Dynamics of Household Joblessness : Evidence from Australian Micro-Data 2001-2007

This paper investigates the persistence over time of living in a jobless household, aiming to disentangle the roles of state dependence and unobserved heterogeneity. In addition, the potential heterogeneity of state dependence is examined through estimation of interaction terms with the lagged household joblessness variable. Finally, the robustness of results is explored through the use of alternative definitions of household joblessness each based on different variables available in our data. Using the two definitions that are most different, we find substantial state dependence which is larger for women than for men under both definitions. That is, being in a jobless household in the previous year increases the probability of currently living in a jobless household by 7.7 to 17.2 percentage points for men and 12.7 to 25.1 percentage points for women. Although state dependence clearly is an important factor, as are a number of observed characteristics, unobserved heterogeneity also plays an important role for men and women: 32 to 40 per cent of the unexplained variance can be attributed to unobserved heterogeneity for men, and for women this is 42 to 46 per cent. A few characteristics (age, disability, student status, living outside of major cities, having a university degree, presence of preschool children) seem to affect the level of state dependence to some extent. However, aside from the age effect, which can increase state dependence by up to 50 per cent for men aged 60 to 64, the level of state dependence seems fairly homogenous amongst men and amongst women