Public Assistance Use Among Two-Parent Families: An Analysis of TANF and Food Stamp Program Eligibility and Participation

Contents

• Data Sources and Methodological Approach
• Key Findings
• Conclusions

In accordance with the goals of the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) of 1996, states increasingly are focusing on family formation and on the role of state policy in promoting and supporting healthy marriages. To understand the role of state policy in promoting marriage, we must first look to existing programs and understand the role they play in the lives of married-parent families, particularly the extent to which low-income married-parent families are eligible for various public assistance programs and the degree to which eligible married-parent families obtain benefits.

Although public assistance programs such as the Food Stamp Program (FSP) and Temporary Assistance for Needy Families (TANF) are available to low-income married-parent families, married-parent families do not use these programs to the same extent as single-parent families. For instance, TANF programs have historically targeted single-parent families. Some research on the FSP suggests that eligible married-parent families are less likely than eligible single-parent families to participate in the program. However, little research has been conducted on married-parent families' TANF and FSP eligibility and participation rates, how these rates may have changed, or how the rates compare with rates for single-parent families. Furthermore, although some research has been conducted on the factors influencing the program participation decisions of single-parent families, little attention has been given to understanding the factors influencing the participation decisions of married-parent families.

To learn about TANF and FSP eligibility and participation of two-parent families, the Assistant Secretary for Planning and Evaluation, U.S. Department of Health and Human Services, contracted with Mathematica Policy Research, Inc. (MPR) to conduct an exploratory study. The goals of this study were (1) to determine appropriate data sources, methodologies, and data definitions for analyzing program eligibility and participation; (2) to document how both TANF and FSP eligibility and participation rates among married-parent families differ from the rates among single-parent families; (3) to explore, for both family types, the factors that are associated with eligibility and participation in TANF and FSP; (4) to examine TANF and FSP eligibility and participation rates for cohabiting families; and (5) to suggest avenues for further research on the program eligibility and participation of married-parent families.

In this study, we therefore sought to answer the following questions:

1. What Are Eligibility and Participation Rates in TANF and FSP Among Married-Parent Households?
   • What fraction of the low-income married-parent population is eligible for TANF and FSP? How does the eligibility rate among low-income married-parent households vary according to subgroups defined by such characteristics as the age of the household head, the ages and number of children in the household, and household income and participation status in other programs? How do eligibility rates in TANF and FSP among low-income married-parent households compare with those of low-income single-parent households?
   • What fraction of eligible married-parent households participates in TANF and FSP? How does the participation rate among eligible married-parent households vary according to the subgroups described above? How do these rates compare with those of eligible single-parent households?
   • How did eligibility and participation rates for married-parent households change during the mid- to late-1990s? How do trends in these rates compare with trends for single-parent households over the same period? How much of the change in participation is due to changes in the number of low-income households, changes in eligibility rates among low-income households, and changes in participation rates among the eligible?
2. What Factors Are Related to Eligibility and Participation in TANF and FSP Among Married-Parent Families?
   • What factors are related to TANF eligibility among the low-income population? While TANF eligibility is clearly a function of demographic and socioeconomic factors, state program policies, and state economic conditions, does the relative importance of these factors in predicting TANF eligibility differ for married- and single-parent families?
   • What factors are related to TANF and FSP participation among eligible married-parent families? To what extent do demographic and socioeconomic factors, state program policies, and state economic conditions predict participation rates among eligible families?
   • What are the differences in factors affecting the participation rates in TANF and FSP among eligible married-parent families versus eligible single-parent families? Can the lower participation rates of married-parent families be explained by differences in observed characteristics of the two family types, or do married- and single-parent families make fundamentally different participation decisions, even among families with very similar observed characteristics?
3. How Do TANF and FSP Eligibility and Participation Rates of Cohabiting Households Compare to Those of Married- and Single-Parent Households?
   • What fraction of low-income cohabiting households are eligible for TANF and FSP? How do eligibility rates in TANF and FSP among low-income cohabiting households compare with those of low-income married- and single-parent households?
   • What fraction of eligible cohabiting households participates in TANF and FSP? How do these rates compare with those of eligible married- and single-parent households?
   • How did eligibility and participation rates for cohabiting households change during the mid- to late-1990s? How do trends in these rates compare with trends for married- and single-parent households over the same period?

Data Sources and Methodological Approach

To address these questions, we used data from the Current Population Survey (CPS), a monthly survey conducted by the Bureau of Labor Statistics, and information on simulated program eligibility and participation from the Urban Institute's Transfer Income Model (TRIM3) and from MPR's Micro-Analysis of Transfers to Households (MATH®) microsimulation models. We supplement these data with state-level information on key program parameters and state economic conditions. All data are from the year 2000, the most recent year for which the microsimulation models were available at the time of the analysis. We used descriptive analytic methods to address the first and third sets of questions, and we used multivariate methods to address the second set of questions. It is important to note that the data, methods, and definitions used for these analyses were chosen to help inform the research questions of this report, rather than to provide point estimates of program caseloads. Therefore, the results presented here differ in numerous ways from official agency statistics released for TANF and FSP.

Because of ASPE's interest in keeping a common sample for determining eligibility in TANF and FSP, we examined participation and eligibility at the household level for our descriptive analysis. This also allowed us to capture characteristics of other individuals who are part of the household (such as a cohabiting partner or the parents of an unmarried mother), but are not classified as part of the family unit. Since the TANF program unit is typically the family, the unit in many cases is smaller than the household, and unit income may be smaller than household income. This is less likely to occur in the case of FSP, since the FSP program unit is typically the household. Because eligibility and participation determinations are made at the program-unit level, we aggregated the units to the household level. For the multivariate analysis, we conducted the analysis at the program unit level, but included both household- and unit-level characteristics as covariates in order to capture the characteristics of other household members that might influence program participation decisions.

To determine program eligibility, we used data simulated by the microsimulation models, as information on eligibility is not directly available from the CPS. We used these simulated data on eligibility for both the descriptive and multivariate analyses. The CPS has self-reported information on program participation, which we used for our multivariate analysis of factors related to participation. Because of underreporting of program participation in the CPS and other survey data, for our descriptive analyses we used simulated participation data from the microsimulation models, which correct for underreporting of participation.

Key Findings

1. What Are the Eligibility and Participation Rates in TANF and FSP Among Married-Parent Households?
   • Eligibility rates in both TANF and FSP are considerably lower for married-parent households than for single-parent households. Among the low-income population (households with incomes below 200 percent of the federal poverty level), only 15 percent of married-parent households were eligible for TANF, compared with 41 percent of single-parent households (Figure 1). Similarly, 33 percent of low-income married-parent households were eligible for FSP, compared with 57 percent of single-parent households.
     
     Figure 1.
     Eligibility and Participation Rates for TANF and the FSP, By Household Type, Year 2000.

     

     

     Source: Calculations from the March 2001 CPS, the Urban Institute's TRIM model, and the 2000 MATH CPS model, conducted by Mathematica Policy Research, Inc.
     Note: Eligibility rates are computed as the percentage of all low-income households (income less than 200 percent of the poverty level) that are eligible. Participation rates are computed as the percentage of all eligible households that participate, and are not limited to the low-income population.
     ^a Includes cohabiting households

   • Participation rates in TANF and FSP are lower for married-parent households than for single-parent households. Only 35 percent of all eligible married-parent households participated in TANF, compared with 57 percent of eligible single-parent households. Similarly, only 42 percent of eligible married-parent households, but 76 percent of eligible single-parent ones, participated in FSP.
   • Even within demographic and economic subgroups, married-parent households have lower eligibility and participation rates than do single-parent households. We examined eligibility and participation rates within subgroups based on the age of the household head, the race/ethnicity of the household head, family size, household income relative to the poverty level, and whether the household had any earnings. Within each subgroup, married-parent households had lower eligibility and participation rates in TANF and in FSP than did single-parent households. This suggests that even within these subgroups, the two household types may differ in ways that affect their TANF and FSP eligibility and participation.
   • While participation in both TANF and FSP decreased considerably between 1996 and 2000 for both married- and single-parent households, the decline was greater for married-parent households and was more strongly linked to a reduction in participation rates among the eligible. Between 1996 and 2000, the number of married-parent households participating in TANF fell by about 277,000, and the number of participating single-parent households by nearly 1.4 million. While nearly all the decline in participation among married-parent households was due to a decrease in the participation rate among eligible households, this accounted for just less than half of the decline among single-parent families. During the same period, the number of married-parent households participating in FSP fell by 729,000, and the number of participating single-parent households fell by about 1.3 million. Among married-parent households, over half the decline was due to a decrease in participation rates among the eligible. The most important factor in the decline in FSP participation among single-parent households was a decrease in the number of low-income households.
2. What Factors Are Related to Eligibility and Participation in TANF and FSP Among Married-Parent Families?
   • The primary factor explaining the difference in TANF eligibility rates between married- and single-parent families is the difference in their financial circumstances. Even within the low-income population that was the focus of the analysis, married-parent families tended have higher incomes than single-parent families, and this factor explains most of the observed differences in eligibility rates across the family types. Several other factors were also significant predictors of eligibility, including citizenship and age of youngest child, however, they were less important in explaining the differences in eligibility rates across family types.
   • Differences in TANF and FSP participation rates across family types are not fully explained by differences in observed characteristics, suggesting that there may be unobserved behavioral differences between married- and single-parent families. Even among families with similar demographic and financial characteristics who live in states with similar policies and economic conditions, eligible married-parent families are considerably less likely than eligible single-parent families to participate in TANF and FSP. The fact that a broad range of demographic characteristics, financial circumstances, and state policies explain so little of the differences in participation rates across family types suggests that the differences may be due to different behavioral responses across the family types. For example, compared with single-parent families, married-parent families might be more sensitive to stigma associated with collecting public assistance, or they might be more optimistic about their future employment prospects. The differences in participation rates may also be due to differences in other unobserved factors that are correlated with both family type and program participation decisions, such as unreported income, knowledge of eligibility, and unobserved variation in how state policies are actually implemented.
3. How Do TANF and FSP Eligibility and Participation Rates Among Cohabiting Households Compare to Those of Married- and Single-Parent Households?
   • Defining and identifying low-income cohabitating households in the data was challenging. These challenges lead to difficulties in modeling eligibility and participation. This is true especially with respect to the TANF program where the family (as opposed to the household) is the unit of observation, and income of cohabitors is treated fairly differently across states. There is considerable scope for future research in this area.
   • TANF and FSP eligibility rates of cohabiting households are closer to those of single-parent households than those of married-parent households. TANF eligibility rates for cohabiting households (51 percent) were greater than those of both single- and married-parent households. FSP eligibility rates for cohabiting households (53 percent) were between those of married- and single-parent households but were closer to those of single-parent households.
   • TANF and FSP participation rates of cohabiting households are between those of single- and married-parent households. TANF participation rates for cohabiting households (48 percent) were closer to those of single-parent households than those of married-parent households. FSP participation rates for cohabiting households (53 percent) were closer to those of married-parent households than to those of single-parent households.
   • TANF and FSP eligibility and participation rates of cohabiting households fell between 1996 and 2000, mirroring trends in the rates for single-parent households. Between 1996 and 2000, TANF eligibility rates for cohabiting households declined by six percentage points, rates for single-parent households fell by a similar amount, and rates for married-parent households increased slightly. TANF participation rates for cohabiting households also declined over this period, by 30 percentage points, mirroring declines of similar magnitude for both married- and single-parent households. FSP eligibility rates for cohabiting households declined by 9 percentage points, and FSP participation rates of these households declined by 17 percentage points, also mirroring similar declines for married- and single-parent families.

Conclusions

This study conducted exploratory research to learn more about factors related to eligibility and participation in TANF and FSP for married-parent families. Our analysis reveals the complexities in conducting such an analysis, including identifying appropriate data for eligibility and participation, defining family types, defining appropriate units for the analyses, and identifying methodological approaches to learn more about why eligibility and participation rates differ among the different family types.

We find that eligibility and participation rates in TANF and FSP are considerably lower for married-parent families than for single-parent families, as shown in Figure 1. Rates for cohabiting families generally lie between those of single- and married-parent families. Demographic characteristics and financial circumstances explain much of the difference in eligibility rates between married- and single-parent families. However, demographic characteristics, financial circumstances, and state program rules explain little of the observed differences in participation rates across the two family types.

This analysis suggests several avenues for further research. For instance, given the large unexplained differences that persist in participation rates between married- and single-parent families, it would be useful to learn why married-parent families have lower participation rates than single-parent families, even after controlling for numerous demographic and financial characteristics. One explanation may be related to differences in state policies for married- and single-parent households. Although we have included several policy variables that vary across states in our models, our models are unable to capture the effects of policies that differ for married- and single-parent families, but that do not vary across states. For instance, the work participation requirement for TANF is 55 hours for two-parent families compared with 30 hours for single-parent families. Although such differences may influence the participation decisions of these family types, we cannot capture them in our models if there is no variation in the rules across states. Additionally there may be unobserved state differences in the implementation of policies that affect married families differently than single families, and it may be useful to talk with key state officials to learn about how these policies are actually implemented for the two family types. It would also be valuable to understand the relative importance of such factors as stigma and families' failure to realize that they are eligible compared with factors that reflect the families' optimism about their future income or employment prospects. To learn more about this subject, as a starting point, it may be useful to conduct interviews or focus groups with small numbers of eligible married-parent families about their reasons for not participating in TANF and FSP. Finally, more research can be conducted on cohabiting households, who formed about 7 percent of all low-income households in the CPS.