Fighting words: The marketplace lending debate heats up
Last year, a community bank client of Christopher Marinac looked over the results of a $100 million investment it made in funding a pool of marketplace loans.
The return was around 5.5% – not bad compared to the investments of the institution’s overall portfolio, said Marinac, director of research for investment banking and research firm Janney Montgomery Scott.
But for the risk it was taking, the bank’s executives were (realistically or not) expecting more. After all, they were pursuing high rewards from high-risk funding into the relatively unproven long-term performance of online, unsecured consumer loans.
According to Marinac, the feeling at the bank was – is that it?
“The bank didn’t lose money on these,” said Marinac. “But the losses were to some extent higher than expected.” Instead of an expected loss level equal of 150 basis points on the investment, the bank “discovered they were losing about two, two-and-a-quarter [percent] on those loans."
“Other banks ... they’ve had a similar experience,” he said. “If they thought they were going to lose 250 basis points on marketplace lending, they wouldn’t have [backed] the loans.”
The small banks’ lukewarm feelings described by Janney were reflected recently in the securitization market for new-issue marketplace loan (MPL) pools. In the most recent tracking study of marketplace securitizations, loan data specialist Peer IQ reported MPL investors have grown concerned about the declining all-in yields of new securitizations, as well as rising levels of delinquencies and charge-offs in outstanding deals.
In one example, Peer IQ reported the 24-month late-payment performance of MPL deals issued in 2017 was lagging that of MPL portfolios issued by lenders in 2015 and 2016.
Although fintech lenders such as Prosper Marketplace and LendingClub have worked to improve credit metrics by adding more prime borrowers into their asset-backed pools, “[w]e don’t see the purported improvement in underwriting just yet” across the industry, the Peer IQ report bluntly stated.
“Consumer credit delinquencies and charge-offs have continued to increase during the fourth quarter.”
Despite these concerns, institutional MPL investors have stayed in the market as deals continue to roll out (22 deals totaling $8.1 billion, per Finsight data). Their confidence is maintained through "robust" protections built into portfolios by sponsors, as well as improved ratings agency confidence, wrote Peer IQ.
When the online consumer-loan ABS market took off in 2015, credit enhancement features like excess collateral in pools or funding reserves were fairly substantial. Prosper, for example, launched its first securitization in 2017 with an initial enhancement level that amounted to a substantial cushion of 43.9% against losses on the senior notes – an amount equivalent to perhaps a pool of deep-subprime auto loans.
These enhancements level have declined the last two years as the fledgling asset class matures, giving ratings agencies more comfort in determining loss expectation performance of the deals. As a result, “[r]ating agencies have played a major role in creating investor confidence in capital markets,” Peer IQ stated.
That confidence, however, is coming under question. And the criticism is coming from within the ratings agency universe itself.
A call out
In May, Fitch Ratings issued a report critical of high investment-grade ratings applied by other agencies to recent MPL bond offerings.
Although the agency did not specify any deal it considered misrated, Fitch argued that the single-A or double-A ratings that some 2018 and 2019 MPL securitization sponsors are undeserved, given dwindling protections for investors and no improvement in underlying loan quality, Fitch said in its May report.
“Credit enhancement levels ... meaningfully declined since 2017, while asset quality remained relatively steady,” the report stated. “As a result, bondholders of more recently issued transactions have less loss protection or the same amount of asset risk.”
Because of the nascent asset class' short history, most lenders' securitizations are capped at a single-A rating by the agencies rating most of the deals, including Kroll Bond Rating Agency, Morningstar Inc.'s DBRS and S&P Global Ratings.
(Only one MPL deal sponsor, Social Finance, has achieved a triple-A rating, due to its collateral focus on debt consolidation loans to high-earning professional borrowers, as well as its extensive experience in student-loan ABS.)
Fitch’s report included statistical comparisons showing how CE levels have declined across the board for deals sponsored by online lenders. Avant, for example, provided an average 43% of credit enhancement to its deals in 2017, but maintained only 31% for its latest transaction in 2019 for the same single-A rating, despite declining weighted average FICO in the new deal.
The fintech sector has “demonstrated notable volatility in its history, including underperformance compared to originator expectations, shifts in underwriting standards and regulatory challenges to the business model. Fitch believes these challenges should be reflected in stable, not declining, credit protection to investors.”
A few weeks after Fitch's report, Deutsche Bank weighed in via a credit research report that Fitch may “arguably have a point” about the lower CE levels, “[but] this seems to ignore the precedent for rating agencies requiring less credit enhancement at a given rating category for an issuer as more historical performance data becomes available due to seasoning and continued issuance.”
But Deutsche consumer ABS research analyst Kayvan Darouian agrees that the benign credit environment has benefited the growth of the MPL space.
"You basically don't have historical performance data through a cycle," Darouian said in an interview. "For some investors that is a sticking point. I think we have a pretty good idea of how these things perform in a benign environment, but we would love to see how they perform in a downturn."
Fitch's report added that while MPL transactions have “rapidly deleveraging” structures that support stable ratings performance, “this dynamic only holds when there is sufficient time for CE to build from initial levels before the ratings are reviewed. Should significant stress occur too soon after deal closing, noteholders are likely to be subject to downgrades across the capital structures."
Fitch itself has not been solicited to rate an ABS deal since a Prosper Marketplace transaction in early 2018. Fitch analysts won't say if the firm’s dormant activity in the space stems from its reluctance to issue high single-A or double-A ratings that lenders are obtaining elsewhere.
But “[w]e do believe the credit enhancement levels are often too low for the ratings being assigned,” said Eric Orenstein, a director at Fitch. “We have declined to rate certain transactions.
"There have been deals we could rate in principal, but the CE levels we would look for in the requested ratings were higher than what the issuer had in mind, so we ultimately didn’t get engaged.”
Fitch named no agencies in the report, but an obvious target was the ratings firm that issues the preponderance of MPL ratings – Kroll. Kroll rates about 75% of outstanding MPL deals issued to date, according to Deutsche Bank, followed by DBRS (34%) and S&P (29%). Fitch's legacy activity in rating MPL deals leaves it with an 8% market share of outstanding deals while Moody's Investors Service has only rated 5% of fintech deals to date.
Kroll has rated 65 transactions since 2016 (totaling $22.4 billion in bonds). It was solicited to rate all 144A exempt (non private-placement) securitizations in 2016 and 2017 and 33 of 35 securitizations issued in 2018 and 2019.
Across most of those deals, Kroll is the source of several high investment-grade ratings. While LendingClub, Avant and Prosper remain at A-level ratings, issuers including Marlette Funding (d/b/a Best Egg) have been able to push their deals into AA-rated status.
In the weeks after Fitch report, Kroll appeared to take umbrage as it quickly issued its own report answering many of the criticisms that "misinformed" industry observers had laid out. Kroll insisted credit quality and performance of MPL securitization platforms have improved enough to support existing ratings levels, despite declining credit support.
“These improvements during the last four years have resulted in more predictable credit performance across the industry and upward rating migration for MPL securitizations,” the report stated. “[Kroll] is better able to forecast base-case losses and, when appropriate, may accept lower loss multiples for a requested rating, which results in lower credit enhancement requirements consistent with the evolution of a maturing sector.”
Underwriting improvements include more advanced internal credit-scoring abilities that make consumer behavior (and expectations for consistent future payments) more predictable that FICO numbers reflecting prior credit history. Avant, for example, had declining FICO averages in its 2017 loan ABS pools, but its delinquency levels on six-month seasoned loans fell to 4.4% from 5.82% by placing more underwriting emphasis on a borrower's resulting debt-to-income from the loan being taken out.
"Other platforms develop similar ability to pay ratios, such as [free cash flow] or net disposable income, to determine maximum loan amounts," Kroll stated.
Kroll also argued that deals are deserving of upgrades as they amortize and have reduced leverage. (MPL deals are analogous to subprime auto securitizations - senior notes are paid quickly, usually within three years, which provide more immediate principal payments to previously subordinate notes).
Though 2018-2019 vintage deals have had decreasing levels of overcollateralization and lower proportions of subordinate notes in a first-loss position, Kroll argues the credit performance of these MPL ABS deals have improved. That improvement supports continued high ratings as well as upgrades as the deals amortize.
Kroll also credits the sector with recruiting more executives with traditional and specialty banking finance backgrounds – as well as with introducing tighter underwriting criteria and advancing operational strength in loan servicing and fraud detection. Veteran executives in the sector include Marlette’s Jeffrey Meiler (a U.S. credit-card executive at Barclays) and Prosper’s David Kimball (a former chief operating officer at USAA, as well as securitization roles as Ford Motor Credit)
MPL lenders are also diversifying their source of funds. Having long cast off the legacy "peer-to-peer" model dependent on individual investors, lenders now underwrite loans (through bank partners like Cross River Bank) with instittuional warehouse lines that maintain stable lines of credit for new loan offerings. Deals are also being issued with lenders choosing to better align their interests with investors witrh more skin in the game.
This wasn’t the first time Fitch has engaged in unsolicited, critical opinions of other agencies’ work. Last September Fitch criticized the triple-A ratings assigned by S&P Global Ratings and DBRS for a middle-market CLO. Fitch also has critiqued residential mortgage-backed securitizations ratings as well as those for an aircraft-lease asset-backed transaction, among 25 such deal commentaries in recent years.
May's report was also the second time in the past year Fitch aired concerns for the MPL space. In April 2018, Fitch issued a report casting doubt that the latest collateral pools from lenders like Prosper and Lending Club would ultimately perform as well as their older asset-backed deals of unsecured consumer loans. Fitch said at that time investors should still be wary of assuming new-issue MPL securitizations are a step up in quality over previous ABS deals, even though lenders were tightening lending standards and boosting the pools with borrowers with higher credit scores.
On that end, Fitch's concerns were somewhat abated. In June, Fitch issued an upgrade to the remaining Class B notes of Prosper’s 2017-2 transaction, following a senior-note payoff that significantly boosted CE.