The Financial Crisis and the Impact on Households

Household balance sheets were severely affected during the financial crisis and ensuing recession. According to the Federal Reserve's triennial Survey of Consumer Finances (SCF)—the most comprehensive examination of household balance sheets—average household wealth declined 15 percent between 2007 and 2010 while median household wealth dropped 39 percent.

More important, however, we must understand who lost wealth and why. Accordingly, we focus on families grouped by age, educational attainment, and race or ethnicity—demographic and other "exogenous" dimensions that are reliably measured, that are not subject to choice or random variation over time, and that are not difficult to interpret due to potential reverse causation.

Although many subgroups experienced large declines, the Fed's survey suggests that families that were younger, that had less than a college education and/or were members of a historically disadvantaged minority group (African-Americans or Hispanics of any race) suffered particularly large wealth losses Figure 1.¹

Even before the crisis, younger, less-educated and historically disadvantaged minority families were known to be among the most economically vulnerable groups because of the particular occupations and sectors in which they were over-represented, such as low-wage service-sector jobs and construction. What was not well-known—but which we document here—is that families in these economically vulnerable groups often also had very risky balance sheets going into the crisis. Our research suggests that both economic vulnerability and risky financial choices may stem from one or more common causes, including low levels of human capital, relative youth and inexperience, as well as the legacy of discrimination in education, employment, housing and credit markets. As we show later, these groups experienced the most-acute balance sheet "failures"—high concentrations of wealth in housing and high levels of debt.

FIGURE 1

Percentage Losses In Mean Net Worth

2007-2010

AAH: African-Americans and Hispanics of any race.
WA: Whites and Asians.
GED: General Educational Development certificate.

SOURCE: Fed's Survey of Consumer Finances, 2007 and 2010.

DEFINITIONS

median: The number that ranks precisely in the middle of a set of numbers arranged in order of magnitude. If the set of numbers has an even number of members, the median is the average of the two numbers that are closest to the middle of the ranking. In contrast, the mean is the average value of a set of numbers divided by the number of members in the set.

reverse causation: A relationship between two variables, each of which may be important in explaining the other, rather than one being clearly causal with respect to the other. For example, income and marital status may be subject to reverse causation. Having a high income may increase the chance that an individual is married, but being married also might contribute to an individual’s having a higher income. Thus, the causal relationship between the variables is ambiguous. Demographic variables such as age and race or ethnicity are not subject to reverse causation in the same way. Being a minority may reduce a family's chance of being a homeowner, due to discrimination in housing or mortgage markets, but not being a homeowner does not "cause" minority status. Causation clearly is one-way only, if it exists.

human capital: A concept meant to capture the potential earning power of an individual. Unlike physical capital, such as a machine, human capital cannot be measured precisely because it is not legal to buy and sell financial claims on a person’s future earnings. The concept is useful, nonetheless, to facilitate discussions of why people make investments in education and what financial benefits this investment might generate.

ENDNOTE

1. Notice that the percent declines in average net worth between 2007 and 2010 for each of the education groups is larger than the overall average decline. This anomaly is due to changes in the number of families in each category and differences in the average wealth losses in those categories. To illustrate how changing cell sizes can produce individual category percentage declines that all are larger than the overall decline, consider a simple example. Suppose that, in 2007, you owned two cats and two dogs. The average weight of your cats was five pounds and the average weight of your dogs was 10 pounds; so, the average weight of your pets was 7½ pounds. Suppose that, in 2010, you had one four-pound cat and three dogs with an average weight of nine pounds. Comparing 2007 and 2010, the average weight of the cats you owned decreased 20 percent and the average weight of your dogs decreased 10 percent. But the average weight of your pets actually increased 31/3 percent, from 7½ to 7¾ pounds. In terms of wealth changes among families of different education levels, less-than-high-school families with relatively large average losses (analogous to cats) decreased as a share of the sample, while college-educated families with relatively small average losses (analogous to dogs) increased as a share of the sample. The number of families with college degrees increased between 2007 and 2010, from 35 to 37 percent of the sample, while the number of families with less than a high-school degree declined from 14 to 12 percent. The number of high-school degree families stayed roughly the same, at about 51 percent.