What do the dynamics of employment, earnings and participation in social programs say about the potential of low income households to achieve self-sufficiency?

In this research we developed a typology of low income households that conceptually classifies them in five mutually exclusive groups in terms of their potential to achieve self-sufficiency.  In developing such typology we took two complementary approaches. On the one hand, we employed latent class analysis to classify households on the basis of how similar they were in their use of social programs (i.e. how alike those are in terms of their use of TANF, SNAP, Medicaid and CCAP subsidies) and their household earnings. On the other hand, we used k-mean clustering to classify those same households according to how alike they were in terms of a selected group of demographic attributes thought to explain their self-sufficiency trajectories. By providing conceptually rich typologies of households, those two complementary approaches allowed us to conceptualize the households’ potential for self-sufficiency as they belong to similar groups of low income families.

For the purposes of this research, we formed a unique longitudinal database of linked administrative records in which the units of analysis were female-headed Illinois households with at least one child under the age of five. The households comprised the population of participants in the Supplemental Nutrition Assistance Program in FY 2004. All households were followed retrospectively and prospectively in terms of their trajectories of use of TANF, Medicaid and CCDF-Childcare Subsidies programs; and they were also monitored in terms of their longitudinal entries and exits in UI-covered formal employment and earnings. We also identified a rich set of neighborhood and family characteristics associated to each one of the households.

Latent class analysis results:       

Our latent class analysis allowed us to classify households in five mutually exclusive types as follows:  1) Zero income (38 percent of all cases); 2) Strivers with income (38 percent); 3) New SNAP cases and stable (20 percent); 4) TANF cases(11percent), and 5) New SNAP cases, moving up (4 percent).

K-means cluster analysis results:       

As the result of K-means cluster analysis, we divided the population into five classes. 1) Non-Chicago heterogeneous income (33 percent); 2)  Chicago low income; 3) Chicago moderate income; 4) Non-Chicago moderate income (9 percent); 5) Non-Chicago high income (1 percent).

We explan the basic attributes of the households in each one of the categories presented above. We show how by classifying low income households in those categories, we gain a deep conceptual understanding of their prospects to achieve self-sufficiency.