Debt and Inequality

Will Wilkinson is responding to commenters on his inequality paper. Ezra Klein brought up worries about debt and households (worries that echo my own), especially the recent dramatic increase in debt/income we see in this graph:

Here is Will’s response (my underline):

That said, I suspect many lower-income households have over recent years increased their level of debt, and I suspect that this has played some small role in keeping consumption inequality in check…So, yeah, lots of people who wrongly thought they were house-rich ran up their credit cards. I think it’s safe to say that there was “too much” debt-financed consumption. But I would hazard to guess that, on the whole, this would tend to widen rather than narrow the income and consumption gaps

That incomes at the top are now so sensitive to aggregate consumption (it didn’t use to be that way, Parker and VIssing-Jorgensen say) would seem to at least partly explain the coincidence of very high average debt-levels and high levels of inequality that Ezra emphasizes later in his post.

Folks were running up their credit cards because they thought they were house-rich. They thought they were house-rich because they were in the middle of a housing boom that made the current and future value of their houses look a lot higher than they really were. Those with compensation schemes highly sensitive to changes in aggregate demand–high income households–saw a disproportionate rise in already high income as consumption boomed. So income inquality went up.

I don’t want to put words in Will’s mouth, but I believe he is saying one of two things here. (1) The increase in debt reflects an increase in debt-fueled consumption at the top 10-20% of the income distribution, and as such will increase the amount of consumption inequality we see as top earner fuel consumption with debt. (2) The debt increase among everyone will fuel the earnings of the top 1% (or top .01%) of earners, increasing consumption inequality between the top 1% and the bottom 99%, because increases in aggregate consumption are linked via Parker and VIssing-Jorgensen’s paper. Even if Will isn’t making these points, they get made by others so I want to examine them.

Now remember our original assumptions: Income inequality is large. Consumption inequality is smaller. So savings inequality must be huge. Money doesn’t spoil if it isn’t spent, it goes towards investments/savings/etc. Fair enough, and the point of consumption inequality is just to focus on consumption. The problem with just focusing on consumption is you lose sight of the ways in which consumption/savings/debt are all part of the same thing.

One brings up that a lot of this consumption is debt fueled. A response is “yes, but clearly those spending more at the top must be taking out more debt to fuel that consumption, because they are spending so much more on consumption.” See how we lost our original assumption, looking at just consumption? We started under the assumption that the rich are saving a lot more in addition to consuming more. Those who make this point say that those at the top of the income distribution should have increased their debt burden, but looking at all inequalities together we know that the debt burden must decline, since the top’s savings inequality is increasing faster than their consumption inequality.

And it does. From the Federal Reserve’s Changes in U.S. Family Finances from 2004 to 2007: Evidence from the Survey of Consumer Finances, two charts:

The first chart is the leverage ratio – debt/assets, by income distribution. From 1998-2007 (as well as 2004-2007) we see the leverage ratios of the top 10% decline, while the leverage ratios of the 20%-90% range of household income range skyrockets. Notice that the average overall doesn’t change much, but the dynamics underneath involves a lot of shifting around, with the biggest percent gains in the 20-60% of the distribution.

The second chart (click through to see the whole thing) is the ratio of debt payments to family income. We are interested in the first two boxes, which are the aggregate and median for debtors. Once again, same dynamic. From 1998-2007 (as well as 2004-2007) we see the debt payment ratio of the top 10% decline, while the leverage ratios of the 20%-90% of household income range increases. Mind you, because of the way these numbers get to us, a small increase can translate as a giant increase in bankruptcies and troubles. The second set of boxes are important, because Will says that the Parker/VIssing-Jorgensen says that the debt consumption should impact the top 1% of earners disproportionately; however the median value, so those at 95% of income, also see a similar decline.

So I still think we have a story about the middle-class and working-class debt consumption. While there was a relatively flat consumption inequality trend, there was a massive difference in leverage, where those at the bottom were leveraging up and those in the top 10% deleveraging. As far as the first graph on debt-to-income goes, the debt-to-income of the top 10% are actually slowing down the increasing trend.