A report from the National Research Council (NRC) of the National Academy of Sciences reviews the concepts and methodology for measuring food insecurity and hunger. It recommends that USDA no longer refer to the more severe forms of food insecurity as “hunger” since hunger is a physiological condition experienced at the individual level and not necessarily at the household level (NRC, 2006). In line with this recommendation, USDA has classified the most severe form of food insecurity as “very low food secure,” which it identifies if a household answers affirmatively to six or more (eight or more if children are present) questions on the CFSM. The NRC (2006) report includes the recommendation that USDA continue to measure and monitor food insecurity regularly in a household survey. It recommends further that, given that hunger is a separate concept from food insecurity, USDA undertake a program to measure hunger, which is an important potential consequence of food insecurity. The report also concludes that exclusive reliance on trends in the prevalence of food insecurity would not be an appropriate measure of the effectiveness of food assistance programs such as SNAP. For program evaluation purposes, it is important to know what effect SNAP has on food insecurity. As discussed here and in Chapter 2, however, a challenge facing evaluation of the impact of SNAP on food insecurity is the prospect of reverse causality; that is, food insecure households may self-select into SNAP. Several authors have used sophisticated econometric techniques to model the self-selection process and, after controlling for nonrandom selection, generally have found that SNAP reduces food insecurity (Gundersen and Oliveira, 2001; Kreider et al., 2012).
Figure 2-5 shows trends in 12-month prevalence rates of food insecurity and very low food security among U.S. households from 1995 through 2011. Prevalence rates for 1996 and 1997 were adjusted for the estimated effects of differences in data collection screening protocols used in those years. The supplements were conducted in various months in the initial years but since 2001 have been fielded in December, which implies that the 12-month recall refers to the actual year of the survey. The fraction of households experiencing food insecurity or very low food security held fairly steady until the Great Recession that began at the end of 2007. Thereafter, food insecurity increased by 31 percent and very low food security by 32 percent, although both indicators fell slightly between 2009 and 2011 as the economic recovery began to gain traction.
These trends in food insecurity must be interpreted in the context of other factors that may impact access to food for low-income households, including changes in income distribution across the low-income range, noncash assistance (e.g., participation in other assistance programs), and other basic household needs (Nord, 2007). Indeed, an apparent contradiction in Figure 2-5 is that as SNAP participation and expenditures accelerated in the latter half of the past decade, food insecurity accelerated as well. In