I want to thank Matt Yglesias for pointing out a fairly significant error in my last column, a theoretical speech I wrote on behalf of Mitt Romney. I wrote the following two sentences, only the first is which is true:
The first is that after a brief uptick in the savings rate after the crisis, the savings rate is once again trending downward. American families are loading up on debt to make up for an economy that isn’t growing fast enough.
That is, I interpreted a decline in the savings rate as an indication that after having made impressive strides in deleveraging, the subject of a recent report from the McKinsey Global Institute, U.S. households were once again taking on an unsustainable level of debt. But at Matt’s post indicates, this interpretation isn’t borne out in the data.
Part of the issue is that there is some disagreement regarding the appropriate level of household debt. I am partial to Neel Kashkari’s take:
[G]iven that consumption represents more than 70% of U.S. GDP, this is an important estimate that will affect our economic growth in the next few years. Figure 3 is a graphic from the McKinsey report that shows U.S. household debt as a percentage of disposable personal income. It indicates that Americans have less debt than they had in 2008. This trend leads to McKinsey’s estimate “that U.S. households could face roughly two more years of deleveraging. … One possible goal is for the ratio of household debt relative to disposable income to return to its historic trend.”
But why should the trend be upward sloping at all? According to the McKinsey chart, in 1955 household debt to disposable income was about 45%. McKinsey says household deleveraging should be completed by mid-2013 when Americans “only” have debt to income of 100%.
Why should Americans each year take on more and more debt? Is this sustainable? Is this a good trend for Americans and for the U.S. economy? McKinsey acknowledges these questions – but doesn’t change their conclusion that U.S. deleveraging is on track and the American consumer is almost back.
To look deeper behind the increased consumption numbers, one must look at how much Americans are actually saving. As Figure 4 indicates, from 1960 to the mid-1980s, Americans saved more than 8% of their income each year. Then savings began to fall – and fell almost to zero just before the crisis in 2008. In fact, by some measures, Americans on average had a negative savings rate in the mid-2000s. They were consuming with borrowed money, often by taking out home equity loans. The shock of 2008 scared many Americans into saving more. They realized their home values wouldn’t climb forever and they needed to save for their futures. The personal saving rate climbed back to almost 6%, still lower than all the years from 1960 to the mid-1980s, but a more financially sound level than the years immediately before the crisis.
Alarmingly, this healthy increase in savings appears to have been short-lived: Consumers have actually started saving less in recent quarters and this has fueled much of the recent uptick in consumption.
This is the landscape I had in mind. But given the space constraints, I managed to mangle the point rather badly. Though I can’t promise I won’t mangle a point or two in the future, I will strive to be more careful. (Moreover, as Mason and Jayadev argue, debt dynamics also play an important role in determining household debt levels, a subject I neglected in that paragraph.)