Zero marginal utility goods
Dec 3, 2011
4 minute read

Standard consumer theory says that rational consumers will select their consumption of goods A and B such that \frac{MU_A}{P_A} = \frac{MU_B}{P_B}, with of course the tiny disclaimer that for some pairs of goods, there will be a corner solution; that is, for some goods A and B, some consumers will optimize not according to the preceding expression, but by consuming zero of either good A or good B.

Now, is that really a tiny disclaimer? Don’t we each get zero marginal utility from most goods? I made this point on Twitter yesterday and got this reply from Alex Tabarrok.

Now, it’s true that in a world of discrete goods, when price is high, that will result in a corner solution. But sometimes marginal utility is just really low, zero, or even negative due to storage and disposal costs. There are some things that even most billionaires do not buy.

I pointed Alex to a squirrel yard statue, which I found by searching “knick knacks” on Amazon.com (it was the fourth item). He yielded.

This resulted in a fun game of finding weird, zero-marginal-utility stuff online. Among other good items, Adam Ozimek found Vanna Speaks, Adam Gurri the Misty Mate Pet Misting System, and Jim Ulbright a life-size Anubis statue.

What can we learn from this exercise? I think there are a few things. First, while it’s fun to have a laugh at some of the weird products on the market, somebody is buying at least some of this stuff at least some of the time. I was memorably reminded of this because when Daniel Lin nominated gold lamé MC Hammer Parachute Pants as a ZMU good, Adam Ozimek vociferously disagreed. He ended up buying a pair for himself.

It’s tempting to think of the economy as supplying goods we all want to everyone (in uneven proportions of course), but it’s important to remember that the economy also supplies goods that most of us would not want to the few people who want them, including ourselves. This is not a trivial problem and it may not be solved smoothly through time. Fischer Black built his theory of the business cycle around the idea that it is difficult to match today’s production to tomorrow’s tastes, but it is also difficult to find the right buyers for today’s products.

Second, there is a direct analogy from the goods market to the labor market. Once you concede that most goods provide zero marginal utility to most consumers, you almost must concede that most workers provide zero marginal product to most firms. The math is the same. The ZMP hypothesis, therefore, is not some extraordinary claim that defies common sense.

Just as matching weird goods to weird people is hard, matching some workers to the right firm is hard. Nominal shocks can make this harder even in the absence of sticky wages and prices. People who have nominal debt change their consumption patterns when a nominal shock hits, and both people and firms can misinterpret nominal shocks in the short run per the Lucas Islands model. In the face of changing consumption and production patterns, Black’s Noise plays a role, and when the consumer-product matching problem is hard, the firm-worker matching problem is that much harder. Again, this is true even if wages and prices are totally flexible, and even if you support NGDP targeting (I do).

Oddly since these ideas owe much to Hayek, some Hayekians are slow to accept the ZMP hypothesis. Don Boudreaux at Cafe Hayek tells a comparative advantage story to rebut ZMP that is completely devoid of firms and therefore employment, in which all parties have perfect information and do not need to discover patterns of production. But ZMP is on the march; Karl Smith, a diehard New Keynesian, is starting to make some pro-ZMP noises.

So when you start to think that ZMP is a weird claim, just remember ZMU goods. I myself am not expecting to have much difficulty remembering ZMU goods. One is on its way to me. Tweeter @fbaseggio has bought for me the squirrel yard statue.

I’m looking forward to installing it in Alex’s office when he isn’t there.