Wirtschaft und Gesellschaft 1 9. Jahrgang (1 993), Heft 4 over 10 per cent in 1980-82, average unemployment in the five EFTA countries remairred below 3 .5 per cent throughout the 1970s and 1980s. Only in the recession that began in 1991 has unemployment in many of the EFTA countries reached levels similar to the rest of Western Europe. According to conventional wisdom, the combination of rising oil pri� ces, increased competition from the rapidly growing economies in the Pacific, and the slowdown in productivity growth all combined to redu� ce the scope for real wage increases. In countries with rigid nominal wages (a category that was often claimed to include the US and Canada), an increase in inflation was required before real wages could fall and employment increase again. In countries with rigid real wages (a catego� ry that was perceived to include many members of the EC), nothing would help other than a weakening of the unions. Yet, it is unsatisfactory from an intellectual point of view and unhelp� ful from a policy point of view to simply label the problern as rigid wages. The wage level should be an endogenous variable in economic analysis. The failure of wages to fall sufficiently to maintain full em� ployment requires explanation, a task that focuses attention on the me� chanisms of wage-setting. Moreover, the existence of European countries with generally strong and often highly centralized unions that main� tained virtually full employment during most of the period suggested that institutional features of collective bargairring might have important macroeconomic consequences. The question of the comparative performance of different systems of wage formation proved to be highly controversial, in part because the debate is highly political. On one side, social democratic governments and trade unions encouraged the explicit coordination of wage setting across different industries and firms in the belief that economic growth is best achieved through cooperation and bargairring among highly cen� tralized organizations of unions and employers. Social democrats argued that the benefits of wage moderation are public goods to an important extent. The wages received by the members of any singly union, it is clai� med, have only a small effect on the aggregate wage level. Individual unions therefore rationally ignore the effects of their wage demands on macroeconomic performance. If the unions can be induced to negotiate jointly, however, the wage agreement affects wages throughout the eco� nomy with visible macroeconomic consequences. Thus, the centraliza� tion of wage setting may prevent individual unions from aggressively seeking to improve their own members' wages at the expense of workers who belong to other unions (or who belong to no union at all). On the other side, conservative governments and employers' associa� tions responded that labor markets require competition and wage flexi� bility, rather than coordination, if they are to function smoothly. Centra� lized wage setting, it is claimed, reduces the sensitivity of wages to con� ditions in the local labor market. In addition, employers have charged that centralized bargairring inhibits microeconomic adjustment by redu� cing their ability to use wage differentials to encourage workers to ob- 424