Wednesday, February 27, 2013

MInimum Wage and the Law of Many Margins

Last November, I pointed out that President Obama had campaigned in 2008 on a pledge to raise the minimum wage, but that this proposal had vanished during the rest of his first term. Now, after the election, Obama somewhat unexpectedly resurrected the proposal in his State of the Union address. For a review of the controversy over the economics of the minimum wage, a useful starting point is
"Why Does the Minimum Wage Have No Discernible Effect on Employment?" written by John Schmitt for the Center for Economic and Policy Research.

While Schmitt's title suggests, albeit in the form of a question, that it is an agreed-upon truth that the minimum wage has "no discernible effect on employment," I would say that his own review of the evidence suggests that there is still a genuine controversy between those who see the employment effects of the minimum wage as nil and those who see it as small. As Schmitt writes in the conclusion: "[W]hat is striking about the preceding review of possible channels of adjustment – including employment – is how often the weight of the empirical evidence is either inconclusive (statistically insignificant or positive in some cases and negative in others) or suggestive of only small economic effects."

There is a difficult problem of inferring causality here. Compared to the overall costs of firms, or even compared to the costs of low-wage labor, the effects of a slightly higher minimum wage are going to be hard to distinguish from everything else that's happening in the economy. The employment prospects for low-skilled workers have been falling for decades, and it would clearly be incorrect to blame that on the minimum wage. Rises in the minimum wage are more likely to occur when the economy is doing well and adding jobs, but it would clearly be incorrect to infer from this correlation that a higher minimum wage causes an increase in jobs. In addition, there are difficult questions of what is sometimes called "publication bias" in the minimum wage literature, in which researchers of different political bents may--surprise, surprise--tend to publish the results that confirm their pre-existing beliefs.

Rather than try to unpick this empirical puzzle here--for those who are interested, Schmitt provides a nice overview of the key paper and their methods--I'd like to focus on a separate issue, which I call the Law of Many Margins. The "law" simply points out that when a rule is imposed, like a minimum wage, there are almost always a wide variety of possible reactions to that law. Schmitt provides a list of 11 possible reactions (!) to a higher minimum wage. They are:

  1. Reduction in hours worked (because firms faced with a higher minimum wage trim back on the hours they want)
  2. Reduction in non-wage benefits (to offset the higher costs of the minimum wage)
  3. Reduction in money spent on training (again, to offset the higher costs of the minimum wage)
  4. Change in composition of the workforce (that is, hiring additional workers with middle or higher skill levels, and fewer of those minimum wage workers with lower skill levels)
  5. Higher prices (passing the cost of the higher minimum wage on to consumers)
  6. Improvements in efficient use of labor (in a model where employers are not always at the peak level of efficiency, a higher cost of labor might give them a push to be more efficient)
  7. "Efficiency wage" responses from workers (when workers are paid more, they have a greater incentive to keep their jobs, and thus may work harder and shirk less)
  8. Wage compression (minimum wage workers get more, but those above them on the wage scale may not get as much as they otherwise would)
  9. Reduction in profits (higher costs of minimum wage workers reduces profits)
  10. Increase in demand (a higher minimum wage boosts buying power in overall economy)
  11. Reduced turnover (a higher minimum wage makes a stronger bond between employer and workers, and gives employers more reason to train and hold on to workers)

The evidence on many of these  points is ambiguous at best, and indeed may vary across industries or geographic areas or employers. But it's worth noting that which of these effects arise, and with what magnitude, can only be settled by empirical evidence, not theoretical assertions.

I confess that I find it hard to get too excited about modest increases in the federal minimum wage every few years, which has been happening for decades. As Schmitt points out, the evidence is that this pattern of minimum wage increases has had at most a small effect on employment and other outcomes. But the minimum wage was $5.15/hour in 2007, when President Bush signed legislation to raise it to $7.25/hour by 2009. Given an unemployment rate that has been stuck near or above 8% for four solid years now, my preference would be to de-emphasize rises in the minimum wage for awhile longer--and instead focus on other methods to help the working poor.