In the study, titled "Do We Know What We Owe? Consumer Debt as Reported by Borrowers and Lenders," authors Meta Brown, Andrew Haughout, Donghoon Lee, and Wilbert van der Klaauw found that the pictures of U.S. consumer debt produced by the Survey of Consumer Finances (SCF), which is a survey of U.S. households focusing on household assets and liabilities, and the Consumer Credit Panel (CCP), which is based on data supplied by credit reporting agency Equifax, were more similar than expected.
"Our most striking finding is that, overall and in most disaggregated debt categories, debt levels reported in the SCF and CCP are quite similar."
For the study, the authors compared consumer debt aggregates as well as metrics of household distribution of the following forms of debt: total debt, mortgage and home equity lines of credit (HELOCs), vehicle loans, credit card debt, student loans, and other debt in both SCF and CCP data, according to the New York Fed. The study also computed household delinquency and bankruptcy rates in both the SCF and CCP samples for four different years (2001, 2004, 2007, 2010), two of which came prior to the implementation of major bankruptcy law reform (2001 and 2004) and two came after.
For the majority of disaggregated debt categories and borrower characteristics, the authors found that debt levels reported in the two sources agreed for the most part.
"Our most striking finding is that, overall and in most disaggregated debt categories, debt levels reported in the SCF and CCP are quite similar," the authors wrote. "Even bankruptcy measures correspond well. The exceptions lie in the unsecured debts. Under our most inclusive assumptions, SCF-implied aggregate credit card debt is 37 percent lower than that implied by the CCP, and SCF-implied aggregate student debt is 25 percent lower."
The debt levels followed similar age, geographic, patterns, and time trends in both the SCF and CCP for mortgages, HELOCs, and vehicle loans. One notable exception in which the debt levels were not similar was credit cards; authors found that in the CCP, household credit card balances were reported to be 37 percent higher than in the SCF.
The similarity in bankruptcy data between the two sources, however, was "unexpected," according to the authors.
"Bankruptcy appears to be reported at similar frequencies in the SCF and the CCP (though differences in available measures of bankruptcy in the two data sets impose qualifications on this claim)," the authors wrote. "We find that, among one- and two-adult households, the CCP’s two-year household bankruptcy rates in 2001, 2004, 2007, and 2010 fall comfortably between the SCF’s one- and three-year bankruptcy rates, and that, if anything, one- and three-year bankruptcy rates in the SCF appear to be a bit high relative to CCP two-year rates."
Click here to read the complete study.