DGActual Market Brief — Week of August 24, 2026

This week's brief leads with a disagreement. The most widely circulated industry report on used EVs says that market is loosening but healthy, with inventory broadly aligned with demand. Our own index says the near-new EV cohort just posted its steepest weekly drop in fourteen weeks of tracking, and that the cars themselves are taking far longer to find a buyer than the industry's supply ratio suggests. Both are looking at the same market. Only one of them is measuring what actually happens on a lot. That story, plus a hawkish turn in the Fed minutes, Hyundai's first full strike in a decade, and a seventh straight week of used retail contraction, below.

Overview: the week the ratio and the cars disagreed

Used EVs are this week's main event, and the disagreement matters beyond EVs. The industry measures used supply with a ratio called days' supply: inventory divided by a trailing 30-day sales rate. We measure the cars themselves: asking prices and how many days each vehicle has actually spent on the market. When the market is stable, the two tell similar stories. When the market turns, they split, and the EV segment is where they are splitting right now.

Elsewhere, the model held its Q3 base at 16.07 million through two scheduled updates, the Fed minutes revealed several participants arguing for a hike, and the used retail contraction we have tracked all summer extended to a seventh consecutive week with the retail-wholesale spread setting another record. The acquisition desk found private-seller Camrys at more than three thousand dollars under dealer retail. A full week.

The used EV read: 46 days on paper, 81 days on the pavement

Industry data published this month put July used EV sales at 36,810 units, up 10.1% year over year, with days' supply at 46 and inventory described as broadly aligned with demand. Prices were reported up 8.3% from a year ago.

Our index tracks the same market with live retail listings, measured weekly, and it disagrees. The near-new EV segment, the one-to-three-year-old cars that lease returns feed, just posted its steepest weekly drop in fourteen weeks of tracking. It sits at 94.6 against a June baseline of 100, the weakest segment we track. Median asking price: $34,500 in mid-June, $32,656 now.

The days number is where the two views really split, because there are two ways to count. Their 46 days' supply divides today's inventory by last month's sales rate, so when demand decelerates, the denominator is always a month behind. We count the actual days a car has spent listed for sale. The median near-new EV on sale today has spent 81 total days on the market, including relistings and transfers between sellers, and about 28 days at its current store. That 81 has stretched by sixteen days since June. A market where inventory increasingly changes hands to find its buyer is a different market from one turning cleanly in 46 days, whatever the ratio says.

On the year-over-year gain: it is real, and it is also not the number that runs a store. Nobody buys inventory in the twelve-month market. Our job is keeping dealers who put real money on the table current on the now market, and the now market says near-new EV asking prices are down more than five percent since mid-June and fell again this week. Both facts are true. Only one of them prices Thursday's appraisal.

Under the surface, this market is not one market at all. Older, cheaper EVs sit slightly above their June baseline; the glut is specifically the $30,000-and-up near-new cohort where the off-lease wave lands. The 2023 Tesla Model 3 has barely moved in five weeks while the 2023 Mustang Mach-E gave up 4.2 percent, thirteen times the decline, in the same segment over the same weeks. And the turn speeds are just as split: a used Model Y finds its buyer in about 34 total days and a used Model 3 in 36, while a Mach-E takes 90, an Ioniq 5 over 100, at nearly identical money. Model Y and Mach-E both carry a median asking price around $32,300 right now. Same price, one-third the wait. Kia EV6 lots are turning the freshest inventory in the segment at 20 days, and the sub-$22,000 Bolt and Leaf are moving in under four weeks. The fast lane and the slow lane are parked next to each other on the same row.

One more split, new to our tracking as of this morning: private sellers. On Camrys and CR-Vs, private sellers undercut dealer retail by thousands. On near-new EVs the pattern inverts: private sellers are asking $34,900 against $31,924 dealer retail, $2,976 OVER the market, and their listings have been up a median 92 days against 28 for dealers. The likely mechanism is the record negative-equity share: a seller two years into an EV loan often cannot price below the payoff, so they anchor above the market and wait. Dealers marked down. Trapped sellers cannot. Every one of those overpriced driveway listings is a future trade-in conversation, and the store that makes the call first gets it.

So where we stand: we remain bullish on used EVs. The buyers are real, the affordable end is proving it, and the off-lease wave is a sourcing opportunity for stores equipped to work it. The success lives in the nuance. Year, model, battery option, hardware version, software generation, charging port. Two vehicles wearing the same nameplate and model year can be different products underneath, and the market prices them like it knows. EV buying strategy means knowing these things unit by unit. EV theory is just generalizing, and generalizing is how a store ends up owning the 90-day car at the 34-day price.

Sales and demand: the model holds while payments set records

The DGActual model ran its two scheduled updates this week and the Q3 base held at 16.07 million both times, with no input moving enough to shift the projection. And a correction from last week, because we hold our own numbers to the same standard we hold everyone else's. We attributed the model's lift from 15.81 to 16.07 to falling rates. This week we decomposed the model's own arithmetic, and that was wrong. The lift came from the sales pace itself: the Fed's confirmed July reading replaced a preliminary estimate that had been running about half a million light, and the model marked demand to the stronger confirmed number. Rates fell in the real world, but the model's rate input never moved. Demand held up better than the early data suggested, which is the same story with a different engine, and the engine matters for what comes next.

The outside numbers moved toward us again. J.D. Power's August forecast calls for a 16.4 million SAAR, with retail sales projected down 6.9% from a year ago. SAAR, for newer readers, is the seasonally adjusted annualized rate: the pace the whole market would sell at over a full year if this month's rhythm held. Our 16.07 against their 16.4 is a gap of 0.33 million, right at the edge of our flag threshold and well inside the model's uncertainty band.

The number inside that forecast that deserves the desk conversation: the average monthly payment is projected at a record $812, after July posted a record $808. Volume is holding because rates eased, but every month the payment record resets, the affordability wall gets a brick taller. Watch hybrids too: 18.2% retail share in the same forecast, another record, and consistent with where we see retail demand rotating in our own segment data.

Credit and the Fed: the minutes just put the rate relief at risk

The rate relief that lifted our model two weeks ago now has an asterisk on it. The minutes from the Fed's July meeting, released Wednesday, showed several participants favored a quarter-point hike at a meeting that ended in a 9 to 3 vote to hold at 3.50% to 3.75%, and many said tightening may be needed if inflation does not come down toward 2%.

For dealers, the chain runs through the payment. The 60-month auto rate falling from 7.53% to 7.14% is part of why buyers kept showing up this summer, and a Fed that resumes hiking pushes payments the other way at a moment when the record is already $808 and rising. Our model tracks the sales pace that results, not the minutes, so nothing moved this week. The first place a rate-driven wobble would show up is the August sales reading, due in early September. Until then, the risk to the demand story is on the table in the Fed's own words, but it is a risk, not a reading.

Still waiting: the Fed's Q2 bank delinquency data, which we check every week, has not posted as of Sunday night. The model continues to run on Q1's 2.64%. That release typically lands in late August and remains the next scheduled input that could move the forecast either direction.

OEM and production: Hyundai goes from partial to full strike

Hyundai's union staged its first full strike in ten years on Friday, roughly 40,000 members, then suspended Monday's planned action as talks resumed over the weekend. Korean reporting puts the cumulative disruption near 55,000 vehicles for 2026 across roughly 120 production-line hours. A suspension is not a settlement: for Hyundai stores, July's thin allocations still extend deeper into Q3, and the affected pipeline is the Korea-built portion of the lineup.

The Japan earthquake's production bill also got its first independent estimate: 46,300 light vehicles lost, per S&P Global Mobility's rapid impact analysis. Stellantis is running Mirafiori at reduced capacity through August 24 on component delays. And the Fed's industrial production data showed motor vehicle and parts output fell 2.1% in July, the aggregate confirmation of a month everyone on this list felt one allocation call at a time.

One quieter number for the ops file: dealership employment reached 2.0501 million in July, up 2,600 from June but down 6,100 over six months. Stores are holding headcount roughly flat into a softening retail tape, which is its own bet on the second half.

Trade and tariffs: the 50% is live, and the exit ramp is being negotiated

The 50% duty on covered Canadian-origin goods took effect Wednesday as scheduled. The same week, Reuters reported U.S. and Canadian negotiators are discussing cutting the automotive tariff to 15% from the current 25% track, which would be the first meaningful de-escalation of the year if it lands. Nothing is signed. Treat it as direction, not policy.

Separately, the metals side widened: more than 400 additional part categories became subject to the 50% metal-content duty last Tuesday, reaching deeper into the replacement-parts bin. Between this and July's parts-dealer inflation print, the parts counter is quietly becoming one of the most tariff-exposed corners of the store. If your parts matrix has not been reviewed since spring, that is this week's homework.

Safety desk: a million GM trucks under the microscope

The week's headline item is not a recall yet, and that distinction matters. NHTSA expanded an engineering analysis covering 997,743 GM pickups and SUVs with the 6.2-liter L87 V8, after nearly 7,000 complaints of engine failure including failures after the recall repair. An engineering analysis is the last step before a recall decision. If it converts, it would be one of the largest engine-related actions in years, and the service capacity math would be brutal: these are engine jobs, not software flashes.

Confirmed recalls this week: Stellantis recalled 955,000 vehicles for radio software that can disrupt the rearview camera display, and BMW recalled 27,720 vehicles for driveshaft wear that can cause loss of rear-wheel power or a rollaway, carrying do-not-drive and park-outside advisories with owner notices planned for October 2. The BMW campaign is small but hot: park-outside guidance means scared customers, and scared customers remember how the lane treated them.

DARPI: seven straight weeks down, and the spread sets another record

DARPI is the DGActual Automotive Retail Price Index: used retail asking prices across 11 segments, tracked weekly, indexed to 100 at our June 8 baseline, with the Manheim wholesale benchmark rebased to the same 100 so the two lines compare directly.

Week 14: retail fell 0.12 points to 96.84, the seventh consecutive weekly decline. Wholesale is still parked near 99.3. Retail now sits 3.2 points below baseline against wholesale at 0.7 below, and the spread widened to negative 2.47, a record for the fourth straight week. The story has not changed, only compounded: retail asking prices keep grinding lower under a wholesale floor that will not move, and every acquisition priced off stale comps absorbs that gap as margin.

Under the headline number, this was actually a split week: seven segments fell, four rose. EV led the decline at minus 1.23, nearly three times the next-largest drop. But fullsize car rose 1.32 and midsize car rose 0.99, the sedan segments that have quietly firmed while SUVs and trucks soften. Pair that with the record hybrid share in the demand section and a pattern is forming: the affordability buyer is rotating toward sedans and efficiency, and the price data is starting to reward the segments that serve them.

Acquisition desk: private sellers this week

Fresh pull from this morning, with a methods upgrade worth naming. Our earlier pulls compared a single model year on both sides; as of today the desk covers the full 2022 through 2025 populations for every nameplate, private sellers against dealer retail, identical years on both sides. The numbers below move more than usual this week because the lens got wider, not because the market whipsawed. One honesty note that applies every week: private-seller samples run a few dozen listings per model, so treat single-week moves as direction, not gospel.

Honda CR-V is the deepest signal on the corrected board: private sellers asking $3,638 under dealer retail. Toyota RAV4 follows at $3,111 under, with private listings actually fresher than dealer stock. Toyota Camry reads $2,277 under, Nissan Rogue joins at $1,190 under with private listings up a median 64 days, and Honda Accord, the deepest spread under the old narrower lens, shows $653 under across the full population.

Tesla Model Y tells this week's EV story in a single row. On price, private sellers have converged with dealer retail almost exactly: $28 apart. On time, private Model Y listings have been up a median 139 days against dealer listings at about 25. Sellers are finally asking the right number and still not finding buyers. That seller takes the call, and probably takes the second appointment too. Given where near-new EV prices are heading, the appraisal you write this week beats the one you write next month.

Still anchored, and further from reality under the wider lens: Ford F-150 private sellers are asking $5,724 over dealer retail, and Jeep Grand Cherokee sellers $3,554 over. Both are patience files, not pursuit files.

Full segment data and model-level breakdowns at data.dgactual.com. Dealer reputation across 17,814 stores at reviews.dgactual.com.

Sources: DGActual Retail Price Index, SAAR model, and FSBO acquisition pull (MarketCheck API). Industry EV market report figures as aggregated by ASOTU Daily, August 2026. J.D. Power and GlobalData U.S. Automotive Forecast, August 2026. NADA Market Beat, July 2026. Federal Reserve FOMC minutes, August 19, 2026. Federal Reserve G.17 Industrial Production, July 2026. FRED DRCLACBS. Reuters (Hyundai strike, U.S.-Canada tariff talks, NHTSA GM probe, Stellantis rearview camera recall). Seoul Economic Daily (Hyundai production impact). S&P Global Mobility (Japan earthquake impact). MarkLines (Stellantis Mirafiori). Bureau of Labor Statistics (dealership employment). NHTSA recall filing 26V525. RVIA tariff update (metal-content HTS codes). DGActual Market Intelligence. Week of August 24, 2026.