The 9/11 terrorist attacks, Hurricane Katrina and the recent financial collapse: These and other recent catastrophes are a wakeup call for the need to better prepare for natural and man-made disasters that have wider effects as the world grows more interconnected.

But do we have the planning tools we need?

Perhaps not, some experts in risk analysis warn. And even if experts could develop better statistical tools, there are human tendencies to be overcome, like the habit of downplaying some risks and overreacting to others. Reason becomes clouded when people assess emotion-laden events like terrorist attacks, making public-policy decisions on risk very difficult, according to participants in a recent Wharton conference titled, “The Irrational Economist.”

One hundred leading scholars — including Nobel Laureates — in the fields of decision sciences, economics of information, political economy, catastrophic risk management and insurance gathered for the December conference, which was organized by Erwann O. Michel-Kerjan, managing director of the Wharton Risk Management and Decision Processes Center, and Paul Slovic, professor of decision research at the University of Oregon.

The conference had several goals. One was to recognize the contributions made to these fields by Howard Kunreuther, Wharton professor of decision sciences and public policy, during nearly four decades of research – and to note the occasion of Kunreuther’s 70th birthday. In addition, the conference was intended to bring together leading scholars to exchange ideas and cross multiple perspectives on some of today’s most pressing issues, and to lay the groundwork for future multidisciplinary research over the next decade. A book, titled The Irrational Economist: Making Decisions in a Dangerous World, distilling the knowledge of these scholars, will be published in early 2010.

Kunreuther says he found the conference gratifying. “The two days interacting with good friends, family and colleagues were very special and ones I will always remember,” he says. “Thanks to the contributions of so many people, the conference provided a broader perspective on ways to increase our understanding of human behavior with an eye to developing policies and programs for improving individual and social welfare. From my perspective, what more could one ask for?”

When people think about risk, they tend to dwell on ones experienced in the past rather than new types likely to surface in the future, said Michel-Kerjan. “We are always surprised. We try to fight the last war over and over again.” He was part of a conference panel titled, “Managing and Financing Extreme Events II: Dealing with a New Era.”

Traditional risk assessment, like that used by insurers, gauges the probability and likely cost of an event by assessing historical data. Insurers know from experience, for example, that an 18-year-old driver is more likely to have a serious accident than a 40-year-old, so the teenager’s premium is higher.

But a number of factors make this type of calculation less reliable in assessing the kinds of catastrophic events that have struck in recent years. Terrorist attacks on the scale of 9/11 are too rare to establish a statistical probability. And although there are many statistics on natural events like hurricanes, the old numbers may be less reliable if global warming is making hurricanes more frequent and severe. Rapid population growth in coastal areas adds to the challenge of assessing future costs.

Michel-Kerjan predicts “an accelerating rhythm of large-scale catastrophes” such as financial crises, problems with food and energy, disease pandemics, terrorism and natural disasters. With costs adjusted for inflation, all of the 20 most-costly disasters from 1970 through 2007 occurred after 1987, and half of them came after 2001.

In a paper titled, “Toward a New Risk Architecture: Welcome to Risk Management 2.0,” Michel-Kerjan writes: “There will be more of these [disasters] as well as brand new ones in the future. This poses a real challenge: Dealing with an average of one or two such catastrophes every 20 years is one thing; dealing with 10 or 15 on many different fronts, as is currently occurring, is a whole different game.”

China’s Quakes Rattle Wal-Mart

A further problem: the world’s growing interconnectedness. “An earthquake in China is not just a Chinese issue any more, it’s a Wal-Mart issue,” Michel-Kerjan said during the panel.  

The financial markets have tied countries together in ways that are not fully understood. The current financial crisis began with a small slice of the home mortgage market in the U.S. and evolved into a worldwide recession. Indeed, the global trade in securities backed by American mortgages and other forms of debt helped spread the contagion, while many experts had expected the opposite — that these securities would dampen financial shocks by diluting risk.

“Conventional thinking holds that risks are mainly local and routine; that it is possible to list all untoward events that could happen, determine their probability based on past experience, measure the costs and benefits of specific risk protection measures and implement these measures for each risk,” Michel-Kerjan writes in his paper, which was presented at the conference. “Many organizations and governments are making decisions using risk and crisis management tools based on these outdated assumptions.”

People can better prepare by recognizing that risks are more common and more serious than previously thought. But it is hard to know what the best preparation is, since overreaction can be a problem, too. “Panic and passivity are the opposite ends of a spectrum. Each is unwelcome,” Harvard professors Cass R. Sunstein and Richard Zeckhauser, write in their paper titled, “Overreaction to Fearsome Risks,” also presented in the panel discussion.

The risks that trigger the strongest responses tend to have extremely low probabilities of occurring, the authors said. Overreaction can lead to excessive spending on remedies, too much anxiety or resorting to counterproductive choices — like keeping the children indoors so they won’t be hit by meteors.

Overreaction is generally caused by a fixation on the possible event while ignoring the fact it is unlikely, a process called “probability neglect.” For example, after a terrorist attack, the public is likely to demand a government response. But excessive airport security may cause more people to drive, leading to more fatalities than if those people traveled on commercial air carriers, which are safer. “The monies spent in recent years on [airport] security seem out of scale with the level of risk reduction produced, particularly since numerous tests have found that the screening routinely fails to find weapons,” the authors write.

The Gruesome Factor

To study probability neglect, the authors asked a large sample of law students how much they would pay to reduce the risk of drinking water containing arsenic. The students were randomly divided into four groups. In the first, students were told their risk of getting cancer from the water was one in one million, while the second group was told it was one in 100,000. The third group also faced a one in one million risk, but the cancer was described as “very gruesome and intensely painful, as the cancer eats away at the internal organs of the body.” The fourth group faced a one in 100,000 risk and also received the graphic description. The study showed that adding the emotional element dramatically increased the amount students would pay to avoid the cancer, even though the graphic statement actually did nothing to increase their risk. Fear made the students ignore probabilities.

Indeed, the study showed that students would pay more to avoid the one in one million risk that was graphically described than to avoid a one in 100,000 risk that was not, even though the one in 100,000 risk is 10 times more likely, said Zeckhauser, who took part in the conference panel. “People are just mesmerized confronting a fearsome risk.”

Turning to a real-life example, Sunstein and Zeckhauser note in their paper that the anthrax scare in the U.S. in October 2001 caused only four deaths and a dozen illnesses, but triggered an enormous government response. “We have suggested that when risks are vivid, people are likely to be insensitive to the probability of harm, particularly when their emotions are activated,” they write.

Irrational about Insurance

The insurance industry has long profited from customers’ irrational behavior. In his paper titled, “Half Full or Half Empty? Economic and Behavioral Explanations for Insurance Demand,” Mark V. Pauly, professor of insurance and risk management at Wharton and a conference panelist, described consumers’ tendency to buy insurance policies with low deductibles and high premiums even though it is more economical to do the opposite — buy policies with high deductibles and lower premiums. If the event being insured against is unlikely to take place, the policy holder will probably never pay the large deductible, but will certainly benefit by paying a low premium. Studies have shown that even wealthy people who could easily pay large deductibles tend to prefer low-deductible policies.

It is unclear why people behave this way, Pauly said. It is possible they place a high value on peace of mind, freedom from the kind of regret they might feel if they did someday have to pay a large deductible. It also is possible that people are looking for a rough way to factor in other losses associated with the event that cannot be insured against directly. By avoiding the high deductible, the policy holder is paid more when a claim is filed, helping to offset some other loss.

A typical example is the high-cost rental car insurance many people buy even when their own auto policies cover rental-car accidents, according to Pauly. A customer might buy the rental insurance for non-financial reasons. “The main reason why I took the rental car insurance is because if anything happens, I don’t want to have to explain it to my wife,” he joked.

The incompletely understood behavior of individuals addressing risk is reflected in the public-policy realm, he said, since politicians share those tendencies, or must satisfy constituents who may react irrationally in the face of risk.

If risk assessment is behind the times, or still struggling to understand human behavior, new approaches are needed. While no one has yet devised an ideal system for addressing new types of risk, Michel-Kerjan offered conference participants six points to consider.

Extreme costs and extreme benefits: Disasters will cost much more than they have in the past, making effective methods of addressing them ever more beneficial. In the U.S., 1989’s Hurricane Hugo was the first catastrophe to cost more than $1 billion, while Hurricane Katrina just 16 years later probably cost more than $150 billion. The population of hurricane-prone Florida was 2.8 million in 1950. It is expected to exceed 19 million in 2010.

Confusion over public- and private-sector roles in catastrophes: Most people assume government is responsible for handling crises, but many key services that keep the country running, such as power and transportation, are private. “We must look at how private actions affect public vulnerability so that we are better prepared.”

Growing interdependence and globalization: In 2003, a problem in an Ohio utility caused a power failure throughout much of the northeastern U.S. and parts of Canada. Similarly, the near collapse of American International Group, the mammoth insurer, was caused by a relatively tiny unit in London.

Change in scale: Greater independencies and other changes mean individual disasters will affect more people. “Dealing with large-scale disasters is much more challenging than dealing with a series of local small accidents.”

Celerity: Rapid transportation and cheap communication allow people and products to move much faster, speeding the spread of contagions like diseases and financial reversals.

Uncertainty: As disasters become bigger and more widespread, it becomes more important to make decisions quickly, often before a situation is fully understood. “We were trained to solve problems with clear questions and clear scientific knowledge,” he writes in his paper, arguing that this traditional approach won’t be good enough in the future.

“As the world is becoming more of a small interdependent village, crises that you decide to ignore because they are 5,000 miles away…will certainly have second-order effects on you, very soon. Given this, we might reflect on an irony: Being selfish today means taking care of others.”