The primary purpose of this paper is to investigate, both theoretically and empirically, the relationship between the socially optimal degree of central bank conservatism and the level of unionization. The theoretical framework suggests a clear and negative relationship between the two, such that the lower the level of unionization, the higher the socially optimal degree of central bank conservatism. This result is derived from the combination of two effects that operate when the level of unionization changes. These effects can be identified as the "competition effect" and the "strategic effect". An increase in the proportion of unionized labor force reduces competition in the labor market and, as a result, induces the central bank to choose higher inflation in order to erode real wages ("competition effect"). Simultaneously, an increase in unionization leads to real wage moderation by the union as it desires to avoid inflation ("strategic effect"). Because this moderation will be greater the greater the weight that the central bank ascribes to unemployment stabilization, a more liberal central bank is socially preferred. In case the union's inflation-aversion is sufficiently high, the combination of these two effects produces a hump-shaped relationship between the level of union density and both unemployment and inflation (a Calmfors-Driffill relationship). The paper also presents an empirical examination based on panel data from over 48 countries. The analysis provides some support for the theoretical model's prediction about the negative relationship between collective bargaining and central bank conservatism as measured by legal independence.

Key words: Central bank conservatism, collective bargaining, inflation, labor market, trade unions, unemployment.

JEL Classification: E50, E58, J50, J51.

The full article in PDF file