We are staring down a hurricane and people are evacuating. Those that are not evacuating are stockpiling and hunkering down. And those who are outside of the era are watching and commenting. Inevitably, the following exchange occurs:
[Photo of bottled water with exorbitant price tag]
Normal Human: Can you believe that people are taking advantage of people in need like this?
One of my fellow economists: [pushes glasses up nose] well, actually prices should go up to prevent shortages so that water and fuel can get to where they are needed most.
So the thing here is that the economist is wrong, but I don’t think they will believe you if you tell them they are wrong because the only counter-argument that ever gets made is “you are a heartless egghead sociopath”. The economist knows that their approach maximizes overall human welfare and so relegates you to the bin of moral intuitionists incapable of stepping out of emotion long enough to make a clear-eyed executive decision to help people.
But the economist is wrong!
And they’re wrong for a classic reason: they are using a comfortable and trusted tool for a purpose for which is neither intended nor well-suited.
There was a great question that we were given on my comprehensive field exam in microeconomics in grad school. In it, we had to find a consumer’s optimal consumption bundle given their utility and a budget constraint. Now, we had done dozens of these problems before, and in almost every case, the solution is very straightforward and uses a small bit of calculus: you take the utility function and the budget constraint and combine them using what’s called the Lagrangian method, which essentially limits you to considering a subset of consumption bundles that satisfy the budget constraint.
For a visual metaphor, imagine that there are two trees on a hill, planted one to the east and one to the west. You want to find the highest point on a hill that is still south of the line made by connecting the trees. So what the math here does is reduce these two goals into one goal: find a flat point (at least in the tree-line-wise direction) directly between the trees—because at a flat point, you’re as high as you can go without starting to go downhill again.
Now you also have to check a set of second-order conditions to make sure you’re at the top of a hill, and not the bottom. It sounds more complicated than it is in practice, but it takes some time to get the hang of it, and it takes some time to do it right. And so we had drilled until we could do them quickly.
There are a few cases when you don’t want to use any of that stuff, though. When one tree itself is on top of a hill for example, or when the whole landscape is flat. And the question we got on our exam was one of those special cases: the tool we have been given is not well-suited for this task, so can we reason from the original principles? Because the whole point of the Lagrangian is to assist us in finding what makes people better off—the mathematics is just instrumental. So consider a new problem: what makes people better off in this case?
And that brings us back to overpriced water during hurricanes. Why do economists think it makes people better off? Because we economists learn in a very stylized application a very stylized truth: prices send information about who needs what. Which is kind of true? Except that “who” and “needs” and “what” are kind of poorly defined. Now economists know, deep in their brain that it’s more complicated than this. The “paradox of value”—the contradiction between the importance of water vs. diamonds relative to the price of water vs. diamonds—demonstrates that the price does not actually communicate need.
So what does the price communicate? At present, most economists will argue that it represents the marginal utility—or marginal value—of a good. This is to say, it captures the well-being provided by the final traded unit of the good. And while we spray water on our lawns, we don’t do the same with diamonds, so the marginal diamond must be more valuable than its weight in water. When people trade things like diamonds and water they make themselves better off—some of my extra water might get me a scarce shiny diamond, while someone else gets the water they want. We are both made better off—and so society is made better off! We can try to measure this by looking at the prices at which we trade. This is convenient, because if we know how much people care about something—about everything—by looking at its price, we can find the distribution of goods and services that makes society as a whole most well-off: the social-welfare-function maximizing distribution of goods.
I don’t think I have lost any economists yet, but here I’m going to: what price actually communicates is the marginal willingness-to-pay for a unit of the good. Which is fundamentally—and morally—different from the marginal utility. It is also fundamentally and morally different from average utility or average willingness-to-pay: important and useful concepts in their own right.
The thing here is that we cannot measure utility directly. Sometimes we economists talk about “utils” as the thing we are trying to maximize. But as we take this theory of individual behavior and choice and try to turn it into an applicable theory of social human behavior, there are certain modeling conveniences we make. One is that we use money as a “numeraire”. Since utility has no natural units, we use money as a natural unit, and utility measured in this way is called money-metric utility. The problem is that we know that utility of money is not linear, because we know that people are generally risk-averse—and certainly not all risk-neutral, which is what linear utility in money would require.
Now it turns out that the nonlinearity of utility of money is fine, not a big deal, until you start trying to compare peoples’ utilities. We economists spend a lot of time handwaving about how interpersonal comparisons are not possible as they require cardinal measures of utility rather than ordinal measures, and then? And then we barrel right along and do it anyhow. If you have ever heard of a cost-benefit analysis being presented as measuring whether society will be better or worse off, then what you have seen is an application of a money-metric utilitarian social welfare function (MMUSWF). The math is very straightforward but the morality is whoa not great. This is because a MMUSWF says society is better off in situation A than situation B if the amount of money everyone combined would be willing to pay for situation A is higher than the amount of money everyone combined would be willing to pay for situation B.
Now imagine I have $1000, and I live in a society with 99 other people, each of whom have $1. Imagine that in situation A, everyone gives me a dollar, but because I have a magic touch, when they give me the money, it is doubled. In situation B, everyone keeps their money—the status quo ante is maintained. So in A, I get $1198. How much might I be willing to pay to be in situation A? Under standard neoclassical assumptions of myopic self-interest? $197.99. How much would everyone else, combined be willing to pay to stay in Situation B? The maximum they CAN pay is $99.
That is the crucial moral dilemma of MMUSWF: every dollar counts as a vote, and so people with more dollars get more votes. Your maximum willingness-to-pay is bounded by your maximum ability-to-pay, and so a bottle of water to someone who is penniless is treated as literally worthless.
Is this okay? Well, of course it’s not okay; what the hell?
For traditional consumer products, in times of peace and harmony, when people have the Maslow’s hierarchy met, or we have sufficient redistribution such that gains do not accrue to the few, it’s a decent approximation. And the math really is a lot simpler. But if we’re using cost-benefit analyses to make policy decisions, then we’re systematically siding with those with more money.
Prices are gorgeous things—they really do coordinate flows of information very well and very cheaply. It is not a surprise that so many of my fellow economists remain infatuated with them as mechanisms for coordinating distribution. But a good tool is not fit for every job. And too often economists who just have not thought hard about other distribution mechanisms fail to notice that there are other tools that exist for distribution.
Rationing exists for a reason. It is an institution for exchange that has certain properties, most of which are incredibly unappealing in peacetime. I drink more coffee than the average person and don’t care much for candy. It would make us all worse off, much of the time, for us all to have to consume the same products. But rationing ensures that distribution is equal when it might not otherwise be possible. There are tradeoffs between these institutions—and as we trust individual consumers to make these tradeoffs internally, so sometimes we have to trust decision-makers to make these tradeoffs on a social scale.
But there are other methods. We can keep prices and means-test things with subsidies—find a way to reliably indicate who has lost a lot and issue an emergency charity card—sales to them will be covered by a FEMA fund. We could institute a pay-what-you-can system in emergency situations. We can legalize looting of grocery stores: take what you can carry. There are better ideas out there just waiting to be discovered, but we need to think about the actual problem of human wellbeing, and not the simplified version of welfare-function-maximization.
Why don’t we do these things? Well, people are allergic to people getting something “they don’t deserve”, which makes no sense whatsoever to any right-thinking economist. But the economists out there defending price increases during emergencies because they haven’t given it enough thought are not helping the situation.
