The minimum wage paper that needs to be written

I have been following the debates on the minimum wage in view of the forthcoming increase of the federal minimum wage in the United States. And there is something that is bugging me to no end: the explanation provided by labor economist for why there’d be a limited employment effect.

Notice that I say employment effect and that I thus limit myself to the number of bodies hired. I am also not going to consider the distribution of elasticities found since the seminal work of Card and Krueger that tends to point to negative effects on employment and even clearer effects on hours of work. I am also not going to account for other adjustment channels like those that Brian Albrecht discusses here. I am especially going to discount issues of minimum wage enforcement.

I am just going to focus on the channel for why I’d be wrong to give weight to the evidence that suggest no employment/positive employment effects (and I could very well be wrong — my wife says it happens all the time). The argument is that labor markets aren‘t competitive. Take this quote from Noah Smith that illustrates the argument best:

Suppose some employers are big and powerful, or they somehow cooperate to keep wages down, or simply that it’s really hard to switch employers (emphasis mine). Then the theory changes a lot. In the extreme case — a “company town” with only one employer — the market outcome is whatever The Company wants it to be (technically, where The Company’s marginal cost equals its marginal revenue). It has nothing to do with supply and demand! And of course, The Company is going to hold wages artificially low, which kills jobs because some people just can’t afford to work for wages that low.

In that case, minimum wage can actually create jobs (emphasis in original). It forces The Company to raise wages, which allows more people to work. It looks like this.

Notice the first bit that is emphasized above. This speaks essentially to costs for employees to switch between employers. It is entirely relevant to mention those but notice that there is a sleight of hand being performed here. What are these costs a function of? That question is just starring us in the face!

There are costs of changing employers that are “natural” (e.g. function of technological setting, mobility, information etc.). But a large share of these costs (I believe) are “institutional”. For example, zoning laws raise rents and make cities less accessible. Cities are more productive and increases in productivity in those cities are harder to access for people in less productive areas. As way of another example, hiring costs are a form of costs to employees as new employers are less interested in seeking them!

All these are a function of X or Y policy elsewhere that governments apply. And this is why there is a paper that needs to be written that estimates labor demand elasticities with and without accounting the effects of already existing policies.

Imagine the following paper: you take low-wage industry employment in all the MSAs of the United States and check the effect of changes of the minimum wage in those MSAs. You get a first batch of elasticities. Then, you introduce the economic freedom index of MSAs produced by Dean Stansel out of Southern Methodist University to that series of estimate. The economic freedom measure (if stripped from its minimum wage component) would proxy institutionally-imposed cost on employees leaving their jobs. You then go a step further and segment the sample between high and low economic freedom (which is something I have done for a paper on Olympics here).

By breaking the sample in two parts, here is I am willing to bet 100$ on: the adverse effect of the minimum wage on employment is greater in high economic freedom states. There, the lower-level of institutionally-imposed costs means that the market was more competitive and thus the minimum wage kills jobs. However, in the low economic freedom states, you may find no effect or a small positive effect. Why? Because these costs to switching jobs for employees are higher there! (Again, I am not even considering other adjustment mechanisms, I am just thinking about bodies with jobs). In those states, the minimum wage may actually generate the outcome described by some.

But notice something here: if I am right, there is nothing to be cheerful about! Indeed, this would suggest that the problems that the minimum wage is supposed to solve have roots in other government policies. If anyone takes up that study, I am willing to pay that person to do it (not kidding, say 500$) on top of the 100$ that I am betting on the results.