On Tuesday we brought you “Car-mageddon: Your Complete Guide To A Bursting American Auto Bubble.

In it, we reviewed our previous posts on the US auto bubble and specifically, on the role subprime has played in helping this $1 trillion+ bubble to inflate.

You’ll recall that way back in February, we highlighted a Morgan Stanley note which took a look at subprime originators, their market share, and the extent to which vintages tied to increased lending by companies with lower standards affects performance in auto ABS. Here’s an excerpt from our post:

• Delinquencies and default rates are climbing in the Auto ABS sector. While these metrics have not yet deteriorated as far as they did during the financial crisis, the YoY increases are meaningful, especially coming at a time when the rest of the consumer balance sheet – outside of student loans – appears to be performing well.
• The cracks in fundamentals can be partly attributed to the shift-in-mix of origination share away from lenders with stronger borrower profiles.
• Some issuers have lower underwriting criteria and the share of Auto ABS originated by these issuers is trending higher
• In the tables below, we identify issuers with lower underwriting standards (shaded in red). The share of outstanding deals originated by issuers with lower underwriting standards has increased from 23.1% and 3.4% in Jan 2010 to 33.9% and 11.2% in Oct 2016, for Prime and Subprime Auto ABS respectively.

Meanwhile, auto ABS makes up a decidedly large percentage of the overall ABS market. As Citi wrote in November, “auto ABS is one of the largest sectors of the ABS market, representing 42–56% of all consumer ABS primary market issuance since 2009.”

Here’s a look at supply and how much of this paper is outstanding:

(Citi)

As noted by Morgan, delinquencies and default rates are climbing and you’ll never guess what the bank found when they checked to see if perhaps there was a correlation between the share of issuance from originators with lower lending standards and cumulative losses. To wit:

• As we dissect the performance of recent vintages, we find divergence in cumulative loss trajectories across various origination quarters. Not surprisingly, those are the quarters with higher origination share from lenders with lower underwriting criteria.
• Given the high proportion of 2016 subprime deals issued by lenders with lower underwriting standards, we should expect delinquencies and losses to climb even higher over the course of 2017.

Ok. So given all of that, you’ll probably find a new note out on Wednesday morning from Goldman to be of interest.

Essentially, Goldman tells us what we already knew. Namely that weaker underwriting is at the heart of the problem. But the interesting bit is about the growth of deep-subprime and how it’s being embedded in the system via Wall Street’s securitization machine. Of course the bank’s analysis comes with the usual caveats that i) this isn’t big enough to be systemic, and 2) delinquency rates are unlikely to climb too high because after all, people need their cars. Here’s more.

Via Goldman

Subprime mortgage lending in the US declined sharply in 2007-2008 in the early months of the global financial crisis, and has not recovered since. In 2006, non-prime private label RMBS issuance volume was $940bn, vs. less than $2bn in 2016. However, while recent non-prime mortgage origination volumes are still quite small, non-prime auto lending in the US has remained relatively robust, with over $110bn in originations in 2016 for borrowers with credit scores below 620, a cutoff commonly signifying “subprime”. The overall auto loan market in the US is now over $1.1tn in debt outstanding, 40% above 2008 levels. By comparison, US mortgage debt outstanding has contracted by 9% over the 2008-2017 period.

As auto lending volumes have expanded, delinquency and loss rates on auto loans have also begun to increase. Exhibit 1 shows, for example, the trend towards rising delinquency rates among outstanding auto loans. Aggregate loss rates for auto asset backed securities (ABS) have also been picking up. Does this deterioration in auto loan performance reflect broader problems with the US consumer and US economy? We think the answer is no; rather, we see rising auto loan loss rates as more reflective of conditions specific to the auto loan sector. Specifically, the higher loss rates are largely a reflection of a loosening of underwriting standards for car lending.

Exhibit 2 shows the distribution of credit scores among US auto loan borrowers. The chart would seem to suggest that the share of borrowers with a credit score below 620 has been stable, if not falling, over the past year. However, the chart masks the growth of “deep subprime” lending programs, targeting borrowers with credit scores far below the 620 cutoff. For example, some recent deep subprime ABS transactions featured pools with average credit score of just 545, with 20% of borrowers not having a credit score at all, a condition known to signal high default risk. Further, lending standards have loosened on dimensions beyond credit score. For example, the maximum term for auto loans has been extended from 60 months to 72 months or more for many recent transactions. When holding collateral quality roughly constant (e.g., by looking at a fixed non-deep subprime issuance shelf), delinquency trends have increased for 2016 vintage deals, but not by a large extent, suggesting that most of the deterioration in aggregate industry performance is explained by a shift towards more deep subprime deals.

Exhibit 3, below, shows the significant variation in performance across ABS deals labelled as “subprime”. The overall cumulative loss rate on 2014 vintage subprime deals has passed 9%. By comparison, the best performing subprime issuance shelf has just 6% cumulative losses, while the highest loss issuer, representing deep subprime, has cumulative losses above 28%. The growth in industry loss rates is explained to a significant extent by the shift towards more deep subprime, high default rate lending.

The increase in auto loan delinquency rates is an important trend to monitor, but we do not think it represents a source of significant systemic risk. For one, Exhibit 1 shows that aggregate auto loan delinquency rates – including prime, subprime and deep subprime loans – are, while rising, still well below delinquency rates on credit cards. The lower default rates on auto loans reflects the fact that borrowers frequently tend to prioritise auto loan payments ahead of other accounts. Also, Exhibit 4 shows that aggregate measures of US consumer debt burden, including debt service ratio and a broader measure of financial obligations including rents and insurance payments, remain low by long-run standards. The low debt ratios are a reflection of the low interest rates on consumer and mortgage loans, and also of the fact that aggregate consumer debt balances remain below 2008 levels. As most US consumer and mortgage debt is fixed rate, we expect these low debt ratios to be sustained in the coming years if, as we expect, interest rates rise.