A new study shows that government aid and World Bank projects are not enough to spur lasting recovery. New Orleans is growing, but is New Orleans back?

Not entirely.

That's the exchange we'll all be hearing in the coming weeks as the city marks the ninth anniversary of Hurricane Katrina. Interesting new businesses have sprung up. Many schools are better than they used to be. New Orleans has more bicycle paths. But the city can't claim the population it had in July 2005. And poverty rates have increased from their level in the first hopeful years after the storm.

This outcome disappoints, and it also challenges received wisdom. Americans nurse their own very private and personal storm of emoto-thoughts when it comes to natural disasters. We want to do something, so we look for theories that support action. One such theory is that restoring old structures or hurricane and flood spending can so stimulate economic activity at a disaster site that the place will emerge better than it would have been prior to the misfortune. Our officials routinely buttress this thesis.

In addition to such arguments for government or private spending, we see a second theory: disaster as a kind of natural selection of businesses. There's yet a third theory, which is related: that disaster spaces can benefit from a specific version of Joseph Schumpeter's "creative destruction." In this view, the same disaster environment that is hostile to humans is especially hospitable to innovators. Technological innovators flock in and drive out old backward technology, yielding productivity gains that also render New Orleans, or any other such unfortunate locale, superior to its old self.

This optimism speaks well of our national character, but not necessarily of our logic. For a systematic and global sweep of the evidence suggests that New Orleans is no exception: We often tend to overrate the quality of post-disaster intervention. Economists Solomon Hsiang and Amir S. Jina recently studied economies of nations that had endured the prototypical natural disaster, the cyclone. Studying 6,700 cyclones that took place around the world between 1950 and 2008, the pair published a National Bureau of Economic Research paper supplying strong evidence that national economies decline compared with their pre-disaster trend and "do not recover." Wrote the authors: "The data reject hypotheses that disasters stimulate growth or that short-run losses disappear." The conclusion: Cyclone-hit countries, rich or poor, experience such losses. Places where very big cyclones hit lose 3.7 years of development over the following two decades. This blow compares to a tax increase of 1 percent of gross domestic product, or a currency crisis.

"There is no creative destruction," Jina told me, repeating what he had said to writers at The Atlantic.

Three findings of the Hsiang-Jina report stand out. The first is that their study includes the contributions of international disaster aid. Since international aid money usually, if not always, buys something at the disaster site, this suggests that our faith in spending might be too strong. "We can't test whether aid and government spending is good or bad, but we can just say it is not doing what some people claim it is," says Jina.

The second finding is that repeat disasters further burden the economies they hit, resulting "in an accumulation of income losses over time." In other words, what is true once is also true the second time, with even greater impact. Perhaps those New Orleans families who have chosen to settle somewhere else have made the correct choice.

But the third point is perhaps the subtlest and most interesting. Economies do experience a jolt of growth when governments or private companies, not to mention international nonprofits and agencies, dump cash and rock concerts in the rush th at follows tragedy. That jolt may include food, bottled water, and blankets that save lives. But economically, a jolt is just a jolt. The growth is not sustained. The true economic picture, and a negative one, comes clear over the long term, the 10- or twenty-year period. The only reason we have not noticed this, as the authors point out, is that "the gradual nature of these losses renders them inconspicuous to the casual observer." Politicians think in election cycles, and so do voters, which explains why we may heretofore have found it expedient to ignore any evidence of long-term weakness that came before us.

The authors' first big takeaway relates to global warming, which, even by conservative estimates, is likely to provoke more cyclones. The authors estimate that to sustain traditional growth patterns, the world's nations will need to find, one way or another, $10 trillion more than they had planned to.

But the greater value of this study is in how it might help sort out what part of the disaster response results from fallacy. Now the evidence is there: Government spending alone simply does not do the trick. Big aid might not do the trick, either. And what about "creative destruction"?

Here one might reach a conclusion that differs from Jina's. The "creative destruction" of which he spoke, and of which the development community speaks, is innovation catalyzed by disaster. The error here might be in believing that disasters foster extensive extra creative destruction. Perhaps the broader force of creative destruction, for which there is plenty of evidence, is not prompted by disaster: E-mail killed the fax without the necessity of any hurricane.

The work of Hsiang and Jina suggests that a fruitful research topic would be whether wars, so similar to disaster, actually foster the level of innovation traditionally assumed. The cyclone study also raises the question of whether authorities ought to focus on delivering humanitarian support exclusively, and abandon aid. After all, many of us, most notably Simeon Djankov, have found evidence of an aid curse."

The perspicacious Joe Carter at Acton Institute writes that this study brought to mind the famous "broken window" thesis of the French philosopher Frédéric Bastiat, who sets forth this scenario: Suppose a window breaks at a shop. The bystanders invariably comment consolingly that at least some good will come of it all, for a glazier will get business, and, they ask, "What would become of the glaziers if panes of glass were never broken?" But this "seen" benefit obscures the unseen loss: that the cash spent to repair the window might have been invested more profitably by the shopkeeper, creating benefit elsewhere. Hsiang and Jina vindicate Bastiat. But they also undermine John Maynard Keynes.

In any case, "The Causal Effect of Environmental Catastrophe on Long-Run Economic Growth," this paper's formal title, provokes more thought than a dozen other scholarly tracts. Any one of its ideas is worth considering, and preferably before the next hurricane.

— Amity Shlaes is the author of four national bestsellers: Coolidge, The Forgotten Man, The Forgotten Man (graphic edition), and The Greedy Hand.