Figure CEO Michael Tannenbaum foresees potential for the fintech he heads that goes beyond its current niche as a blockchain-based marketplace for consumer loans, most of which have been home equity products.

"We're kind of like an Encompass or Ellie Mae plus Fannie Mae, but we're doing it in a modern way," he said in an interview.

Tannenbaum acknowledges that the current overlap between what Figure does as a fintech and what Fannie does has limits.

Fannie does not currently buy second liens, which are a loan type that can be conducive to more streamlined workflows. Its competitor, Freddie Mac, only has authorization for a limited pilot in that area.

However, Tannenbaum says Figure does compete with the GSEs and the major industry systems that Intercontinental Exchange now owns in other ways.

Edited excerpts from an interview about how his company is doing this follow.

Competition with the GSEs

TANNENBAUM: We are competing with Fannie Mae in that loans that are going through our marketplace. We offer essentially the same thing that Fannie does, a standardized capital market which offers liquidity.

Most of our loans are home equity products. That doesn't necessarily mean the second-lien position. Around 20% of what we do involves the first lien, and that's very important in the context of the direct comparison for Fannie Mae.

Only 10% of our loans were first liens last year, but they're growing faster than our overall marketplace. A first lien loan made to an average 740 FICO borrower who has got around $180,000 in income is absolutely a conventional borrower. If they're borrowing in the first lien position that loan could go to Fannie Mae. 

We have a bunch of the same customers as Fannie: lots of independent mortgage companies, traditional banks, and credit unions that are using our technology. These are companies who actively sell loans to Fannie Mae, who are, in some cases, directing loans to us instead. 

We've had primarily nonbanks on our platform but in our Q1 earnings back in May, we announced Flagstar signed up. They still originate into their footprint to their customer base. I think we have about 20 different banks on the platform now.

We have a few subservicers we work with, but generally we are the servicer. We are a private label player, so we're not using servicing for customer acquisition. It's more about providing nuts-and-bolts servicing, statements and collections, on behalf of our customers.

We're providing an origination system as well, so we're kind of like an Encompass or Ellie Mae plus Fannie Mae, but we're doing it in a modern way on blockchain. There are some significant benefits in doing it that way.

We've also launched a fully prefunded securitization. That replicates a lot of what the TBA style structure that Fannie is known for does.

The bonds were pre-sold, and therefore the originators we partnered with had certainty of both price and execution prior to even funding their loan, which puts much more liquidity in the market. We have done partially prefunded securitizations before. This was one where everything was prefunded.

Efficiencies key for small loans

TANNENBAUM: The average cost to originate a loan in our platform is around $1,200 vs. $12,000 for the industry. We see an average of nine days on our platform, including a rescission period, versus the 45 day average. It can be as fast as five.

We're providing real-time data and transparency into performance as part of the platform, and I think it's become a faster, better, cheaper way to do things.

Our average loan balance is around $100,000. That's where the cost to originate being $1,000 vs. $12,000 is so relevant. Spending $12,000 to manufacture a $100,000 loan is 12%. 

The Home Ownership and Equity Protection Act deters fees above 5%. It's really hard to make that work, and that's one of the reasons why there are pay ups on the specified pools. Otherwise, people just won't do these small balance loans.

The biggest money-saver is the income approach that we take. 

Traditionally, you're going to have an underwriter that is looking at income. We underwrite just based on the bank account, so we're pulling cash flows and coming up with an estimate of income based on bank account history. We're also connecting with the payroll vendor. 

We're using an automated valuation model approach, or a waterfall of AVMs, rather than a traditional appraisal. We're using an estimate of home equity based on data at the county level, rather than focusing on title insurance as a way to get the equity.

All that's happening automatically without a traditional underwriter's back-and-forth inquiries. So you're saving the borrower and the process lots of time and money. Fannie Mae does offer some appraisal and title waivers, too. If they took that further, they'd be more like Figure, but without blockchain.

We have credit performance that's disclosed in our financials. It has been around 80 basis points total for loans that are over 60 days-plus late across all of our securitizations. That includes loans in the second lien position, which are traditionally considered riskier. We have a minimum 65% combined loan-to-value ratio after the loan closes.

It is a very low risk product. The upshot is we have been able to lower costs while obtaining the same high credit quality results. 

Why blockchain matters

TANNENBAUM: Years ago, to bring up tokenization in the context of loans, banks or mortgages would have been considered totally crazy. Now it's crazy if you're not bringing it up. This is the future of the capital markets. 

We're using blockchain as a way to, for example, prevent the double sale or pledging of loans. There was a double pledging incident recently involving corporate receivables at a bank in Colorado called First Sun Capital. There was another example of this involving mortgages at England's MFS.

If you go back to the financial crisis, those types of things were a big issue. That and third-party due diligence reviews are big opportunities for blockchain to produce real savings.

Post crisis mortgages have had 100% loan file reviews but at Figure, we reduce that by 80% because we're taking attributes of the loan like the credit score, loan-to-value and home value and putting those on blockchain on day one. So multiple firms don't need to check and recheck the same information.

Our Digital Asset Registry Technology is an alternative capital market like everything that we're building versus what it is in the Fannie Mae/ICE Mortgage and MERS world.

Fannie Mae only allows MERS. If Fannie Mae accepted DART, it would be compatible. It's cheaper, and it actively listens to trades in addition to doing more to prevent double pledging because loans are tracked on blockchain. 

We have a lot of investor support. Different securitization and whole loan buyers, with some bonds rated AAA by S&P, and Moody's. We did $1.5 billion of loan volume in June alone.

We do charge a transactional fee on all this volume. It generally is a set amount, but for partners that do more with us, there's a sliding scale where fees go down with volume. In order to buy loans on the platform, you have to sign up. It typically takes maybe a week or so.

Future plans

TANNENBAUM: We're a business that does well in rate environments where people are more interested in second liens because a cashout refinance doesn't make sense when you want to keep an older, lower-rate loan. But with our low costs and growing first liens, we also intend to do more in a lower interest-rate environment.

Almost everything that we do is home equity, but it can be in the first position, so about 20% of what we do is in the first lien space. We are purchasing a company called Kiavi, which is the market leader for residential transition loans. That is about to be 40% of our revenue mix when the deal closes.

We talked about going into the purchase space in the near term. We broadly do non-QM, we also have new debt-service coverage ratio capabilities for the acquisition of Kiavi. 

We've had a marketplace running on blockchain now for two years. That gives us an opportunity to marry the data with the decision engine through AI, and I think that's something that we will begin to talk about more. We've been launching some new features lately. You'll see us do more there.