Modelling dynamic choice: Private health insurance in Australia, CHERE Research Report 24

This study investigates the role of dynamic, discrete choice modelling in the context of private hospital insurance in Australia. This is achieved with the use of a unique panel data set of young Australian women ? The Australian Longitudinal Study on Women?s Health (or ALSWH). Very few (if any) private health insurance studies in Australia have used panel data due to the limited availability of longitudinal data sets. Yet panel data allows two important innovations ? it allows a researcher to control for unobserved heterogeneity across individuals and facilitates dynamic modelling that would otherwise require long time series data. Both these innovations are a feature of the dynamic, random effects probit model I propose here. Using the ALSWH data set I find that the choice to purchase private hospital insurance is strongly determined by income, access to hospitals and inertia in choice. I also find family formation, pregnancy, education, exercise levels and country of birth to be significant drivers of choice. Interestingly, I find little evidence of adverse selection in this sample of young women, as those more likely to be insured have higher self-reported health and few chronic conditions. Overall, I find the dynamic specification with state dependence effects provides an important insight into consumer behaviour, with young women exhibiting statistically and economically large amounts of inertia in choice. Women who were in cover in 1996 or 2000 were more likely to be covered in 2003. Conversely, women without cover were unlikely to move into cover despite a wide range of Australian Government incentives. As policy makers consider the future of private health care, researchers must consider dynamic studies as way to fully gauge consumer behaviour.