Pagaya Technology’s lenders are leaning on the expansion of their dealership networks for growth as credit quality worsens and credit access remains mixed. 

Credit access climbed 3.6% year over year and 0.8% month over month in June according to the DealerTrack Credit Availability Index published July 10. This marked the second consecutive month of credit access expansion following a dip in April as consumers rushed to purchase vehicles ahead of expected tariff-induced price hikes.  

The index ended the month at 97.3. 

However, June’s expansion follows mixed reports of credit access as many consumers entered the market with FICO scores up to 100 points lower following resumption of student loan delinquency reporting in the first half of the year. 

These market trends prompted lenders to look for ways to grow without loosening credit standards, Sanjiv Das, president at Pagaya, told Auto Finance News. Pagaya purchases loans that meet its underwriting criteria from lenders and securitizes the loans to fund further originations. 

“Our lenders are gradually starting to lend more,” he said. “They are spending a lot more of their efforts on dealers and building their networks, as opposed to expanding their credit box.” 

Pagaya sees volume growth 

Pagaya reported a 50% quarter-over-quarter increase in the first quarter in its auto annualized run rate, which surpassed $1.1 billion, according to a May 7 letter to shareholders. However, auto volume decreased 7% YoY.  

“Last year was a relatively tough year for the entire auto industry,” Das said. “When you’re in the public markets as a public company, you’re always being pushed for growth, until one day that growth story cracks and falls on the other side.” 

Slowed growth pushed Pagaya to focus on maintaining consistent yield for investors, he said. Simultaneously, its lenders reduced volume through the end of 2024, when consumers appeared to be in better shape.  

Consumers seem to be in good shape “through the middle of 2025, and so we have significantly opened up our pipes into our lenders,” Das said, noting that Pagaya increased its volume with lenders because investor appetite has strengthened. 

Tune in to Weekly Wrap to hear Das’ conversation with Auto Finance News Senior Associate Editor Truth Headlam.  

Subscribe to “The Roadmap Podcast” on iTunes or Spotify or download the episode. 

Auto Finance Summit, the premier industry event for auto lending and leasing, returns Oct. 15-17 at the Bellagio Las Vegas. To learn more about the 2025 event and register, visit www.AutoFinance.live/AFS. 

Editor’s note: This transcript has been generated by software and is being presented as is. Some transcription errors may remain. 

Truth Headlam 07:00:51
Foreign Hi everyone, and welcome to The Roadmap. From Auto Finance News since 1996 the nation’s leading newsletter on automotive lending and leasing. It is Monday, July 28 I’m Truth Truth Headlam 07:01:17
Headlam. Today is a special episode of the weekly rap joining me to discuss trends and the auto finance industry is Sanjv, das president of Pagaya. Sanjiv, thank you so much for joining me today before we dive in. Can you tell us a bit about yourself and your organization?Sanjiv Das 07:01:35
Sure. Hi, truth. How are you? Thank you for having me on the on the call today, super quick. I am, as truth said, the President of here, of Pagaya Technologies. Pagaya is one of the leading fintechs in artificial intelligence. We I we lend, we provide approvals and loans on behalf of our lending partners in personal loans, auto loans and point of sale loans using artificial intelligence. And as I said, we supplement and augment the growth of our partners by approving consumers that would have been traditionally left behind by mainstream, traditional lenders. And then we very effectively securitize these loans as well, almost simultaneously, same time that we approve the loans, so that it’s very capital efficient for our lending partners. We lend roughly eight to $10 billion a year, and we get almost a trillion dollars worth of loan flow that we get to see every year. And we have about 31 partners on our platform, including some of the major names like us, bank, Clarno, Ally, Westlake, stellantis, which I’m sure a lot of your listeners are familiar with. Truth Headlam 07:02:51 Thank you. So my first question, I would love to hear your thoughts on how the use of AI by auto finance series has evolved post pandemic and more specifically, during the past year. Sanjiv Das 07:03:05 Sure truth. So typically, people have been somewhat cautious about the use of AI for a number of reasons. While everybody understands the power of AI regulation, I’ve been a former banker myself, so I understand the power of regulation really well, but regulation has also basically limited the full use of AI in finance, as it has, you know, as you know, in other industries, AI has done a lot more and been very effective. So people know about the power, the positive power, of AI, but I would say that AI has traditionally been used in, I’d say Fraud Management, in compliance management, and to a limited degree, in customer service and managing sort of customer calls coming in on the back end. So it’s traditionally been used in areas where there hasn’t been a lot of underwriting decisioning since during covid, a lot of it, you know, was doubled down in the areas I just talked about, so more Fraud Management, more customer service management and more compliance management. But since, since covid, I would say that it has evolved a lot more in its acceptance towards credit underwriting and its acceptance towards what I call offer personalization. In other words, every customer gets a set of options that are much more personalized than manual or human underwriting, in the case of underwriting decisions where the explanatory power is most important, and by the way, the explanatory power is more important when a loan is rejected, as opposed to when a loan is approved, because you want to show to the regulators, and rightly so, that the loan has been approved for all the right credit reasons, as opposed to any discriminatory reasons. That has improved. The adoption of that has improved a lot more. And I would say that Pagaya, which is the company I represent, is one of the leading lights in demonstrating how to use AI to not only approve more customers the right way, but also make sure that customers that were left behind by mainstream lenders are now part of the mainstream, as opposed to them having to go either without getting financing for a car, which is such an important part of their life. Lives, or having to go to lenders who charge very extreme rates to buy has been able to use AI very effectively in a regulatory compliant way, along with the lenders, on behalf of the lenders, to get more consumers into mainstream. Yeah, Truth Headlam 07:06:10 and I might have you repeat some things you just said here. But my next question, you know, I’m curious, how has AI adoption and lenders underwriting processes changed, specifically, like, are you seeing more acceptance and use of AI in terms of lending to more borrowers and capturing a clearer picture of customers credit worthiness and debt levels? Sanjiv Das 07:06:32 Yes. Yes. So I would like to make a distinction most traditional lenders themselves have not fully adopted AI or even to the extent that one would expect, because most traditional lenders tend to use rule based credit models. If your FICO is a certain number, you get approved or you don’t. If your income is a certain number, you get approved or you don’t. If your assets make a certain are a certain number, you either get approved or you don’t, but they are very strictly rules based, and that has been the domain of traditional lending. Traditional as I said, traditional lending hasn’t changed that much. However, what firms like ours, leading edge technology firms like Perga, use vast amounts of data. So for example, I mentioned to you we have 31 lenders. So at any point in time, we are not just seeing the data from one lender, we are seeing the data from 31 lenders across the same asset class. So let’s say auto we have about 10 to 12 lenders we’re looking at in real time we’re looking at the data so we can tell how loans are performing in a cross sectional basis across 12 lenders at the same time, the power of that and then the ability to see almost a trillion dollars worth of flow, out of which we are very selective. We offer loans only to eight to $10 billion out of that trillion dollars worth of flow. But the power of looking at the data of a trillion dollars worth of loans. And the power of being able to see it cross sectionally is very, very powerful. So even before you get to AI, just the power of that data is so huge that you are able to approve loans on a much more data driven, intelligent, informed way, as opposed to the rules based way of the past. And so that itself, in itself, is very, very powerful. Now we could use on top of that, we have models which consistently keep fine tuning to what are the right levels at which a loan, a customer will find a loan attractive, the right fee, the right APR, the right loan amount. These are loans that we manage very systematically through data in a way that the offer is likely to be accepted by the customer and is likely to work well for the customer in a financially responsible way, and so that has been adopted extremely well by lenders who see this as complementary to what they do, as opposed to changing their own traditional systems. They look at pagaia as something that augments their growth, as opposed to replaces what they do today, and that has been extremely well accepted. As I said, we have 31 lenders. We are talking to pretty much all the major auto lenders across the country, and it’s growing very rapidly. It has a huge impact on the number of borrowers that are getting approved. It has a huge impact on the number of dealers who find their customers getting more approvals, and because these loans get, as I said, immediately, securitized by us. So it’s taken off the lenders balance sheet. The lender can keep it if they want to, but we give them the option of securitizing it on their behalf. It also is extremely capital efficient for the lender, and by the way the lender earns a fee from us. So what’s not to like about it? Right? It’s the lender earns a fee. The lender approves more customers. The dealers are happy, and the lender then takes the loan off of their balance sheet. So sorry to be somewhat long winded about it, but the value proposition is so comprehensive that I couldn’t have done it in a shorter way. Truth Headlam 07:10:32
Hope that helps. No, that’s totally okay. Listen, journalists, we love the details. It’s always better when you have details. So thank you for that. I did want to dig into that just a little bit more to really get into whether or not or how much alternative data is, you know, taken. Into, you know, the success rate of approval that you’re you know, speaking of Sanjiv Das 07:10:59 the truth, alternative data is helpful. Transactional data, for example, is extremely helpful. Vantage type data is extremely helpful. And for you know, a former banker like myself who’s done a lot of lending for the last 2530 years, sort of across mortgages and personal loans and credit cards and all that kind of stuff. For me to see the value of alternative data outside of the traditional banking system has been just mind blowing. However, because we lend on or rather approve a loan on behalf of our lenders, the use of data and the ability to have explanatory power on why a loan was rejected is extremely important, so that we meet all our fair lending requirements. The use of data is done in a very, very regulatory compliant way, with huge explanatory power and the ability to a ensure that there is no discriminatory lending. And this is done at multiple levels. So the use of data is we use it effectively, but very carefully and cautiously, Truth Headlam 07:12:17 switching gears, just likely to focus on pagaya A little bit more specifically, I’m curious as to how pagaya Auto volume has been trending. More recently, I understand that auto volume surpassed an annualized run rate of 1.1 billion in the first quarter, and even though that was down about 7% year over year, it was up 50% compared with the previous quarter. So can you tell me a little bit more about what drove this growth and where does the volume stand year to date? And sorry to throw another question in there, where you see it landing by the end of the year. Sanjiv Das 07:12:59 So you’re 100% right. We have grown volumes 50% quarter on quarter 2024. Last year was a relatively tough year for the entire auto industry. It went through a lot of transition. And you know, when you’re in the public markets as a public company, you’re always being pushed for growth, growth, growth, growth, growth, until one day that growth story cracks and falls on the other side. And as I said, as a former mortgage banker, I’m way too familiar with people having done that during the financial crisis, and then sort of growth story suddenly changes into a crash in crisis management story. But pragav, you’re very, very, very disciplined about making sure that it’s not just about growth, but it’s about risk adjusted growth. And so in 2024 we were very careful, protecting our investors and making sure that the yield that we gave them was a consistent yield, which has been our focus for a very long time, and we’ve delivered time and time again. So in 2024 when the market went sideways, in fact, somewhat, we were not sure where the consumer would land. We were very cautious about growth. Our investors were very cautious about growth. So the ABS markets were the spreads on auto. Abs, was substantially higher because of a lot of risk that was factored into it. So obviously, the execution capital markets, execution of those loans was not very efficient, and so our lenders reduced volumes. We reduced volumes for all the right reasons. That is the right way to conduct business across cycles. By late 2024 we found that the auto businesses, the investors, were starting to feel comfortable about the risk in the ABS business, we found that the consumers were in a far better shape than we had expected to and we are seeing that knock on wood happen through till the middle of 2025, and so we have significantly opened up our credit models around, I wouldn’t say so much of our credit models as much as we’ve opened up our pipes into our lenders. So now we have ally Westlake stellantis, one main a bunch of auto lenders with whom we are doing a lot more volume, because there’s appetite on the other side with the investors. And the execution has been extremely good with the investors. The demand from the investors extremely strong. We did our largest RMBs. Sorry, auto MBs. Auto abs. I’m sorry, it’s still in the mortgage world. In the auto ABS world, we did one of our largest ABS issuances. We just, you know, closed another very successful ABS round, and we expect that to continue through the year. We had a little bit of, you know, spotlight during liberation day timeframe, but we found that because car prices generally went up, our recoveries also went up, but they are all trending back to normal right now. So we expect to finish the year very strong.
Truth Headlam 07:16:27 And is Pagaya seeing more activity from lenders that it works with and if so, like, what is driving that increased activity? Sanjiv Das 07:16:37 Yes, we are seeing pretty much what I described for Gaia and how we saw 24 and 25 our lenders are thinking about it pretty much the same way. We are all in tandem the delinquency and the credit performance has been pretty stable. It has ticked up a little bit, but not a lot, but very stable across all of our lenders. So our lenders see the same thing. Our lenders are gradually starting to lend more, I would say that they are spending a lot more of their efforts on dealers and building their networks, as opposed to expanding their credit box. I think we are being cautiously optimistic about the credit box and but the short answer is yes, lenders are increasing their volume, but not by increasing the credit box. It’s more by doing more at the dealers and things like stipulations and you know, which are basically verifications, and making the customer experience a lot less friction, a lot more friction free, is the right way to do it. So think about it. Truth Headlam 07:17:52 And going back to some general market questions, how has the resumption of student loan debt reporting impacted the auto finance market today, both generally and in terms of AI, adoption and underwriting? Sanjiv Das 07:18:08 So the short answer is, we are all watching it. We know that it will have some kind of an adverse effect on debt levels and repayment levels, so I’m not going to brush it and brush it aside and say it will have no impact. Don’t worry about it. So far, it has had minimal impact, but we are watching student debt levels as an important input into our loans, underwriting of our loans, number two in terms of AI acceptance, the fact is that there is very little data historically based on which Our AI models are based that would say that you know the models will tell us exactly how these loans will behave, because historically, the data is somewhat sparse. And as you know, in the last four years, modeling has been a bit of a challenge, a because of the huge fiscal injection that happened post covid, and then rates going up very sharply. A lot of things happened that was somewhat extreme in the last three to four years. So to model based on just the data in the last three or four years can be tricky, and can give you outcomes that are false positives. And so we we look at that by basically doing manual overlays and making sure that we we cut off loans where student loan levels, debt levels we think are too high. Again, we are cautious about it, but we the data has limited ability to tell us based on the model, so we do it based based on manual intervention Truth Headlam 07:20:03 and and a few of your previous responses, you mentioned that you know Being regulatory compliant is super important, not just for pagaia, but in general with the use of AI as far as lending is concerned. And so I’m curious, given the shifts in the compliance landscape this year, with a pullback at the federal level, but scrutiny still remaining on AI practices and lending. How do you foresee AI adoption being impacted, and what do lenders need to look out for? Sanjiv Das 07:20:41
Yeah, I think that’s a great question. First, my first response to the change in the regulatory ratio. Team is that it really doesn’t matter. At the end of the day, we will do what we believe is in the spirit of what regulators always want, which is ensuring that consumers are not discriminated against in in in the underwriting criteria. So those rules follow regardless of the regulatory regime. And I think that’s what regulators want in any regime. And so we make sure at the guy that our models are extremely regulatory compliant and to the highest standards. Our lenders, lending partners, hold us to those standards, and we drift towards the highest common denominator as opposed to the lowest common denominator, so to speak, right the highest common standards, as opposed to the lowest common standards. And so our models go through very rigorous testing in terms of Fair Lending and very rigorous model governance in terms of Fair Lending. So we, internally, as well as through third parties, validate that our models are not discriminating against the criteria that I mentioned around fair lending. The other thing is, we have very strong model risk management. So for example, our chief risk officer is an ex city banker worked at several banks before. So the background is banking with your brother, you will agree that the regulatory standards were the highest our compliance Chief Compliance Officer, likewise, from very solid background of banks and very strong regulatory standards. So they are independent, and they hold us to those standards. And in fact, work very closely with the chief risk officers of our lending partners to make sure that the model governance process and the model rules meet the regulatory requirements in regardless of the regime. And finally, I will say this, that in some ways an excellent mission, the AI models need to be somewhat dumbed down so that they have greater explanatory power. And to us, that’s far for the course. Our models have a very we articulate the reasons for rejection, and then we send it back to our lenders. Our lenders look at the overall rejections relative to their own standards and hold pagaya reasons for rejection within the compliance standards of the regulator of the lender itself. And so it goes through a double verification, so to speak, before we say no to a customer and give them the reason for it and ensure that it meets all the fair lending standards. So that’s how we have built a very strong, regulatory compliant system, and that’s one of the reasons why adoption some of the biggest banks in the US has been so high. I mean, US Bank and several other banks that we are talking to right now. US Bank is actually a partner of ours, Ally, bank. A lot of banks are our partners, and so we meet those red standards pretty well. Truth Headlam 07:24:22
And what are some best practices for lenders adopting AI, especially those who are looking to adopt this moving forward, especially as it pertains to the underwriting and collections processes. Sanjiv Das 07:24:36
Yeah, so I would say that there are two parts to this. To this answer, there’s best practices in terms of just the quality of underwriting. I would say that people get too caught up in AI and have spent very, very little time on the power of data. The power of data, especially for a guy where you have a network of 31 lenders, is extremely, extremely powerful. It’s more powerful than the data of any one lender, who are basically, basically seeing just their data, but because we are able to see a cross section of data across 31 lenders, and we bring that power back to our individual lenders, is extremely powerful, but I will re emphasize it came from data, Not from modeling. It came from effective use of data. Now the models that are built on top of that obviously are more powerful because they were built on this cross section of data, and the ability for the models to really understand what works best for a single consumer is expressed. Powerful because it has the ability to approve more customers in a responsible way to the customer and in a powerful way for the dealer and the lender. That, in my opinion, is best practice in terms of just how you manage your business from a commercial standpoint, the second part of it, which is equally important, because there is no commercial business without the right regulatory standards. Best practice also comes from ensuring that what you’re doing is not only right, but it’s also responsible. You have to be responsible to the lender, you have to be responsible to the consumer, you have to be responsible to the dealer. You have to be responsible to the whole financial system, including your investors, on whose behalf you are essentially underwriting those loans. We take that responsibility extremely seriously. And my view is that while the market is consumed with growth, the market needs to be equally consumed about the right quality of growth. And that’s what Pagaya represents. And I would say that those two constitute best practice.Truth Headlam 07:26:58
Well, thank you for that. And is there anything else that I didn’t ask that you would like to touch on before we close out? Sanjiv Das 07:27:07 Well, I think, frankly speaking, you’ve been pretty comprehensive with some great questions. I will say that the fact that we are able to improve loans across the system and improve more loans, it’s not just about augmenting the growth of our lenders, but getting more consumers into mainstream lending, really, in my opinion, bridges the gap between Wall Street and Main Street in a very interesting way. And I think the power of AI, the power of good capital markets execution, and the ability to do all of this in a very responsible way, in a regulatory framework, is, in my opinion, a very powerful value addition to the mainstream financing and lending industry. And I think that’s where I would really sort of ask you and your listeners to focus on and see the power of it, because I think that there is something to be said about where this is going and and I thank you for the opportunity to to help me tell the pagaya story, which, in my opinion, is a very powerful story.Truth Headlam 07:28:18 Well, you’re so welcome. And thank you again, Sanji for a wonderful discussion. So that wraps up today’s episode. Thanks for joining us on the road map, and be sure to follow us on x and LinkedIn. Registration is also open for our upcoming auto finance Summit, 2025 and power sports finance Summit, 2025 in the fall, as always, we will see you online at auto finance news.net, see you next time you.