Summary of “Wage Dispersion and Productive Efficiency: Evidence for Sweden”
Hey everyone, hope all is well! Today I’ll be summarizing this piece by Douglass A. Hibbs Jr. and Hakan Locking from 2000, analyzing how different union policies in Sweden effected productivity and growth. Swedish wage bargaining was centralized and produced very equal outcomes, but broke down somewhat in 1983. This allows us to examine how this breakdown effected certain economic measurements. My own notes, as usual, will be in brackets [like this].
From 1956–1983, the Swedish Trade Union Confederation (abbreviated LO in Swedish) covered and negotiated for basically all blue-collar workers, while the Swedish Federation of Salaried Employees in Industry and Services (PTK) represented white-collar workers. They negotiated with the Swedish Employers Association (SAF). We’ll be focusing on the LO, since their data is more extensive and complete.
Swedish unions had tremendous success at compressing wages, reducing variance by 75%. In 1983, a 30% increase in wage for a blue-collar worker was enough to bring them from the bottom to the top decile of wages. For comparison, at this time that number would be 200% for the UK and 400% for the US. [Today in Sweden, this number is about 130%]. These numbers were likely similar for the PTK as well.
Broadly speaking, the approach to union bargaining from economists at the time took one of 3 forms. The first, from trade union economists, focused on the growth and productivity improvements from decreasing inequality between different sectors [formalized here] and firms. The second, the fair wage hypothesis, states that decreased inequality within firms would increase cohesiveness, effort, and therefore productivity. The last, the neoclassical view, saw the imposition of wage bargaining as an inefficient economic distortion. All these theories will be examined here.
2. A Brief History of Postwar Swedish Wage Formation
Broadly speaking, there were 3 phases of wage bargaining.
Phase 1, from 1956–1969, focused on wage leveling between industries and plants. The goal was “equal pay for equal work”, regardless of a company’s profitability or ability to pay (thus the decreased inequality between firms). If this sounds familiar, it’s because this is the story right-libertarians tell us about how the free labor market works: that you get paid what you’re worth. It’s simple to see why this doesn’t happen in the real world. Compare a very productive and profitable firm, let’s call it firm A, to a less productive and profitable firm B. Firm A obviously has a greater ability to pay workers. So a worker will be paid more (a wage premium) in firm A than firm B, even if they’re equally productive either way. You might respond that workers would move to A, but various barriers prevent workers from freely doing so in real life. Under union bargaining in Sweden, this wage premium was nearly eliminated, while it remained large in places like the US.
Phase 2, from 1970–1983, LO focused on wage leveling within plants. The goal became closer to “equal pay for all work”, focusing on raising wages of the least paid especially. PTK followed suit, as part of a deep, ideological commitment to equality. This intensified conflict between the SAF and union confederations.
Phase 3, starting in 1983, saw the all-encompassing and highly centralized wage bargaining system break down. This began when employers in metal industries pried metalworkers away from the LO with promises of new wage agreements with higher wage differentials. These workers represented a third of the LO workforce. As expected, evidence of rents accruing to more productive and profitable firms increased.
Now we turn to an analysis of the data. Phase 1 policy increased productivity and output by up to 8% and 20% respectively, while phase 2 policy counteracted those productivity gains entirely, and halved the output increase. This is reflected by debates in the LO at the time about the tradeoff between equity and efficiency. This supports the trade union view, and “the regressions yield no support of the thesis… that wage leveling within workplaces enhances productive efficiency”, though the thesis may hold true for less extreme equalization of wages. Once centralized bargaining broke down, productivity and output did not change much overall, as higher inequality between firms and industries decreased output and productivity, while higher inequality within firms increased it.
Meanwhile, wage dispersion increased by 50% overall, including a 76% dispersion increase between different industries.