same time, it has serious questions about the operational and methodological issues. The panel concludes that until better methods to survey the homeless are developed, continuing to limit the target population to the household population seems appropriate. The panel, however, urges USDA to undertake research as part of its long-term agenda, leading to obtaining estimates of food insecurity in periodic or a one-time survey to get a sense of the negative bias of excluding this population in the household survey.
However, even an appropriate measure of food insecurity or hunger using appropriate samples would not be a definitive performance indicator of food assistance programs because their performance is only one of many factors that result in food insecurity. Consequently, changes in food insecurity could be due to many factors other than the performance of the food safety net. For example, if food insecurity declined because the price of food declined, the decline in food insecurity may not indicate better performance of the food safety net, because these programs would not have been responsible for all or part of the change. Conversely, if the reverse were to occur—that is, if food prices were to rise steeply or household income were to fall—the result might be an increase in the number of food-insecure households. But this, too, would not be the result of a decline in the performance of the food safety net. Developments like these could result in errors in assessing the performance of such programs as food stamps that are intended to reduce food insecurity and hunger.
USDA staff and colleagues have studied issues of this kind in a number of ways. They have compared the food security data with information collected in other surveys, such as the Survey of Income and Program Participation (SIPP) and the National Health and Nutrition Examination Survey (NHANES). They are currently using the panel feature of the CPS to look at the food insecurity of households as they approach the beginning of a food stamp spell—a period of one or more months during which a household received food stamps every month. Analysis is still in progress. A recent paper by Wilde and Nord (2005) used the food security data collected in 2002 and 2003 to estimate the effect of Food Stamp Program participation on food security. They used CPS Food Security Supplement data for December 2001 and December 2002 to determine the change in food insecurity with hunger status for food stamp participants who were in the survey both in December 2001 and in December 2002. They found that only 41 percent of the food-insecure households with hunger in 2001 had become food secure in 2002. They concluded that “it appears that unobserved hardships strike from time to time, with large effects on both program participation and food security. These hardships are sufficiently severe to swamp the presumably beneficial direct effect of food stamps on food security” (Wilde and Nord, 2005, pp. 430–431).
To better attribute changes in food insecurity or hunger to the food