Last week, the Federal Reserve released new figures for its household debt service ratio. This ratio represents the share of aggregate after-tax household income accounted for by required principal and interest payments on household debt. As shown in the chart below, the debt service ratio declined to 12.8 percent in the third quarter of 2009 and now stands a little more than a percentage point below the highs seen in late 2007 and early 2008.
The decline in required debt service payments since the beginning of 2008 is largely attributable to a reduction in the amount of debt that U.S. households are holding. Outstanding mortgage debt has been shrinking for six consecutive quarters. Non-mortgage debt has also been falling, particularly credit card balances, which have declined nearly 10 percent since their peak in September 2008. Households are borrowing less both because they have cut back their spending in the face of tough times and because lenders are far less willing to lend than they were during the credit boom. The reduction in outstanding household debt also reflects the high recent rates of defaults on household debt—not an ideal way for households to deleverage, but one just the same. ##1## Whatever its causes, the decline in household debt will leave Americans in a more solid and sustainable financial position over the long run. With lower required payments on debt, households will be less stretched should their incomes decline unexpectedly or their wealth not reach levels as high as they had expected. Over the short run, however, reduced borrowing is likely to be associated with softer consumer spending. And, more modest increases in consumer spending mean a slower economic recovery so that it will take longer for the economy to return to full employment.
Measures like the debt service ratio can be a useful way to gauge whether the economy is on a sustainable track, as I noted in a discussion for a Brookings/Heritage conference on economic data and turbulent economic times last month. In the years leading up to the financial crisis, we saw the aggregate debt service ratio climb well outside of its previous range.
However, as I argued in the same discussion, we need to be mindful that aggregate measures like the debt service ratio are fundamentally limited. The measure treats the nation as if it were one enormous household, whereas the debt experience of Americans varies widely, from those households that carry no debt to those who have very high leverage. It was the most indebted households that played a key role precipitating the trouble in mortgage markets that spurred the financial crisis. Based on analysis of mortgage loans and credit records, we now know that the fraction of households with very high mortgage leverage soared during the credit boom—both because of an increase in the number of home buyers that were making little or no downpayment and because many existing homeowners were using “cash-out” refinancings to extract large amount of home equity to fund other spending. However, such information was not widely available to analysts in the period leading up to the financial crisis, in part because the underlying data sources are expensive and proprietary.
To anticipate the next crisis, analysts need to be watching more than aggregate measures like the debt service burden—they need to be watching distributional measures like the fraction of households engaging in particularly risky financial behavior. And, as we think broadly about what role government can play in reducing the likelihood of future crises, some part of the discussion should concern whether statistical agencies should be publishing information about the distribution of debt use in order to facilitate vigorous discussion within the large group of analysts that follow these issues.