Lenders Are Taking More Risk

April 2, 2014

Lenders are showing an increasing willingness to take risk, especially over the last two years. One measure of risk tolerance that economists typically track in corporate bond markets is the credit spread–the difference between interest rates for risky debt versus safe debt. But for the household sector, it is much more informative to track the quantity of credit extended.

The Federal Reserve Consumer Credit Statistical Release has shown strong growth in consumer debt (excluding mortgages). But this doesn’t necessarily mean that lenders are taking on more risk. They could be extending more credit to very high credit score individuals who are unlikely to default.

But the microeconomic evidence suggests that the opposite is true: much of the growth in auto and credit card debt is among individuals most prone to default. We can see this using zip code level data.

We split up zip codes in the United States into four groups based on their 2009 default rate. The groups each contain 25% of the population. The highest default rate zip codes tend to be those with the lowest credit scores. Lending in these areas is almost by definition more risky. The lowest default rate zip codes are the safest. We then track the growth in auto debt and credit card debt originations in the riskiest and safest zip codes.

Here is the chart for auto debt:


Auto debt originations have increased by almost 70% from 2010 to 2013 in the riskiest zip codes. They have increased by only 30% in the safest zip codes. The difference is especially pronounced in 2012 and 2013. Lenders are extending much more credit in the riskiest areas.

We see the exact same pattern in credit card originations:


Credit card originations increased strongly in 2011. But after 2011, they more or less leveled out for zip codes with a low default rate in 2009. Instead, they continued to grow strongly in 2012 and 2013 in the riskiest zip codes.

This evidence is consistent with the view that lenders are taking on more risk in search for yield and earnings. Some, such as the Federal Reserve Governor Jeremy Stein, have openly worried about easy monetary policy lowering the risk premium too far.

Should we worry about this credit expansion? Some would say this is a sign of a healthy credit market, returning to normalcy. Others would say that another unsustainable credit boom is in the works.

In our view, the critical question is: do the income prospects of individuals in riskier areas of the country warrant an increase in borrowing? Will they be able to pay back the debt when it comes due? During the subprime mortgage boom, our research shows that the answer to this question was clearly no. But what about now?

Unfortunately, we do not have up-to-date data on the income of the riskiest borrowers. But we doubt that their income prospects have improved significantly–we haven’t seen the kind of strong job and wage growth that we would need to make sure that the riskiest borrowers have adequate income going forward. We hope we are proven wrong.

Bookmark and Share

9 Responses to Lenders Are Taking More Risk

  1. rob urie on April 2, 2014 at 3:14 pmi

    Absolute yield levels are the measure of credit risk for corporate bond investors, not ‘spreads’– high yield portfolio managers can explain this to you.

    Emmauel Saez’s work on income distribution and the mix of jobs created since 2008 strongly suggest that the subprime auto and credit card loans are both predatory and will end poorly.

  2. MrGotham on April 2, 2014 at 6:07 pmi

    Subprime auto debt issuance contracted quite a bit in 2008 and 2009, so the starting point of 2010 kind of skews the data, and perhaps the conclusion. What does the data look like if you start in 2006 so you capture the full credit cycle?

    • Squeeky Wheel on April 3, 2014 at 12:44 ami

      “Subprime auto debt issuance contracted quite a bit in 2008 and 2009″ That is probably an understatement. Loans to high credit risk borrows went to almost nothing, so it is entirely expected that the rate of growth would be higher than for low risk. The low risk segment had less drop in the first place.

  3. Lukas on April 2, 2014 at 6:26 pmi

    I’m curious how much of the faster growth rate is just a bounce back from the lows in ’09. How does the data look if you index vs. the prior peak?

  4. SN on April 3, 2014 at 7:52 ami

    I echo the comments of others. I would like to see the data from the past decade.

    • Atif Mian and Amir Sufi on April 3, 2014 at 8:25 ami

      Thanks all for comments. We only have the data 2010 and after. We wish we could track it back! That being said, notice that rise is pretty strong from 2011 to 2012 and again to 2013. So it may have been very low in 2009, but the rise in 2012 and 2013 is occurring quite a bit after 2009.

  5. ReturnFreeRisk on April 3, 2014 at 8:00 ami

    That will warm Yellen’s heart. She wants to make sure free money reaches every corner of the nation. After all, the mandate is to make everyone happy.

    • Eduardo Elias on April 5, 2014 at 4:44 ami

      That was the aims of Sapatero goverment in Spain, even they naned it. The equivalente in english of the “the well being state”. It was moré of a goverment spending than central bank money. We all know how it ended. It is going to be iteresting to see how it goes with unlimited resorses for central bank.

  6. joelw on April 3, 2014 at 3:25 pmi

    Still with others. If, say, the base index was 50 for the high risk and 100 for the low risk, then a 60% increase for the high risk and a 25% increase for the low risk are basically the same absolute increase.

    Moreover, while income might not have improved that much, interest rates are low. http://research.stlouisfed.org/fred2/series/TDSP Debt service as a percent of disposable income is historically low.