Objectives: Firstly, the study seeks to consider how factors at both individual and area level impact outcomes of underemployment (Baum, Bill, & Mitchell, 2008a, p. 200, 2008b, p. 5; Baum & Mitchell, 2008, p. 191, 2010, p. 13; Wilkins, 2004a, p. 366). My second objective is to contribute to the literature on underemployment by making my first objective more suitable for a single level model and better able to address the challenges of ecological validity (Baum & Mitchell, 2010, p. 20). Finally, I will test for a positive correlation between adverse labour market outcomes and social capital which is in accord with Australian findings (Baum & Mitchell, 2008, p. 199).
Design: Using 2016 HILDA data, this study runs two logistic regression models to investigate whether the dependent variable of underemployment is impacted more by the inclusion of an area level predictor, remoteness, net the individual level predictors age, sex, remoteness, social capital, mobility of residence and education (Baum et al., 2008a, pp. 198, 201; Baum & Mitchell, 2010, pp. 18–20; Flynn, 2003, p. 315; McCulloch, 2001, p. 673).
Subjects: Australians aged 15-44 who are currently in the labour market.
Results: By controlling for the area level variable, the predictors age, sex, education and remoteness all had strong odds ratios for underemployment outcomes of respondents, with education having the greatest influence. When interaction effects were tested for, strong gender inequalities were uncovered for outcomes of underemployment. Notably, completion of post-secondary education leads to heightened likelihood of underemployment for female youth but outcomes for male youth who achieve this educational level remain the same. Finally, metropolitan areas were found to have a greater likelihood of underemployment than non-metropolitan areas which could be due to youth migration out of rural areas to major cities.