Prior to the Covid-19 pandemic, the U.S. economy was on the verge of extending the longest recorded expansion in history. Consumer spending made up the largest portion of the economy, supported by a healthy employment picture and free flowing credit. The Bureau of Economic Analysis reported that personal consumption expenditures made up 67.7% of GDP in the 1st quarter 2020. Unfortunately, the Covid-19 pandemic abruptly halted the expansion as shelter-in-place orders shut down businesses and curtailed economic activity across the country. The impact could be felt immediately as the unemployment rate increased to 13.3%, with approximately 40 million individuals finding themselves out of work.

In this analysis we look to see the damage that the pandemic has inflicted on the consumer. Using monthly remittance data from both asset-backed securities (ABS) and non-agency mortgage-backed trusts, we can see how borrowers have been able to weather the storm through their ability to service their consumer debt. Thus far, it appears that consumers have been faring better than expected, particularly in light of the historic unemployment figures.

During the 1st quarter, the Federal Reserve Bank of New York reported that total household debt increased by $155 billion to $14.3 trillion. It was noted that even with the increase in household debt, there was an unusually large decrease in credit card balances ($34 billion), possibly an early sign that consumers have started retrenching.

Source: FRBNY Consumer Credit Panel / Equifax

Credit Card ABS

Credit card trust remittance data can give investors a plethora of information as ABS master trusts are made up of pools of receivables from a large group of individual cardholders. When looking at performance of a master trust, one often looks at metrics including excess interest, portfolio yield, monthly payment rate, delinquency, and charge-off rates. Given the high level of unemployment, expectation for a surge in delinquency rates would be the most logical assumption. Yet, April rates have only increased marginally.

As banks have established Covid-19 assistance programs allowing borrowers to defer payments without penalty, many of the traditional metrics have been distorted. Thus far, delinquency and charge-off rates in the various credit card trusts have not meaningfully increased. This can be attributed to the reporting guidelines of the various assistance programs that freeze the delinquency status of the borrower at the initiation of the payment deferral option. Many borrowers have also had access to the deferment option in addition to enhanced federal unemployment programs, which have allowed them to stay current by paying the minimum monthly payment.

Source: Osterweis and company filings

As delinquency and charge-off metrics have stayed relatively stable, it is noticeable that excess interest (the interest remaining after paying investors and expenses) and portfolio yield (generated from finance charges and fees) of the trusts have decreased. It is difficult to ascertain how much of this decrease can be attributed to borrowers not paying interest as part of Covid-19 assistance programs. Also factored into these metrics is the move in the prime rate, which is used to set credit card annual percentage rates (APR). The prime rate has dramatically fallen given the Federal Reserve rate cuts, resulting in lower excess interest and portfolio yields. Thus, the fall in these metrics is not clearly indicative of borrower stress.

For now, monthly payment rates (MPR) give an investor the best sense of the overall health of the pool of borrowers. Thus, a falling MPR is an indication that the rate of principal paydown has slowed and could be a precursor to credit stress within the trust. This measure can be used as a proxy of the borrowers’ willingness and/or ability to pay down existing credit card balances. During the 2008 financial crisis, falling payment rates were a leading indicator as financially stressed borrowers were unable to pay the required monthly payment or could only pay the minimum payment.

Source: JP Morgan

The monthly payment rate will serve as a useful indicator of the health of the borrower as shelter-in-place restrictions are lifted across the country. The recent slowdown of the rate could be indicative of borrowers looking to increase their financial flexibility by only making minimum monthly payments and could also be driven by higher usage of payment deferrals. In both scenarios, the monthly payment rate suggests that borrowers are looking to mitigate financial pressure presumably as a result of the shutdowns driven by Covid-19. Moving forward, it will be telling to see if MPRs continue to fall and will remain the best indicator of the health of the credit card borrower.

Auto ABS

As credit card borrowers seek relief through payment deferrals, a similar trend has emerged in auto loan ABS. Auto loan lenders are offering Covid-19 assistance to financially stressed borrowers through loan modifications – more specifically, loan extensions. Loan extensions allow the borrower to skip payments while extending the maturity of the loan by the same number of postponed payments. This grace period gives borrowers temporary relief allowing them time to recover and avoid default. As with previous experiences, such as hurricanes Harvey and Irma in 2017, servicers use extensions as a loss mitigation strategy, which has been shown to preserve value during temporary shocks.

The most recent remittance data of various auto loan ABS deals are showing increased rates of loan modification. Such elevated levels have been observed in pools collateralized by loans made to both prime (higher credit score) and subprime (lower credit score) borrowers.

Source: JP Morgan

Increases in loan modifications are not surprising as the unemployment rate has hit historically high levels. In addition, loan modifications may have been the only remedy for lenders as various states and local jurisdictions have established regulations to limit or even prohibit vehicle repossessions for the time being. Additionally, repossessions have not been the optimal course of action as used car prices had been depressed with auction activity severely reduced and even shut down in the months of March and April.

Similar to credit card trust data, auto loan delinquency data has not been indicative of the increased strain facing borrowers. Across a sample of both prime and subprime ABS pools, delinquency status has been steady over the last three months while the percentage of loan modifications has increased dramatically. Increases are particularly noticeable in the subprime cohort (deals with FICOs lower than 650) as loans in these deals have lower average credit scores, higher percentages of used cars, and higher loan-to-value ratios (LTV - amount of the loan as a % of the value of the collateral). In deals such as DRIVE 2019-1 and SDART 2017-1, the percentage of modifications in April increased by over two times the March levels.

Source: Osterweis and Bloomberg

Looking at the same sample deals and targeting the modified loans within the pools, one can compare the characteristics of modified loans in relation to the overall pool. Although not conclusive at this point, on average, the modified loans had higher LTVs and lower FICOs. Historically, loans with high LTVs and lower FICO scores have performed less reliably, so it is reasonable to conclude that borrowers of loans with these factors are more financially stretched, increasing their need for extensions during the Covid-19 pandemic.

Source: Osterweis and Bloomberg

Going forward, it will be necessary to continue to monitor the modification rates within auto ABS. Although the level of these rates remains elevated, there are reasons for some optimism as used car prices have rebounded. The J.D. Power’s weekly wholesale auction price index continued to recover after dropping as much as 16% and as of the week of May 24th remains only 1.9% below expectations. It is important to reiterate that extensions are a temporary solution to a sudden shock to the economy and the effectiveness of these modifications will depend on the speed of the recovery from the Covid-19 pandemic. As such, extension rates can be an indicator of worse things to come as in the 2008 financial crisis or can be an effective loss mitigation tool, as it was in 2017 for the Harvey and Irma hurricanes, when extension rates quickly reversed after surging.

Source: AmeriCredit

Mortgages

With the dramatic increase in the unemployment rate, the government was quick to respond to assist mortgage borrowers. With the passing of the CARES Act, mortgages backed by the government or government-sponsored enterprises (GSEs) became eligible for forbearance up to 180 days with another 180-day extension available if needed. More recently, the GSEs introduced a payment deferral plan that shifts up to 12 months of deferred monthly payments to the end of the mortgage loan with no interest or fees, effectively giving the borrower a 0% second lien loan that is due at the time of refinance or sale of the property. This payment deferral option will be available to borrowers on July 1, 2020.

Mortgage remittance reporting classifies forbearance as a late payment, so investors can observe increased use of forbearance through delinquency data. Although forbearance practices are standardized in government and GSE backed mortgages, in the non-agency space, it is left to the servicers to decide the appropriate course of action. An example of this can be seen through the large divergences in performance within the RMBS 2.0 cohort as the change in 30-day delinquencies ranged between 1% and 5%. Much of this was driven by bank servicers actively offering forbearance to borrowers, so the magnitude of the change may not be a pure indication of borrower stress. In addition, servicers are reporting that many borrowers in forbearance are making monthly payments. According to Black Knight Inc., approximately 46% of homeowners who had mortgages in forbearance made a full payment in April, suggesting that many borrowers are taking the forbearance option as a precautionary step rather than one that is necessary.

Due to the variety of mortgage loans available to homeowners in the marketplace, the dispersion of the performance of various mortgages has been quite wide as the effects of the Covid-19 pandemic have started to appear in mortgage remittance data this month.

Source: JP Morgain

Non-QM mortgages were hit the hardest with the largest uptick in 30-day delinquencies, increasing from 4.5% in April to 14.3% in May. Non-QM mortgages are loans that do not adhere to the underwriting guidelines of the Consumer Financial Protection Bureau's Qualified Mortgage standard yet still need to be compliant with ability-to-repay (ATR) rules. The largest cohort that accesses this mortgage product is made up of individuals who are self-employed and cannot produce full documentation (W2s and 1099s), relying instead on alternatives such as bank statements, P&Ls, and CPA letters. In addition, other collateral types in Non-QM pools include investor properties, which are underwritten using metrics such as debt service coverage ratio, rather than the financials of the borrower and fully documented loans with borrowers who have had previous credit events. Given the high exposure to self-employed borrowers, Non-QM mortgages have been adversely affected as small businesses have been disproportionately hit during the pandemic. As more data comes available, it will be telling to see how Non-QM performs as the economy reopens.

On the other end the spectrum, RMBS 2.0 (residential mortgage backed securities) had an increase in 30-day delinquencies of just 2.4%. The differential can be explained by the type of borrowers who access these different mortgage products. Collateral in RMBS 2.0 consists of jumbo prime loans which are mortgages that can’t access agency delivery due to the large balance size. Thus, these loans have balances that are larger than the conforming limit of $510,400 and $765,600 in high-costs areas. Borrowers in jumbo prime mortgages have pristine credit metrics including high FICOs, low LTVs, and low DTIs (debt-to-income ratio) and provide full documentation to the lenders in the form of W2s and 1099s. As such, one can assume that borrowers in this cohort are in better financial standing and have more flexibility to withstand the current economic shock.

Finally, we will take a closer look at the performance of Freddie Mac’s CRT (credit risk transfer) STACR (structured agency credit risk) mortgage product which currently has the most detailed loan level May remittance data available.

The May STACR remittance report shows the 30-day delinquency rate at 4.15% in April, which was the second lowest monthly increase after RMBS 2.0 according to JP Morgan. As Freddie Mac issues both HLTV (high loan-to-value) and LLTV (low loan-to-value) STACR deals, it is easy to compare effects on performance based on LTV. LLTV deals had a 30-day delinquency rate of 3.92% which was better than the 4.52% in HLTV deals. When layering in other credit metrics such as debt-to-income (DTI) ratios, it is apparent that the pandemic had a larger adverse effect on borrowers who were more financially stretched (higher DTI) and used more leverage to purchase their homes (higher loan-to-value). The cohort with high LTVs and DTIs over 44 had a delinquency rate of 7.62%. On the other end of the spectrum, the low LTVs and DTIs less than 30 group only had a delinquency rate of 1.89%. A similar pattern is seen when layering in FICO scores. The worst performing group had high LTVs and low FICOs.

There are reasons for some optimism even as the Mortgage Bankers Association (MBA) estimates 8.46% of servicer’s portfolio volume is in forbearance, which equates to approximately 4.2 million homeowners. In the same report, the MBA had noted that the increase in forbearance volumes has slowed and requests/call volumes for forbearance is the lowest since March 2nd. In addition, the latest release of the MBA Purchase Index showed mortgage applications rebounding to near pre-crisis levels.

Source: Mortgage Bankers Association

Conclusion

Going forward, remittance data from ABS and non-agency mortgage deals will continue to allow investors to monitor how borrowers are weathering the disruption from the Covid-19 pandemic. As the country starts to reopen and economic activity restarts, the data will show whether borrowers continue to access assistance through payment deferrals, loan extensions, and forbearance. This will help us get a better understanding of the damage the lockdowns have inflicted on both the economy and more specifically the consumer.

It is clear from the data that many borrowers have taken the opportunity to access Covid-19 assistance programs. As noted previously, these programs are a temporary bridge for borrowers to weather current economic stresses and should allow them to recover if the economy rebounds in short order. If there is a robust recovery, data should show borrowers becoming current on their mortgages, credit card MPRs increasing, and the percentage of auto loan modifications decreasing. If the economy remains anemic for a prolonged period, one could expect delinquencies and modifications becoming credit events. In that instance, we will start to look for divergences in performance among the different consumer loans as borrowers start prioritizing which type of loans (auto vs. credit card vs. mortgage) they try to keep current while letting others fall behind.

Daniel Oh

Vice President & Portfolio Manager