The Supplemental Nutrition Assistance Program (SNAP) aims to help low-income families obtain an adequate and nutritious diet; there is also ample evidence that it functions as an income support.  From 2006-2009 (pre-recession to full-blown recession) California’s SNAP caseload grew by 41%, while the number  in poverty grew by 16%.  At the same time, only an estimated 52% of SNAP eligibles in California were enrolled in 2009. 

This paper decomposes the expansion of the SNAP caseload in California in order to identify the sources of growth and areas of likely under-enrollment.  The research is built around merged individual-level, longitudinal administrative data for adults age 18-64 in SNAP between 1/2006 – 12/2009.  Data include SNAP receipt (from the state's Medi-Cal Eligibility Determination System (MEDS)) and quarterly UI-covered wages (from the Base Wage file). 

Prime-age adult SNAP recipients are diverse, some with recent job history, some currently employed, and some chronically without a job; some with children in the household, and others “able-bodied adults without dependents” (ABAWDs), and the analyses decompose caseload growth into multiple subgroups.  The varying experiences of counties in the recession are also taken into account: the 58 counties are grouped into three by pre-recession unemployment rates and unemployment rate growth.

We quantify the sensitivity of the six SNAP caseload subgroups (defined by recency of UI-covered earnings) to the economy.  The coefficients on current and lagged unemployment are generally substantively large and are significant.  However, results to-date suggest that the response is nonlinear, with the caseload increasing proportionately less for larger increases in the unemployment rate. 

We then pursue the relationship of work and SNAP receipt by decomposing the caseload expansion into the two primary mechanisms: due to slower pace of exits or due to increased numbersentering.  To address under- enrollment, both might matter.  We find increased entrances to have been twice to three times as important in the recession as reduced exits, across types of adults and across county context. 

The number of entrants in each county type grew by 60 to 80 percent. Entrants to SNAP almost always have either no earnings in recent quarters, or have lower earnings in the quarter of entry than in the previous quarter.  Among parents, in the fourth quarter of 2006 55% of entrants had earnings in the prior quarter and only 18% had no earnings for at least two years; by 2010 these shares were 40% with recent earnings and 30% with no earnings in two years.  The absolute numbers of entrants with earnings in the prior quarter declined steadily from mid-2008 onwards.  ABAWD recipients likewise show a substantial reduction in the share of entrants who had prior earnings. 

Finally, dividing exiters into those with and without earnings, we see that exits among the former declined sharply, but the same proportion exited each month without subsequent earnings:  about 2% of parents and 6.5% of ABAWDs left each month without earnings. This steady level during the recession is puzzling and, from the perspective of under enrollment, troubling.