The Effect of the Economic Crisis on American Households

01/19/2011
Featured in print Bulletin on Aging & Health

Over the past several years, the American economy has experienced the most severe recession since the Great Depression of the 1930s. The unemployment rate has risen by more than 5 percentage points since the crisis began, while the stock market and housing market have tumbled - from late 2007 to late 2008, the S&P 500 Index fell by about 40% while the Case-Shiller home price index fell by about 20%.

These dramatic and nearly simultaneous shocks to labor, stock, and housing markets have undoubtedly affected American households. But quantifying how big the impacts have been and how households have responded to the crisis is difficult. Many of the data sets economists typically rely on to examine consumption and wealth are fielded only every two or three years, so researchers may have to wait several years until data collected during the recession is available. Moreover, the fact that many surveys do not follow the same households over time makes it difficult to track the recession's impact.

In Effects of the Financial Crisis and Great Recession on American Households (NBER Working Paper 16407), researchers Michael Hurd and Susann Rohwedder take up this question using data from a series of surveys developed for this specific purpose and fielded in the American Life Panel (ALP).

The ALP is an ongoing survey of about 2,500 people run by RAND Labor and Population. The survey is conducted via the internet, with Web TV provided for those without internet access; results are weighted to be representative of the population. For assessing the effects of the financial crisis on American households respondents were first interviewed in November 2008, just after stock prices had fallen sharply and following a somewhat longer period of housing price declines but before most of the rise in the unemployment rate. Respondents were re-interviewed in February and May 2009 and have been interviewed monthly since that time. The paper reports results covering the 17-months period from November 2008 through April 2010.

The authors first quantify the share of American households who have experienced financial distress, which they define as being unemployed (respondent or spouse), being more than two months behind on mortgage payments, or having a home valued at less than the mortgage. They find that back in November 2008, shortly after the onset of the financial crisis, roughly 13 percent of households were in distress and that this fraction increased rapidly by June 2009 to about 17 percent where it has remained. The cumulative measure shows that nearly 40 percent of households have experienced financial distress at least some time during the 17-months period.

Spending changes provide one measure of the recession's impact on households' well-being. At the first survey, nearly three-quarters of respondents said they had reduced spending due to the economic crisis; at the next survey, 30 percent said they had reduced spending further since the first survey. Among those reporting a decrease in spending, many cited the need to reduce debt (80 percent), a reduction in income (70 percent), a change in employment status (45 percent), or a decrease in the value of their home (45 percent) or stock holdings (35 percent) as reasons for the decline.

It is difficult to quantify changes in spending, especially retrospectively. Because reductions in spending turned out to be the most wide-spread response to the economic crisis, the authors instituted a monthly survey schedule and included detailed questions about the amounts that households spent across 25 categories. Even though this effort missed the initial reductions in spending, it documents further declines of 5 to 10 percent that occurred between April 2009 and March 2010.

Changes in specific categories of spending may shed some light on how households achieved the reductions in spending and the resulting impact on well-being. Food spending fell by a similar percent as overall spending; most of the decline occurred in spending on food purchased away from home, as households substituted "food in" for "food out." Spending on prescription drugs and health care services dropped more sharply than total spending; to the extent that these expenditures are protective against future health declines, these cuts may have long-term negative impacts on health.

Some households responded to the crisis by borrowing more. While the number of households with credit cards fell by 3 percent over the sample period, perhaps reflecting a tightening of access to credit, among those households with credit card debt the amount of debt carried over from month-to-month rose by 25 percent. This added debt could translate into additional interest payments of nearly $1,000 per year.

Another way for individuals to respond to a fall in stock or housing assets is to retire later. However, the share of respondents who indicated that they intend to work past age 62 actually declined by 3 percentage points during the sample period. To put this in perspective, labor force participation of men ages 60 to 64 increased by 3 percentage points between 2003 and 2008. The decline in expected work past age 62 suggests an increase in pessimism among workers about their labor market prospects at older ages.

The ALP illuminates both how households have been affected by falling stock and housing markets and how expectations about these markets may be changing. By early 2009, households who had at least some retirement savings reported that those savings had dropped in value by about 30% on average. One in six households responded by reallocating existing retirement account balances away from stocks, nearly three times as many as shifted balances towards stocks. In terms of expectations, in November 2008 respondents estimated the chances of a positive gain in the stock market over the next ten years to be 60 percent on average, although historically the odds of this are over 90 percent. These long-term expectations became even more pessimistic as the economic crisis went on, reaching their low point in April 2010. Respondents were similarly pessimistic about the chances that house prices would increase over the next five years, estimating the chances to be 56.4 percent in November 2008, and lower yet in April 2010 (51.3 percent), even though national housing prices have never failed to rise over a five-year period since the index began.

A somewhat different way to measure the recession's impact is to explore the effect on health and subjective measures of satisfaction. While only 7 percent of survey respondents report being dissatisfied or very dissatisfied with their lives in mid-2009, this fraction increased to 10 percent by April 2010. One-third expressed dissatisfaction with their income situation, showing little trends through the period of study while the fraction dissatisfied with their overall economic situation declined from 45 percent to 34 percent from November 2008 to April 2010. Some measures of self-assessed well-being show little trend (about 28 percent report feeling worn out, 25 percent feeling happy, 17 percent reporting depression problems), however the fraction with difficulties sleeping was highest at the onset of the crisis (34 percent) and dropped to about 24 percent by the end of the study period (April 2010). The fraction with fair or poor self-rated health dropped only a little from 16 to 14 percent over the same period.

The authors conclude by pointing out that the recession officially ended in June 2009. They note that according to their data, "the economic situation of the typical household is no longer worsening, which is consistent with the end of the recession defined as negative change. However, when defined in terms of levels rather than rates of change, from the point of view of the typical household, the Great Recession is not over."


The authors are grateful to the National Institute on Aging (under grants P01 AG008291, P01 AG022481, P30 AG012815, and R01 AG20717) for funding of data collection and research reported herein The Social Security Administration also supported part of the data collection.