Food Stamp Utilization Patterns in Nonmetropolitan Counties in Texas: A Multilevel Analysis of the Micro- and Macro-level Determinants of Caseload Dynamics

Year: 2000

Research Center: Southern Rural Development Center, Mississippi State University

Investigator: Swenson, Tami, Steve White, and Steve H. Murdock

Institution: Texas A&M University

Project Contact:
Tami Swenson
Center for Demographic and Socioeconomic Research and Education
Department of Rural Sociology
Texas A&M University
College Station, TX 77843-2125
Phone: 979-862-3060


In this report, Swenson, White, and Murdock examine and contrast food stamp caseload changes occurring in metro and nonmetro areas in Texas. Their primary objectives are to identify economic, sociodemographic, and policy factors associated with the dynamics of food stamp utilization and to examine the effects of such factors on the decline in the food stamp caseload. Much of the research on the dynamics of the food stamp caseload in the welfare reform period has used aggregate caseload or national survey data. These findings demonstrate that marginal economic and policy effects on the food stamp caseload decline nationally. Yet few regional studies or comparisons of metro and nonmetro areas within a State have been completed. Because nonmetro areas have distinct demographic and socioeconomic characteristics in Texas as elsewhere, an examination of economic and policy effects on nonmetro caseloads in Texas may provide useful information for rural areas throughout the South.

The authors use descriptive and analytical methods to evaluate the micro- and macro-level factors associated with the food stamp caseload in Texas from September 1995 through December 1999. They examined monthly administrative food stamp data by county, along with data on county socioeconomic conditions.

Using multilevel models, they estimate the effects of micro- and macro-level factors on exit probabilities and caseload decline.

Changes in the food stamp caseloads of metro and nonmetro areas in Texas suggest that both economic and policy factors may be affecting the rates of decline. From September 1995 to December 1999, caseload decline was substantial, dropping 47.6 percent in metro counties and 37.0 percent in nonmetro counties. Both metro and nonmetro caseloads experienced relatively greater rates of decline following the passage of welfare reform legislation. The nonmetro caseload had a larger relative increase in its rate of decline following welfare reform. A decrease in the number of entries was the primary cause of caseload decline in metro counties. In nonmetro counties, the decline has been a function of an increase in the number of exits from the Food Stamp Program. The metro caseload began to drop prior to welfare reform, and the number of entries has dropped faster in the post-reform period, suggesting that economic conditions may be more important than policy in reducing metro participation rates. In contrast, though the nonmetro caseload declined slightly prior to welfare reform, the substantial increase in the exit rate following reform suggests a larger role for policy-related processes in these areas.

Demographic differences between metro and nonmetro food stamp recipients mirror those of all metro and nonmetro residents. The nonmetro caseload is older and ethnically more Anglo, has a lower average level of education, a larger percentage of working recipients, more long-term recipients, and fewer able-bodied adults without dependents (ABAWDs) than the metro caseload. After reform, the proportion of long-term recipients, recipients with severe work impediments, and ABAWDs decreased in both metro and nonmetro areas. There were increases in the percentage of recipients able to obtain work without assistance, recipients with a medical incapacity, disqualified heads of household, and food stamp-only cases (those not also receiving TANF). The level of employment in the metro caseload increased after reform, but remained relatively constant in the nonmetro caseload. In metro counties, recipients remaining after welfare reform on average had lower education levels and less work experience. In contrast, the demographic and socioeconomic characteristics of the nonmetro caseload remained relatively constant.

Swenson et al. found that the probability of an individual’s exiting the Food Stamp Program was affected by different economic factors in the pre- and post-reform periods in nonmetro areas. The probability of exit was associated with declines in unemployment in 1995 and increases in wages in 1999. By contrast, declining unemployment and rising wages affected the probability of exiting the metro caseload in both time periods. In 1995, the effect of declining unemployment on the probability of exit was about 2.7 times greater in metro than in nonmetro counties. In 1999, increasing the frequency of recertification, which is required to continue receiving food stamps, increased the probability of exit by twice as much in metro as in nonmetro counties.

The overall rates of decline in the food stamp caseload were affected by both economic and policy factors, but the magnitudes of these effects were also significantly different in nonmetro and metro areas. Reducing unemployment decreased the size of the food stamp caseload by almost twice as much in metro as in nonmetro areas. Similarly, increasing the frequency of recertification for food stamp receipt resulted in much larger caseload declines in metro than nonmetro areas. Even with the 37-percent decline, the nonmetro caseload experienced little change in demographic composition and probability of employment. Because of this, the factors examined explain less of the decline in the nonmetro caseload. The authors conclude that the lower rates of decline observed in the nonmetro areas of Texas are best explained by differences in the demographic characteristics of the recipients and the economic conditions they face.