June 2020 performance showed improvement on a month-on-month basis across U.S. auto asset-backed securities (ABS) as losses and delinquencies declined and recoveries improved after hitting record lows in April. On a year-over-year basis, delinquencies decreased substantially, but these were masked by higher-than-normal extension rates. Prime and subprime extension rates, although declining for the second month in a row, remained elevated. Based on Reg AB II loan level data, June extensions for public issuers of prime and subprime auto loans were 4.3x and 5.0x, respectively, February's levels, the last month before COVID-19 (See "To Pay or Not To Pay: Navigating Through U.S. ABS Auto Loans As They Roll Off Extension Status," published Aug. 14, 2020). Taking into account 144a issuers, subprime extensions in June were only 2.4x February's level.

Improved performance and fewer extensions in June were partly due to enhanced unemployment pay of $600 a week, which ended July 31 but was renewed for $300 with the President's executive order on Aug. 8, 2020. June's generally positive results were also attributable to 4.8 million jobs returning and the unemployment rate declining to 11.10% from 13.40% in May.

S&P Global Ratings acknowledges a high degree of uncertainty about the evolution of the coronavirus pandemic. The consensus among health experts is that the pandemic may now be at, or near, its peak in some regions but will remain a threat until a vaccine or effective treatment is widely available, which may not occur until the second half of 2021. We are using this assumption in assessing the economic and credit implications associated with the pandemic (see our research here: www.spglobal.com/ratings). As the situation evolves, we will update our assumptions and estimates accordingly.

Subprime Losses Reach Lowest Level Since 2015

Prime net losses decreased to 0.50% in June 2020 compared with 0.55% in May 2020, but increased from 0.48% in June 2019 (see table 1 and chart 1). A higher recovery rate in June relative to May supported lower losses. Nonetheless, recovery rates remained below last year's levels and contributed to the 2 basis points (bps) increase in losses on a year-over-year basis.

Subprime net losses decreased by 202 bps to 4.20% in June 2020 from 6.22% in May 2020 and by 273 bps from 6.93% in June 2019. June 2020's losses of 4.20% were the lowest in this segment since April 2015's level of 4.31%. Lower losses were due to a substantial improvement in recoveries. The use of extensions also suppressed delinquent accounts charging off, which generally occurs when they become 90- to 120-days late.

Subprime shelves reporting lower losses were Drive Auto Receivables Trust (DRIVE), Santander Drive Auto Receivables Trust (SDART), Exeter Automobile Receivables Trust (Exeter), and Flagship Credit Auto Trust (Flagship), which combined, comprise 53.56% of the total subprime composite. After netting out three deep subprime issuers (American Credit Acceptance, Exeter, and DRIVE), modified subprime losses decreased by 142 bps to 4.04% in June 2020 from 5.46% in May 2020, and 143 bps from 5.47% in June 2019.

Table 1

Subprime Recoveries At Highest Level Since April 2016

Prime recoveries increased 218 bps to 56.78% in June 2020 from 54.60% in May, but declined 340 bps from June 2019's 60.18% rate (see table 2 and chart 2). June's recoveries were mostly influenced by higher recoveries from CarMax Auto Owner Trust (CarMax), Toyota Auto Receivables Owner Trust (Toyota), and GM Financial Consumer Automobile Receivables Trust (GMCAR), which comprise of 39.21% of the total prime index.

Subprime recoveries increased nearly eight percentage points to 50.26% in June 2020 from 42.28% in May 2020. June's recovery rate of 50.26% was the highest since April 2016. Year-over-year, rates increased over six percentage points from 44.10%. Subprime modified recoveries increased 4.7 percentage points to 45.48% in June 2020 from 40.79% in May 2020, and increased 108 bps from 44.41% in June 2019.

For two months in a row recoveries have trended upwards improving from their record lows in April. Several factors have helped strengthen used vehicle values:

• Lack of new and used vehicle supply. New vehicle inventories were down due to a large number of incentives early in the COVID-19 pandemic, which cleared out significant inventory; supply chain issues, which have limited the production of desirable vehicles; more lessees keeping their vehicles when they came off-lease earlier this year; and increased demand for vehicles among city dwellers, some of whom have switched to driving to work instead of public transportation or have moved to the suburbs where they need a vehicle.
• The economic relief checks of $1,200 an individual, enhanced unemployment benefits, and low interest rates have made vehicle ownership more attractive for many, thereby increasing demand.

Table 2

Subprime Delinquencies At Their Lowest Level Since April 2014.

The prime 60-plus-day delinquency rate remained essentially flat at 0.33% in June 2020 versus 0.34% in May 2020, but it decreased six bps from 0.39% in June 2019.(see table 3 and chart 3).

The subprime 60-plus-day delinquency rate decreased to 3.34% in June 2020 from 3.70% in May 2020 and from 5.00% in June 2019. June's 3.34% level was the lowest subprime delinquency rate since April 2014. Delinquencies are lower than they otherwise would be due to higher-than-normal extension rates. Accounts that receive deferrals or extensions are usually listed in the servicer's accounting records as being current.

On a modified basis, after netting out three deep subprime lenders, the subprime modified 60-plus-day delinquency rate decreased to 2.64% in June 2020 from 2.92% in May 2020, and from 3.60% in June 2019.

Table 3

Extensions Decline Month-On-Month

Extensions declined in June 2020 for public prime issuers to 1.39% from 2.16% in May 2020 and for subprime public shelves (DRIVE, SDART, AmeriCredit Automobile Receivables Trust, and World Omni Select) to 7.66% in June 2020 from 8.90% in May 2020, based on their Reg AB II loan level data (see table 4). When we account for the outstanding 144a subprime transactions that we've rated, the subprime monthly extension rate fell to 5.69% from 7.62% in May. For transaction-specific extension data for subprime transactions, please see Appendix II.

When we account for several months of elevated multi-month extensions, the percentage of loans in extension status (on a dollar basis) in the prime sector was 3.65% and 16.29% for public subprime shelves as of June 30. This clearly shows that a fair number of consumers are not yet willing or able to resume their monthly scheduled payments. In addition, extensions could increase in August due to the time lapse between when the enhanced unemployment benefits ended (July month-end) and when President Trump's executive order is implemented. Additionally, the order to maintain them, at a lower federal amount of $300, could prove insufficient for many to pay all of their living expenses. For more on the performance of these loans as they come out of extension status and roll through to the delinquency buckets, see "To Pay or Not To Pay: Navigating Through U.S. ABS Auto Loans As They Roll Off Extension Status," published Aug. 14, 2020.

Table 4

Collateral Characteristics: The First Of The Post-COVID-19 Deals

First half of 2020 consisted of loans that were originated prior to the pandemic and during the pandemic, when the unemployment level peaked at 14.70% in April (although it dropped to 11.20% in June). Issuance volume through June dropped on a year-over-year basis. Prime deals in our index equaled $20.5 billion in collateral across 16 deals compared with $26.7 billion for the first six months of 2019 across 16 deals. Subprime deals equaled $11.1 billion across 19 deals compared with $13.1 billion a year earlier across 20 deals.

Prime

In our view, prime collateral characteristics remained strong in the first half of 2020. The weighted average FICO increased to 758, the highest since we started to track collateral characteristics, up from 754 in 2019; and the percentage of loans with greater than 60-month loans declined to 59.51% compared with 60.51% in 2019. A slight negative was that the weighted average loan-to value (LTV) ratio increased to 96.83% from 95.64% in 2019, but it remains below 2015 and 2016's levels. Those years were marked by intense competition and many lenders subsequently tightened their credit standards in 2017.

Prime securitizations completed in second-quarter 2020, after the COVID-19 pandemic began, appear marginally stronger in some respects than those in first-quarter 2020. The weighted-average FICO increased to 760 (from 758 as of first-quarter 2020), the weighted-average original maturity declined to 65.7 months (from 67.1), and the percentage of loans with 73-75 month maturities declined to 4.71% (from 10.86%).

Table 5A

Subprime

In our view, subprime collateral credit quality improved from the first quarter of 2020 to the second quarter. The weighted-average FICO increased to 592 from 575 and the weighted-average original maturity declined to 67.2 months from 68.4 months. In addition, the weighted-average LTV declined to 112.53% from 116.22% as the following reduced their LTVs relative to the first quarter: American Credit Acceptance, Flagship, Exeter, CPS Auto Receivables Trusts (CPS), and DRIVE. Only GLS Auto Receivables Trusts reported an increase in their weighted-average LTV quarter-to-quarter (to 121.31% from 119.68%). Additionally most companies increased their emphasis on income and job verification processes. Other noteworthy collateral changes include:

• Reflecting the high demand for used vehicles in the post-COVID-19 pandemic environment, used vehicle financing increased to 77.80% in the first-half 2020 up from 70.55% in 2019;
• To account for the weaker economic outlook and potentially higher losses, the weighted average annual percentage rate (APR) increased to 18.72% from 17.84% in 2019; and
• Loans with original maturities of 73-75 months represented 5.13% of the loans securitized in the first half of 2020, down from 8.14% in 2019.

Table 5B

Auto Loan Static Index: Vintage Analysis-–Too Early To Determine COVID-19 Pandemic Impact

Prime

Prime auto loan ABS issuers' tighter underwriting standards and more selective criteria for their 2017 and 2018 securitizations are still generating positive results. The 2017 vintage with 32 months of performance is experiencing lower losses than 2016 (0.79% for 2017 compared to 0.85% for 2016). Also, the 2018 vintage, with 19 months of performance, has incurred 0.53% of cumulative net losses (CNLs) compared with 0.55% at the same seasoning point for 2017 vintage and 0.60% for 2016.

The 2019 vintage with seven months of seasoning has incurred 0.19% of CNLs, which is in-line with the 2018 and 2017 vintage at the same seasoning point. However, it's too early to draw conclusions from the limited data, especially since higher-than-normal levels of extensions have kept delinquencies and losses lower than they otherwise would have been.

Subprime

Subprime CNLs, in aggregate, have been fairly stable for the 2015-2017 vintages, and the 2018 and 2019 vintages are demonstrating a slight improvement relative to the 2015-2017 vintages. The 2018 cohort has CNL of 7.47% at month 19 compared with 7.88%-8.00% for 2015-2017. The 2019 cohort is trending below the 2015-2017 vintages with 2.46% in losses at month eight compared with 2.52%-2.72% for the earlier vintages. Currently, the 2019 cohort is tracking even with the 2018 vintage.

As noted above, it's too early to draw conclusions from 2019's limited track record, especially given subprime's elevated extension levels. In addition, we have noticed that the 2019 cohort of deals are paying down more slowly than earlier vintages, likely due to high deferments, and these deals may experience more back-loaded loss curves than earlier transactions.

Issuer-Specific Cumulative Net Loss Index Data

We track CNLs by vintage for a number of issuers and compare those losses to the prime and subprime Auto Loan Static Index (ALSI) (see tables 6A and 6B).

A number of prime issuers reported higher losses on their 2016 securitizations at month 36 than their 2015 deals. CarMax's 2016 vintage experienced 2.06% losses, whereas its 2015 pools averaged 1.91% at month 36. Similarly, World Omni Auto Receivables Trust's (World Omni's) losses rose to 2.33% for its 2016 vintage compared with 1.70% in 2015. However, World Omni's prime pool performance has recently improved due to the following: beginning in 2017, it added a minimum FICO score of 650 to its prime pools; beginning in 2018, it excluded non-Toyota vehicles; and in 2019, it created a separate issuing shelf, World Omni Select Auto Trust, to securitize its nonprime/subprime auto loans.

Most prime issuers are reporting satisfactory results for their 2017 through the first-half 2019 pools. However, Bank of the West's 2017 vintage shows weakening in performance with losses of 1.52% at month 31 compared with 0.52% at month 36 for its 2015 vintage, and its second-quarter 2019 deal at month 12 is performing worse than 2018 at month 18. Similarly, Nissan's first-quarter 2019 vintage at month 15 with 0.75% in losses has already surpassed its 2018 vintage's losses of 0.68% at month 18. However, Nissan's second-quarter 2019 vintage at 12 months is trending in line to slightly better than first-quarter 2019 at the same point.

Most of the subprime issuers' 2016 vintages performed worse than those from 2015. Their 2017 vintages are generally performing in-line-to-better than their respective 2016 vintages. The 2018 deals are generally performing within satisfactory bounds. It's difficult to gauge at this point how the 2019 deals will perform given their young status and because high levels of extensions has limited the amount of losses that they would otherwise have incurred at this point. Expected high unemployment levels through the rest of the year could also negatively impact these transactions, as well as the reduction in unemployment benefits.

Table 6A

Table 6B

Recent Auto Loan ABS Rating Activity/Revised Loss Expectations

In July, we revised our loss expectations and took the following rating actions:

• Raised ratings on three classes and affirmed 25 ratings across eight Toyota Auto Receivables Owner Trust backed by prime auto loans (see "Three Ratings Raised And Twenty Five Ratings Affirmed On Eight Toyota Auto Receivables Owner Trust Transactions," published July 30, 2020).
• Raised ratings on seven classes and affirmed three ratings across two United Auto Credit Securitization Trust backed by subprime loans (see "Seven Ratings Raised, Three Affirmed On Two United Auto Credit Securitization Trust Transactions," published July 28, 2020).
• Affirmed ratings on six classes from two Fifth Third Auto Trust deals backed by prime loans (see "Ratings Affirmed On Six Classes From Two Fifth Third Auto Trust Transactions," published July 14, 2020).
• Raised our rating on one class and affirmed 15 ratings across four Canadian Pacer Receivables Trust transactions backed by Canadian prime loans (see "One Rating Raised, 15 Affirmed On Four Canadian Pacer Auto Receivables Trust Transactions," published July 13, 2020).
• Affirmed our ratings on nine nine classes across four Securitized Term Auto Receivables Trust transactions backed by Canadian prime loans (see "Ratings Affirmed On Nine Classes From Four Securitized Term Auto Receivables Trust Transactions," published July 13, 2020).
• Raised ratings on five classes and affirmed five ratings across three Santander Drive Auto Receivables Trust deals backed by subprime auto loans (see "Five Ratings Raised, Five Affirmed On Three Santander Drive Auto Receivables Trust Transactions," published July 13, 2020).
• Raised our ratings on 17 classes and affirmed seven ratings across six American Credit Acceptance Receivables Trust transactions backed by subprime loans (see "Seventeen Ratings Raised and Seven Affirmed On Six American Credit Acceptance Receivables Trust Deals," published July 8, 2020).

Out of the 29 transactions that we reviewed, we lowered our revised expected cumulative net losses (ECNL) on 12, raised them on 12, and maintained them on five. The upwardly revised ECNLs are primarily on the second-half 2018 and more recent transactions, and take into account potential deterioration resulting from high unemployment levels. In fact, most of the transactions are not currently trending toward the higher loss levels, but their higher-than-historical extension levels and future economic uncertainty led us to increase our base case. Even with these upward adjustments, we upgraded 33 classes and affirmed 67 in July.

Table 7

Table 8

Table 9

Table 10

Table 11

Table 12

Table 13

Historical Ratings Activity

Through July 31, 2020, our reviews resulted in 127 upgrades and no downgrades. Subsequent to July 31, we took rating action on six CPS transactions. These actions consisted of lowering our ratings on seven classes, raising them on five, and affirming them on five (see "Various Rating Actions Taken On Six CPS Auto Receivables Trust Transactions," published Aug. 7, 2020). These actions are not reflected in the table below. On Aug. 10, 2020, we renewed the CreditWatch negative action that we took on May 12, 2020, keeping most of those classes on CreditWatch with negative implications (see "Twenty-Seven Ratings From 20 U.S. Subprime Auto ABS Transactions Remain On CreditWatch Negative," published Aug. 10, 2020). All of the class are nonIG, and most are from the second-half 2019 to the first-quarter 2020.

Table 14 shows S&P Global Ratings' historical ratings activities for U.S. ABS auto loans.

Table 14

Appendix I: ALSI Performance Data

Table 15

Table 16

Table 17

Table 18

Table 19

Table 20

Table 21

Table 22

Table 23

Appendix I: Subprime Extensions

Table 24

Appendix II: Auto Tracker Methodology And Definitions--Frequently Asked Questions

How do you define prime auto loan ABS?

We generally categorize prime auto loan ABS transactions as those backed by loan pools with initial expected CNLs of 3.00% or less, average FICO scores of 700 or higher, and APRs of 0.00%-5.00%.

How do you define subprime auto loan ABS?

We generally categorize subprime auto loan ABS transactions as those backed by loan pools with initial expected CNLs of at least 7.50%, average FICO scores of less than 620, and APRs that exceed 14.00%.

How do you calculate the monthly net loss rate?

The monthly net loss rate is annualized. It equals each transaction's net loss rate weighted by the transaction's ending pool balance for the current month over the aggregate ending pool balance of all transactions included in the index.

We only allow a transaction to enter the composite starting in its fourth month outstanding. Transactions usually have zero or low losses during their first three months, which dilutes the composite figures.

How do you calculate the monthly recovery rate?

We calculate recoveries by taking the recovery amount reported (which typically includes all recoveries, including disposition proceeds, post-disposition proceeds, and any other reported recoveries) over the gross loss amount for the current month. We then weight each transaction's recovery percentage by the transaction's ending pool balance for the current month over the aggregate ending pool balance of all transactions included in the index.

We only allow a transaction to enter the index starting in its fourth month outstanding. During a transaction's first three months, unusually high or low recoveries are reported, leading to a spike in the composite figures.

How do you calculate the monthly 60-plus-day delinquency rate?

We calculate delinquencies by taking each transaction's 60-plus-day delinquency amount over the ending pool balance for the current month. We then weight each transaction's 60-plus-day delinquency percentage by the transaction's ending pool balance for the current month over the aggregate ending pool balance of all transactions included in the composite.

We only allow a transaction to enter the composite starting in its fourth month outstanding. During the transaction's first three months, zero or lower delinquencies are reported, which dilutes the composite figures.

What is the ALSI?

Our ALSI monitors the credit performance of securitizations that were originated in the same year on a weighted average basis. The number of months displayed for each vintage is generally determined by the last point that all securitizations for that time period have a data point. We calculate the prime and subprime ALSI CNLs by taking the weighted average of the CNLs of the transactions that were completed in the same time period (generally a year). Each transaction's CNL is weighted by its initial pool balance over the aggregate initial pool balance of all the transactions included in the index for that period. In the subprime ALSI, transactions from Byrider Finance LLC (doing business as CarNow Acceptance Corp.), Credit Acceptance Corp., and DriveTime Automotive Group Inc. are excluded because they do not have the typical indirect auto loan business model.

Which transactions are included in the prime, subprime, and modified subprime composites and indices?

For a list of the transactions included in our prime, subprime, and modified subprime composites and indices, see "U.S. Auto Loan ABS Tracker: March 2019," published May 23, 2019. However, note that we subsequently added S&P Global Ratings-rated transactions that have since closed, most prime transactions that closed and were not rated by S&P Global Ratings from 2016 through the present, and most Santander and AmeriCredit transactions non rated by S&P Global Ratings.

Related Research

• To Pay or Not To Pay: Navigating Through U.S. ABS Auto Loans As They Roll Off Extension Status, Aug. 14, 2020
• U.S. Biweekly Economic Roundup: The Jobs Recovery Loses Momentum, Aug. 7, 2020
• U.S. Real-Time Economic Data Signals A Faltering Recovery, Aug. 3, 2020
• Three Ratings Raised And Twenty Five Ratings Affirmed On Eight Toyota Auto Receivables Owner Trust Transactions, July 30, 2020
• Seven Ratings Raised, Three Affirmed On Two United Auto Credit Securitization Trust Transactions, July 28, 2020
• Ratings Affirmed On Six Classes From Two Fifth Third Auto Trust Transactions, July 14, 2020
• Five Ratings Raised, Five Affirmed On Three Santander Drive Auto Receivables Trust Transactions, July 13, 2020
• Ratings Affirmed On Nine Classes From Four Securitized Term Auto Receivables Trust Transactions, July 13, 2020
• One Rating Raised, 15 Affirmed On Four Canadian Pacer Auto Receivables Trust Transactions, July 13, 2020
• Seventeen Ratings Raised and Seven Affirmed On Six American Credit Acceptance Receivables Trust Deals, July 8, 2020
• The Potential Effects Of COVID-19 On U.S. Auto Loan ABS, March 26, 2020
• How The Wave Of Negative Rating Actions On Global Automakers Has Affected U.S. Auto ABS Ratings, Feb. 13, 2020