A shared parameter mixture model for longitudinal income data with missing responses and zero rounding

The analysis of longitudinal income data is often made challenging for several reasons. For example, in a national Australian survey on over time, non-negligible proportion responses are missing, and it believed the missingness mechanism non-ignorable. Also, there large number reported zero incomes, some which may be true zeros (corresponding to individuals who legitimately do not earn an income), while false choosing round their zero). We propose new shared parameter mixture (SPM) model analysing semi-continuous data, addresses two challenges non-response rounding. This accomplished by jointly modelling individual's underlying together with probability rounding zero, where both probabilities permitted vary smooth manner non-zero income. Applying SPM reveals that average, older female long-term health condition considerably less likely income, tended highest male fixed-term/permanent job contracts between ages 50 60. Furthermore evidence rounding, conditional assumed mechanism, incomes at higher lower ends more report