It was inevitable that with the arrival of Hurricane Sandy, various economic pundits would speculate on its effects on “the economy.” Needless to say, some were saying that the hurricane would boost spending—both at the retail and then reconstruction level—and in that sense might actually provide a lift to GDP. The whole episode is yet another reminder that old fallacies die hard in economics. The commonsense notion that a natural disaster is bad is correct; only “sophisticated” analysts could think otherwise.

The basic problem here is what Henry Hazlitt called the “Broken Window Fallacy,” following the famous exposition of Frédéric Bastiat. The essential insight is that it is shortsighted to focus just on the employment given to workers who must rebuild after some act of destruction. In the present case, it is certainly true that glaziers, producers of telephone wire, and various construction crews will see more demand for their services following Hurricane Sandy. Their higher earnings in turn may lead them to spend more on restaurants, luxury items, and so forth, boosting employment in those sectors as well. This is the genesis of the notion that an act of destruction can actually have a silver lining.

But what this overlooks, explained Bastiat and then Hazlitt, is that the original people spending money on reconstruction could have spent their money on other things. Thus, instead of paying $1,000 (say) to replace his shattered store-front window, a shopkeeper might instead have spent $1,000 buying a new computer. Thus the extra employment given to the glazier et al. because of the hurricane, would merely be redirected from employment that otherwise would have gone to the software industry and all of the restaurants and so forth on which its employees would have spent their earnings.

Furthermore, Bastiat and Hazlitt pointed out that it’s not simply a wash: In the case of the hurricane, there is no new wealth created; people have to expend labor and other scarce resources just to get back to the status quo. In the absence of the storm, the extra expenditures would have led to the creation of new items of wealth, or of an additional flow of enjoyments.

Now a sophisticated Keynesian economist might respond to my above arguments along the following lines: “Yes, generally speaking a natural disaster confers no economic benefit whatsoever. However, in the case of idle resources and high unemployment—such as we currently face—the artificial jolt to demand really does raise GDP, rather than simply changing the composition of what is produced. The extra people put to work repairing storm damage don’t simply come from other lines of production, because there was a giant pool of unemployed people on the eve of the storm.”

Even on its own terms, this proposition is debatable. The people unemployed across the country on the eve of the hurricane’s landfall were not perfectly equipped to repair damage on the Eastern seaboard. In other words, much of the labor, glass, wood, steel, rubber, and other resources used in the wake of Sandy will indeed come at the expense of other employments. These resources will be used to merely replace what the storm destroyed, when even in our current recession they would have otherwise been used to produce new flows of services or tangible wealth. To the extent that this condition holds empirically, then the storm truly confers no economic benefit whatsoever.

However, let us concede for the sake of argument that there is, say, $10 billion in total reconstruction spending, and that this money sucks only unemployed workers back into production; no other output suffers. Just for a specific example, suppose that the $10 billion ends up going to 1 million previously unemployed workers, who each earn $10,000 over a few months repairing the damage. Official GDP goes up, because (by assumption) the output on other items follows the same path it otherwise would have, and now in addition we have the “finished goods and services” of the new window panes, shingles, telephone lines, etc. being produced over the next few months. Can we say in this case that the storm “helped the economy”?

We might say this, if we take “the economy” to be the same thing as “official GDP,” but in so doing we have totally severed the connection between conventional metrics of economic health and actual human welfare. For in this hypothetical case, the boost to GDP would go hand-in-hand with a demonstrable reduction in aggregate economic well-being. In particular, the people spending the $10 billion would be out $10 billion. By construction in this example, their consumption and accumulation of other durable goods is the same, but they are also spending an extra $10 billion just to repair storm damage. So their savings is necessarily lower by $10 billion, and they have nothing to show for it.

In contrast, the previously unemployed workers are up by $10 billion. Yet it’s not a simple transfer or redistribution—these people had to work for a few months to earn that money, so they didn’t actually gain a full $10 billion, as if they had just been handed the money for free. Thus, to the extent that we want to engage in aggregate measures of human well-being, the only sensible conclusion is that “society” or “America” or “the economy” is poorer on net. To repeat, this is because one group of Americans (those suffering storm damage) are down $10 billion and have nothing new to show for it, while another group of Americans (the 1 million previously unemployed who now get hired to fix the damage) are up, but not the full $10 billion, because of the value of their forfeited leisure.

The “macro” case of an economy with idle resources, suddenly being jolted out of its rut by a hurricane, is analogous to a “micro” case of a man who was laid off, agonizing over what to do with himself. Should he go back to school, apply to work at fast food restaurants, start his own lawn-cutting business…? Then, in the midst of his indecision, he realizes his house is on fire! The man suddenly knows exactly what he needs to do with himself—he has to run to the kitchen and grab the fire extinguisher. Yet would anybody dare argue that the fire, notwithstanding the property damage to the house, at least solved the man’s problem of idle labor?

In conclusion, under any reasonable sense of the term, a natural disaster such as Hurricane Sandy is not “good for the economy.” People who argue otherwise are simply demonstrating the problems with the conventional obsession with official GDP statistics.

Robert P. Murphy is author of The Politically Incorrect Guide to Capitalism. His blog is Free Advice. Follow him on Twitter.