Auto Loan Securitizations – Mid-year Update
Earlier this year, we explored data on securitized auto loans, the data having been made publicly available for the first time following implementation of the SEC’s asset-level disclosure requirements. As more auto loan issuers have been reporting data and as more time has passed, the quantity of available information has increased, and we now have the opportunity to conduct additional analysis. In addition to the overall market, we’ll provide a glimpse into the subprime segment, with further detail to be explored in a later post.
Since January, the quantity of auto loan data in the Orchard database has increased substantially, now totaling $67.4 billion as of June 2017. While in our earlier post, we analyzed only the data from 6 issuers, we will analyze 11 issuers in today’s post, across 44 deals. With significantly more data in our system, we can explore the auto loan industry in greater depth.
Volumes & Population
In the chart below, we show the history of auto loan originations over time, displayed in terms of the original loan amount and remaining balance. As we can see, most of the reported volume was originated in recent quarters, as the asset-level reporting rules only took effect in late 2016.
Stats by Issuer
There are currently 11 issuers in our dataset, ranging from Toyota Auto with $11.54 billion in loan volume to California Republic with $429 million in loan volume.
In the graph below, we visualize the distribution of borrower credit scores for each issuer. As we can see, each issuer addresses its own particular, though in many cases wide, part of the credit spectrum. For instance, Honda Auto has the highest median credit score, while Santander and Americredit are clearly focused on a less-prime segment of the market.
We can see that issuers also differ in the lengths of loan term they tend to offer their borrowers. Interestingly, the more prime issuers tend to offer shorter-duration loans, while the more subprime issuers tend to offer longer-duration loans. This makes intuitive sense, considering that borrowers who are less financially secure may wish to have a lower monthly payment, which would correspond to a longer loan term, all other things being equal.
Stats by Vehicle Make
The asset-level data made available by the new regulations contain a wealth of vehicle-specific information, which is all the more interesting if you are both an automotive and financial enthusiast. The downside of this data is in its lack of consistency and prevalence of spelling or transcription errors, requiring a decent amount of text-processing logic to overcome. For example, the brand name “CHEVROLET” appeared in the combined dataset under 16 separate variations, which included “CHEVRLET”, “CHEVRO”, “XHEVROLET”, and “CHEVROELT”. We were able to use text processing and word-similarity algorithms to generalize these names in order to explore vehicle data more cleanly.
In the graph below, we see the average loan amount by vehicle make. What immediately stands out is the bar for Maserati, which is nearly double the average for many of the other automakers, even those who tend to focus on luxury vehicles.
Next, we see the distribution of credit scores for each major manufacturer, along with a vertical line on each violin plot to denote the median. It is quite interesting how different the distributions are by car type, with Lexus owners having a median score of 757 and Suzuki owners having a median score of 607. Also notable is the difference in credit profiles among owners of various luxury brands. While brands such as Lexus and Infiniti tend to attract borrowers with very high credit scores, Mercedes and BMW’s median scores are substantially lower, at 650 and 649 respectively. Of course, it’s entirely possible that this phenomenon is due to some other factor, such as a large proportion of high-credit-score BMW and Mercedes buyers either paying cash or financing with an entity that does not issue a public securitization of the assets.
Of course, not every aspiring car owner possesses strong enough credit to qualify on his own. The chart below shows the prevalence of co-signers on auto loans for each different vehicle make.
If you’ve spent a lot of time driving in different parts of the United States, you’ve likely noticed some difference in the cars people drive across the country. We’ve visualized this effect in the animation below.
Credit and Performance
As one might imagine, borrowers with better credit scores are able to finance their vehicles at more attractive rates. The graph below visualizes the distribution of annual interest rate by credit score band. What is striking is the availability of financing across the entire spectrum of consumer credit, although clearly at prices that compensate for increased risk at the lower end.
In the graph below, we show the percentage of balances greater than 30 days past due, which have been increasing over time. Part of this phenomenon may be due to a greater proportion of subprime loan origination (more on that below), but it is important to note that as many of these loans are recent originations, it is still too early to measure long-term performance, and we will need to monitor this over time.
When borrowers experience financial difficulty, they sometimes negotiate loan modifications with their issuer or servicer that allow them to more easily make their payments. These include term extensions, rate changes, payment forbearance, and other programs. While modifications help borrowers to stay current, the need for them does reflect some underlying weakness in a customer base and should be monitored, even if these loans are technically marked as “current”.
There have been news stories as of late pointing out the low rate of income verification in the auto loan market. In the linked article, the author discusses the discrepancy between Santander and Americredit, which is backed up by our graph below. Beyond those 2 particular issuers, the striking takeaway from this graph is just how few borrower incomes are verified across the industry, with most issuers not verifying income on nearly anyone.
While the graph above may be alarming to some, the graph below presents a more nuanced view. It is clear that income verification is substantially more prevalent among borrowers with lower credit scores and nearly absent among those with higher scores. It seems that issuers are making a cost-benefit calculation as to whether the expense and time associated with income verification are worthwhile and generally opting to do so only on riskier applicants.
Within the automotive lending market, the subprime segment has recently gotten a lot of attention. From stories about the speed of growth to concerns about underwriting standards and performance, there are plenty of eyes watching this space, and we can gain some insight from exploring our dataset.
The graph below shows origination volumes, broken out by prime and subprime. As we can see, subprime lending has gone from being a relatively small piece of the market to being even larger than prime in the first half of 2017. Of course, we should keep in mind that there is likely some bias in our sample, as we are only looking at securitized loans and not including those that are funded privately or sitting on a bank’s balance sheet.
While we will delve deeply into the characteristics of these loans in a subsequent analysis, we will focus here on performance. Below is a graph showing the delinquency status of subprime vs. prime auto loans, displaying a large difference between the two populations. As of July 2017, 0.6% of prime loans are past due, while 6.0% of subprime loans are past due, a difference of 544 bps.
The great American love affair with the automobile is legendary, and the auto loan market provides a helpful window into consumer behavior and the financial health of American households. As we’ve seen through this analysis, the appetite for credit has increased quite consistently over the past decade, and the availability of credit has risen to match that demand. While growth can often be a good sign, it also raises important questions about the credit quality of the loans, the value of the collateral, underwriting, verification, and servicing practices. It’s worthwhile to question if this level of growth is sustainable and to closely monitor the industry for signs of weakness or deterioration. Thanks to this vast and constantly growing dataset, we can now conduct this monitoring with greater precision.