Thesis (Ph.D.)--Aberdeen University, 2009. / Title from web page (viewed on Apr. 1, 2010). Includes bibliographical references.

This thesis uses personnel records from a large UK based financial sector employer to aid investigation into questions relating to both wage determination and employee turnover.  The thesis begins by turning to the origins of the human capital earnings function (Mincer, 1974).  The emphasis of the discussion is that other specifications can also accommodate estimates that are consistent with the basic components of Mincer’s analysis.  Motivated by these theoretical considerations as well as existing empirical evidence, I model the earnings function as a semi-parametric partial linear model.  The estimation results reveal that conventional parametric specifications result in biased estimates of the growth in wages with within-firm experience.  In addition, conventional specifications do not adequately serve as control variables when the effects of other covariates on wages are of interest. Interpreting the relationship between tenure and wages forms the focus of the remainder of the thesis.  Several theoretical models predict a causal relationship between these variables and estimates of the causal effect of tenure on wages have ranged from negligible to substantial (Garen, 1988).  I begin the investigation by assessing the employee exit decision.  Utilising information on the exact states reason for each employee’s departure and available data on supervisor performance ratings, I also form and estimate models of employment duration.  Estimation results are consistent with the job matching model presented by Jovanovic (1979).  Results also show augmenting hazard models to incorporate reasons for departure is important for understanding gender differences in turnover.  Utilising the econometric methods developed by Altonji and Shakotko (1987) and Topel (1991), as well as evidence from the turnover investigation, I find notable effects of tenure on wages.  Interestingly, both estimators deliver similar estimates and there are small gender differences in wage – tenure profile.  Arguments are presented in order to explain these findings.