The September 1 Wall Street Journal contained an article by David Wessel (page A1in the print edition) that documented average GDP growth over the half-year following quarters in which a major hurricane hit. This picture sums up the main lesson:
Perhaps unsurprisingly, there is likewise scant evidence of much lasting impact -- if any -- on employment growth following major storms past. The following set of pictures depict monthly job growth for the top ten most costly U.S. hurricanes (in terms of material damage) since 1954. The vertical lines identify the month of the event, and the costs are expressed in 2004 dollars.
It may turn out, of course, that the scale of Katrina is so far off the map that these historical precedents are not much precedent at all. Although I'm not entirely sure how anyone can possibly know at this point, at least one analysis has an early estimate in the neighborhood of $100 billion, which would make it twice as much as Andrew, the current record holder (which it itself was well over twice as expensive as any other single storm in the period under examination). On the other hand, there is the hypothesis -- promoted in today's Wall Street Journal (page A1) by Jon Hilsenrath and Greg Ip -- that the U.S. economy is better suited to withstand major shocks than in the past.
I guess, unfortunately, we will know soon enough. The best we have now is plenty of informed speculation, from the likes of Brad Setser, Econbrowser, Angry Bear, The Big Picture, The Capital Spectator, The Eclectic Econoclast, Environmental Economics, Peter Gordon's Blog, the NABE blog, and many more.