This thesis investigates the public sector wage premium in the UK over the last decade using both econometric and economic modelling methods. A comprehensive literature review is conducted to summarise the four popular types of methods adopted by the existing microeconomics studies, which are weakly derived from some labour economic theories. A common problem of the economic methods is the difficulty in dealing with selection bias when valid instruments are not available. All four types of econometric methods are then applied to estimating the public sector wage premium, resulting in an overall average of 6.5% when a relatively higher female's premium. In particular, propensity score matching method provides the most robust estimate against mis-specification. As a bridge between microdata and macrodata in the labour market, the wage premium is shown to be counter-cyclical. Indirect inference is then introduced as a new method of testing and estimating a micro-founded economic model in the microdata analysis context. All four types of econometric methods are used as auxiliary models to summarise the data features, based on which the distance between the actual data and the model-simulated data is assessed. A calibrated model passes the test only when the propensity score matching method is used as the comparison criterion. To focus on the key properties of the model, the OLS coefficients are grouped into a smaller dimension, and the estimated model can also pass the test. The selection bias can be tested in a straightforward way under indirect inference, and we find no evidence for selection in the data. A Monte Carlo experiment is designed to verify the high statistical power of indirect inference test. Finally, a normative analysis is carried out and there is no evidence of unjust factors behind the observed public sector wage premium.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:716017 |
| Date | January 2016 |
| Creators | Wang, Yi |
| Publisher | Cardiff University |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://orca.cf.ac.uk/100994/ |
Page generated in 0.0044 seconds