Efficiency wages, monopsony and labor demand

Again Krugman writes about something I know something about. This time I enthusiastically agree (yeah I know big surprise there).

he noted that, in efficiency wage models, it can be relatively easy to convince firms to raise wages. His presentation is (as usual) extremely, brief, clear, unwonkish and excellent.

Then he linked to Arindrajit Dube who notes that the same is true if there is monopsony (like monopoly but with one big agent on the demand side — in this case a firm big enough to affect the prevailing wage in its local labor market).

As usual, I stress that your time is much better spent by clicking the links and reading Krugman (and in this case Dube) than by reading on. In any case, even if you insist on reading this post, for some reason, you shourd read Krugman first.

I note that the two stories actually become the same if the local area market is very local (that is if each firm including even each McDonald’s franchise has its own “local market” of people who look for a job at that particular firm”. Quits which decrease gradually as a function of the wage can be seen as just another form of labor supply which decreases gradually. So the two models are the same. Now in the real world, the cost of a quit is mostly not the brief period when the job is vacant (pretty consistently about 1 month on average in the USA) but the cost of training a new worker. But a model of turnover costs can be made to work even without training costs so long as it takes a while to fill a vacancy. This point is barely worth making.

But with much more enthusiasm, I call on George Borjas in support of Krugman and Dube (and yes I did enjoy typing that) .

The reason is that there are two litartures with results which appear contradictory if one assumes no monopsony and no efficiency wages.
It is alleged that increases in the minimum wage have a small to tiny effect on employment. This corresponds to a very steep almost vertical demand curve for labor (similar L and very different wages — by convention labor demand is on the x axis). On the other hand, it is also alleged that immigration has a small effect on wages. This fits an almost flat demand curve for labor with very different labor demand at similar wages.

Borjas says the literatures are contradictory (because he insists on assuming perfect competition with no efficiency wage effect). But both results are consistent with some models of efficiency wages.

1) turnover depends on the wage and the local unemployment rate (and not on local employment). A higher wage causes lower turnover which can imply only a slight reduction in employment (or even an increase in employment). Immigrants arriving imply the same unemployment rate for higher employment. Employment can increase (if the no turnover labor demand curve is close to flat) provided unemployment stays high enough to keep people from quiting. The turnover model reconciles the two apparently contradictory results.

This also works in a work/shirk that is moral hazard efficiency wage model in which high wages are needed so that workers are sufficiently afraid of being fired that they actually work even if the boss isn’t always watching. The wage required to make workers actually word depends on the local unemployment rate (and not on employment) an increase in labor supply (due say to immigration) means higher employment is consistent with the same wage and sufficient fear to make workers work. Again if the no shirking problem demand for labor demand curve is almowst flat, a helicopter drop of workers (with parachutes so they can work after being dropped) has a small effect on wages.

In both cases, an increase in the minimum wage can have a small effect on employment.

With monopsony, labor demand can increase one for one with labor supply (if the firm has constant returns to scale and no market power in the goods market — that is if the no monopsony labor demand curve is flat). The wage depends on the elasticity of labor supply, not the level of labor supply so a doubling of the number of workers with each reservation wage causes employment to double.

In such models a minimum wage can actually cause higher employment.

The alleged results that the arrival immigrants has a small effect on wages and that increases in the minimum wage have a small effect on employment together suggest that efficiency wages and/or monopsony are important.