Einblicke
Alternative Lending Through the Cycle
|
|
Insight Article
|
• |
Oktober 13, 2022
|
Alternative Lending Through the Cycle
|
Oktober 13, 2022
|
Alternative Lending Through the Cycle
Alternative Lending Through the Cycle |
|
As fintech alternative lending grew into a new asset class, investors wondered how these loans might perform in an economic downturn. Each recession is different, but we now have instructive data.
The asset class was not immune to the havoc wreaked during the acute initial phase of the pandemic, but unsecured consumer alternative loans originated before the onset of COVID-induced shutdowns broadly exhibited fundamental resilience. Furthermore, impaired loans rebounded rapidly as economic conditions stabilized. For consumer loans originated after COVID’s onset, the marketplace model that alternative lenders use for matching consumer borrowers with loan investors drove rapid recalibration in underwriting standards, reflecting changed economic realities. Tighter credit conditions and higher interest rates on new loans compensated investors for both heightened risk and heightened risk aversion.
We believe three metrics are of critical importance as investors evaluate fintech alternative lending’s performance through the pandemic: how much the loan impairment rate increased relative to the sharp change in the U.S. unemployment rate, how long the loan impairment rate remained elevated, and what percentage of impaired loans ultimately resulted in charge-offs—defaulted loans removed from the books or marked down to low expected recovery values. On these metrics, alternative lending availed itself admirably. Display 1 illustrates the resilient performance of consumer unsecured loans in the face of a spike in unemployment.
Source: Consumer Unsecured Benchmark CDR: dv01, as of June 30, 2022
Quarterly Unemployment Rate: https://fred.stlouisfed.org/series/UNRATE#0, as of June 30, 2022
Conditional Default Rate (CDR) represents the amount charged-off in a given month divided by the start-of-month balance and then annualized.
Source: Consumer Unsecured Benchmark CDR: dv01, as of June 30, 2022
Quarterly Unemployment Rate: https://fred.stlouisfed.org/series/UNRATE#0, as of June 30, 2022
Conditional Default Rate (CDR) represents the amount charged-off in a given month divided by the start-of-month balance and then annualized.
How Do Changes in the Unemployment Rate Affect Changes in the Consumer Loan Charge-Off Rate?
When evaluating the expected performance of consumer alternative loans through a downturn, we use an intellectual framework that models the expected change in the consumer alternative loan charge-off rate as a lagged function of the unemployment rate change. This framework is informed by historical credit card data that tracks consumer charge-off trends over extended periods of time, including multiple prior recessions. Of course, behavior of the American consumer through prior recessions was no guarantee of how consumers would behave post-COVID— particularly given that the pandemic lacked recent historical precedent. Furthermore, while credit card lending is also a form of unsecured consumer lending, we recognize that it provides an imperfect proxy for consumer alternative lending. Credit card loans might sit higher in borrowers’ payment priority hierarchies than do loans facilitated by alternative lenders. Conversely, alternative loans generally repay via automated pulls directly from borrowers’ bank accounts, and they typically amortize rather than revolve like credit card loans.
Display 2 shows that, pre-COVID, changes in consumer credit card delinquencies broadly tracked changes in the unemployment rate—both up and down. For example, both spiked during the Global Financial Crisis (GFC) but normalized soon thereafter. However, the trend was much less clear through the COVID period, with changes in delinquencies remaining much more stable than would have been expected given rapid changes in the unemployment rate.
Similarly, while the explanatory power is limited, regression of pre-COVID historical credit card data demonstrated roughly a 1:1 relationship between unemployment rate change and credit card charge-off rate change on a three-month lag. The regression suggested that any expected increase in charge-off rate following a rise in the unemployment rate typically would only apply to the period immediately following the unemployment rate increase. Afterwards, the charge-off rate would be expected to stabilize. Likewise, the charge-off rate would be expected to fall following an unemployment rate decline, which is typical after a recession. With this framework, it is important to remember that changes in the charge-off rate are annualized, so the duration of unemployment rate increase is critical when thinking about the prospects for alternative loans underwritten in advance of a downturn.
For reasons we will discuss in the following section, changes in delinquencies and charge-offs proved less correlated with changes in unemployment through the COVID period. As illustrated by Displays 1 and 2, consumers showed strong resilience despite the dramatic swings in unemployment as the pandemic unfolded.
Changes in Unemployment Rate (1991-2022)
Source: Based on quarter-over-quarter absolute change in credit card delinquency rate from 03/31/1991 - 03/31/2022 per Federal Reserve data (https://fred.stlouisfed.org/series/DRCCLACBS), as well as quarter-over-quarter absolute change in average quarterly unemployment rate derived from each end of month datapoint calculated using a simple average for the three months included in each calendar quarter from 03/31/1991 - 03/01/2022 per Federal Reserve data (https://fred.stlouisfed.org/series/ UNRATE). Derived from analysis by Upstart Network, Inc.
Changes in Unemployment Rate (1991-2022)
Source: Based on quarter-over-quarter absolute change in credit card delinquency rate from 03/31/1991 - 03/31/2022 per Federal Reserve data (https://fred.stlouisfed.org/series/DRCCLACBS), as well as quarter-over-quarter absolute change in average quarterly unemployment rate derived from each end of month datapoint calculated using a simple average for the three months included in each calendar quarter from 03/31/1991 - 03/01/2022 per Federal Reserve data (https://fred.stlouisfed.org/series/ UNRATE). Derived from analysis by Upstart Network, Inc.
Charge-Off Rate Change Expectations Based on History Did Not Reflect Positive Factors at Work Post-COVID
Pre-COVID historical credit card data stemmed from prior recessionary periods that lacked the truly unprecedented fiscal and monetary stimulus measures put in place starting in Spring 2020. Those stimulus efforts likely reduced the pass-through from unemployment rate changes to lagged changes in the loan charge-off rate. Furthermore, U.S. households significantly reduced their aggregate debt relative to GDP in the aftermath of the GFC. They also entered the COVID crisis with debt service ratios at low levels not seen in decades.1
Industry-level data from the lending analytics firm dv01 highlights the U.S. consumer’s resilience during the pandemic. Display 3 shows that from the beginning of March 2020 to the end of May 2020, change in the rate of unsecured consumer alternative loan impairment (including both delinquent and modified loans) was in line with change in the official U.S. unemployment rate. However, the impairment rate change was well below change in Morgan Stanley Research’s estimate of the adjusted unemployment rate, after correcting for undercounting in the official rate.2 Furthermore, both the unemployment rate and the loan impairment rate trended down as the economy began to recover.
Unemployment Rate
See Glossary for definition of loan impairment. Source: dv01 – Loans in the dv01 dataset had a weighted average original FICO score of 703, weighted average coupon of 15.2% and weighted average balance of approximately $12,900 as of March 31, 2022; https://www.bls.gov/news.release/empsit.t15.htm, and https:// ny.matrix.ms.com/eqr/article/webapp/495476fc-93b4-11eb-be2a-2ded72795adb?ch=rpint&sch=sr&sr=58.
Unemployment Rate
See Glossary for definition of loan impairment. Source: dv01 – Loans in the dv01 dataset had a weighted average original FICO score of 703, weighted average coupon of 15.2% and weighted average balance of approximately $12,900 as of March 31, 2022; https://www.bls.gov/news.release/empsit.t15.htm, and https:// ny.matrix.ms.com/eqr/article/webapp/495476fc-93b4-11eb-be2a-2ded72795adb?ch=rpint&sch=sr&sr=58.
What Could This Mean for Investing in Alternative Loans Going Forward?
Not surprisingly, the performance of alternative lending was influenced by the pace of employment recovery, with divergent outcomes across different groups of consumers. Consumers’ willingness and ability to service debts was influenced by their financial histories, balance sheets, incomes, types of employment and states of residence— with COVID-19 likely amplifying the effects of all these variables. Alternative lending’s differentiated underwriting may take these factors, and many more, into consideration when determining whether a loan should be extended and at what risk-adjusted pricing.
Performance also may differ across loan vintages. In certain cases, seasoned, partially amortized loans may prove to be less sensitive to acute changes in the economic backdrop than similar but less seasoned loans underwritten just before a downturn. Contractual monthly principal amortization may naturally reduce a borrower’s propensity to default as the loan progressively pays down. Furthermore, newly originated alternative loans underwritten after the onset of the COVID downturn typically reflected changed economic circumstances through tightened credit standards and higher borrower interest rates.
From an investor’s perspective, reinvesting cash flows from alternative loans already in a portfolio into new alternative loans may facilitate vintage diversification that may reduce recession sensitivity. Over time, vintage diversification also may allow portfolios to reflect changing economic conditions in terms of varying credit standards and borrower rates across portfolio vintages (Display 4). In our view, loan vintage diversification provides a powerful risk management tool for investors in alternative loans that may not be as available to investors in traditional corporate credit portfolios which repay principal only at maturity.
Diversification does not eliminate the risk of loss.
Source: AIP Alternative Lending Group. The chart assumes a simplified illustrative $100MM portfolio of loans with a 7% coupon rate. The chart also assumes no losses or prepayments, and $20 million of additional investments each month. It further assumes that the first investment was made on 12/31/19 and that all interest and principal payments from the loans were re-invested into new vintages on a monthly basis.
Diversification does not eliminate the risk of loss.
Source: AIP Alternative Lending Group. The chart assumes a simplified illustrative $100MM portfolio of loans with a 7% coupon rate. The chart also assumes no losses or prepayments, and $20 million of additional investments each month. It further assumes that the first investment was made on 12/31/19 and that all interest and principal payments from the loans were re-invested into new vintages on a monthly basis.
Conclusion
We view the performance of unsecured consumer lending through the COVID pandemic as evidence of the asset class’s fundamental resilience. In up and down markets, we believe that diversified exposure to alternative loans across regions, sectors, platforms and vintages remains the best way to access the asset class. Alternative lending may diversify both traditional and private credit allocations.
|
Managing Director
|