A National-Level Study of Food Stamp Program Participation Among American Indians: 1988-2004

Year: 2006

Research Center: American Indian Studies Program, The University of Arizona

Investigator: Gundersen, Craig, Dean Jolliffe, and Laura Tiehen

Institution: Iowa State University

Project Contact:
Craig Gundersen
Department of Human Development and Family Studies
Iowa State University
74 LeBaron Hall
Ames, IA 50011
Phone: 515-294-6319


The economic status of American Indians is substantially worse than that of the general U.S. population. For example, per capita income of American Indians is 40 percent below the national average, large numbers have earnings in the lower end of the income distribution, and in counties with a high proportion of American Indians, per capita income is much lower than in other counties. This worse economic status is also seen in direct indicators of well-being. For example, one in five American Indians was food insecure compared with about one in eight for the entire population in 2003. Further, the rate of food insecurity with hunger was twice as high for American Indians than for the rest of the population.

A large share of American Indians, given the high poverty rate among them, would be eligible for food stamps and presumably would benefit from receiving this assistance.

A series of public policies are being used to address the hardships facing American Indians. One program with the potential to have a major impact is the Food Stamp Program. In light of the importance of food stamps, understanding food stamp use by American Indians in relation to the general population seems relevant. This paper represents the first examination of food stamp use among American Indians at the national level.

The participation rate among American Indians in the Food Stamp Program is compared with the rate among the general population by using data from the 1989-2005 March Supplements of the Current Population Survey (CPS) (calendar years 1988-2004). From these data, samples are constructed consisting of (1) gross-income-eligible households and (2) gross-income- and asset-eligible households. The use of multiple years allows the study to ascertain how the participation rate may have changed over time. Results indicate that, from 1988 to 2004, the participation rate of American Indians was higher than the rate of the general population in every year. This result holds for both the gross-income-eligible sample and the gross-income- and asset-eligible sample. This difference does not generally hold, however, once the sample is broken down by presence of children in households and by whether the household is in a metro or nonmetro area.

The probability of American Indian households receiving food stamps is greater than other households in both metro and nonmetro areas.

Next, the study considers, after controlling for other factors, the effect of American Indian status on the probability of food stamp participation. For the full sample under the gross-income test, American Indians have a 30.1 percent higher probability of receiving food stamps. For the gross-income- and asset-eligible-sample, the difference is 20.3 percent. When the sample is restricted to households with children and households without children, American Indians again have higher food stamp participation rates, contrary to the bivariate results discussed above. All else being equal, in the sample of all gross-income-eligible households, the probability of an American Indian household receiving food stamps is 41.3 percent in nonmetro areas and 33.6 percent in metro areas. In contrast, households in the rest of the population in nonmetro areas have a 35.1 percent probability of receiving food stamps and in metro areas a 27.8 percent probability.

To examine the manner in which the unequal distribution of American Indians across States affects the results, the sample is further limited to States where more than 3 percent of the food-stamp-eligible population is American Indian. For all households and for households with children, the restriction to fewer States does not have much of an influence on the estimated coefficients. For households without children, however, in the gross- and asset-income-eligible sample, the effect of being American Indian on food stamp participation is insignificant. Also, the joint effects of American Indian status, nonmetro residence, and their interaction in the gross- and asset-income-eligible sample is insignificant.