Distributive justice and the welfare state

Abstract: This thesis is an attempt to generate knowledge about the justification, measurement and consequences of public redistribution. The first paper poses the following question: How much redistribution from working to non-working can be justified by arguing that external resources are owned jointly by all, or that ownership should ideally be divided equally? This question is analyzed using bargaining theory in a setting with two types who differ in production skill and consumption-leisure preferences. The bargaining concerns the proportional tax rate in a constitution where all resources are allocated to the high-productivity type and taxation redistributes to the type who does not work. It is shown that the tax constitution Pareto dominates the initial distribution only if preferences of the two types are sufficiently similar. Using the Nash bargaining solution and the Kalai-Smorodinsky solution, it is shown that under joint ownership, the justified tax rate is independent of the productivity difference between types, but under divided ownership, the tax rate is decreasing in the productivity difference. In the second paper, it is examined how inter and intra-individual redistributions depend on characteristics of the welfare state such as the degree of tax progressivity, the choice between flat rate and income-related benefits and the proportion of the lifecycle spent working. A model of a welfare state is used to calculate the difference between gross income inequality and net income inequality among workers as well as inequality of lifetime income. A simulation based on the Swedish system indicates that the redistributive effect of the Swedish system is higher in terms of lifetime incomes than in terms of disposable incomes among workers. The share of inter-individual redistribution in Sweden is approximated to 30 percent of total redistribution. When individuals are allowed to choose labor supply freely, the redistributive effect of the welfare state can not be correctly described by comparing inequality of gross and net income. The third paper deals with welfare state typologies, i.e. the categorization of different types of welfare states. Literature using the term universal welfare state is analyzed and it is shown that the term is ambiguous. Different ways to alleviate the problem are discussed, including the distinction between group universality and income universality. This paper also contains an empirical application illustrating how universality can be studied over time using the case of Sweden during the 1990s. Following a number of indicators during the 1990s reveals that universality was roughly constant for the Swedish welfare state during this period. The final paper studies social insurance against short-term income losses, in a setting where low-income earners are more likely to suffer from income loss. With a standard adverse selection model, non-universal social insurance is likely to be better for low income earners. If private insurance firms can observe risks using a costly technology, non-universal social insurance is likely to be better for high income earners. The paper also demonstrates that under this alternative insurance technology, risk aversion does not unambiguously increase the political sustainability of social insurance.