Low gas prices have fueled a shift in consumer preferences to larger, more expensive vehicles, such as crossovers, SUVs, and pick-up trucks. This has caused the average transaction price, according to Kelly Blue Book, to increase approximately 15% to $36,442 in 2018 from $31,616 in 2012. Because down payments haven't kept pace with vehicle price inflation, the average financed amount increased 18% to $30,494 as of fourth-quarter 2018 from $25,843 in fourth-quarter 2012, according to The Federal Reserve's G.19 report. At the same time, the average annual percentage rate (APR) for a 60-month bank-originated auto loan increased to approximately 5.4% as of fourth-quarter 2018 from an average of 4.2% in 2014-2017. Holding loan term constant at 60 months, the higher APR and financed amount would cause the average monthly payment to increase roughly $100 to $581 from $478.

This combination of higher financed amounts and more expensive borrowing costs have created an affordability issue, which many lenders are addressing by lengthening loan terms. According to Experian, the most common loan term for a new vehicle is now 72 months, with loan terms in the 73-84 month range now representing 30% of new vehicle loan origination volume as of fourth-quarter 2018. Over the past decade, S&P Global Ratings has observed an increase in longer-term loans in rated U.S. prime retail auto loan asset-backed securities (ABS). The percentage of loans with original terms greater than 60 months increased to 59% through first-quarter 2019, from 42% in 2009 (see chart 1). As a result, the weighted average original term of prime pools has increased to approximately 66 months from 62 months.

Chart 1

Borrower Repayment Risks Increase As Loan Terms Lengthen

Credit risk could increase as the amount of longer-term loans in these pools rises. Because new vehicles depreciate rapidly, especially in the first three years, the loan balance often exceeds the depreciated value of the vehicle for the first few years. Therefore, absent a large down payment, the severity of loss upon a repossession is generally substantial (40%-50%). To the extent that loan terms lengthen, the severity of loss could increase due to the slower amortization of the loan balance.

Additionally, because longer-term loans delay the point at which a borrower starts to build equity in the vehicle (and could sell it to repay the loan), the borrower may have less incentive to continue making payments on the vehicle in the event of financial distress. Moreover, longer loan tenors increase the likelihood of the borrower experiencing a job loss or a medical emergency while the loan is outstanding. These and other such events jeopardize the borrower's ability to make payments on time and in full.

In addition to increasing credit risk, longer-term loans cause securitization losses to become more back-loaded. We analyzed the paid-off transactions from a group of seven auto issuers and found that, as the average term of the pool lengthened to 64.9 months from 62.5 months in 2012, total losses incurred by month 18 contracted to 55.7% from 62.2% (see chart 2). The pools we examined for this part of our analysis were the paid-off 2012 and 2014 transactions from Ally Bank, CarMax, Ford Credit, Honda, Hyundai, USAA, and World Omni. We excluded Toyota from our loss-curve timing analysis due the high level of seasoning in its transactions: approximately 14 and 15 months in 2012 and 2014, respectively.

Chart 2

Our analysis reveals that loss-timing curves can change--in this case due to the lengthening of loan terms--and that projecting cumulative net losses (CNLs) using stale loss curves could underestimate potential losses. For example, if we projected CNLs on a pool that had 18 months of performance and a CNL to date of 1.5% using the 2012 loss timing curve above (62.2%), we would arrive at a loss forecast of approximately 2.4%. However, if we used the 2014 loss timing percentage (55.7%), the loss forecast would be 2.7%.

We also tested the relationship between the weighted average original term of the pool and the percentage of losses taken at month 24 for the 2012 and 2014 securitizations for the above-mentioned issuers, plus Toyota. We found that there was a significant (at the 0.5% level) negative linear relationship (with R-squared statistics of 0.45 and 0.39 for the 2012 and 2014 vintages, respectively) between the weighted average original term of a pool and the percentage of losses through month 24. As the loan terms lengthened, the percentage of losses taken through month 24 generally declined, meaning the losses were incurred later (see the trend lines in charts 3 and 4). The pools with the longest weighted average original maturities, including Ally, CarMax, and World Omni, incurred the lowest percentage of total losses through month 24.

Chart 3

Chart 4

We also observed that greater pool seasoning (i.e., a higher average number of payments that have been made) at closing generally leads to a more front-loaded loss timing pattern (see charts 5 and 6), which potentially offsets the back-loaded effect that longer maturities have on the loss curve. For example, Toyota's 2012 and 2014 transactions had weighted average original seasoning of 14 and 15 months, respectively, which is the highest of all prime securitizers in our study. The Toyota transactions also experienced the bulk of losses earlier than any of the other originators, incurring approximately 90%-95% of the losses by month 24. In contrast, Ford's 2012 and 2014 pools, which had average respective seasoning of five and eight months at closing, incurred approximately 72% and 66% of their total losses through month 24.

Loan maturity and loan seasoning appear (at least in the sample we studied) to have a negative linear relationship (i.e., lower maturities are generally associated with higher seasoning). This collinearity is likely the reason why a multivariate regression (not presented here) between these two covariates and losses taken at month 24 (as dependent variable) failed to yield meaningful, significant results.

Chart 5

Chart 6

Credit Enhancement Levels Could Rise

We believe longer-term loans will become increasingly popular among the auto manufacturers' captive finance companies, and these entities will join certain banks in including 84-month loans in their ABS. The continued shift to larger vehicles and new vehicle technology, including more driver assistance systems, may push median transaction prices higher over the coming months, thus creating demand for longer loan terms. To offset the potential risk that longer-term loans will have higher expected loss rates, lenders may seek to offer these products only to their higher-credit-quality obligors, thereby limiting default frequency. In addition (or alternatively), lenders may try to contain the severity of loss by limiting the loan-to-value ratios on longer-term loans. Higher levels of seasoning could also help to dampen the potential increase in losses associated with these loans.

As the percentage of longer-term loans increase in U.S. prime retail auto loan securitizations, S&P Global Ratings will continue to analyze static pool vintage performance by loan term and adjust its expected loss levels on the pools accordingly. To the extent that these loans represent greater risk for an issuer's pools, our expected CNL is likely to increase, which could result in higher credit enhancement levels.

The authors would like to thank Aaron Dalal for his contribution to this report.