Publication type
Thesis/Degree/Other Honours
Author
Publication date
June 1, 1998
Abstract:
This thesis is concerned with the measurement of horizontal inequity in the income tax and with the dynamics of earnings.
I present new methods to measure horizontal inequity and assess the behaviour of existing indices to changes in the income tax. In the first method I present equals are defined as those with the same pre-tax equivalent income. Inequality of post-tax income among pre-tax equals is evaluated and aggregated to form a global index of horizontal inequity. The vertical performance of the tax is measured by the difference between pre-tax inequality and the inequality in the hypothetical income distributions which would obtain if each individual had had the mean post-tax income of her equals group. Putting the two together, horizontal inequity measures loss of vertical performance. The identification problem is addressed by a procedure validating the banding of income units into 'close equals' groups. The horizontal and vertical effects of a major Spanish income tax reform are evaluated.
Building on the first methodology I present a model in which the tax breaks appended to the income tax are classified into two types, those with deserving social goals and those regarded as undeserving from an equity standpoint. This model provides the means to assess the cost of each type of break in terms of foregone equity. US income tax data for 1990 is used to illustrate the model.
The second part of the thesis uses the BHPS to study the level of mobility of British males within the earnings distribution from one year to the next and to analyse the extent to which earning differences are permanent or transitory.
I find that earnings mobility is low and that the more mobile male earners are the young, the least educated, the part-time workers and the self-employed. The process generating earnings transitions can be modelled as a second-order stationary Markov chain (assuming homogeneity) and as a second order non-stationary Markov chain (allowing for some heterogeneity).
Subjects
#496907