A New York wholesaler posted a help wanted ad. The only qualification that really mattered was a good credit score.
Whoever answers an ad like that walks into a store like yours with clean credit, a real license, and a name that checks out everywhere you look. They skip the test drive, take the add-ons without negotiating, and sign in their own name. Nothing your team does that day is wrong.
Then the first payment never comes. The car is already on a Turo listing or a container ship, and the lender is calling you about a deal that looked like the cleanest one on the board.
That posting was a straw borrower recruitment pitch, and pitches like it are landing better than they have in years. Point Predictive's Master Straw Borrower Index is up 37.6% since August 2023 and set an all-time high last December. Call it roughly $600M a year in lender losses, and the company expects it to stay that ugly through year-end.
It works because nothing on the application is fake. Real name, real Social Security number, real credit, real signature.
The only thing that isn't true is who ends up with the keys.
Point Predictive
What is a Straw Borrower?
Somebody uses their own identity and credit to finance a vehicle for someone else. That's the whole thing.
In organized rings, whoever signs gets $500 to $7,000 a car. Recruiters work ads like that one plus social media, and they are shopping for clean credit, not employees. Every straw who brings a friend collects a referral fee, because a friend is another credit file.
Your lender funds a clean deal, the car leaves with somebody who was never on the contract, and by the time payments stop there's usually nothing left to repossess. The loss runs close to the full balance, and the charge-off lands on the straw.
Point Predictive
Why is it On the Rise?
Four things are driving it:
Cars got more expensive while credit got worse, so borrowing somebody else's file started looking like a solution.
Straw-financed cars turned up in peer-to-peer rental fleets and online sublease operations, where rental income covers the note right up until it doesn't.
Export rings figured out they can ship a straw-financed luxury car overseas at up to 3x markup, and title washing handles the ones that stay home.
And recruiting moved to social media, where a broker can line up twenty people with 700 scores before lunch.
Point Predictive
Not every straw deal blows up. A parent signing for a kid who can't qualify usually pays to term.
The real damage sits at the profit end of the spectrum: sublease fleets, exporters, and bust-out crews who stack loans across several lenders inside a 30 to 45 day window, because that's roughly how long credit bureau tradelines take to catch up.
Every lender in the stack thinks it's the only one. Bust-out by itself runs an estimated $250M a year.
Point Predictive
How Does This Affect The Store?
Every one of those 90 cars was sold by a real dealership that thought it had a clean deal.
Straw deals come back to you as unwinds, chargebacks, buyback demands, and contracts in transit that suddenly aren't. And your lenders are already keeping score.
Point Predictive's risky-dealer pattern carries a 10x first payment default multiplier, the biggest number in the whole report, and DealerCheck tracks more than 68,000 dealer profiles against consortium benchmarks, flagging risky stores an average of 180 days before traditional monitoring catches up.
A ring can run four deals through your store in a month, disappear, and leave your rooftop attached to the losses in a database you'll never get to read.
The good news, and there is some, is that the strongest signals are things a lender can't see and your people can. None of what follows touches a credit file. All of it happens in front of your desk, your F&I office, or your delivery coordinator.
Point Predictive
What Your People Can Actually Catch
Somebody else is running the deal. The buyer defers on payment questions, skips the test drive, and lets a companion do the negotiating. Sometimes that person knows the stock number better than the buyer does. Get the buyer alone for five minutes and ask what they liked about the car.
The money comes from somewhere else. The down payment arrives from an account that isn't on the contract, or a companion hands over a cashier's check in a different name. Ask where the funds came from and listen to how long it takes to get an answer.
The credit tier and the deal don't match. A 780 score takes a packed subprime structure without a word of pushback, accepts every add-on, and finances near max advance. Prime buyers negotiate. That's the whole reason they built the score.
The credit file got dressed up recently. Authorized-user tradelines added in the last few weeks, or a score that jumped 80 points since the customer's last inquiry. Rings buy tradelines the same way they buy signatures.
The employer doesn't survive a phone call. The employer number rings to a personal cell, or the LLC was formed four months ago while the applicant claims six years of tenure. Compare the formation date against the stated tenure before you send the deal.
The car doesn't match the buyer. Point Predictive's example is Dodge Chargers financed by borrowers in their nineties, where early payment default hits 17.71 percent against 9.49 percent for borrowers 75 to 79. You know your traffic. When the vehicle is that far outside the applicant's history, look at the rest of the file.
They bought a long way from home, repeatedly. One out-of-state purchase has plenty of innocent explanations. Early payment default runs 4.03 percent across all funded loans and 20 percent once a borrower has more than five out-of-state deals.
The co-applicant vanished between attempts. A deal that came in joint somewhere else shows up at your store with one name on it. That alert's firing rate has more than doubled in two years, from a low of 1.5 percent to 4.4 percent, and it's riskiest when the solo version lands at a different dealer.
The contact information belongs to other deals. The same phone number, email, or address appears on other applications under other names. References that loop back to the co-applicant or to someone at your store are not references.
The paperwork points at a third party. The trade is titled to somebody who never walked in, or the insurance binder lists a different primary driver or a garaging address two hours up the interstate. Confirm the binder before you fund.
None of these is proof on its own, but two or three together on the same deal deserve a deeper dive.
Where AI Can Help
Build the deal review card. Feed the 21 red flags into whatever model you use, along with how your store actually delivers cars, and have it produce a one-page checklist your desk and F&I managers can run before a deal goes to funding. You want something that fits on a clipboard, not a policy binder nobody opens.
Run the cross-check nobody can do by eye. Export the last 90 days of deal jackets and have AI hunt for repeated phone numbers, emails, addresses, employers, and references showing up across different applicant names. Rings reuse infrastructure because it's easier. Three deals, three names, one phone number isn't a coincidence.
Use the phone system you're already paying for. If your BDC calls get transcribed, and they probably do, pull the same 90 days and look for deals where somebody other than the buyer negotiated terms or scheduled delivery. A name on three calls and zero contracts is worth asking about.
Audit what you already own before you buy anything else. Does your ID verification vendor do live phone-based presence checks at the point of sale? Is your lender sitting on consortium alerts nobody at your store looks at? Can your CRM flag a second application from the same household inside 30 days?
⭐️ One guardrail before you go build something:
The report shows first payment default climbing with every age band, and it puts the heaviest clusters of high-velocity shoppers in Houston, San Antonio, Las Vegas, and four Florida metros.
Age and geography are not decisioning criteria, and AI will happily help you build a fair lending problem if you ask it to.
Use these patterns to decide what gets a second look, never to decide who gets approved.
Point Predictive
3 Prompts to Run This Week
1️⃣ For the F&I manager to build the pre-funding review card:
Act as a dealership F&I compliance advisor.
I want a one-page straw borrower screening card my F&I office can run before a deal goes to funding.
My store:
[NEW / USED / BOTH]
[FRANCHISE OR INDEPENDENT, BRANDS]
[STATES WE REGULARLY DELIVER IN]
Our current delivery and funding process:
[PASTE THE STEPS FROM CREDIT APP TO CONTRACTS IN TRANSIT]
Known straw borrower red flags to work from:
[PASTE RED FLAG LIST]
Build me:
1. A checklist of 10 to 12 yes/no questions, written for a busy F&I manager, not a compliance department
2. For each item, what my manager should physically look at or verify
3. Three questions to ask the customer directly, alone, without the person who came with them
4. A clear escalation path: what gets a second look, what gets a phone call, what stops the deal
5. A short list of things that must never be used as a reason to decline, including age, national origin, and where the customer lives
Keep it to one page. Plain language. No legal boilerplate.
At the end, tell me the single item most likely to catch a straw deal in my process as it exists today.2️⃣ For the sales manager or office manager to scan 90 days of deals and calls":
You are helping me look for fraud patterns across recent deals at my dealership.
I am uploading our last 90 days of deal data. Columns include:
[PASTE OR LIST YOUR COLUMNS: applicant name, co-applicant, address, phone,
email, employer, references, lender, amount financed, vehicle, delivery date]
Find and rank:
1. Any phone number, email address, physical address, or employer that appears
on more than one deal under different applicant names
2. Reference phone numbers that match another applicant, a co-applicant, or
one of our own employees
3. Clusters of deals funded within a 30 to 45 day window that share any of
the above
4. Applicants who purchased out of state, and anyone with more than one
5. Deals where the same person appears as primary on one and co-applicant
on another
I am also uploading call transcripts from the same 90 days:
[PASTE OR UPLOAD BDC AND SALES CALL TRANSCRIPTS]
From the transcripts, flag any deal where:
- Someone other than the buyer negotiated price, payment, or terms
- Someone other than the buyer scheduled or took delivery
- The buyer never asked about the vehicle itself
- A third party answered questions the buyer could not
- The same voice or name appears across multiple unrelated deals
Then connect the two. Tell me which flagged deals appear in both the data
and the transcripts, because that overlap is where I should start.
Output a table: what you found, which deals, how confident you are, and the
one thing I should check next on each.
Rules:
- Flag for human review only. Do not tell me to unwind or report anything.
- Do not use age, race, national origin, or neighborhood as a risk factor.
- If a pattern has an innocent explanation, say so.One housekeeping note on this one. You're handing customer PII to whatever platform you use, so run it inside your enterprise account or the AI already built into your CRM, not a free consumer tier.
3️⃣ For the GM or controller to audit the tools you already pay for:
Act as a dealership technology advisor.
Before I buy another vendor product, I want to know what my current stack can
already do to catch straw borrower and identity fraud.
Tools we currently use:
[PASTE EXISTING TOOLS: CRM, DMS, desking, credit platform, ID verification,
phone system, e-contracting, insurance verification, Microsoft 365 or
Google Workspace]
The fraud checks I want covered:
- Repeated phone, email, address, or employer across different applicants
- Employer verification and employer formation date
- Live identity and presence verification at point of sale
- Title and prior owner check on trade-ins
- Duplicate or near-duplicate applications inside 30 days
- Call recordings that show a third party running the deal
For each check, tell me:
1. Which tool we already own that can do it, and how to turn it on
2. Which ones need configuration or a support ticket, and what to ask for
3. Which are genuine gaps
Do not recommend buying anything until you have ruled out the tools listed
above. If a gap is real, tell me the category to shop and three questions to
ask that vendor.
Finish with the first thing I should do Monday morning.Run the 90-day cross-check this week. If the same phone number shows up on three deals under three different names, you've already got your answer.
