The AI Journal

In the week after Hurricane Fran tore through North Carolina in 1996, a bag of ice that had sold for a dollar on Tuesday was going for ten by Thursday, and the ten-dollar bag drew more public anger than the storm itself. Investigators opened price-gouging cases within days; most American states now have anti-gouging statutes on the books, triggered automatically by a declared emergency and carrying fines or jail time for anyone who raises a price on water, fuel or generators by more than some fixed percentage once the governor has signed the order. The laws assume something specific: that the ten-dollar bag of ice is a moral failure, evidence of a seller exploiting people who have no choice but to pay. Economists have spent four decades arguing the opposite, and on the narrow question they were asked, they are largely right. They have mostly stopped there, which is the problem.

The economists’ case is informational before it is moral. A price is not only a number a buyer pays; it is the only signal that travels faster than the disaster itself. When ice at ten dollars a bag is more profitable to truck into a flooded county than to sell at home, trucks move, and they move within hours rather than the days a relief convoy takes to organise. Michael Munger’s account of Fran makes the mechanism concrete: the entrepreneurs who drove rented refrigerated trucks loaded with ice from South Carolina were arrested on arrival, not because anyone was short of ice afterward but because they had charged eight times the pre-storm price to people queuing in the heat. The ice that didn’t spoil in those trucks is the counterfactual the moral objection never prices in. Ban the signal and you do not get the dollar bag of ice distributed by fairness; you get no bag, because nobody outside the flood zone has a reason to find one.

Matt Zwolinski pushed this further than the logistics point. He asked why the moral intuition against gouging survives a comparison its own defenders rarely make. If a shopkeeper’s price was high before the storm — because her supply costs are high, because she is a monopolist in a remote town — nobody calls her corrupt for refusing to lower it once a hurricane arrives. The shopkeeper who was charging two dollars and raises it to ten is treated as having done something the shopkeeper who was already charging ten was never asked to answer for. Zwolinski’s point is not that gouging is therefore fine; it is that the condemnation is aimed at the wrong feature of the transaction. What offends is the spike, not the level, and a spike is exactly what a market clearing a sudden shortage is supposed to produce.

The usual reply is that the law isn’t banning information, it’s banning extraction — and if the price is suppressed, something has to do the job prices do, which is where the empirical literature on rationing by waiting becomes relevant rather than academic. Robert Deacon and Jon Sonstelie found their natural experiment in the US gasoline price controls of 1979–80, where some stations sold at the capped price only to drivers willing to queue, while others sold at an uncontrolled price with no queue at all. Comparing the two let them price the line itself: the hours spent waiting were a real cost, borne disproportionately by whoever could least afford to be elsewhere, and in their later estimate the deadweight loss from rationing by waiting — the queuing time plus the wasted larger-than-needed purchases people made to avoid returning to the line — ran to roughly one and a sixth times the money transferred from sellers to buyers by the price cap. That is not a rounding error against the transfer; it is the price control destroying more value than it moved from one pocket to another. A queue is not a fairer allocator than a price. It is a different allocator, priced in a currency — time — that falls hardest on whoever cannot pull a relative off a shift to stand in it, and the accounting rarely gets done because the cost never shows up on a receipt.

So far this is the standard defence, and it is sound within its own terms: price spikes allocate scarce goods faster and with less total waste than the alternatives anti-gouging laws actually produce. But the defence answers an allocative question with an allocative answer, and the strongest version of the objection to gouging was never allocative. Jeremy Snyder’s reply to Zwolinski argued that the wrong in price gouging is a failure of respect — that pricing people out of water or shelter during the one window when they cannot shop around treats their need as leverage rather than as a claim on anyone’s consideration, independent of whether the total ice supply ends up larger. Kaitlin Caruso’s 2023 survey of anti-gouging law across fifty-six jurisdictions gives that intuition a number: poorer communities face measurably higher exposure to the kind of price spikes gouging statutes target, precisely because they are more likely to live in the housing stock and infrastructure that fails first, so a law that flattens the spike is doing real distributive work that the efficiency calculation leaves out entirely. This is the objection in its strongest form, and it does not dissolve under the Fran story or the Deacon–Sonstelie numbers: a bag of ice efficiently allocated to whoever can pay ten dollars for it is not the same achievement as a bag of ice reaching whoever needs it most, and the first can come at the second’s expense in exactly the households Caruso is describing.

The honest answer is to stop asking the price mechanism to do two jobs at once. Letting prices spike is the right way to get ice, fuel and generators to a flooded county quickly; it is the wrong tool for deciding which households in that county can afford to buy them, because it was never built to ask that question and a price cap does not ask it either — it just hides the failure to answer inside a queue. The alternative that actually meets Snyder’s and Caruso’s objection is not suppressing the signal but decoupling it from the distribution problem: disaster vouchers, means-tested cash transfers, or FEMA-style direct provision aimed at the households a price spike would otherwise exclude, funded and triggered by the same emergency declaration that currently triggers the price cap. That keeps the ten-dollar bag of ice moving toward the flood zone on the first day, while putting the compensation question where it belongs, with the people who can actually answer it, rather than with a law that solves distribution by making sure nobody gets the ice as efficiently as before.