Nearly 3 in 10 Trade-Ins Are Underwater. Most Stores Find Out Too Late.

July looked like the best month of the year. The deal jackets tell a second story.

Cox Automotive's July forecast, published July 27, puts the seasonally adjusted annual selling rate near 16.7 million units - the strongest pace of 2026 and a step up from June's 16.5 million. Volume is forecast at 1.395 million, up 1.2% month over month and down 0.4% from last July, on 26 selling days. Charlie Chesbrough, Cox Automotive's senior economist, attributed the strength to a market that "is being driven by more affluent buyers," who may be less exposed to inflation and weak consumer confidence than the average household.

That attribution is the part worth reading twice. A record-for-the-year SAAR carried by affluent buyers is not the same thing as a healthy retail funnel, and the clearest evidence of the gap showed up eleven days earlier, in Edmunds' quarterly transaction data.

The myth: Negative equity has never been worse than it is right now. The data: The share of underwater trade-ins is not at a record. Edmunds put it at 29.6% in Q2 2026 - up from 26.6% a year earlier, but still below the 34.6% recorded in Q2 2019 and well below Q2 2020's 37.2%. What is at a record is the size of the hole. The average amount owed on an upside-down trade-in reached $6,884, a Q2 high, versus $5,317 in Q2 2019. The problem did not get more common than it was pre-pandemic. It got about 29% deeper.

What $6,884 does to a deal

Rolling that balance forward does not vanish into a longer term. It shows up in two places a customer feels immediately, and Edmunds quantified both for Q2 2026.

Both the $944 payment and the $16,270 interest figure are the highest Edmunds has on record. Jessica Caldwell, Edmunds' head of insights, described the mechanism plainly: buyers who financed at 2022's peak prices are returning with thousands in old debt, and "relying on longer loan terms as a coping mechanism to keep monthly payments down only causes total interest charges to be higher in the long run."

The credit market is accommodating that stretch rather than resisting it. Cox Automotive's Dealertrack Credit Availability Index reached 104.6 in June, its highest level in roughly a decade, with the overall approval rate at 73.8% and the share of loans longer than 72 months hitting an all-time high of 31.1% - up from 27.0% a year earlier. We covered that loosening in Auto Credit Just Hit a 10-Year High. The two datasets read together give you the shape of the 2026 deal: approvals are easier to get, and the structures required to get them are longer and thinner.

This is a financing problem, not a vehicle problem

The instinct is to assume negative equity clusters around fast-depreciating models. Edmunds' Q2 model-level analysis says otherwise. The largest average shortfalls sit on trucks and mainstream sedans with strong reputations for holding value.

Ivan Drury, Edmunds' director of insights, drew the conclusion for the industry: "When historically safe residual value bets are showing up underwater, it's clear this is a financing problem, not always a vehicle choice problem."

For a store, that has a practical consequence. You cannot screen for equity risk by looking at the vehicle on the appointment. A 2023 Tundra owner and a 2023 Camry owner can both walk in $7,000 to $9,000 upside down, and neither of them looks like a problem on the calendar.

Equity gets decided on the phone, long before it reaches the desk

Edmunds also reports that 46.2% of new vehicles purchased in Q2 involved a trade-in. Put the two figures together and the arithmetic is unremarkable but useful: roughly half your new-vehicle opportunities arrive with a payoff attached, and close to three in ten of those payoffs exceed the vehicle's value.

Now look at what a typical inbound sales call captures. Name, number, vehicle of interest, appointment time. Payoff amount, current payment, and months remaining are almost universally treated as desk questions - things the sales manager will figure out once the customer is in the building.

In 2022, when the average underwater trade was $4,487 and only 14.7% of trades were upside down, that sequencing cost you very little. In a quarter where the average is $6,884 and the resulting payment lands 21% above market, arriving blind is expensive in three specific ways:

• Desk time. A structure that needs $7,000 absorbed takes multiple lender submissions and multiple pencils. That time comes out of the same manager's day as your other appointments.
• Show-rate quality. An appointment that cannot be structured produces a "be-back" that rarely comes back. Your appointment count looks fine; your closing ratio does not.
• Experience. A customer who learns at the desk that their trade is $8,000 short - after being told on the phone that their vehicle is in demand - has been set up to be disappointed. That is a retention problem, not just a gross problem.

None of that is a new insight in isolation. What has changed is the magnitude, and the fact that the data is now specific enough to act on.

Where AI actually helps here - and where it does not

Most of the dealership AI conversation in 2026 has been about coverage: answering calls after hours, cutting hold times, following up on leads nobody had time for. Those gains are real and well documented, and we have written about the size of them in The Real Cost of Missed Calls. But coverage by itself does not touch the equity problem. An AI agent that books the same unqualified appointment faster has improved your response time and left your close rate where it was.

The part of AI that is relevant to negative equity is duller and more valuable: consistency of data capture. A human BDC agent asks the payoff question when they remember to, when the customer seems willing, and when the call is not already running long. A scripted agent asks it on every call, records the answer as a structured field, and writes it to the CRM before the appointment exists. That is not a smarter conversation. It is the same conversation, captured the same way, every time - which is the precondition for routing deals by equity rather than by vehicle.

The direction of the broader dealer-software market points the same way. When Cox Automotive closed its acquisition of Fullpath on June 1, Fullpath CEO Aharon Horwitz said that "deeper shopper, inventory, service, trade-in valuation and equity data" would progressively move into the platform's customer data layer so dealers could pursue "more informed next-best actions." That is a vendor's roadmap rather than an independent finding, and it should be read as one. But it is a useful signal about where the largest player in dealer software believes the next competitive axis sits: not in generating more conversations, but in knowing more before the conversation happens.

Two cautions belong here. First, a payoff amount and a monthly payment are financial data about a consumer, and they should be handled with the same care as anything else in that category - captured with a clear purpose, stored where your existing consent and retention rules apply, and never restated to third parties casually. Second, an automated agent should capture equity information, not interpret it. It should not quote a trade value, imply an approval, or suggest a payment. The compliance boundaries around automated outreach in auto retail are narrow enough already; we walked through them in TCPA and AI Calling.

Five changes worth making this week

1. Add three fields to the inbound sales call. Payoff amount, current monthly payment, months remaining. Ask for approximations - most customers know their payment even when they do not know their payoff.
2. Route by equity, not by vehicle. A known-underwater opportunity should reach a finance-capable person before it reaches the appointment book. That is a different queue, not a lower priority.
3. Pre-desk the hard ones. If a payoff is known and the gap looks material, run the structure before the customer arrives. The goal is to walk in with an answer rather than to discover the problem in front of them.
4. Work your own service lane first. The average underwater trade-in is 4.0 years old, which points squarely at 2022 purchases. A meaningful share of those buyers are already in your service drive with a repair order open. That is a warmer, cheaper equity conversation than any third-party lead, and it is the same argument we made in Grow Fixed Ops Revenue.
5. Change the scoreboard. Track the share of sales appointments that arrive with a known payoff on file. Appointment count tells you your phones are working. That percentage tells you whether the appointments are workable.

The read

The July sales pace is genuinely good news, and Cox's own economist is clear about who is producing it. For the customers outside that affluent cohort, the constraint in 2026 is not whether they can get approved - approvals are at a decade high - but whether the deal can be structured around a payoff that averages $6,884 and a payment that lands at $944.

That constraint is knowable early. It is a question, asked on a phone call, recorded in a field. The reason most stores answer it late is not that the technology is missing; it is that nobody made it part of the first conversation. Whether that first conversation is handled by a person or by software matters less than whether it reliably ends with a number in the CRM.

If you are looking at AI for your phones, that is a fair test to apply: not just whether it answers, but what it captures, and whether the answer shows up in the deal before the customer does. You can see how we approach it at Carbuki.

Sources

• Q2 New-Vehicle Purchases with Negative Equity Trade-Ins Hit Record Monthly Payments and Interest Costs - Edmunds, July 16, 2026. Link
• Cox Automotive Forecast: New-Vehicle Sales Pace Strengthens in July as Buyers Shrug Off Economic Headwinds - Cox Automotive, July 27, 2026. Link
• Credit Availability Index Hits 10-year High as Lenders, Consumers Take on More Risk - Cox Automotive (Dealertrack, June 2026 data). Link
• Loan approvals rise to 73%, while negative equity on trade-ins hits $6,884 - Car Dealership Guy News, July 18, 2026. Link
• Edmunds pinpoints top 20 trades with negative equity in Q2 - Auto Remarketing, July 16, 2026. Link
• Cox Automotive Completes Acquisition of Fullpath - Cox Automotive (vendor announcement), June 1, 2026. Link
• Cox Automotive Report: Dealer Sentiment Improves on Current Conditions in Q2, Outlook Weakens for Months Ahead - Cox Automotive, Q2 2026. Link