Monday, December 28, 2009

Are wages really that rigid?

The New Keynesian theory has come under increasing assault because its basics assumptions, price and wage rigidity, look more and more tenuous in the data. Now that we have much better data sets, for example, we realize that prices of everyday goods fluctuate a lot. This may be a recent phenomenon, as these data sets come from the databases that retailers use nowadays for scanners, and all this information technology has considerably reduced the cost of changing prices. So much for the now obsolete menu costs.

What about wages, then? Wages are typically set for at least a year, and collective bargaining sets wage schedules for about three years on average. But people change jobs, get promoted, get bonuses, so it is difficult to measure well wage rigidity. Still, there is some consensus that there is at least some nominal downward rigidity, as it is believed that there is some psychological barrier to reducing nominal wages. I am not that convinced, as the current recession has shown plenty of examples of nominal wage decreases. But I need more than anecdotal evidence.

There now also a study of wage rigidity across European firms, where one would suspect the most evidence of this given the labor market institutions. Martina Lawless, Jan Babecký, Philip Du Caju, Theodora Cosma, Julián Messina and Tairi Rõõm show that there is rigidity, and that it can be related very closely to the various theories that have been put forward to justify these rigidities: efficiency wage theory, insider-outsider theory, and collective bargaining. What I take from this: wage rigidity is for real, at least in Europe, but you cannot just assume it in a theory, you need to properly model its origin. Indeed, as this paper shows, rigidity is flexible as economic conditions change.

1 comment:

Agent Continuum said...

What does the Logician make of the claim that sales periods "don't count"? If you drop sales you're basically smoothing the data and you get more price rigidity than you'd have otherwise.