It is often stressed that unemployment insurance may lead to less employment and output and, consequently, creates efficiency losses (see for example, Danziger et al, 1981, Björklund and Holmlund, 1991and Layard et al, 1991. In Vijlbrief (1992), we focus on the discouraging effect of benefits and taxation on labour supply and demand, although other decisions, like those to save and invest, may also be influenced by unemployment insurance, or more in general, by social security 1 . The first question we want to answer in this paper is why an institution that causes efficiency losses, social security, is so widespread in developed economies. The most frequently heard explanation for the existence of social security is the combination of risk, aversion to risk, and failures of private insurance markets to cover this risk. A second reason for the existence of social security may be that people like a certain amount of income redistribution, either because of utility interdependence or because they just prefer equality to inequality. Since income redistribution is a public good, state intervention is needed to supply the optimum amount. Two additional arguments for social security are the administrative efficiency of the system (social security covers multiple risks at the same time) and paternalism (left to themselves, people may have a tendency to underinsure). Section 2 surveys the reasons for the existence of social security more extensively. Hence, this section also indicates why the adverse effects of unemployment insurance on efficiency do not necessarily imply that the optimal unemployment benefit level is smaller than the current level.Although there are motives for state intervention, social security can only be a second-best solution, since it involves efficiency losses. A trade-off occurs between the most important advantages of higher unemployment benefits, more equality and more risk reduction, and the disadvantages, efficiency costs, for example measured by the loss in output (see Okun, 1975). Section 3 discusses some literature on this trade-off in social security, with special attention to the optimal unemployment insurance models by Baily (1978) and Flemming (1978). The balance between equity (equality and risk reduction) and efficiency in unemployment insurance is further elaborated in Section 4, by means of the 'output possibilities curve ' (see Breit, 1974). Firstly, this curve, that gives feasible combinations of output and equity, is confronted with indifference curves that reflect typical opinions on social security. Secondly, the framework is used to identify the determinants of optimum benefit levels, like preferences regarding equality and efficiency, the degree of See for a survey of this topic, Danziger et al, 1981. Unemployment insurance may also have an efficiency-increasing effect by improving job-matches (see Albrecht and Axell, 1984). 1 risk aversion and conjectures about the functioning of the labour market.The final part of section 4 is devoted to the construction of empirical out...
