The 2023 outlook for auto lending
Warnings of an auto loan bubble, or new subprime auto crisis, have been popping up every three months or so for the last 20 years. There is a long-standing and very consistent pattern in the media’s coverage of the auto lending space, particularly as it relates to subprime lending. First, raw statistics are released by the Federal Reserve, a credit reporting agency or a bond rating agency. Second, major financial news outlets will use selected statistics as a pretext for writing a sensational piece warning of imminent danger. Finally, the tabloid media (i.e., short sellers, bloggers and others who make their living pulling the fire alarm) pick up on the story and hyperbolize it. Some recent examples of this are:
• Barron’s, July 8, 2022: Car Repos Are Exploding. That’s a Bad Omen.
This recent gem, which has been recycled by others, features the scandalous headline that repos are up from 2020, when no one was repossessing cars due to COVID-19. The article also quotes a law professor who predicted an auto blow up in 2021.
• SeekingAlpha, April 17, 2022: Auto Loans: The Next Debt Bubble?
An alarming graphic from the Federal Reserve on the level of motor vehicle loans outstanding is presented in this article. One of the reasons the graphic is so alarming is that the x-axis (baseline) starts at $650 billion, making the balances appear as if they are skyrocketing from nothing. Higher amounts financed for more recent vehicles and longer terms (slower runoff) easily explain this. It should be noted that the Federal Reserve data is not consistent with what is reported from the credit reporting agencies for those same periods of time. This may may have something to do with the inclusion of securitized data, which would already be included in the credit bureau data (potential double counting).
• MarketWatch, March 24, 2022: Past-due subprime auto loans climb to highest rate since April 2020, a potential sign of trouble
This article used information reported by Deutsche Bank, which was corrected by S&P on May 26, 2022. In the S&P article, Amy Martin points out that the artificially high delinquency figure included charged-off accounts in addition to delinquent accounts. Of course, that didn’t stop a flurry of reports that were written on the spurious data.
• Consumer Reports, January 22, 2022: The Big Business of Bad Car Loans: How major auto lenders profit at the expense of the most vulnerable consumers
This one is my personal favorite. The author somehow was able to get on a private expert network presentation I gave to equity investors. He reached out to me with several quotes that were taken grossly out of context. I spent over an hour explaining to him the mechanics of auto lending, which were completely disregarded for this hit piece.
Many of these writers are merely trying to sell papers. Some wish to be perceived as clairvoyant should a disaster occur, while others have nefarious motives such as betting on a stock to plummet. The auto lending sector is not free of risk, and there are negative economic headwinds that deserve attention; however, for those who wish to understand where this market is headed in the next twelve to eighteen months, the “Chicken Little” press is not the answer. The following items represent key facts that help to put both emerging trends and potential threats in perspective.
The more sensational headlines point to the Federal Reserve’s G-19 report, which tallies auto loans owned and securitized at approximately $1.3 trillion, up from $1.2 trillion in 2018. The implication is that more people than ever are getting cars that they can’t afford while lunging toward a potential recession.
Experian’s State of the Auto Finance Market Report for Q1-2022 states that the average amount financed for used vehicles has increased from $21k to $28k since the pandemic, while the average payment has only increased approximately $100. Loan terms over that same period of time have increased from 65 months to 68 months. The Bureau of Labor Statistics reports that the average vehicle price has gone up over 40 percent in the last year, far more than loan amounts have increased.
The Federal Reserve reports that the seasonally adjusted sales rate for new vehicles is trending between 13 and 15 million. Historically, in non-recessionary times, that value has been in the 18 million rate for the better part of 23 years (refer to Graph 1 – Total New Vehicle Sales).
Cox Automotive recently reported that the seasonally adjusted annual sales rate for used vehicles is trending in the 35 to 37 million range, down from historic levels of 41 million. Experian reports that roughly 80 to 85 percent of new car transactions have financing attached to them, whereas 50 to 60 percent of used car transactions have financing. These facts suggest:
• Fewer loans are being made due to inventory issues
• Consumers are financing a greater amount
• Loan terms are increasing, keeping payments within $100 of historic levels
• Longer terms cause the runoff in loan volume to slow down
• Consumer are staying in their current loan longer due to lack of inventory and pricing
• Balances increasing by $100 billion at an aggregate level since 2018 is not a sign that the market is overheating, but rather is the natural result of an inventory shortage combined with inflation
Any time there is strong growth in auto lending the journalist’s assumption is that it all comes from subprime, but this is not the case. Prior to the Great Recession, subprime loans generally accounted for around 30 to 33 percent of the auto finance market. Following that recession, prime rebounded quickly, but it took several more years for subprime consumers to return to the market. Since that time, and in spite of the constant drumbeat of a subprime auto meltdown, these consumers have been paying their bills. A funny thing happens when consumers pay their bills – their credit scores increase!
In December of 2016, Experian issued a release that reported the percent of subprime and deep subprime auto loan consumers was at the lowest level since they began tracking those figures in 2007. Since that time, deep subprime decreased from 5.18% to 2.12% and subprime decreased from 19.8% to 15.01% from 2017 to 2022. Over that same period of time, prime loans increased from 38.77% to 45.45% (refer to Graph 2 – Total Auto Loan and Lease Distribution).
There is a very good publication produced by Fitch Ratings titled In the Auto ABS Driver’s Seat. For many years, this report has presented the delinquency and net loss rates of securitized auto loans in prime and subprime (published quarterly). For nearly 25 years, the annualized net loss rate of subprime pools has fluctuated very reliably around 6.5%. In good years, the rate trends down to 4% and in stressed periods it may rise to 8 or 9 percent. This stability was observed during the subprime auto lender shakeout of the late 90’s, the recession of 2001, the Great Recession and the COVID-19 pandemic. Ironically, the best quality paper tends to be originated during a crisis as lenders tighten up to control credit quality (a sign of rational lending, no less).
There are a few important items to take note of when looking at bond performance, which are:
• There are approximately $700 billion in new auto loans funded on an annual basis in non-crisis periods. Subprime and Deep Subprime represent around $150 billion of that. Only 29% of subprime loans are securitized. Compare this to the multi-trillion dollar subprime mortgage market pre-2008, where almost all of it was securitized. The actual volume in auto is far too small to throw the economy into a crisis. For more on the source data, I refer you to Asset-Backed Alert, Asset Securitization Report, the Auto ABS League Tables, the bond rating agencies and many other publications where these figures are readily available.
• Three large issuers account for nearly two thirds of the total subprime abs volume. This is important because a shift in collections strategy by any of these players can materially impact the aggregate delinquency that is reported for the entire contingent of subprime bond issuers. Numerous times over the past five years, journalist have reported that delinquency is higher than ever for these bonds. What the reporters fail to point out (which is readily observable in the presale reports of the issuers) is that many subprime lenders have pushed out their repossession assignment dates in order to give the customer more opportunity to resolve their delinquency issue. This has the effect of increasing delinquency due to the inclusion of accounts that would have previously been charged off. If these lenders pulled their repossessions forward by a month, their delinquency would go down while losses would increase. It is important to track both trends together in order to put things in context.
• Delinquency and losses for subprime bonds are influenced by the mix of lenders. As some of the well-established, deeper subprime players increase issuance, you will see delinquency and loss statistics increase where nothing actually changed for the individual lenders. Amy Martin, Senior Director at S&P, has done several very good presentations regarding the effect of lender mix on bond performance at several major industry conferences over the years, some of which are available online.
The broader pool of loans (including those not in securitized pools) have also shown significantly improved performance since 2019. Delinquency and losses have still not returned to baseline levels from the pre-pandemic period, although they are headed that direction. This is a return to a normal range of performance, as opposed to being caused by loose lending – as is often reported (this information is publically available both from the reports produced by the credit reporting agencies as well as the quarterly releases of publicly traded subprime lenders.
The Outlook for 2023
The data clearly demonstrates that performance is strong, volume in terms of number of loans is down and vehicle recovery rates are at record levels (driving down net losses). Does this mean there are no threats to the industry? Of course not. Lenders should be careful to watch the following trends:
• Vehicle Values – From 2009 to 2011, used vehicle values increased 30 percent. They remained high for many years because of the lack of inventory and pent-up demand associated with the Great Recession. We are in a similar situation today, as it is likely to take 18 to 24 months for new inventory to return. It is also probable that pent-up demand will keep values elevated for some time after that. When new inventory returns, there will still be a dearth of 2020-2022 inventory that will help to sustain more recent used vehicle pricing as well. Of course, values will come down at some point. Lenders must pay attention to auction trends and make modest adjustments in real time to avoid a surprise in net losses as portfolio growth resumes. I highly recommend subscribing to the rich data and indices provided by the value guide publishers.
• Default Rates – Default rates are likely to increase over the next six months as the forbearance from the pandemic period has almost completely ceased (with the exception of student loan relief). The concern I have is that with the price of fuel, food and other staples, the consumer’s debt load is worse than it appears on paper. A consumer with a 15 percent PTI, for example, is likely to behave like a consumer with similar credit quality at a 20 percent PTI. The same goes for other measures of debt. Lenders are likely to see modest increases in delinquency and loss at the lower credit score ranges for those who are most debt-stressed. It would be wise for credit managers to make modest PTI and DTI limit adjustments for their lowest program tiers.
• Recession Risks – I generally take most economic predictions with a grain of salt as many pundits cannot wait to be the first one to call the next downturn (particularly if they do not like the current political administration). Having said that, there are some key indicators that are cause for concern. The Institute for Supply Chain Management’s PMI index is on a consistent trend downward and is nearly below 50 for the servicing and manufacturing sector. This is an indicator the economy is shrinking. In addition, we are experiencing a negative yield curve which is coincident with recessionary periods. Finally, early predictions suggest we will have a second consecutive quarter of negative GDP growth. A growing number of analysts have proclaimed that we are already in a recession. It is unlikely, in my opinion, that a possible forthcoming recession will be worse than what we saw in 2007-2009 period. If you look at the static pool bond data from 2000-2003 and 2006-2010, you will observe that the losses came from originations that occurred prior to the downturn during periods where the market was highly competitive and lenders had loosened standards somewhat (i.e., modestly higher LTVs, less down payment, high debt loads). Performance deteriorated within a manageable range, and generally speaking, lenders managed through that period quite well. At present, auto lending is not marked by lowered standards and rapid growth – but is actually showing better quality than prior baseline periods. I say these things not to warn about credit performance risk (although there will likely be some deterioration), but rather to draw lender’s attention to their capital market risk. If the economy enters a full-blown recession, there could be contraction in the capital markets (both securitization and warehouse lending). It is well advised that auto lenders ensure they have excess capacity in their lines of credit, and possibly work to have more than one line where the facilities have staggered maturities. In addition, sacrificing a portion of lower tier paper now, ahead of a recession, will produce improved static pool delinquency and loss rates within 6 to 12 months, which will support the argument for lower enhancement levels and better pricing execution for those who securitize.
The auto lending industry is large and fragmented. There are hundreds of lenders, over 50 thousand dealers (franchise and independent) and more than 40 million auto purchases annually in the United States. As the second largest consumer loan class, it is a natural focus area for those in the financial press. With a subprime lending component, it becomes an irresistible target for disparagement.
Unlike student loans and mortgages, auto loan portfolios are very short-lived assets. Most runoff significantly within a 30 month period, even with 72 month terms. As lenders adjust standards for new originations, they can course-correct quickly in order to ensure performance stays on target. Through multiple downturns, auto lenders (including subprime) have shown tremendous stability and resiliency as they help people of all backgrounds and income levels acquire basic transportation to get to work.
With perpetual talk of a bubble, I took the time to look the word up in Merriam-Webster’s dictionary. The second definition reads “something that lacks firmness, solidity, or reality”. For those who resort to fearmongering in the auto finance market, I would like to say that I’m sorry to have to burst your bubble.