The ghost of Congressman Chisolm (who died in January) in an interesting comment states that the rhetoric of cost-benefit analysis "dangerously backgrounds challenging ethical ramifications," by which she means the disproportionate impact of Hurricane Katrina on the poor people (mainly black) of New Orleans. Many poor do not own automobiles and so found it difficult or impossible to comply with the initial evacuation order; later, of course, evacuation became impossible. What this shows, however, is not that cost-benefit analysis ignores the fact that people and groups are different in their access to various resources, but that an evacuation plan that assumed that everyone has a car would flunk a cost-benefit test, because the benefits of providing public transportation for those people would exceed the costs. The City of New Orleans could have had a fleet of buses cruising the city street before the hurricane struck and offering free transportation from the city to people who didn't have a car.
Katrina does underscore a point that should be obvious but tends to be neglected: poverty does not mean just having fewer physical goods than the nonpoor. It means living a riskier life, because among the resources that the well to do acquire are resources for protecting one's life and health.
It is not correct, as another comment states, that Becker and I think that New Orleans should be abandoned to teach the City a "nasty lesson." The question is simply whether rebuilding is worth the cost. The same question would arise if Beverly Hills was destroyed by a North Korean nuclear missile. Or consider that the World Trade Center is not being rebuilt--is that to teach New York a "nasty lesson"?
Another comment suggests that the port of New Orleans should be preserved. That may well be true. The question should be what parts of New Orleans should be preserved (say, the historic districts, for their tourist value, and the port) and how much of a residential hinterland would be required for the persons employed in the preserved areas. Some, undoubtedly, but it should be located on high ground. I certainly agree, by the way, that in comparing the costs of relocating the former New Orleans population elsewhere and rebuilding the city so that they can return, the attachment to place that many people have should be factored in as a cost of the relocation option. I don't want to truncate cost-benefit analysis, but only insist that it has an important role to play in guiding so immense an allocative decision as whether or to what extent to rebuild New Orleans.
I agree emphatically with the comment that people's thinking about the abandonment option is fogged by familiar cognitive errors--failing to treat sunk costs as bygones and inability to understand the most elementary statistical theory. So many people think that since New Orleans cost a great deal to become the city it was before the catastrophe, we should pay whatever it takes to restore it; and that now that the hurricane has come and gone, we are not due for another one of comparable magnitude for a long, long time--in fact the probability of another such hurricane tomorrow is the same as the probability of Hurricane Katrina before it was formed. A value of cost-benefit analysis is in combating these cognitive errors.
A subtler mistake is to assume that the issue of abandonment can arise only after disaster strikes. In principle, if the expected cost of disaster is great enough, a city should be abandoned before the disaster occurs. Indeed, the case for abandonment could be stronger, because the benefit would include averting the first disaster. Cutting the other way, however, is the fact (related to the sunk-costs point) that once disaster has struck, the cost of recovery becomes another cost of not abandoning, along with the expected cost of a future disaster that would nullify rebuilding efforts.
A few brief responses. It was suggested that we should not compensate for loss of property, but only for humanitarian assistance. In some situations that is reasonable, but in otheres it would be misleading since property assistance and humaritarian assistance are sometimes hard to distinguish. For example, owners of small business may have all their wealth invested in property, or an elderly couple may own their home and nothing else. Property losses to these and other groups would be catastrophic too.
Perhaps the government could catalogue the risks in different locations. It would be a mighty undertaking however, since, for example, New Orleans on its past record did not seem very risky. But something along these lines is required if compulsory insurance is to be required in selected areas.
The Good Samaritan problem does not arise only when the compensation paid more or less fully offsets the loss. In past disasters, the federal government often paid victims a substantial, although still incomplete, part of their losses. Such victims would generally have their decisions affected even though they are not fully compensated for losses.
The poor of all colors and ethnicities do move across regions less than others. But over the years actually a very significant number of both poor white and African-American did move out of the South.
The strong emotional bond within many families cuts both ways. Parents may be more moved to help children in distress than governments are to help citizens, but at same time children may be more likely to refrain from bad behavior because they care about the effects on their parents.
I certainly do not oppose helping those enormously harmed by Katrina and other disasters. However, that does not mean that the Good Samaritan problem is unimportant, or that policies cannot make it less serious. Even if the choices of many persons are not altered by prospects of help if disasters strike, as long as significant numbers do look ahead, it is important to implement policies that push their decisions into more efficient directions. By implementing sensible public policies regarding disaster assistance, more resources would be available to help people truly in need.
Although the death toll from Hurricane Katrina is not yet known, the loss in life and property makes this one of the worst natural disasters in American history. Coming less than a year after the Great Tsunami that killed about 300,000 people in Asia, this experience has many wondering whether the world is facing much more erratic and violent weather, possibly due to global warming.
I do not know the answer to this question, but important public policy issues are raised by major disasters, even if they will not increase in severity over time. One major question is whether individuals, companies and local governments have the right incentives when they determine where to locate plants, oil drilling rigs, refineries, factories, homes, roads, levees, and bridges.
Location decisions would be optimal if those making these decisions had to bear the full social cost of any damages to their property and person from a disaster. Under these conditions, greater insurance premiums in areas that are prone to hurricanes, earthquakes, tsunamis, and other disasters would reflect the greater risk to life and property in these areas. The expected loss for those not insuring would rise in proportion to the greater risk. People, companies, and governments would then build homes, roads, businesses, and the like in disaster-prone regions only if the benefits exceeded the full risk of damages.
However, generous private and public help to victims of terrible disasters, while highly desirable, distort such rational calculations. Congress just voted over $10 billion of relief help to victims of Katrina, the Red Cross and other private groups have already had pledges of over $200 million of private help, other nations have offered generous assistance, and the United States has the Federal Emergency Management Agency that provides substantial assistance to people and businesses in areas that are declared to be disasters. Presidents are making greater and greater use of this Act to declare regions in need of emergency assistance.
Such public and private assistance in the event of disasters make it more likely for persons, companies, and public activities to locate in high-risk areas because they will often be spared much of the losses. They also may not take out insurance against risks that would inflict large losses; for example, rather few New Orleans homeowners had flood insurance. Studies have shown a small propensity to insure against low probability natural disasters that cause great damage- see "Paying the Price: The status and Role of Insurance Against Natural Disasters In the U.S.", Ed. by Kunreuter and Roth). So private and public generosity to victims of disasters help distort many pre-disaster decisions.
This distortion goes under the name of the "Good Samaritan" paradox in philosophy and economics. To illustrate this problem, consider the behavior of loving parents toward their children. Such parents would come to the assistance of their children if they get into financial trouble, have serious medical problems, or experience other difficulties. At the same time, they want their children to use their money wisely, work and study hard, prepare for future contingencies, and lead healthy life, so that they can avoid personal disasters.
Unfortunately for the parents, children can distinguish reality from lectures, and threats that will not be backed up by parental behavior. If they anticipate that their parents will help them out if they get into trouble, and if they are not so altruist to their parents, they would consume and possibly gamble excessively, and they might quit good jobs to "find themselves". Parents might then be indirectly encouraging the very behavior by their children that they want them to avoid.
The federal government and private philanthropy are in a similar Good Samaritan situation with respect to families, businesses, and local governments that build where there is likely to be flooding, landslides, hurricanes, earthquakes, and other major natural disasters. The federal government and others may wish they did not build so much in these areas, and the government may hope for diversification elsewhere. Yet if the government's advise is ignored, and if there is terrible suffering from a disaster, all humane and politically sensitive governments and philanthropic organizations would help, even though they wish the victims had made more socially efficient decisions before disaster struck.
A particularly important location decision facing the U.S. is whether it should encourage a greater decentralization of its oil, refinery, and natural gas facilities. Katrina shut down about 1.5 million barrels of oil production, 16 per cent of American natural gas production, and about 10 per cent of U.S. refining capacity at a time when there is little slack in worldwide refining capacity, and the region’s electric and natural gas distribution system has also been knocked out. "Altogether, about 800 manned platforms, plus several thousand smaller unmanned platforms, feed their water and gas into 33,000 miles of underwater pipelines…" (Daniel Yergin, The Wall Street Journal, 9/2/05).
Much of the cost of this shutdown will be borne by the country as a whole, partly already reflected in the decision to release some of the 700 million barrels of oil in the Strategic Oil Reserve- the U.S. government has no gasoline reserves, although some other countries do have such reserves. The Gulf energy complex seems especially vulnerable not only to a future natural disaster, but also to well-planned terrorist attacks.
The ability to decentralize natural gas and to some extent also oil production is limited by the federal moratorium on natural gas and oil production in the Outer Continental Shelf. The recent energy bill includes requires a requirement to prepare an inventory of these reserves, and some effort is being made in Congress to relax the moratorium. The damage from Katrina indicates that such steps are more urgent than they appeared even a few weeks earlier.
Can anything generally be done to weaken before disasters strike the inefficient incentives caused by the natural and laudable tendency of governments to help the victims of terrible disasters? To start, it would help to have tougher zoning restrictions in disaster-prone areas to reduce construction and raise their capacity to withstand major shocks. These restrictions would apply, for example, to areas subject to bad flooding, perhaps because buildings and roads would be below sea level, as in New Orleans, and to buildings on large faults that are particularly subject to earthquake damage. After the great destruction caused by the 1989 California earthquake, that state toughened its building code to make it more likely that buildings could survive major earthquakes. Perhaps they did not toughen them enough since California can expect a large amount of federal assistance once again in the event of a future serious earthquake. Asian nations have restricted rebuilding in some areas devastated by the Great Tsunami.
In addition, anyone who does build in designated high-risk areas should be required to carry insurance that covers most of their losses, the way many states mandate liability car insurance. This requirement would place the bulk of the cost of damages on the individuals and businesses that locate in risky areas, and the companies that insure them. Provisions have to be made for persons considered too poor to have adequate insurance, and insurance companies would need to have sufficient resources in the event a major disaster strikes. Many states already require insurance companies to have minimum levels of liquid capital.
Given the need to help victims of disasters, there is no perfect way to induce individuals, businesses, and local governments to incorporate more fully into their decisions the risks of living and building in disaster-prone areas. But more can be done, and Katrina proved that despite 9/11, the U.S. is still terribly ill prepared to handle a major disaster. It is scary to contemplate how well the country would respond to even greater disasters, such as one induced by a future and even more deadly terrorist attack.
The Speaker of the House of Representatives, Dennis Hastert, got into trouble, and had to apologize, for suggesting that maybe New Orleans should be abandoned rather than rebuilt. He raised a valid issue; that he got into trouble for doing so just proves the adage that, in politics, the phrase "to tell the truth" is synonymous with "to blunder."
Not that it can yet be said that New Orleans should be abandoned; that conclusion could emerge only from a complex analysis. The point is rather that the analysis should be undertaken. The broader issue is the role of cost-benefit analysis in the analysis of disaster risk. In addition, the disaster to New Orleans is a timely reminder of the risk of terrorism.
Because New Orleans is both below sea level and adjacent to a sea (Lake Pontchartrain is connected via another lake and a strait to the Gulf of Mexico), the city is extraordinarily vulnerable to just the sort of flooding that occurred when the levees broke as a result of Hurricane Katrina. To decide whether to rebuild or abandon the City, the cost of reconstruction, plus the expected cost of a future such disaster, should be compared to the cost of either building a new city or, what would be cheaper and faster, simply relocating the present inhabitants to existing cities, towns, etc., a solution that would require merely the construction of some additional commercial and residential facilities, plus some additional infrastructure. Of course New Orleans has great historic and sentimental value, and this should be factored into the analysis, but it should not be given conclusive weight. Perhaps it should be given little weight, since the historic portions of the city (the French Quarter and the Garden District) might be rebuilt and preserved as a tourist site, much like Colonial Williamsburg, without having to be part of a city.
The decision to abandon or not cannot be left to the market. It could be if federal, state, and local government could credibly commit not to provide any financial assistance to the city’s residents, businesses, and other institutions in the event of another disaster--but government could not make such a commitment. Or if government could require the residents, businesses, etc. to buy insurance that would cover the complete costs of such a disaster. But again it could not; insurance in such an amount, to cover so uncertain a set of contingencies, could not be bought in the private market.
So the decision would have to be made by government, and, ideally, it would be based on cost-benefit analysis. In such an analysis, the expected cost (that is, the cost discounted by the probability that it will actually be incurred) of a future disastrous flood would probably weigh very heavily and could easily tip the balance in favor of abandonment. The reason is only partly that constructing levees and making other improvements that would provide greater protection against the danger of flooding would be very costly (such a program was proposed in 1998 that would have cost $14 billion, according to Mark Frischetti, "They Saw It Coming," New York Times, Sept. 2, 2005, p. A23); it is also that the levees, seagates, etc. would remain highly vulnerable to terrorism. Breaches similar to those that caused the recent flood, but created without warning by terrorist bombs, would cause much greater loss of life because there would be no time to evacuate the population, whereas with the warning of the approaching hurricane 80 percent of the New Orleans population left the city before the flood. The expected cost of a terrorist attack on rebuilt levees cannot actually be calculated because the probability of such an attack cannot be estimated. But it should probably be reckoned nontrivial given the wide publicity that the vulnerability of the city to flooding has received and the fact that a port city is more vulnerable to terrorism than an inland one because terrorists approaching from the sea are less likely to be detected before they attack, since they would be spending little time on U.S. territory. (Analytical techniques for adapting cost-benefit analysis to situations in which risks cannot be estimated with any precision are discussed in Chapter 3 of my book Catastrophe: Risk and Response (2004).) Of course, massive amounts of money could be devoted to protecting the vulnerable rebuilt city from a terrorist attack, but that would be just another substantial cost that abandonment would avert.
New Orleans is becoming more vulnerable not only because of the terrorist threat, but for three other reasons as well. The city is sinking because (paradoxically) flood control has prevented the Mississippi River from depositing sediment to renew the subsiding silt that the city is built on. The wetlands and barrier islands that provide some protection against the effects of hurricanes are disappearing. And global warming is expected to increase sea levels and also to increase the severity and frequency of storms--all factors that will make New Orleans more vulnerable to future floods.
It might seem that if, as current estimates have it, the cost of the damage inflicted by Hurricane Katrina will prove to be "only" $100 billion, the expected cost could not have been too great, since the probability of such a flood as occurred presumably was low. But while the annual probability was low, the cumulative probability over a relatively short period, such as one or two decades, was probably quite high. Moreover, $100 billion is almost certainly a gross underestimate, because it ignores the loss of life (economists are currently using a figure of $7 million to estimate the value of life of an average American), the tremendous physical and emotional suffering of the hundreds of thousands of refugees from the flood, and the lost output of the businesses and individuals displaced by the flood. These are real social costs, as an economist reckons cost. What is not a social cost is certain purely pecuniary losses that will be made up for elsewhere in the economy; for example, the loss of convention business by New Orleans will be a gain to other cities.
Another hidden cost of rebuilding rather than abandoning the city is the uncertainty concerning how much time it will take to rebuild and what the former residents will do in the meantime. If they expect to return to the city in several months, they will find it difficult to obtain remunerative employment in the meantime.
For simplicity, I have assumed that the choice is between rebuilding New Orleans and abandoning it. Realistically, given politics and the typical (and on the whole commendable) American reflex refusal to accept defeat, the choice is the scale of the rebuilding. I urge that careful consideration be given to rebuilding on a considerably reduced scale from what the city was before the flood.
Speaking, as I did earlier, of terrorism, an article in the Washington Post this morning (Susan B. Glasser and Josh White, "Storm Exposed Disarray: What Went Wrong," p. A1), provides support for those who claim that the slow response to the New Orleans flooding shows that the nation has not made adequate preparations for responding to a terrorist attack by means of weapons of mass destruction--a significant (though again unquantifiable) and growing danger. An attack with nuclear, radioactive, or biological weaponry could easily require the evacuation of an entire city without warning and with much greater loss of life.
It seems that, four years after the 9/11 attacks, we are still not taking the threat of terrorism seriously. There are four basic counterterrorist tools: (1) Threat assessment, which means conducting cost-benefit analyses designed to identify the targets that are most vulnerable to terrorist attack, having in mind the goals of the terrorists (so far as we can determine them), the value of the target, and the cost of hardening (defending) it. (2) Hardening at least the most vulnerable targets. (3) Warning intelligence (which of course failed us on 9/11), designed to detect impending attacks. (4) Emergency response measures if an attack occurs, designed to minimize human and property damage.
(1) has made little progress, in part because of political obstacles; all elected officials except the President and Vice President have geographically circumscribed constituencies and naturally resist efforts to devote proportionately more resources to defensive measures that would benefit only outsiders. (2) has made very little progress, because of cost. (3) has improved, though not as much as it should have. (See my book Preventing Surprise Attacks: Intelligence Reform in the Wake of 9/11 (2005), and my just-published monograph Remaking Domestic Intelligence (Aug. 2005).) And judging from the New Orleans disaster, (4) remains completely inadequate. One possible response would have been for the President to declare martial law and place a general who had combat experience (i.e., someone who knows how to coordinate a large number of people in circumstances of urgency and uncertainty) in command of all federal, state, local, public, private, military and civilian response agencies and personnel. The article in the Washington Post this morning that I mentioned notes the bureaucratic logjams that delayed the response; martial law would have overcome them. My idea about how to respond to such a disaster may be excessively dramatic and quite unsound; I am no expert. But ever since 9/11 it has been known that there could well be a terrorist attack utilizing weapons of mass destruction and that, if so, the correct emergency response might involve the evacuation of a city. It is disheartening to think that after four years there are still no plans, preparations, or command systems for dealing with such an eventuality.
A number of valuable points raised by the comments.
To Li Huafang, yes you can use my blog entry on Chinese ownership. On neuroeconomics, it may be promising, but so far it has delivered almost nothing of value to an economist. But it is still a young field, and may do better in the future.
I like the distinction between narrow power and general power. Narrow power might be more easily be corrupted. On the other hand, it may be easier to expose a corrupt use of narrow than of general power.
I believe politicians are like most of the rest of us, and are mainly looking out for their own interests, broadly defined. That is why it is necessary to have a competitive political process and competitive media to keep politicians in check. This issue goes back to Edmund Burke, and perhaps earlier. By the way, I do not believe the Rockefellers, Corzines, Bloombergs, Perots, etc are more praiseworthy political figures than many others who are much poorer. Rich politicians also want to be elected, and cater to various political interests. But I agree very rich politicians are less likely to take bribes.
I do believe the evidence is that better paid officials are less corrupt. Rank both corruption and the pay of officials (relative to the average in their countries). I believe the correlation between these ranks would be strongly negative, although I have not seen such an approach carefully done.
I like the comment that bad laws that induce widespread corruption may have a very negative effect on basic values. The former Soviet Union may illustrate this. But it is still true that these countries would have been devastated if all these laws were followed exactly. To take a different example, it is unfortunate that many American cities have numerous ridiculous laws concerning materials, etc to be used for construction. Yet without corruption that enables builders to get around these laws, construction would come to a virtual halt in many large cities like Chicago. Unfortunately, they may also corrupt officials to ignore valuable building laws!
One does not have to know who to bribe in order to engage indirectly in bribery. Some people pay lawyers to take care of their traffic tickets. They do not bribe the lawyers, but the lawyers might bribe officials.
Honest persons may be at a disadvantage compared to corrupt persons if the latter bribe officials to overlook bad (or good) laws. So industries where corruption has great value will come to be dominated by individuals willing to corrupt officials. This happened during prohibition where the liquor industry became dominated by gangsters- it is happening now with respect to illegal drugs.
There were many good comments, raising a number of issues that require clarification on my part.
First, there is a significant distinction between public and private corruption, but the latter is a source of inefficiencies. Indeed, although the aggregate economic impact of private corruption is no doubt much smaller than that of public corruption, there is actually less ambiguity about the inefficiency of private corruption. In a standard case of commercial bribery, a buyer for a firm accepts a bribe from a supplier to buy some input from that supplier rather than another. The result is to increase the buying firm's costs, which is both a private and a social harm. Some public corruption, however, can be efficient because it circumvents inefficient laws. Private firms will rarely impose inefficient rules.
This may clarify the sense, confusing to one of the commenters, in which I used "victimless" to describe the crime of bribery. I placed the word in quotation markets in my original posting as here because there are of course victims. But the victims tend to be members of society at large, none of whom has an incentive to complain to the authorities. The bribe giver and bribe taker are unlikely to complain; in that respect the bribe giver is different from the victim of a crime like robbery, who is eager to complain.
A cost of public corruption that I neglected to mention is that it advantages insiders, i.e., people who know whom to bribe, how much to pay and in what form, etc. So people invest in becoming insiders, and the investment represents an added cost of transactions that has on average no economic benefit. This moreover can help us to see why, as I pointed out in my original posting, bribery tends to be more common among immigrant groups. Newly arrived immigrants have difficulty assimilating to their new society so they fall back on personal and business networks within their immigrant community, where they do not face a language or other cultural barrier. In effect they bypass many of the established transactional systems in their new society in favor of systems more suitable to their circumstances. Instead of dealing comfortably with strangers, they base transactions on trust, personal and family contracts, and ethnic solidarity--in other words on methods of transacting that are different from that of impersonal markets--and this facilitates bribery when some members of the network become officials.
Some comments suggest that Becker and I neglect the role of personal values of honesty and integrity in limiting the amount of bribery in a society. Such values are indeed important influences on behavior. But the interesting question is where they come from. We believe they come largely from social and family structures strongly influenced by economic factors, such as immigrant status.
One commenter suggested that bribery is egalitarian, because contracts favor the rich and bribery does not. I think this is mistaken. Contracts often involve asymmetric resources and information, as in the typical consumer transaction, but in a competitive market the net transactional benefit is to consumers because competition forces price down to (or at least close to) cost. Since the cost of a good may be much less than its value (i.e., how much the consumer could be forced to pay for it in the absence of competition), consumer transactions generate consumer surplus, which is to say a net value, above cost.
What is true is that bribery tends to be more common in poor societies. But that is not because bribery benefits the poor, but because bribery is one of the factors that makes a society poor!
Finally, some comments raise the question whether donations by business or other organizations or groups to political campaigns should be considered a form of bribery. There is certainly an analogy, since these donations involve giving money to politicians, many of whom are already officials, in the hope of obtaining some favor to which the donor would not otherwise be entitled. But the differences between this type of "corruption" and conventional bribery are sufficiently great to warrant treating the political donations as a separate issue. The differences include: (1) the pecuniary benefit to the recipient of the donation is indirect, as the money is not being used for his personal consumption but for his political campaign; (2) the recipients are nonincumbent candidates as well as officials, and without access to private donations it would be difficult to challenge an incumbent; (3) more broadly, the alternatives to private financing of political campaigns are fraught with problems; (4) the donations go largely to finance political advertising, which has some social value; and (5) there is no direct quid pro quo.
