英语论文网

lies groundless assumptions of labour market separation into male and female markets. In addition, Oaxaca himself notes the limitation of this model in that it ‘does not take into account the effects of the feedback from labour market discrimination on the male-female differences in the selected individual characteristics’, (Oaxaca 1973, p.708). It might be that females have less incentive to invest in human capital as their returns are lower, which already is an impact of discrimination. Lastly, the great challenge of almost any microeconometric model is the issue of omitted variable bias. Oaxaca-Blinder decomposition is regrettably no exception. It may be impossible to truly detach discrimination from other effects as it might be merged with unobserved factors. However, over the time, more and more productivity parameters have been included in wage gap regression in order to try and overcome this drawback, and in cases when such variables are not available, proxy indicators are substituted in order to achieve a more fully specified model (as we will see in this paper).

On the contrary, the reason the Oaxaca-Blinder technique was selected for this case study is straightforward. Despite its limitations the method has many benefits and is relatively easy to understand at an undergraduate level. The main reason for adopting this particular methodology in my study was its scientific prominence and applicability to a wider extent. Its foundations are built upon a solid neo-classical human capital theory, and is an excellent tool in explaining labour market phenomena through a conventional framework. It can also be employed as a tool in policy making and improvement since it brings all the relevant factors into the picture, leaving the controls in the background. Finally, the main advantage of Oaxaca-Blinder decomposition is that the gender component is clearly identified and the importance of it can be understood and compared with other elements. Since the pioneering works by Oaxaca (1973) and Blinder (1973) the decomposition methodology had developed significantly and can now be used with much more credibility.

The British Household Panel Survey 2002/2003 (later referred to as BHPS 2002/03) is the only source of data in this study. It contains information about the representative individual in the UK followed and annually interviewed. As a panel survey, it is a form of longitudinal study which is especially beneficial in terms of tracking the same individual observations throughout different periods. The advantages of the BHPS are that it is extensive, up-to-date, and implicates data on employment experience (if work history is collected). A more relevant advantage of using BHPS is that its data is extremely rich in productivity related characteristics useful in discrimination studies. Therefore, this paper looks into the gender wage gap in year 2002 taking a sample of individuals from BHPS.

The dataset analysed in my study is taken from wave twelve (dates of fieldwork 1 September 2002 to 30 April 2003). This paper refers to BHPS data from year 2002 so that the findings could be compared to results obtained by other works. Only one record (LINDRESP; containing 1348 variables and 16597 cases) from this wave was used to design a dataset for the following estimation. LINDRESP record is consisted of mainly job related and income related information. The data was cleansed from irrelevant and missing