Because (Behavioral) Science!

A few days ago in Slate, Barry Schwartz posted an article on how the psychology of anchoring means that raising the minimum wage from X to X’ won’t just benefit folks whose current wage is less than X’, but will probably end up cascading through the lower tier of the labor market. That is, if somebody is making $10 and we raise the minimum wage to $9, they might still get a raise if they and their employer think of their wage less as $10 than as 133% of the minimum wage. That is, the piece is a mix of gee whiz social psychology and political mood affiliation flattery that together we can call “because (behavioral) science.”

Since Jim Manzi doesn’t post much around here no more, let me take this one.

Let’s put aside all the usual minimum wage debate shtick about how to value transfers, what year to use as the baseline, appropriate deflators, whether it’s all teenagers and ex-cons, etc., and focus on Schwartz’s main point about anchoring. Let’s furthermore stipulate that we can take the upshot of Schwartz’s argument as “behavioral science proves that raising the minimum wage is even better than you thought.”

Okey dokey, but are there any ways that behavioral science proves that a minimum hike is worse than we thought?

Of course there is, and not only that, but the micro-mechanisms are closely related to anchoring. There’s a general problem for why companies pay more than they really have to and the assumption is that they have to be getting higher productivity out of it which is why this issue is called the efficiency wage. One of the models for an efficiency wage proposed by Akerlof 1982 is that the firms pay more than they have to as a gift to the workers. Gifts imply a debt of reciprocity (this is why charities send you return address labels before asking you for money) and the workers reciprocate through greater work effort. Basically, high wages are good for morale and morale is good for productivity. Or rather higher wages are good for morale because it only works if the employer could be paying less and the employees know that with a less munificent employer they would be making less. That is, it only works if the employees see themselves as having a boss who is more generous than those poor souls working for the skinflint across the street. If the boss is required by law to pay X then you don’t feel grateful for getting X. Likewise if the law changes from requiring X to requiring X’ and you had been making 1.5X but the anchoring effect is only partial so your new wage is only 1.4X’ you might actually feel less grateful to your boss since you’re now making closer to the minimum. The upshot is that a minimum wage hike would probably undermine the efficiency wage and so decrease productivity. There you have it, a minimum wage hike is bad because (behavioral) science says so. QED.

Now, to be candid with you, while I like Akerlof and I’m really really really really into gift exchange, I think gift exchange is an implausible micro-mechanism for efficiency wage. Partly this is an intuitive hunch and partly it’s because Gneezy and List 2006 have shown that the gift exchange efficiency wage effect demonstrated in experiments only lasts as long as the typical experiment and if you follow-up even a few hours later people aren’t grateful (or hard workers) anymore. (Hey, look, a hedonic treadmill, that’s a social psych concept too). (For broader critiques of generalizability of economic games with college students see Levitt and List 2007 and Henrich et al 2010). However, skeptical as I am of my own argument about gift exchange and efficiency wage, is it really any less plausible than this anchoring business? Moreover, there are a lot of models for the efficiency wage and many of them are a lot more plausible than gift exchange. The thing is though that pretty much all of them involve micro-mechanisms that would be undermined by wage compression from a price floor.

The really broad point is that when you’re dealing with really high causal density and subtle mechanisms, it’s pretty easy to pick out things that make your side’s policy pitches look good, and not only good but good in the fashion of distinctly #aspenideas / #slatepitches kind of counterintuitive erudition. However it is ultimately pretty selective and less about what the science shows than how we can draw on the literature to make ourselves feel more sophisticated than those troglodytes so ignorant as to disagree with us.