Minimum Wage Wizardry

Ezra Klein over at TAPPED, the American Prospect blog, takes William Niskanen to task for his opposition to the minimum wage below. "[W]hile reasonable people can disagree on the impact of minimum wage laws," Klein writes, "it's time they stopped."

Wow! Why? What's the debate stopper?! Klein says, "Just crosscheck this list of state minimum wage laws with this rundown of state unemployment rates." Turns out that there is an inverse eyeball correlation between high minimum wage and high unemployment rates. QED? Well, no. This factoid might help Klein's case if it wasn't totally meaningless in isolation from auxiliary assumptions.

That pattern is perfectly consistent with Niskanen's claim, which is, after all, just an application of the bedrock Economics 101 principle that if the price of something goes up, consumers will tend to buy less of it. In fact, Klein's pattern might be evidence in favor of Niskanen's claim. Here's some more Economics 101 to explain why high minimum wages and low unemployment rates might be expected to go together.

A high unemployment rate indicates a significant oversupply of labor relative to available jobs. In that case, you expect the price of labor to be low, since it is so abundant. If there is already a minimum wage---a lowest legal price---high unemployment will tend to drive wages toward that floor. Let's say it's $4 an hour. Now, if there is already high unemployment, and you raise it to $5 an hour, lots of people will have to get a raise, since lots of workers are probably being paid something close to the lowest legal wage. Employers will not be able to afford to give all those people raises. So unemployment would increase further. Now, the effect is quite different in places that have low unemployment rates. In a tight labor market, wages will be higher. So fewer people's wages will be near the price floor. And so if you raise the floor, fewer workers will be affected. If the labor market is tight enough, and almost no one is getting a wage even close to the floor, raising the floor a little may have no detectable effect at all---like a law mandating breathing.

Now, suppose legislators more or less understand this (or that key constituencies pressure them to act like they do). You'd then expect that states with high unemployment rates and low wages to be least likely to raise their minimum wage, since it would have a relatively large adverse effect for them. And you'd expect states with low unemployment and high wages to be most likely to raise their minimum wage, since it is least likely to make a difference for the worse. And so you end up with high unemployment states with low minimum wages, and low unemployment states with high minimum wages.

Now, I have no idea whether this reasoning in fact explains the pattern Klein observes. But then, neither does he. He's just a victim of confirmation bias, seeing what he wants to see in an inkblot of ambiguous data. But the pattern he points to might be evidence in favor of the idea that minimum wages increase unemployment. Hardly a debate stopper, is it? Perhaps Klein will grant reasonable people the privilege to continue disagreeing.

It's worth nothing that Klein admits "And yes, if you jack the wage up to $16 an hour, jobs will be lost. But up to $7 over a period of years?" So what weird science reveals the "no effect" point between $7 and $16? $16 an hour? Unemployment for sure. But not at $7! So what about $8? How about $12? $15.75?

Of course, a bump up to $7 will push fewer people out of the legal labor market than a bump up to $16. But why Klein thinks that a bump up to $7 will push zero people out, when he has already conceded the general point, is mysterious.