The union wage premium: Is it real?
Author: Alex Bryson
Date: December 5, 2016
Very few areas of labor economics have attracted as much attention as unions’ impact on wages. The literature goes way back and has a well-established pedigree. There is a general consensus that unions do raise wages — at least in English-speaking countries which are characterized by fragmented collective bargaining. This is the received wisdom, even though the standard method for estimating the “effect” is basic regression, with little or no attempt to tease out a clear causal impact. Some might argue that a credibility gap opens up between the received wisdom and state-of-the-art research methods: How do we really know that the association between union membership and wages is causal when we have no exogenous variation in unionization that could deliver such a result?
It is perfectly reasonable to assume that unions will have a causal impact in raising wages relative to workers’ outside market options. There is the direct impact of wage bargaining: unions target wages in their negotiations with employers, seeking to raise them as high as possible while being mindful of the potential negative consequences on their members’ employment if high wages mean employers cut jobs. You can actually see these effects in the US by comparing workers’ first contracts after a successful organizing drive with the non-union contracts they worked under prior to unionization.
There are also indirect effects that are plausibly causal, such as those arising because unionization reduces quit rates, increasing both employers’ and workers’ incentives to invest in the current job which, in turn, leads to higher wages. It also makes sense that the higher the percentage unionized (union density), the stronger the union’s bargaining hand, thus raising the wages the union can elicit from the employer. If unions behave as cartels, threatening to cut off the supply of labor to employers, this is what you would expect. So the theory is sound but perhaps the empirics leave something to be desired.
The challenge, then, is for economists to come up with more credible efforts to identify the true causal impact of unions on wages. They should also be going a stage further to establish what the implications for any such wage hike are for workers’ and society’s overall well-being. After all, unless unions can also raise productivity to compensate for wage rises, those wage premia will come at the cost of job loss and potential firm closure. This is where the literature is right now. One example is our work with Norwegian data where we exploit exogenous variance in the price of union membership (state subsidies through the tax system) to identify the causal impact of unions on productivity and wages. We find that firms’ productivity rises with union density, and so too do wages, as workers share in the improved performance of their firms. These effects are causal. What we need now is more research in this vein to see whether these results are valid generally.
We recognize that IZA World of Labor articles may prompt discussion and possibly controversy. Opinion pieces, such as the one above, capture ideas and debates concisely, and anchor them with real-world examples. Opinions stated here do not necessarily reflect those of the IZA.