UBS: How idiosyncratic are auto loan borrowers (and related credit stress)?

Our core thesis is that while rising auto loan delinquencies are not yet a financial stability concern, they do expose broader consumer stress which will manifest itself in structurally higher consumer (non-mortgage) loan defaults ahead. In corporate credit, this risk is not yet fully priced in.

Pivotal questions: Is the demographic profile of auto loan borrowers unique? 

Yes and no. The highest frequency of auto loans is in the 21-34yr olds (27%), followed by 35-44 and 55-64yr olds (21%, 22%). By income the highest proportions of borrowers are middle income (23% in 25-49k, 25% in the 50-74k), Overall, the occurrences of debt do not deviate significantly versus that of other consumers. However, by value student and auto debt are more skewed towards Millennials. 

Do auto loan borrowers state they are more likely to default? 
Yes, but not by a statistically significant margin. 19% of auto loan borrowers strongly agree or agree that they are likely to default on a loan over the next year, compared with 17% for peers. 

What are the key reasons why auto loan borrowers say they are likely to miss payment? 
Top factors cited include reached credit limit (23%), medical expenses (19%), growing family (19%), reduced income (18%) and nonmedical expenses (16%).

How much overlap is there among borrowers between auto and consumerdebt? 
The highest concentrations are in mortgage (49%) and credit card loans (41%), followed by bank (27%), student (17%) and medical (17%) loans. However, while student debt overlap is low, debt loads are material (e.g., 41% over $40k in debt).

Are auto lenders and investors over-reliant on credit scores? 
Perhaps. Easing of auto underwriting standards and layering of risks have been well documented. Model dependence on credit scoring is high, but there are limitations. The worse than anticipated performance in auto lending should raise questions around model efficacy. One strand of academic literature argues that underwriting processes fail to incorporate strategic behavior of borrowers, leading to underestimation of losses.

Is there evidence of greater manipulation of risk estimates this cycle? 
The evidence is growing. First, the explosion of technology makes gaining access to information to improve credit scores very simple. Internet searches for 'credit score' are at record levels. Second, our survey finds 21% of auto loan borrowers admitted to some form of inaccuracy in their loan applications. Third, there is growing concern reported among auto lenders around fraud, which is the extreme case of this behavior.

Figure 1: Respondents describing the factual accuracy of loan applications as inaccurate (mostly or partially accurate)

Den vollständigen Artikel finden Sie hier.

powered by