Virtual Roundtable: The Level-Up: Fintech Insights for Lending Innovation

Discover how modernizing your lending tech stack can reduce operational costs and lead to an increase in approvals.

Snippet of Virtual Roundtable Transcript

 

Amias Gerety:

Rudy, if you could, just give a sense of the landscape of what has really changed in the last six months. 

 

Amias Gerety:

From where we are right now, what’s going to be different going forward, and what feels like it’s starting to return to normal in the mortgage industry?

 

Rudy Orman:

There’s nothing normal in the mortgage industry, it’s always changing, and you just have to jump on that roller coaster without the seat belts and just hold on tight. So, obviously COVID’s changed everything, especially on the self-employed borrowers. From the agency perspective, everyone read, Where FHFA, they’re going to add refinance. Because there’s an add-on now for refinance because of all the run-off. I’ve been on both sides, I’ve been on the originations side, I’ve been on the Capital Market side, I worked on the desk at Coleman back in the sub-prime days. I mean, I’ve seen it all, and everything… These events seem similar, they all have a little bit of a different take on it, like COVID obviously, really effecting self-employed borrowers, and that’s got everybody obviously concerned.

 

Rudy Orman:

So, before COVID, we were doing a lot of non-QM, and that’s I’d say 40, 50% of a non-QM borrowers are self-employed. So for those that quickly non-QM would be… We’re verifying their income. So there’s an ATR component to it and we’re using bank statements from it. So Fannie and Freddie won’t buy that loan but it’s got safe harbor to it. So, those loans go into our RBS structure kind of similar to the MAVA structure back in pre 2008. And so the good borrowers they perform well, the loan to value, the fact those are like high 700 CLTV they’re like 70. So it’s like a lot of these people don’t even need the money. They’ve got all the assets, there’s a lot of asset verification, a lot of the per reserve requirements.

 

Rudy Orman:

So, its great [inaudible 00:09:25]. The problem is if self-employed, because they don’t show all their income on their tax return. And generally after 2008, those people couldn’t get a loan from anybody, unless they went to portfolio, bank or whatever, there wasn’t like a national program. So, things were going well and then COVID hit. And that really obviously hit the self-employed borrowers. We stopped buying for three to four months because liquidity was zero and pricing was nonexistent. We actually put that park, that stuff on the bank’s balance sheet, it has come back. Thank goodness.

 

Rudy Orman:

But there are still a lot of scrutiny on the self-employed. And then if you look how Fannie and Freddie are doing it too, tons of scrutiny on self-employed. It’s a shame because there’s a lot of good borrowers that are self-employed, but there are some that are affected by COVID as such they’re going to have a hard time getting a loan. So that has changed from a technology perspective. It’s funny because the mortgage business we’re still doing stuff in the nineties, right?

 

Rudy Orman:

The mortgage industry is not gone very far from technology. So this COVID thing happens. There’s still a lot of manual stuff happening, which is crazy insane because look, we’re on Zoom right now. Right? And we all have smartphones, and yet we’re still manually underwriting self-employed borrowers. But that’s the reality of the mortgage industry. We’re still fighting with rocks and sticks. So that’s interesting to me too, because of COVID, you think, well, there should be a technology solution. It’s not. It’s so a lot of manual process. So, I think the underlying statement is if you’re self-employed the mortgage business got hard for you and expensive.

 

Libby Morris:

I would add to that now more than ever, that bank data is super important. A business that was doing an amazing job their last time they filed taxes, their whole world could have changed due to the COVID crisis and businesses that maybe weren’t doing great, have been able to pivot quickly. What we’re seeing is the data we would normally gather from a business tax return is nothing. We have no idea if it’s any reality anymore. So those banks on the data we can get from them and the sooner we can get it, the sooner we can decide if this borrower is eligible to really move forward.

 

Amias Gerety:

Yeah. Libby, it’s so interesting, because sometimes we think of mortgage and small businesses, two worlds that don’t connect, but what Rudy has just outlined is, these self-employed mortgage borrowers have incredibly high overlap with your population of small business owners.

 

Libby Morris:

That’s right. Yeah. I mean, they’re the same borrower just in a different life experience and maybe they could be doing both at the exact same time, really, depending on the status of their business and their personal life.

 

Amias Gerety:

So Kate, turning to you a little bit, talk about how you’ve seen Plaid obviously market leader here, polling that real time bank data into these borrowing experiences. From the perspective of what’s changed over the last five years, zoom in on what’s changed for the last six months. What are you seeing in your day to day as people grapple with the things that Rudy and Libby are already implementing?

 

Kate Adamson:

Yeah. So, there’s, I think, a couple of buckets of adaptation that we’re seeing from our customers who respond to this new age. But specifically, I think, a general theme is the importance of that recent data. And bank statement data can be the most recent credit related data available. And it is a source of truth. It’s the ledger of truth, of where activity is happening for either a business or consumer who may be a small business owner. So, I think there’s a couple different buckets. There’s one it’s operationalizing some of the manual work that Rudy was just describing. If there’s a way to better detect the income streams related to a self-employed borrower that is GSE guidelines standards, have a way to verify their employment and for that type of borrower.

 

Kate Adamson:

And so as Rudy mentioned, so much of it is, it is hard to generalize. Banking data is hard to work with. It’s not simple and frankly, the complexity is also about how banking data is reflected in a transaction string. So they may say Venmo because a hairdresser may get paid via Venmo. And that is for one person could be rent that one person and another person that could be their business income. And so it is tough to work with. And so to an extent there’s automation that’s possible, but it’s about marrying the automation and the programmatic sort of filtering of data with a human review process.

 

Kate Adamson:

Which means you can’t ever fully automate specifically for this borrower segment. It comes down to interactivity with the borrower, as well as a human underwriter, actually going through a process and making sure everything is valid and maybe a consumer even saying, “Oh, this was rent. This was my hairdresser income.” So being able to identify the distinct income streams will involve some human interaction, whether that’s the borrower or the underwriter. I [inaudible 00:15:17] the second bucket beyond operationalizing, where there are increased costs is the identification of more recent information.

 

Kate Adamson:

And this doesn’t necessarily apply to mortgage and maybe one day it will. But the way that the system works, it doesn’t so much. So, as it does to other industries, where a personal lender might be able to make a credit decision on an ongoing basis or create a credit line for a consumer and continually reevaluate that consumer’s credit worthiness. One of the beauties of the Plaid product is that we do allow with a single authentication, the developer or the lender, in this case, the ability to refresh over time with the consumer’s consent.

 

Kate Adamson:

And so that doesn’t just allow for a more recent upfront decision about that business or consumers credit worthiness, but also on an ongoing basis, being able to detect changes sort of in a way that traditional credit underwriting data does not support. And I think that’s a broader theme we could talk a whole session around, is where there are deficiencies in the credit system today and helping underwrite, but also potentially misleading on the credit worthiness of potential consumers, given FICO it’s a lagging indicator.

Kate Adamson:

You need to have not delivered on a payment in a credit product for that to be reflected in your FICO score versus leading indicators, like no longer employed, no longer managing your budget, gotten Zoom fatigue and COVID fatigue, and now gone on a spending spree and taking that trip that they needed. So there are a lot of ways to detect earlier than the credit score. And I think we’ve seen some of our savvier customers incorporate that cost initial underwriting into an ongoing servicing or portfolio management perspective.

 

Amias Gerety:

Yeah. And I think this just goes to this theme of just tracking the day to day situation, right? I mean, this is a crisis that came on us faster than any recession ever in history. And similarly tracking, opening and closing. And when it’s warm, maybe that you can put tables outside, then there’s a cold day, no tables, right? There’s just so much complexity in the day-to-day situation. 

 

Amias Gerety:

Sipho, you were going to jump in here and just talk a little bit about how you’re translating this day-to-day situational analysis into customers who might not have done it before.

 

Sipho Simela:

Furthering on really what everyone’s brought up, is these problems compound with scale. So if you think about the idea that you have this many consumers, this many Americans running to the same small door at the exact same time, these problems become far compounded and far less easily managed. What we’re seeing in Ocrolus, is we’re seeing things around what happens when you have these problems at scale? What are some of the tools that are available to, let’s say, detect fraud? Fraud becomes something that in an environment like we’re operating in right now, something that lenders have to be mindful of.

 

Sipho Simela:

The costs right now are too high with the cost of reputational risks, the cost of monetary risks, really putting good technology in front of that just further helps those operations teams be able to manage scale. A lot of what we’re seeing right now, as far as our lender partners, we’re seeing people take that dual approach that Kate mentioned, where you do have this intertwining of both human work as well as machine work. Making the machines work could be more efficient for you.

 

Sipho Simela:

That’s a mantra at Ocrolus. We’re focused on the intertwining of our human in the loop. How can we give lenders that dual path? Where, not only is the combination of the machine processes. Whether those are automating data streams and data flows, but also for consumer lending, which, let’s face it, is still a highly paper intensive industry right now. So giving that dual approach is something that we’re really focused on.

 

Libby Morris:

That’s something that we use both Plaid and Ocrolus for is that alternate documents check. It’s very easy to catch the… Well, first of all, with Plaid, they’re lagging around their bank. So there’s not an alternate document. But then with Ocrolus, where they’re able to flag to us, “Oh, the font looks a little off here.” And we have a human eye on it at that point, but it’s nice to have the flag up front to kind of ensure that there’s no manual miss.

 

Amias Gerety:

I mean, I think this combination of the fraud element, there are other Ocrolus partners like SentiLink, who’ve talked about synthetic identities. You build up these synthetic identities for years, and then you wait for a moment to catch them in. And so you have this huge spike in fraud attempts during COVID, because it’s a cash in for organized frames of criminals. I mean, this is a experience that I think companies, both Plaid and Ocrolus, were real leaders on over the spring. I mean, never have we seen the day to day change in lending as severe as what we saw with PPP, right?

 

Amias Gerety:

Where people were both needing to move incredibly fast. Because the money felt like it did run out and then got replenished. And the rules were one way and then they got changed a little bit. And so just that day-to-day situational awareness is not just a underwriting problem. It’s also an operational problem. 

 

Amias Gerety:

One of the things that I would be interested in, Sipho, maybe you helping us transition into the next question, which is sort of, how are you helping a lender who maybe isn’t as forward looking, maybe isn’t as tech-savvy as the ones we have with us today.

 

Amias Gerety:

And they’re saying, “Yes, I get it. But there’s a lot changing. I don’t know if I can also change my tech stack. I understand what you’re promising, but I’m just trying not to break right now under the pressures.” How do you help people work through the natural human tendency to stop and just wait?

 

Sipho Simela:

Definitely. Change is scary in general. I think that in general people being changed reversed the second that regardless of the kind of lift that you could see downstream it really is hard to see the forest, the trees, and there’s a metaphor in there somewhere. But, the reality is that, especially when you’re working day-to-day to kind of just keep things in line. That’s why what we offer is this infrastructure solution. What we’re trying to do is not to introduce the new solutions that are going to need to be borrower facing, new solutions that are going to need to be trained upon. We are infrastructure.

 

Sipho Simela:

So that’s something that message-wise, has really picked up for us. And it’s the idea that, that transition, regardless of what channel the borrower chooses to go down, the lender experience is optimized in this scene. So when we think about that Ocrolus lift, it’s the idea that really the way that your borrower chooses to interact with you, the downstream efficiency is the same whether that’s in the form of reports of workflows, we’re really focused on just operationalizing pieces of that process.

Subscribe to our newsletter