The Rise in the Number of Children in $2-a-Day Poverty Evident in Three Large Nationally Representative Surveys and Administrative Records

$2.00 a Day began with observations by one of us, Kathryn Edin, who by 2010 had spent more than twenty years canvassing poor communities all over the country. Back in the field to study the deeply poor, she began to encounter something markedly different from anything she had seen before: families with no visible means of cash income from any source. Some had SNAP (food stamps), a small group had a housing subsidy, and most had a household member on government-funded health insurance. But as we write: “[W]hat was so strikingly different from a decade and a half earlier was that there was virtually no cash coming into these homes. These families didn’t just have too little cash to survive on…. They often had no cash at all.”

Had there been a spike in the number of children going to periods with virtually no cash in the United States? We began to test this with the Survey of Income and Program Participation (the SIPP) because 1) we knew that the SIPP captured the most income among poor families; and 2) we cared about periods without cash that lasted for less than a year and other surveys didn’t allow us to look at shorter periods. But other surveys can be used to substantiate what we found in the SIPP. Two in particular are the Current Population Survey (CPS) and the American Community Survey (ACS). These surveys are collected using different samples and by different means. They suffer from the same problems that the SIPP does—some people may not want to reveal all their sources of income, they may forget, or they may misunderstand the questions. Yet if numerous surveys tell identical stories, it is less likely that they are all wrong.

Here we examine the number of children in CPS and ACS in households reporting cash incomes under the $2-a-day threshold. ACS has only been collected since 2001, CPS goes back further. Below in blue is the CPS trend line going back to 1995; in orange ACS. We draw data from the IPUMS series, which is publically available and preferably because questions are integrated over time for comparability. We adjust dollar values to 2011$ using the CPI Research Series.

 In CPS we find that the number of children reporting annual cash incomes below the $2-a-day threshold rises from 908,000 (1.28% of children) in 1995 to a high of 1.89 million (2.55% of children) in 2011, and remained at 1.77 million (2.39%) in 2014. That represents nearly a doubling at the high point in 2011 and still elevated as of 2014.

ACS estimates appear going back to 2001, and map on closely to the CPS estimates.  In many years, the estimates are virtually identical. In 2014, for example, CPS yields 1.77 million children in $2-a-day poverty (2.39%) while ACS yields 1.75 million (2.38%). CPS in a few years appears to spike higher than ACS (2008; 2010), but the trend and scale is quite consistent between the two surveys.


Evidence from Administrative Records

Because survey data are imperfect, we sought out administrative data to test our hypothesis that the cash incomes of the poorest families had declined precipitously in the aftermath of welfare reform. 

As reported in $2 a Day, we first examined SNAP administrative data for the total number of SNAP assistance units with children in the U.S. receiving SNAP who report no source of cash income. Households receiving SNAP must verify their income eligibility every three to twelve months, depending on the state they live in and their status (i.e. singles versus parents with dependent children). Note that it is a felony to knowingly engage in SNAP fraud, and the USDA warns applicants that they can “be fined up to $250,000 and put in prison up to 20 years or both.”

In 1996 some 316,000 SNAP households with children reported no source of cash income. That number began to rise in 2002 and by 2005 had increased to about 599,000. By 2014 that figure had grown to roughly 1.4 million—a figure that is close to the total number of $2-a-day poverty households based on SIPP estimates provided in $2.00 a Day. This represents roughly a quadrupling of households reporting zero cash income.

We continue to seek out ways to test the strength of our arguments made in $2.00 a Day. With evidence that there has been a rise in the number of families experiencing periods with virtually no cash, we turn to new questions. One, has this increase been related to the decline of TANF? And even further, does cash matter? If one has food stamps, access to charity, and public health insurance, does cash matter? In our book we found that it clearly did for the families we followed. But we will continue to seek out further evidence of this and report our findings to you.

 - Luke Shaefer