In their 2Q 2017 survey, UBS found that, for the first time since at least 2014, the trajectory of financial health of low-income households has started to diverge from that of more affluent households. Per the graph below from UBS' credit strategy team led by Matthew Mish, while a firming job market has helped households making over $100,000 feel more confident about covering their monthly expenses, spiraling debt balances has left low-income families even more vulnerable to the slightest monthly surprises with 70% reporting that their income just barely covers monthly expenses.

Overall US consumers report a lower likelihood of defaulting on a loan payment in the next year (15% vs. 17% in Q1), but lower income households cited an increase in their default probability (13% vs 9% in Q1). Other responses suggest consumers are incrementally more confident in their ability to pay given labor market stability, but optimism on the outlook is moderating. However, replies from lower income households paint a pessimistic outlook, partly due to lackluster real wage growth, higher financing rates and highly regressive policy.

Of course, with the entire auto and student loan bubbles being fueled by a massive expansion in subprime credit to the most at-risk American households, it's not terribly surprising that the lowest-income folks (i.e. those least prepared to absorb things like rate increases) are suddenly starting to feel the pain of their reckless balance sheet expansions.

There is a fairly strong relationship between income and credit scores, which implies the implications of this divergence will be primarily felt in non-prime and subprime consumer credit markets. Based on credit scores alone (VantageScores below 600), the three largest consumer loan markets in terms of subprime debt outstanding are the US mortgage ($570bn), student ($370bn) and auto loan markets ($180bn, Figure 4). We use the term 'subprime' lightly as some regulators characterize subprime credit scores as those below 660 (vs. 600) and risk-layering in consumer loans (e.g., autos) has been prevalent (which increases the riskiness of loans), both of which would in theory increase the subprime debt balances shown.

We believe the auto loan market best illustrates the fulcrum of credit quality trends in the US subprime consumer sector – one key 'canary' in the coalmine for US consumer credit. Rising collateral values and easy lending conditions supported by federal agency financing do not make the residential mortgage market a leading indicator this cycle, although US subprime mortgage delinquencies proxied by FHA delinquency rates will be important to watch. And US student loan delinquencies are heavily manipulated by the fact that many loans are not yet in payment status and deferral/ modification programs.

So, who is most at risk in the inevitable consumer deleveraging wave that is due any moment?  At least in the auto world, it's all the usual suspects including auto captives and private finance companies that rely almost entirely of the subprime securitization market for their financing.

What are lenders' aggregate exposures to auto loans and how have they responded to recent stress? In terms of the auto loan stock, there is $1.08trn in auto loan debt outstanding, of which 4% and 16% are deep subprime and subprime debt, respectively (or roughly $220bn combined). Among lenders banks originate 35%, credit unions 26%, and finance companies (captive, independent) the remaining 40%.

However, some of this exposure is securitized and moved off balance sheet, particularly for finance companies. ABS auto loan securitization comprises $192bn (18% of the auto loan debt), of which $42bn (22%) is subprime auto loans (Figure 11). In total, we estimate approximate nominal exposures to auto loans for banks, credit unions, finance companies and ABS investors of $360bn, $285bn, $235bn and $195bn, respectively (Figure 12).

We estimate nominal exposures to auto loans for banks, credit unions, finance companies and ABS investors of c$360bn, $285bn, $235bn and $195bn, respectively. Aggregate auto loan originations have remained steady in Q1; the share of loans in the lower quality tiers has fallen incrementally (32% in Q1 '17 vs. 35% in Q1 '16), but risk layering continues with loan terms lengthening.

In Q1 '17 the auto loan finance market split 55% used/ 45% new, but the proportion of subprime and deep subprime loans in the used market exceeded that of the new by a factor of 3x (31% vs 9%, respectively). Used car loan originations were led by finance companies (38%), followed by banks (36%) and credit unions (36%), with finance companies originating a greater share of subprime loans. ABS auto loan securitizations only fund $42bn (or 22%) of the over $200bn in subprime/ deep subprime auto debt, with captive fincos semi-reliant and independents wholly dependent on auto ABS markets.

As such, it's probably not a good sign that subprime auto losses are already soaring in 2016 vintage securitizations even as equity markets constantly confirm that 'everything is awesome.'

What are the current credit trends in autos? Based on the latest performance data for auto ABS loan securitizations, delinquency and recovery trends continue to deteriorate on balance. We have previously highlighted that portfolio seasoning alone will increase auto (and credit card) defaults going forward as each vintage since 2010 has performed worse than the year prior4 (Figures 5, 6). In our view, however, the more forward-looking indicator of credit risk is the change in NPLs for recent vintages. Here we focus on the change in NPLs for 2016 vs 2015 subprime (and prime) vintages, noting for subprime we utilize a modified subprime cohort as per S&P (which normalizes for the lack of a few deep subprime loans in the '16 vs '15 pools)5.

So, what now? Well, UBS suggests that a good start might be dumping all your exposure to low-quality lenders ahead of that forthcoming disruption in the securitization market that, much like 2009, will render their business models pretty much useless for a couple of years.

Consumer loan delinquency rates should continue to rise, specifically in autos but with residual effects in other loan categories amid rising subprime consumer stress. Some lenders (e.g., banks) are clearly tightening the supply of auto loan credit; it will be critical to watch the magnitude of the aggregate credit squeeze and degrees of contagion. At this point we believe financial stability risks are contained, but rising. We are cautious on non-bank lenders, but with our latest update would focus on reducing credit exposure to subprime (e.g, Capital One, Ally1) vs. prime consumer lenders. On a total return basis our preferred core US credit holding is 7-10yr US investment grade debt.