The first comment I would like to respond to is Professor Becker's. It flags an issue that I did not discuss adequately in my posting. He argues that the fact that an individual loses property or for that matter life in conjunction with similar losses of other individuals in an event reasonably classified as a disaster or catastrophe is no reason to stretch the social safety net beyond its usual dimensions; the individual should be treated identically to a person who sustained the identical loss in a noncatastrophic setting--say as a result of a pipe's bursting while the owner of the house was away for several weeks and when he returned the house was so flooded as to be completely ruined. I think Becker is basically right, and indeed I made a similar point in my posting but then wandered away from it. But I also think that there are some differences between the disaster and nondisaster settings. In my hypothetical bursting-pipe case, there would be no occasion to evacuate the owner to another city. As in this illustration, the disaster setting is likely to involve costs that would not be incurred in the nondisaster case, and the social safety net may not have been designed with that possibility in mind. Of course it would be better to alter the net to take account of the possibility, rather than, as we are doing in the aftermath of Katrina, responding ad hoc, excessively, and probably wastefully. My posting distinguished between insurance against loss and damages for a wrongful act; damages are more generous, including for example monetary compensation for pain and suffering. Several comments question the distinction. Let me offer two responses. First, if there was culpable negligence in the handling of the catastrophe by officials or public employees, it is conceivable--no stronger word is possible--that victims of the hurricane and of the ensuing flood may have a legal claim. I say "conceivable" rather than "possible" because there are a number of limitations on suits against public officials, especially suits based on their discretionary acts or their policies, as distinct from execution (usually by lower-level officers such as policemen). Second, the demand for insurance and the (social) demand for legal liability are quite different. The demand for insurance is based on risk aversion, which is based in turn on the declining marginal utility of money. Ask yourself whether you would pay $10 to avoid a one in ten thousand chance of having to pay $100,000. These are actuarial equivalents ($10 = $100,000 x .0001). So if you would prefer the riskless alternative (pay $10), you're risk averse and a potential customer for insurance. So if, as in the usual case, pain and suffering would not reduce your money income, there would be no point in your buying insurance. But since pain and suffering are a real loss, optimal deterrence of acts that cause such a loss require the courts to try to put a price tag on the loss and include it in damages. I was intrigued by the suggestion in several comments to make the purchase of flood insurance mandatory in areas where the risk of flooding is significant. It sounds like a good idea, provided those areas can be defined (the alternative of making flood insurance mandatory for everyone, even in deserts where the premium presumably would be close to zero, is objectionable as being administratively cumbersome) and that the insurance is not subsidized by the government. If it is not subsidized and therefore represents the insurance industry's best estimate of the expected financial losses from flooding, then requiring it will discourage construction in flood-prone areas (by making it more costly to live in such areas) and by doing so will reduce losses from floods and reduce demands for government bailouts--demands that appear to be leading to extravagant federal programs for compensating victims of Katrina. Compulsory insurance is a second-best solution from an economic standpoint; first best would be letting people choose whether to insure or "go bare" but not compensate them if they chose not to insure and then suffered a big loss. But it is unrealistic to expect government to take such a hard line in the disaster setting, where the number of victims is so large as to place the government under irresistible pressure to compensate--unless the victims are privately compensated through insurance.