volume weekend moved out of the month. Cox attributes much of the year-over-year drop to exactly this kind of calendar shift, and J.D. Power quantifies the selling-day effect directly: the same retail number is a 6.9% decline adjusted for selling days and a 10.3% decline unadjusted.

That 3.4-point spread between two readings of the identical month is the whole problem in miniature. Depending on which lens your vendor, your factory rep, or your own DMS report happens to use, August looks meaningfully different - and none of those readings tell you whether your store executed well.

Source: J.D. Power / GlobalData U.S. Automotive Forecast, August 21, 2026.

What is actually true underneath the noise

Strip out the distortion and the August data is not a story about demand collapsing. It is a story about demand changing shape. Three shifts stand out, and all three are real.

Hybrids kept taking share, against tight supply. Hybrid share of retail sales is expected to reach 18.2%, up 4.8 percentage points year over year - roughly 35.5% growth. J.D. Power notes August results would have been stronger still if hybrid availability had not been unusually low. EV share, by contrast, softened to 7.2% following the loss of the federal credit. Your mix is moving toward the powertrain you have the least of.

Payments hit a record, even as rates eased. The average new-vehicle transaction price rose 2.0% to $45,563, and the average monthly finance payment climbed 3.7% to $812 - the highest ever recorded for the month of August - despite the average new-loan interest rate falling six basis points to 6.55%, its lowest August reading since 2022. A cheaper rate did not produce a cheaper payment.

Trade equity is the quiet culprit. J.D. Power attributes much of the payment increase to weaker trade-in equity: 28.8% of trade-ins carried negative equity in August, up 0.6 points from a year ago. Buyers who purchased at peak prices are now returning underwater. To manage the payment, 13.9% of loans now run 84 months or longer, up 2.1 points, and subprime penetration rose 2.0 points to 10.8%.

None of that is calendar noise. It is a durable change in who is shopping and what they need solved before they will sit down.

Which means the call has changed, not just the comp

Here is the operational point a GM can actually use. Each of those shifts lands, first, on the phone - before a salesperson ever gets a walk-around.

Figures: J.D. Power / GlobalData, August 2026.

The last line is the one worth sitting with. Total retail consumer spending on new vehicles is projected to fall 7.6% to $49.8 billion, a $4.1 billion decline. The pool of money in the market got smaller. In that environment, the deals do not go to whoever advertises hardest; they go to whoever answers, qualifies, and books the highest share of the traffic already coming at them. We have put numbers to that leak before in what missed calls really cost and why speed to lead decides so many deals.

One more signal, from an adjacent aisle

There is a related development worth flagging, with a caveat attached. On August 27, J.D. Power released its inaugural U.S. AI Insurance Experience Study, based on 8,352 customer evaluations fielded from June through July 2026. It found that 29% of auto and home insurance customers have used AI tools to research coverage, service accounts, or shop a policy. Among those who used AI to research, 37% changed their policy as a result; among those who used it to shop, 42% purchased.

That is insurance, not car buying, and it should not be read as an auto-retail statistic. But it is a directional signal from the closest adjacent purchase most consumers make: for a considered, high-dollar, comparison-heavy decision, roughly a third of customers are now routing their research through an AI assistant, and a meaningful share are acting on what it tells them. Shoppers who arrive having already narrowed their options elsewhere show up with sharper, more specific questions - and less patience for a call that goes to voicemail. We looked at the dealership version of that behavior in the agent era of car shopping.

Put something else on the board this month

If the year-over-year comp is structurally unreliable in August, the responsible move is not to argue about it. It is to grade the store on things the calendar cannot distort.

1. Answer rate and after-hours miss rate. Selling days moved; your phone coverage did not have to. This is the cleanest month-over-month number you own.
2. Appointments set and appointments kept, not calls handled. A call that was answered but never became an appointment is not a win - a distinction we unpacked in answered but not booked.
3. Leads logged to the CRM, as a share of leads received. If it is not in the CRM, it did not happen, and no market condition caused that.
4. Revenue per opportunity. With the spending pool down 7.6%, what you extract per contact matters more than raw count.
5. Sequential, not annual. Compare August to July, and to a rolling three-month average. The distortions in this comp are specific to last August, not to your process.

Two cautions belong here. First, treat forecasts as forecasts: these are projections published before the month closed, and the actuals will move. Second, if you are leaning on automation to cover more of the phone, the compliance floor does not shift with the calendar - outbound calling and texting still live under the TCPA and state rules, as we covered in our rundown on AI calling and compliance.

The measured takeaway

August 2026 is a month where the headline number and the underlying reality point in opposite directions, and both major forecasters said so unprompted. The rate of sale is essentially flat. The comp is distorted by a tax-credit deadline and a holiday weekend that moved. Meanwhile, the mix genuinely shifted toward hybrids, record payments, and underwater trades - which changes the first ninety seconds of nearly every inbound call your store takes.

The stores that handle this month well will be the ones that refuse to litigate a broken comparison and instead ask a narrower question: of the customers who tried to reach us in August, how many did we actually reach back?

If that question is harder to answer than it should be, it is usually a measurement problem before it is a staffing one. Carbuki builds AI voice agents for U.S. dealerships that answer every call, book the appointment, and log the lead - and we are glad to help you pressure-test the idea against your own numbers.

Sources

• J.D. Power / GlobalData, "New-Vehicle August Sales Rate on Track for 16.4 Million; Year-Ago Dynamics Skew Annual Comparisons," August 21, 2026: https://www.jdpower.com/business/press-releases/jd-power-globaldata-u-s-automotive-forecast-august-2026/
• CBT News, "Cox finds August new-vehicle sales remain resilient as economic pressures mount" (Cox Automotive August forecast; Charlie Chesbrough), August 26, 2026: https://www.cbtnews.com/cox-finds-august-new-vehicle-sales-resilient/
• J.D. Power, "AI Begins to Wield Significant Influence on Consumers' Auto and Home Insurance Decisions" (inaugural U.S. AI Insurance Experience Study; 8,352 evaluations fielded June-July 2026), August 27, 2026: https://www.jdpower.com/business/press-releases/u-s-ai-insurance-experience-study/
• Car Dealership Guy News, "August new vehicle sales to hit 1.35M," August 25, 2026: https://news.dealershipguy.com/p/august-new-vehicle-sales-to-hit-1-35m-2026-08-25
• Cox Automotive, "Cox Automotive Forecasts 2026 New-Vehicle Sales at 15.8 Million" (full-year outlook): https://www.coxautoinc.com/insights/cox-automotive-2026-outlook/