In a recent Bloomberg View column, Megan McArdle gives us “the talk” about efficiency wages. Sure it’s pretty funny, but it is also pretty wrong.
Megan correctly notes that there are a variety of specific efficiency wage models and lists some of their specific mechanisms. I want to focus on her second mechanism, but my point is more general. All the macro models of efficiency wages “work” in equilibrium while the whole point of her article rests on the claim that they do not work in equilibrium:
Here’s her reason #2:
“Workers know that if they lose this job, they are likely to end up with a job that pays less. They are thus highly motivated to keep this job.”
When she applies this mechanism broadly she concludes that”
“Efficiency wages only work because the workers are getting more than they could make elsewhere. If everyone was paying the same wages, all the benefits to the employer would disappear.”
But this is simply not true!
Consider the famous Shapiro – Stiglitz model (American Economic Review 1984), which is the best known model that works using Megan’s mechanism #2 (it’s been cited over 4700 times according to Google Scholar).
The very title of the piece gives the game away: “Equilibrium unemployment as a worker discipline device”
If one firm raises wages its workers will shirk less because another job will pay less. If all firms raise wages all workers will shirk less (other things held constant) because the resulting unemployment from the economy-wide higher wage for the same set of workers, means that a job separation will create a period of unemployment for the shirking worker.
So in equilibrium (when all firms pay the higher wage) fear of unemployment substitutes for fear of a lower wage in motivating workers.
Again, I’m just talking about her mechanism #2, but rest assured that efficiency wages “work” in equilibrium in macro models for a wider variety of mechanisms.