Ok, I want to get in on this. Ezra’s right that it’s good to reevaluate these things every once in a while.

First, two quick quibbles. One, the wage-employment connection may be an article of faith on the right, but it also has a lot of currency with moderate and liberal economists. Two, there may be very good points to make about the SF experience, but anyone should be very reluctant to conclude anything from that graph (Post Hoc Ergo Propter Hoc fallacy, anyone?). We could say that one axis is connected to the other, but we have absolutely no way to tell just by looking at that chart. Now what we’d like to have is a good sample of many cities that have tried this, the better to do some regression stuff on. But we don’t.*

Ok, let’s start with what we know. If the price for something goes up, people will want to supply more of it and demand less of it, other things equal. This means that an isolated change in the wage rate should lead to an increase in those seeking work and a decrease in hiring, which generates a higher level of unemployment. I don’t think anyone, including those who think that increasing the minimum wage is a good idea, would dispute this relationship.

But there are a number of situations where an increase in the wage might not have that particular effect. Off the top of my head, here are a few:

1) The change in the wage is just too small to matter. Economic theory says that if two gas stations are right next to each other, and one charges a penny more per gallon than the other, the cheaper station will get all the business and the more expensive station will get none. In fact we don’t observe this, in part because there are other factors (like the presence of a convenient store, for instance), and in part because a penny is too small to represent real value to consumers at that level. Two cents, or four cents, or even ten cents might not steer business away from the expensive station in any real way, because the amounts are just too small to concern buyers.

This might be the case in the wage situation. Employers might just have so few minimum wage employees, and 2 bucks more an hour might be so little, that the increase just doesn’t affect their labor cost calculations. At some level, this is certainly true. Think about it in a different way: two clerks almost certainly differ in their productivity levels at least a little (one might be a little nicer to customers, for example, generating fractionally more sales than the other), but the firm doesn’t bother with paying them differently; it’s not worth the effort, the difference is too small to matter.

Larger changes in the wage rate certainly would matter, however. In that sense, the government should be wary of relying too much on minimum wage increases as an anti-poverty tool.

2) Labor supply is, for all intents and purposes, unlimited. Let’s say that when making hiring decisions, firms face a labor market that is effectively infinite. That is, there is so much excess unskilled labor that firms can feel free to push down wages to the statutory limit without experiencing an applicant shortfall. This would mean that workers’ price elasticity of supply is near perfectly inelastic–no matter what the prevailing wage rate is, workers choose to supply their labor. This might not be a terribly unrealistic description of labor markets for unskilled workers in some places.

In that case, firms could pay a wage rate below market and could therefore capture for themselves part of the surplus generated by work that ought to go to workers. An increase in the minimum wage might then push labor markets back to the proper equilibrium, restoring that surplus to workers without reducing employment. This framework also carries the proviso that excessive increases in the minimum wage would increase unemployment.

3) There are shifts in labor allocation. This one is fairly easy to imagine. If you increase the minimum wage, then employers that wouldn’t normally find it in their interest to compete for more skilled (and more expensive) labor suddenly would. In that case, employers of the cheapest labor might increase their employment, because the pool of workers from which they can recruit suddenly has a higher average productivity. So, say clerks make $6 and hour and accountants make $9 an hour, and the minimum wage is suddenly increased to $9 an hour. At that level, stores may as well recruit accountants to work as clerks, and because those workers are potentially more productive at clerking, they might hire more of them, even though they’re more expensive. This could have an overall negative effect on employment, but it certainly would have the effect of creating a less than optimal allocation of labor.

Another way this might work could be germane for the SF/Alameida example. Higher wages in SF could shift workers from Alameida to SF. This would deepen the labor pool in SF, potentially improving the skill level of hired workers and thus increasing SF employment. In doing so, however, it might increase unemployment in both areas. It’s important to remember that higher wages encourage some people to enter the labor force that otherwise wouldn’t.

4) The wage increase is expansionary. This strikes me as the least possible of the four. By increasing the wages of workers (who have a greater propensity to spend marginal increases in income than owners or managers) there is a small boom in the local economy, which causes an increase in hiring.

 

In any case, it seems to me that small increases in the minimum wage would probably have very minor effects on employment. In that sense, no real harm, and probably some to considerable good could come of it. This should not be taken to mean, however, that larger or more regular increases are wise.

 

* Second Life, guys. Economists always complain that we’re not allowed to conduct actual policy experiments, so we have to wait for natural phenomena to duplicate what we’d like to do. But with these MMPRPG or whatever they’re called, don’t we have lots of nice little fake worlds to play with? If I were a rich and slightly mad economist, I’d contract some software companies to create lots of fun little Second Life type games. Then I’d change economic variables around to see what happened. Then I’d sit back and wait for my Nobel.