This paper investigates the public sector wage premium in the UK using a microfounded eco-nomic model and indirect inference. The neoclassical wage determination model is tested and estimated without introducing any gap between the theoretical and empirical models. To test if the model is true, four types of econometric methods are used to summarise the data features, based on which we can evaluate the distance between the observed data and the model-simu-lated data in the test. When the distance is minimised, we estimated a public sector wage pre-mium between 6% and 7% using both traditional microeconometrics and indirect inference. In addition, selection bias test can be incorporated into the indirect inference procedures in a straightforward way, and we find no evidence for it in the data. Finally, in a simulation based on the estimated model, we show that it is not the non-market factors, but the total costs and benefits of working in different sectors and the pure market force, that create the public sector wage premium. There is no inefficiency or unfairness in the labour market to justify govern-ment intervention.