U.S. auto loan asset-backed securities' (ABS) performance continued to improve in August as losses decreased month over month for prime, and to an all-time low for subprime, for the second month in a row. Prime recoveries increased to their highest levels seen in over eight years, and subprime recoveries rose to their highest observed level. Delinquency levels increased month over month, their first such increase in five months, while extensions remained higher than pre-COVID-19 levels despite a modest drop month over month. We believe the increase in delinquencies reflects the reduction of enhanced unemployment benefits to $300 per week beginning August from $600 per week, consumers emerging from extension status, and a resurgence of COVID-19 cases in certain states; this could potentially lead to higher losses in the months ahead.

S&P Global Ratings acknowledges a high degree of uncertainty about the evolution of the coronavirus pandemic. The current consensus among health experts is that COVID-19 will remain a threat until a vaccine or effective treatment becomes widely available, which could be around mid-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.

Prime And Subprime Losses Decrease Further

Prime net losses decreased to 0.25% in August, their lowest level since May 2012, from 0.38% in July 2020 and 0.57% in August 2019 due to higher recovery rates (see table 1 and chart 1). Several issuers reported lower losses in August relative to July, including Toyota Auto Receivables Owner Trust (Toyota), CarMax Auto Owner Trust (CarMax), Honda Auto Receivables Owner Trust (Honda), World Omni Auto Receivables Trust (WOART), Ford Credit Auto Owner Trust (Ford), Ally Auto Receivables Trust (Ally), and Capital One Prime Auto Receivables Trust (CapOne), which together comprise approximately 66% of the total prime composite.

Subprime net losses decreased to their lowest observed level, at 2.93% in August, from 3.16% in July 2020 and 8.75% in August 2019. The continued decline in losses was due largely to the rise in extensions, which has since abated, albeit not to pre-COVID-19 levels, and to higher recovery rates resulting from demand for affordable used vehicles outstripping supply. This is reflected in the rolling seven-day estimate of used retail supply at 38 days compared with a peak of 115 days on April 8, 2020, and a normal average of 44 days' supply based on Manheim's mid-August market report.

Approximately 73% of transactions (by volume) comprising the subprime composite experienced lower losses in August. Among these were Drive Auto Receivables Trust (DRIVE), Santander Drive Auto Receivables Trust (SDART), and AmeriCredit Automobile Receivables Trust (AmeriCredit), which together comprise approximately 51% of the subprime group. After netting out three deep subprime issuers (American Credit Acceptance, Exeter Automobile Receivables Trust [Exeter], and DRIVE), modified subprime losses also decreased by 25 basis points (bps) to their lowest levels since February 2012, at 2.71% in August, from 2.96% in July 2020, and by 388 bps from 6.59% in August 2019.

Table 1

Chart 1

Recovery Rates Increased Across The Board

Prime recoveries increased 17.76 percentage points to 80.54% in August (their highest level since March 2012) from 62.78% in July 2020, and 21.13 percentage points from 59.41% in August 2019 (see table 2 and chart 2). August's recoveries were mostly influenced by Toyota, CarMax, Honda, WOART, Ford, Ally, and CapOne, which comprise approximately 62% of the prime composite.

Subprime recoveries increased 4.93 percentage points to 65.02% in August (their highest observed level) from 60.08% in July 2020. Year over year, rates increased 22.07 percentage points from 42.95%. Subprime modified recoveries increased 4.37 percentage points to 60.33% in August (their highest rate since April 2015) from 55.96% in July 2020, and increased 17.92 percentage points from 42.41% in August 2019.

Table 2

Chart 2

Delinquencies Increased As Elevated Loan Extensions Abate

The prime 60-plus-day delinquency rate increased to 0.36% in August from 0.32% in July 2020 (which was its lowest level since March 2015), but remained below August 2019's 0.41%(see table 3 and chart 3).

The subprime 60-plus-day delinquency rate increased to 3.72% in August from 3.22% in July 2020 (which was its lowest level since March 2014), but remained below August 2019's 5.25%. On a modified basis, after netting out three deep subprime lenders, the subprime modified 60-plus-day delinquency rate increased to 2.76% from 2.45% in July 2020 (which was its lowest level since April 2012), but remained below August 2019's 3.82%.

Table 3

Chart 3

Extensions Declined Month Over Month But Are Still Higher Than February's Pre-COVID-19 Levels

Loan extension levels contracted further in August relative to July's levels, but remain higher than February's pre-COVID-19 levels (see table 4 and Appendix I). Prime loan extensions, based on Regulation AB (Reg AB II) data, decreased to 0.63% in August from 0.82% in July, and from the 5.76% peak in April 2020, but still remained higher than February's level of 0.32%. Meanwhile, loan extensions for the four public subprime platforms that file Reg AB II loan-level data (DRIVE, SDART, AmeriCredit, and World Omni Select Auto Trust 2019-A) declined to 3.12% in August from 4.89% in July, and 15.75% at their peak in April 2020, but they also remained higher than February's level of 1.53%. Including 144a subprime transactions, subprime extensions decreased to 3.26% in August from 4.27% in July, and from the 13.26% peak in April 2020, though they continued to track higher than February's 2.40% level. CPS Auto Receivables Trust (CPS), Exeter, and Westlake Automobile Receivables Trust (Westlake) extensions increased considerably month over month.

Table 4

Recent Auto Loan ABS Rating Activity/Revised Loss Expectations

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

• Raised our ratings on 15 classes and affirmed four ratings on six DT Auto Owner Trust Transactions (see "Fifteen Ratings Raised, And Four Affirmed On Six DT Auto Owner Trust Transactions," published Sept. 2, 2020).
• Raised our ratings on five classes and affirmed four ratings on four Drive Auto Receivables Trust Transactions (see "Five Ratings Raised, Four Affirmed On Four Drive Auto Receivables Trust Transactions," published Sept. 11, 2020).
• Raised our ratings on 14 classes and affirmed 17 ratings on eight Flagship Credit Auto Trust Transactions (see "Various Rating Actions Taken On Eight Flagship Credit Auto Trust Transactions," published Sept. 11, 2020).
• Affirmed our ratings on six classes of the Prestige Auto Receivables Trust 2019-1 transaction and removed one rating from CreditWatch Negative (see "Six Prestige Auto Receivables Trust 2019-1 Ratings Affirmed; Class E Removed From CreditWatch Negative," published Sept. 23, 2020).

Of the 19 transactions we reviewed, we raised our expected cumulative net losses (ECNLs) on 11 and lowered them on eight (see tables 5-8). Most of the deals for which we increased our ECNLs came from Flagship, and their deals were already performing worse than expected before the pandemic. Our revised ECNLs also considered each issuer's recent extension history, our expectation that remaining losses could increase approximately 15% over current projected levels (based on performance through July) due to the ongoing recession, and high unemployment levels.

Table 5

Table 6

Table 7

Table 8

Historical Ratings Activity

Through Sept. 30, 2020, our surveillance reviews resulted in 187 upgrades and eight downgrades. The downgrades are related to our rating actions on CPS-related transactions dated Aug. 7, 2020 (see "Various Rating Actions Taken On Six CPS Auto Receivables Trust Transactions" and "Nine Ratings Raised, One Lowered, 11 Affirmed On Five CPS Auto Receivables Trust Transactions," both published Aug. 7, 2020). On Aug. 10, 2020, we extended the CreditWatch 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 classes are non-investment grade, and most were issued from the second half of 2019 through the first quarter of 2020.

Table 9 shows S&P Global Ratings' historical ratings activities for U.S. ABS auto loans.

Table 9

Appendix I

Table 10

For prime extension rates, see "Pace Of New U.S. Auto Loan ABS Extensions Continued To Decline In July, But Delinquencies Inched Up," published Sept. 22, 2020.

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 cumulative net losses (CNLs) of 3.00% or less, average FICO scores of 700 or higher, and annual percentage rates (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 S&P Global Ratings' auto loan ABS 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 Auto Loan Static Index (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 the Santander transactions S&P Global Ratings did not rate.

Related Research

• U.S. Real-Time Data: The Economic Recovery Decelerates, Oct. 8, 2020
• A Double-Digit Rebound Has Begun, But It's No Time To Celebrate, Oct. 6, 2020
• U.S. Biweekly Economic Roundup: The Coast Is Not Clear For Jobs, Oct. 3, 2020
• Economic Research: The U.S. Economy Reboots, With Obstacles Ahead, Sept. 24, 2020
• Six Prestige Auto Receivables Trust 2019-1 Ratings Affirmed; Class E Removed From CreditWatch Negative, Sept. 23, 2020
• Pace Of New U.S. Auto Loan ABS Extensions Continued To Decline In July, But Delinquencies Inched Up, Sept. 22, 2020
• Various Rating Actions Taken On Eight Flagship Credit Auto Trust Transactions, Sept. 11, 2020
• Five Ratings Raised, Four Affirmed On Four Drive Auto Receivables Trust Transactions, Sept. 11, 2020
• Fifteen Ratings Raised, And Four Affirmed On Six DT Auto Owner Trust Transactions, Sept. 2, 2020
• Twenty-Seven Ratings From 20 U.S. Subprime Auto ABS Transactions Remain On CreditWatch Negative, Aug. 10, 2020
• To Pay or Not To Pay: Navigating Through U.S. ABS Auto Loans As They Roll Off Extension Status, Aug. 14, 2020
• Various Rating Actions Taken On Six CPS Auto Receivables Trust Transactions, Aug. 7, 2020
• Nine Ratings Raised, One Lowered, 11 Affirmed On Five CPS Auto Receivables Trust Transactions, Aug. 7, 2020
• U.S. Auto Loan Extension Congestion Partially Clears Up In May, June 22, 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

The analysts would like to thank Sundar Ganesh for his contribution to the report.