Essays on the distributional impacts of government

This thesis consists of three independent essays, unified by the common theme of the distributional impacts of government.The first paper estimates the price elasticity of demand for pharmaceuticals amongst high-income older people in Australia. It exploits a natural experiment by which some people gained entitlement to a price reduction through the Commonwealth Seniors Health Card (CSHC). The preferred model is a nonlinear Instrumental Variable (IV) difference-in-difference regression, estimated on repeated cross sectional survey data using the Generalised Method of Moments. No significant evidence is found for endogenous card take-up, and so cross-sectional estimates are also considered. Taking all of the results and possible sources of bias into account, the ?headline? estimate is -0.1, implying that quantity demanded is not highly responsive to price.The elasticity estimate is a key input into the second paper which analyses the distributional impact of the CSHC. I consider the trade-off between moral hazard and risk pooling. There have been few previous attempts internationally to address this trade-off empirically for any health insurance scheme. The utility gain through risk-pooling is found to be negligible. However, the deadweight loss through moral hazard may be considerable. I also use an illustrative model to demonstrate the possible effects of the CSHC on inter-temporal savings behaviour. While the CSHC may induce some people to save, it may have the opposite effect on others. The net impact was not determined.The third paper estimates the Australian public sector wage premium. It includes a detailed critical review of the methods available to address this issue. The chosen approach is a quasi-differenced panel data model, estimated by nonlinear IV, which has many advantages over other methods and has not been used before for this topic. I find a positive average public sector wage premium for both sexes. The best estimates are 10.0% for men and 7.1% for women. The estimate for men is statistically significant (p < 0.04) and borders on significance for women (p < 0.07). No evidence is found to suggest that the public sector has an equalising effect on the wages of its workers.