The extent to which workers are threatened by unemployment differs considerably e.g. across branches of industry, age education, and regions. Nonetheless, it is customary to levy obligatory contributions to unemployment insurance UI) regardless of a worker´s specific risk of becoming unemployed. Branches of industry or regions characterized by a relatively high rate of unemployment are favoured by this practice because workers and / or employers pay less than the actuarial fair insurance premium. This leads to a distorsion of the decisions regarding where or what kind of labour is being supplied. The purpose of this paper is to discuss the effects of placing the UI under the opligation to equilibrate its budget in each region respectively, as opposed to on the whole. In the former case, the conditions of UI reflect the risk of unemployment within each region. The results may nevertheless be applied to other characteristics systematically influencing the probability to enter employment. Labor markets in both regions do not clear because the wage rate, bargained between trade unions and employers (right-to-manage-approach), is binding. Unions maximize the expected utility of a representative worker. This expected utility is affected, among other things, by the UI contributions employed workers have to pay, and by the UI benefits unemployed workers receive. In this paper it is assumed that the UI adjusts contributions to equilibrate its budget. In this case, contributions themselves will depend on the rate of unemployment, and thus on the bargained wages. The extent to which the unions take the interplay between wages and contributions into account, crucially depends on the size of a union. For atomistic unions (decentralized bargaining), the effect of the bargained wage the unemployment rate is negligible. Because every union neglects it, the total size of this externality however is considerable. The larger a union is with respect to the total workforce, the more of the effect is internal. It is interesting to analyze the effect in the context of regionalized UI because the way UI is organized affects, how elastic UI contributions react on variations of unemployment. Another important aspect is the role of migration. Protagonists of the regionalization of UI often point to the allocative advantages it has. The line of reasoning is that workers migrate to regions which are less concerned by unemployment and which are thus characterized by lower UI contributions. The regionalization thereby serves as a substitute for higher differences of wages, which do not perfectly reflect the relative regional scarcity of labor. This point of view is rather superficial. If there is unemployment in both regions, there is no direct allocative gain in reallocating workers from one region to the other. On the other hand, migration affects the UI`s parameters. If wages react on these, the efficiency may be affected, too. It is not quite clear, however, if it is improved or worsened by the regionalization of UI. Also, workers from the poor region may be better off, because the reinforced emigration ameliorates the probability of becoming employed, and vice versa. Since the effects of regionalization of UI on the economic agents, as well as on efficiency, are subject to the complex interplay of wages, UI contributions, and migration, a formal model is an adequate mean to identify the effects. In this paper, an analytical framework is established to analyze the effects of a regionalization of UI. Since many effects are ambigous in general, the model is calibrated with reasonable parameter values. Our results confirm the importance of the degree of centralization of unions for the assessment of the regionalization of UI by workers and employers, and for the efficiency of the measure.
