Food stamp participation rates plummeted from 75 percent in 1994 to 59 percent in 2000 (Cunnyngham 2002, p. 3).1 In response to plummeting participation rates, and with the new flexibility brought about by the 1996 federal welfare reforms, many states are re-engineering their programs to improve accessibility (Rosenbaum 2000; Bell, et al. 2001). States are extending office hours, establishing automated call centers, and improving outreach, among other changes. But not all program changes are geared toward increasing participation rates. States have strong financial incentives to keep Food Stamp Program certification error rates low, a goal that often runs counter to improving participation rates. States are making policy decisions -- which have strong implications for Food Stamp Program participation decisions -- without the benefit of knowing the factors that make some eligible working persons choose to participate and others choose not to participate. This study will examine how low-income households' employment characteristics influence Food Stamp Program participation. The relationship between employment and Food Stamp Program (FSP) participation is of special interest for two reasons. First, characteristics of the food stamp caseload and the food stamp eligible population have changed to include more working low-income households. As Gleason et al. (2000) note, there has been a large increase in the proportion of food stamp participants with earnings. Among food stamp recipient families with children, the percentage working increased from 27 percent in 1993 to 42 percent in 1999 (Center on Budget and Policy Priorities 2001a, p. 1). Second, along with the declining participation rate has come a growing concern that eligible working low-income families are not participating in the Food Stamp Program. "Food stamps are crucial to helping low-wage working families make ends meet. A family of four supported by a full-time, year-round minimum wage worker will fall short of the poverty line by 25 percent (even after counting the earned income tax credit) if the family does not receive food stamps. Food stamps increase the typical monthly purchasing power of such a family by 39 percent" (Center on Budget and Policy Priorities 2001a, p. 4). Food stamp participation may reduce the chance that families are unable to financially meet basic needs and so use other forms of public assistance. It is important to understand how the Food Stamp Program works for the large fraction of the caseload that is employed, but it is even more important to understand why the Food Stamp Program does not work for low-income working persons who do not participate. The Food Stamp Program structure, with its numerous application rules, program requirements, and administrative practices, may be one reason that working low-income persons choose not to participate. As Besharov (2000) has argued, the Food Stamp Program was built around the non-working poor and the program for the working poor looks like an afterthought. Indeed, important aspects of the program do make participation difficult for the working low-income persons by effectively raising the monetary and nonmonetary costs of participation. For example, many individuals are required to appear in person at their local food stamp office to apply for food stamps and, in most cases, for periodic recertification. In-person application and recertification are more costly for the working low-income persons because the opportunity cost of their time is higher and they may have less available free time. It may be especially costly for people who work during traditional hours (for example, from 9 a.m. to 5 p.m.) because they have a smaller time window to get to the food stamp office and may need to be absent from work to apply or be recertified for benefits. Certification policies provide another example of the increased cost of participation for the working low-income persons. In the late 1990s, many states shortened the certification period for households with a history of earned income to reduce the number of errors in the Food Stamp Program (Gabor and Botsko 2001).2 As a result, working food stamp recipients were required to return to the food stamp office for recertification even more often than non-working persons (Dion and Pavetti 2000). Furthermore, since food stamp benefits decline with income, working low-income persons face higher costs to participation for a smaller benefit amount. To provide an understanding of the relationship between employment and FSP participation, this analysis examines the employment characteristics and patterns of the working food stamp eligible population. In particular, we address three research questions: What are the detailed employment characteristics of low-income, working food stamp participating and non-participating households? Do household members work traditional hours? Are there multiple jobholders in the household? How many hours do household members work? Do they change jobs frequently? How do detailed measures of employment characteristics affect food stamp participation? Does labor force attachment affect participation? Are persons who work non-traditional hours more likely to participate? Does holding more than one job decrease the likelihood of participation? Does working more hours decrease the likelihood of participation? Do frequent job changes decrease the likelihood of participation? How has the relationship between employment factors and Food Stamp Program participation changed since federal welfare reform? Understanding the factors that affect participation decisions among working low-income individuals is necessary to ensure access to program benefits. Identifying these factors will shed light on how the Food Stamp Program is currently operating for the working low-income individuals and how it might be changed to better accommodate these individuals.