Equality of opportunity, heterogeneity and poverty
Abstract: Paper [I] studies equality of opportunity in Sweden. The distinction between circumstances that constrain an individual’s opportunities and the individual choices also affecting a particular outcome is the main idea of theories of equality of opportunity. In this study, equality of opportunity is analyzed for Swedish data using a large set of variables indicating different circumstances likely to affect an individual’s opportunities. A semiparametric model is estimated to allow for a possible nonlinear relation between parental income and the income of the adult child. The reason is a hypothesis that a constrained investment behavior would make the relationship nonlinear. The results indicate significant inequality of opportunities. However, they do not indicate a nonlinear relationship between parental income and the income of the adult child. Thus, the hypothesis that low income families will have a constrained investment behavior in human capital formation is brought into question as the explanation of intergenerational income correlation in Sweden.Paper [II] focuses on the persistence of poverty in Sweden. The purpose is to distinguish between two different reasons why poverty could persist on an individual level. By using a sample of identical twins, this study takes advantage of the similarity within pairs of twins to separate family specific heterogeneity from true state dependence, where the experience of poverty leads to a higher risk of future poverty. The results, based on a four variate probit model, show the importance of true state dependence in poverty. When using the information on whether an individual received social assistance as a measure of poverty, family specific heterogeneity explains between 24 and 31 percent of the poverty persistence in the sample.Paper [III] analyzes the consequences of unemployment for a Swedish sample of couples. The purpose is to estimate the possible income replacement that a spouse can provide. Unemployment can also affect the probability that the couples split up. Since not all couples remain in the analysis, a potential selection problem can occur. To deal with this problem, and also to take care of unobserved heterogeneity, a sample selection model for panel data is estimated. The results indicate that it is necessary to take into account the selection problem. A period in unemployment is found to be correlated with a higher female income only in the case of men who earned a fairly high income before becoming unemployed. Women who earned a fairly low income and were subject to a long period of unemployment are found to be compensated by a higher male income.
This dissertation MIGHT be available in PDF-format. Check this page to see if it is available for download.