For the past 20 years, an increasing number of families are using food pantries and food banks as a way to cope with lack of access to adequate and sufficient food. Since 1995, the U.S. Census has been measuring food pantry use among American households through the Current Population Survey Food Security Supplement (CPS-FSS). In 2014, over 6.70 million households, received emergency food from a church, food pantry, or food bank, representing a near doubling of food pantry user rates since 1999, 2.3 percent to 5.5 percent, respectively. Moreover, use of food pantries is more prevalent among households with children. For instance, in 2014, household with children were more than twice as likely to obtain emergency food from food pantries as households with no children, 7.3 percent and 4.5 percent, respectively.
Prior to 1999, food pantry user rates declined from 1995 to 1997 by nearly 30 percent, from 4.7 percent to 3.3 percent. We find strong evidence that some of this drop may be the result of systematically different survey screening processes used during the first five years of CPS-FSS. For example, relative to other survey years, 1995 estimates of food pantry use may be biased upward due to the high sampling of households with income below 185 percent of poverty. In 1995 over 80 percent of household respondents to the food pantry questions had income below 185 percent of poverty, about 30 percent more than the average ask rate for households below 185 percent of poverty. Since households below 185 percent of poverty are more likely to use food pantries, 1995 food pantry user rates may actually overstate the prevalence of food pantry use among American households at the time. Conversely, 1997 food pantry rates may understate the true use of food pantries. In the 1997 CPS-FSS some households with income below 185 percent of poverty were not asked the food pantry question, resulting in a significant undersampling of poor households relative to the average ask rate of households below 185 percent of poverty, 55.9 compared to 63 percent, respectively.
Overall, the inconsistent screening processes and resulting sampling bias throughout the first five years of CPS-FSS not only make food pantry user rates from 1995 to 1999 incomparable to later annual estimates, but also to one another. As a result it is difficult to attest to the validity and accuracy of the noted decline in food pantry user rates from 1995 to 1997 and make the claim that American households were less in need of emergency food services. In reality, by adjusting for the potential overestimate of food pantry use in 1995 and underestimate in 1997, we suspect that food pantry user rates actually remained relatively constant between 1995 and 1999, in spite of the economic growth at the time.
- Jacqueline Barocio and H. Luke Shaefer