We'll soon know the magnitude of the tragic decision made by 100,000 or so people not to evacuate New Orleans (New Orleans begins ...).
The larger picture of death was just as murky. No one could say how many had died in the hurricane or were waiting to be rescued after the city's levees burst. One morgue at the St. Gabriel Prison near New Orleans was expecting 1,000 to 2,000 bodies. Hundreds were missing in nearby Chalmette. In Baton Rouge, state officials said the official Louisiana death toll stood at 59, but most said that thousands was a more realistic figure. More than 125 were known dead in Mississippi.
Does a tragic decision mean that the choice process itself, to ride out the storm, was irrational? Rational decision makers compare the subjective benefits and costs of an activity, with imperfect information, and pursue those that yield expected benefits greater than costs. So, no. The ex-ante decision to not evacuate New Orleans, for many, was a rational decision -- at least in the sense that economists use the word.
Instead of condemning people, ex post, for making a tragic decision (after all the information is available) let's try to figure out why they made that decision, using a catalog of benefits and costs, and then use rational choice theory to try to avoid a repeat of a catastrophe on the scale of Katrina.
The hurricane evacuation decision has been studied extensively by geographers, psychologists, sociologists, and, to a lesser extent, by economists. Here are some things we know.
The benefit of hurricane evacuation is the increased safety that is achieved by avoiding the wind and flood risks. The safety benefits are filtered through the information, experiences, etc that at-risk residents have available to them. In 2001 (a quote that is all over the blogosphere):
... the Federal Emergency Management Agency ranked the potential damage to New Orleans as among the three likeliest, most castastrophic disasters facing this country.
The objective risk information, New Orleans is a dangerous place with a category 5 storm in the gulf, was available to anyone who thought much about it, so the people of New Orleans should have known to get out, right? The problem is that the subjective risks are a function of the objective risks and a bunch of other factors.
Voluntary and mandatory evacuation orders increase the perceived risk and households tend to respond strongly to these, especially mandatory orders.
But "crying wolf" evacuations (i.e., households evacuate in response to orders and the storm misses or is weaker than expected) reduce the perceived risks and the benefits of future evacuations. Hurricane warnings may take on the same significance as the weekend weather from the local newscast.
Which brings us to the costs of evacuation. These include the travel and time costs of leaving one's home and staying. These costs are mitigated by the availability of hurricane shelters. Low-income people tend to choose this option more often. Time costs are an important factor. People don't like to spend 10 hours driving one mile.
Even if rational decision makers do a good job of balancing the benefits and costs, they face constraints that might make a tragic decision more likely. People with pets tend to not evacuate since hotels and shelters do not allow them. How many New Orleans residents in rooftop squatting communities had dogs? I sas a lot on TV. People with disabilities, the elderly and others with special medical needs have a more difficult time evacuating. Today's news had a story about 32 elderly found in a nursing home.
What can be done?
Anything that would increase the benefits of an evacuation would increase the chances that househols will evacuate when they should. In other words, subjective risks must be close to objective risks. Public education programs, especially for those with less education and poor risk perception formation, are important. But these have been the focus of emergency management for years.
Anything that reduces the costs of evacuation would increase the chance that those who should evacuate will evacuate. This is the part of the equation that seems to have more promise since it has not been a focus of emergency management. Some policies that haven't been tried include:
- Evacuation rebates. This past year, a state legislative aid from Louisiana called and asked me what the appropriate rebate for a hurricane evacuation might be. The idea was to encourage evacuation by reducing the cost. I don't know what happened to this proposal in LA but it sounds like a good idea for those low-income residents that can't afford a hotel room or basic transportation. But it is too expensive for all income levels, especially with several evacuation events.
- Matthew Kahn at the Environmental and Urban Economics blog suggests an automobile program for the urban poor. Commenters raise several concerns, including its expense, increased pollution and urban parking constraints. But the logic is right: how to transport those without transportation. We're read about the idle school buses, they are put in use elsewhere, why not everywhere?
- Would it be too much to ask hotels to allow pets during extreme circumstances? And maybe a shelter with kennel facilities? Government subsidies to hotels and local governments might encourage this.
Finally, natural disasters and emergencies have some public good characteristics. Effective government emergency management is important. Households need enough lead time to develop a cost-effective evacuation plan (emergency managers recommend that you have one before a hazard event but not enough people do so). Susan Cutter, a professor of Geography at the University of South Carolina and a well-respected hazards researcher, says (via NYTimes News Service and papers all over the country) said the:
... New Orleans' mayor should have ordered a mandatory evacuation earlier on Thursday or Friday. She said it would take at least two to three days to fully evacuate New Orleans, a city surrounded by water with few exit routes. The city's hurricane plan said it might take 72 hours for residents to leave.
And emergency responders should feel some urgency.