United States Senate
Committee on Agriculture, Nutrition, and Forestry

April 23, 1998
 

Chris Hamilton
Abt Associates Inc.

What Makes Caseloads Grow or Shrink in the Food Stamp Program?

Food Stamp Program (FSP) caseloads rose dramatically between 1988 and 1994, and have since fallen sharply.  Although unemployment rose and fell during the same period, previous economic cycles have not produced such dramatic movements in the caseload.  This prompts the question of why, and whether some important change has occurred in the relationship between the economy and food stamp caseloads.

Understanding caseload patterns and trends is important for two reasons.  Food stamp caseload growth and shrinkage affects the federal budget and, to a lesser degree, states' budgets for administering the program.  In addition, caseload patterns, taken together with other information on household circumstances, can provide signals about the well-being of American families.

This paper reviews briefly some of the factors that affect the size and composition of food stamp caseloads.  It suggests that the economy, demographic patterns, and policy changes are and will continue to be the main determinants of the number of households potentially eligible for food stamps.  It argues, however, that additional factors must be considered in order to know how many of the potentially eligible households will actually participate.  These factors are currently only partially understood and not well measured.  In the near term, moreover, welfare reform is likely to affect household participation decisions, and therefore food stamp caseloads, in ways that will not necessarily correspond to changes in the economy.

The economy is important, but it's not the only thing.

Any discussion of factors affecting food stamp caseloads has to begin with the economy.  Food stamp eligibility is determined mainly by low income.  Common sense dictates that, since more households have low incomes in economic bad times than good ones, bad times will see more people eligible for food stamps.  If more people are eligible, common sense goes on to say, more will participate.

The relationship between the economy and food stamp caseloads is strong enough to be readily visible in time graphs such as that in Exhibit 1.  Over the 22 years covered in the graph, the peaks and valleys of the national food stamp caseload have occurred in close proximity to peaks and valleys in the US unemployment rate.  Even in the recent period, when forecasts have not been able to predict caseloads accurately on the basis of economic factors, one can still see corresponding patterns of rise and fall in unemployment and the food stamp caseload.

Research aimed at predicting or explaining food stamp caseload changes has consistently found the economy to be an important factor.  The forecasting models used by the Congressional Budget Office rely heavily on economic variables.   Similarly, models developed to forecast caseloads in the main cash assistance programs — formerly Aid to Families with Dependent Children (AFDC), now Temporary Assistance to Needy Families (TANF) — usually include such factors.
 

If unemployment explained changes in caseload fully, the story would end here.  But employment, although a very important part of the picture, is only one part.  This is illustrated in Exhibit 1:  the food stamp caseload trend misses the sharp unemployment spike in the early 1980s, declines less than unemployment in the late 1980s, and then seems to exaggerate the unemployment pattern in the 1990s.  It is necessary to consider what else is going on.

Separate forces determine the pool of potential eligibles and the number of eligibles who participate.

To understand the limits of the relationship between unemployment and the food stamp caseload, it is useful to look broadly at the factors that determine how many people participate.  Exhibit 2 suggests that a two-step process is involved.  The first step determines how many households meet the program's income and other eligibility criteria, and the second determines how many of those households actually receive food stamp benefits.
 

The size and composition of the population, the eligibility rules established by legislation, and the economy jointly determine how many households might be eligible for assistance.  In most years, demographic patterns change slowly and steadily and there are no major changes in eligibility policy.  Over any given period of a few years, then, the economy is likely to be the main cause of fluctuations in the number of potentially eligible persons.  In any event, caseload analyses typically take all three factors into account.

Low-income persons whose characteristics would make them eligible for food stamp benefits do not receive benefits automatically.  Rather, they must take positive action to do so.  They must apply for benefits or, if they are already receiving food stamps, must meet the requirements for continued eligibility, such as appearing for recertifications.

In deciding whether to take the necessary actions, households implicitly compare the value of the benefits they could expect to receive to the cost or difficulty of doing what is necessary to obtain and use the benefits.  Based on their own understanding of the Food Stamp Program (or lack of understanding), many households do not take the actions necessary to participate.  The number of households receiving food stamp benefits is therefore always substantially less than the number who are potentially eligible.  In August 1995, for example, about two-thirds of the households potentially eligible for benefits actually participated.

If people in similar circumstances always made the same decision about whether to participate in the Food Stamp Program, the participation rate — that is, the percent of potentially eligible persons who actually participate — would remain fairly stable over time.  Changes in the number of potentially eligible persons would translate directly into changes in the number of food stamp participants.  History indicates, however, that the participation rate does not remain constant.

In fact, much of the 1988-1994 growth in the food stamp caseload reflects a sharp climb in the participation rate.  The number of potentially eligible persons grew 16.1 percent from 1988 to 1993, which would translate into a 16.1 percent caseload increase if the participation rate remained constant.  But the participation rate also increased substantially.  As a result, the caseload grew 48.2 percent, a much steeper climb than the growth in potentially eligible persons.

To understand caseload changes, then, one needs to know the reasons for changes in the participation rate of potentially eligible households.  Unfortunately, this piece of the puzzle is much less well understood than role of demographic, economic, and policy factors.

Households respond to the potential food stamp benefit amount, but with uncertainty and with varying valuations.

Households with larger potential food stamp allotments have more to gain from participating than those who would qualify for smaller amounts.  They may also have greater need, because they have lower incomes and/or larger families.  Not surprisingly, then, participation rates have been found closely related to the potential benefit amount.  Groups that tend to qualify for higher benefit amounts, such as households without earned income and large households, tend to have higher participation rates.

Potentially eligible persons do not necessarily have a clear understanding of their eligibility or the amount of food stamps they can receive, however.  Surveys that ask apparently eligible persons why they do not participate have typically found large proportions of respondents saying that they do not believe they are eligible for benefits.  Uncertainty about eligibility leads households to discount the value of potential benefits.

Even households who fully understand their potential food stamp amount may differ in how much they value the benefit.  Because the benefit can be spent only for food, households who place a high priority on covering their food expenditures may be more likely to participate than households whose biggest worry is, for example, paying the landlord.  Research on food insecurity and hunger has shown that households in apparently similar financial circumstances can experience quite different levels of food insecurity, although little is known about what causes the differences.

Households take into account the time, hassle, and psychological effects of participation.

The low-income household weighs the expected food stamp benefit against the expected costs of participation.  Two kinds of cost are pertinent:  the time, money, and "hassle" required to apply for benefits and comply with subsequent program requirements; and the psychological negatives associated with depending on public assistance and being known to do so.

Applying for food stamp benefits takes an average of about five hours of the applicant's time and about $10 in out-of-pocket expenditures, according to one study.  The process can also involve uncertainty and confusion.  Some people who begin the application process fail to complete it because they become uncertain about what to do next or how to comply with requirements such as providing documentation on eligibility factors.  Households who are approved for benefits must be recertified periodically, which involves a somewhat streamlined repetition of the application process.  Some types of recipients face a further requirement to participate in employment or training activities.

The psychological cost or stigma of receiving food stamps, though not well measured in empirical research, is almost certainly a factor in some households' participation decisions.  Surveys asking about reasons for non-participation typically find a small percentage of respondents mentioning embarrassment or pride, and these factors may be especially important among the elderly.  Issuing food stamp benefits as cash or electronic benefits, which involve less public identification of the user as a food stamp recipient, leads to some reduction in the psychological negatives.  Little is known, however, about whether the psychological cost stays relatively constant or changes over time.

Participation in AFDC/TANF plays a special role in food stamp participation patterns.

The Food Stamp Program has been closely coordinated in local offices with cash assistance programs, especially AFDC and TANF. This can dramatically affect the household's assessment of the costs and benefits of participation.  Participating in both programs has only marginally higher costs than participating in one, but cash assistance adds substantial dollar value to the benefit.

As a result, FSP participation is very strongly associated with cash assistance participation.  The vast majority of single-parent households who receive cash assistance also receive food stamps, as indicated by the lighter bars in Exhibit 3.  But food stamp participation is much less common among single-parent households who do not receive cash assistance (darker bars in Exhibit 3), even when households are grouped by income levels.  The result is that, for a substantial part of the food stamp caseload, food stamp participation closely parallels AFDC/TANF participation.

Somewhat surprisingly, even though more than half of all food stamp cases do not receive cash assistance, the national food stamp caseload has closely tracked the national AFDC/TANF caseload since the early 1980s.  Referring back to Exhibit 1, note that the food stamp and cash assistance caseload trends parallel each other more closely than they parallel the unemployment rate.  This may mean that, beyond the administrative linkage, both programs are influenced by some underlying pattern in households' assessments of the costs and benefits of participation. This might reflect economic factors, such as perceptions about job opportunities, or non- economic factors.  One can speculate, for example, that the negative portrayal of "welfare as we know it" in the public debate of the 1990s may have increased the psychological costs of participation, which might contribute to a lower participation rate than would otherwise occur.  Or perhaps a period of increasing participation in any one program leads to a broader community effect, increasing people's expectations about the likelihood of being able to obtain benefits from other programs.  Such speculations are plausible, but not substantiated in empirical research as yet.

Welfare reform is likely to affect food stamp participation patterns.

Apart from the direct legislative changes to the Food Stamp Program, welfare reform seems likely to have indirect effects on food stamp caseloads through changes to cash assistance.  Most importantly, welfare reform is expected to reduce cash assistance caseloads, either by moving recipients into employment or by ending benefits as recipients reach time limits.  To the extent that families' participation in TANF influences their FSP participation, food stamp caseloads will decline accordingly.

This phenomenon is illustrated in Exhibit 4, which is based on data from on-going welfare reform demonstration in Indiana.  The exhibit tracks households' rates of receipt of cash assistance and food stamps for about the first two years after they entered the demonstration.  The lines reflect the experience of the group that was subject to the welfare reform provisions being tested.  The proportion of families receiving cash assistance declined steadily over the two years, reflecting normal patterns of welfare exits in a strong economy as well as the effects of welfare reform.  The proportion receiving food stamps declined somewhat more slowly, indicating that some families continued to receive food stamps after they had stopped receiving cash assistance.  The gap between the cash assistance and food stamp lines seems small, however, in light of the fact that over 95 percent of the families still had gross incomes below 130 percent of the poverty line at the end of the study period.  In other words, most of the families that stopped receiving food stamps were still potentially eligible for benefits.

Welfare reform is expected to reduce cash assistance caseloads, and the early results indicate that this occurred in Indiana, with a reduction of about six percentage points after the first five quarters.  Preliminary analyses of food stamp data suggest roughly comparable reductions in the food stamp caseload.  Such patterns indicate that welfare reform reduces food stamp participation without a corresponding reduction in the number of potentially eligible families.

This "copycat effect" of welfare reform on the food stamp caseload reflects the kind of cost- benefit consideration discussed earlier.  Having left cash assistance, some households probably feel that continuing to receive food stamps is not worth the time and effort of being recertified or the psychological costs of participating.  In addition, some households may be uncertain of their continued eligibility for food stamps.  A survey of families in Delaware's welfare reform demonstration revealed considerable confusion about cash assistance rules, and such confusion might well carry over to the FSP.

Concluding Observations

Food Stamp Program caseloads have been and will continue to be strongly influenced by the economy, demographic patterns, and eligibility rules.  These factors determine the number of persons potentially eligible for benefits.  They are normally considered in analyses of caseload changes, but they do not fully account for the changes.

Caseload is determined not only by the number of potentially eligible households, but also by the percentage of those households who take the actions necessary to receive benefits.  Fluctuations in this percentage can be — and recently have been — as important as changes in the economy, and they can either magnify or counteract economic changes.  But the process through which potentially eligible households decide to whether or not to participate is not well understood or measured.  We do not know to what extent changes in the participation rate reflect, for example, changing expectations about future employment opportunities, policy or administrative changes in the FSP and related programs, changes in the community understanding of the programs, or changes in the social acceptability of receiving assistance.  Until such factors can be understood and taken into account, we cannot expect to know exactly how economic changes will affect caseloads.

In the near term, welfare reform is likely to affect food stamp caseloads in ways that would not be predicted from economic and demographic factors.  It is reasonable to guess that, to the extent that welfare reform reduces cash assistance caseloads, the food stamp participation rate will also decline.  While such an effect persists, food stamp caseloads will probably fall below the levels that would be predicted by economic changes alone.