Managing Risks in Public Policy: Impact vs. Probability?

If you were never born, you would never die. Therefore birth is not simply the leading cause of death, but the only one. Everything else is a statistical detail of where and when, with each day bringing to us a new set of end-game probabilities. Fortunately, we have the ability to influence those daily odds, however slightly, by how we assess and manage the risk of living. History teaches us that we could do much better, individually and collectively, in a dangerous world of loose nukes, terrorism, massive oil spills, hurricanes, volcanoes and tsunamis.

Tough Choices

From the moment you get out of bed, every day, you begin to assess and manage the myriad risks that threaten your peaceful existence. You brush your teeth to minimize the risk of tooth decay. Perhaps you eat a light breakfast of fruit and whole-grain cereal to reduce the chance of heart disease. You might arm your home security system at night to counter the risk of theft. Hopefully you diversified your investment portfolio prior to the most recent economic decline to offset the risk of a market slide.

We all make hundreds of these daily decisions. We attempt to make rational choices in a world of uncertainty and unknowns. We calculate risk by forecasting the potential future outcome of today's action, but with no reliable models on which to base those forecasts. The task is quite difficult.

In fact, risk management is so challenging that we all do rather badly at the game. If given a choice between two scenarios in which you have a 10% chance of dying in one case or a 90% chance of living in the other, most of us choose the latter, even though the two options are identical. A proposal can be presented as having either a 70% chance of success or a 30% chance of failure; the former appears more appealing but no difference exists between the two. We bring that same irrationality to our risk management decisions. All of us are vulnerable to the trap of believing our decisions are rational while succumbing to emotional whims. That is simply the human condition. Nothing is rational about our emotional instinct that a 70% chance to win is better than a 30% chance to lose. Our decisions are deeply impacted by how problems are presented to us.

We can overcome our vulnerabilities to some extent by better understanding what risk assessment and management really mean. Let's start by first breaking risk down into two main categories: inherent and operational. Then we will discuss the difference between risk and danger, and how that distinction can inform national policy planning and our daily lives.

Inherent Risk

Inherent risk is that caused by our interactions with the infrastructure of the physical world. Walking down the sidewalk on a sunny day has inherent risk; injury or death can come suddenly from a car jumping the curb, a brick falling off the face of a building, or a falling tree. We can do little to avoid this risk because our own actions, such as strolling down the sidewalk, are not immediately causative. Aviation is a perfect example of this type of risk.

Being transported in an airplane creates an inherent risk, which we endure for the associated benefits of inexpensive and convenient travel over long distances. Even under the most ideal and benign conditions, leaving terra firma and cruising along at 600 knots is not equivalent to sitting home on the couch wearing a helmet. Flying is inherently dangerous. But we have the power to make aviation safer by recognizing, and then carefully managing, the risks inherent to fighting gravity. On the strength of proper training, diligent maintenance of the aircraft and continuing education of our pilots we can largely overcome the underlying dangers. The same logic and approach apply directly to how we manage the inherent risks of a drive on the freeway, nuclear power, oil exploration, terrorism and life in the proximity of natural disasters like earthquakes, floods or volcanoes.

After every accident, whether a plane crash or exploding oil rig, we rightfully seek to assign blame to punish negligence and to learn from others how to avoid future mishap. Our assumptions however when investigating a bad outcome are often wrong, leading us to inappropriate conclusions. Again, aviation provides a good example. Most airplane accidents are portrayed as the result of pilot error. The statement may be correct but contains an embedded fallacy. The false idea is that the pilot was doing something inherently safe and then crashed by taking an unsafe action. But in reality the pilot was doing something inherently risky, and failed to manage the risk properly. Those two explanations are not at all equivalent. Which brings us now to operational risk.

Operational Risk

Operational risk is that caused by our own decisions and actions in pursuit of a goal or objective. We willingly take on operational risks by measuring the anticipated consequences of our actions against potential gains. We can choose to take an action that is more dangerous than not taking that action. I might risk a slip on ice if I were walking to an important meeting, but avoid that risk if I simply intended to go snag a cup of Joe. Or I could decide to risk the ice for a latte, but go a few blocks out of the way to avoid the iciest areas. We have a much greater degree of control over operational risks. Unlike with inherent risk, here our decisions and actions are indeed potentially causative.

Operational risks represent the greatest threat to public safety. We face these threats every day. Our decision-making itself is what creates the risk. To understand this deeply, how our decisions create risk, we need to distinguish between risk and danger. The two are often used interchangeably as synonyms, but they represent two completely different concepts.

Risk versus Danger

Let's say we know a hungry tiger is lurking in a jungle that we must traverse to get to the golden city on the other side. We can take a shortcut to beat the competition, which will increase our chances of encountering the beast; or we can go the long way around and significantly reduce any likelihood of seeing the mad cat. As usual, the greater the risk we take the greater the potential reward. But note that in this equation of risk and reward, the danger represented by the tiger remains constant. What changes is our approach to the danger. Our risk of becoming cat food will increase or decrease depending on what path we choose to follow, but the ravenous tiger is dangerous no matter what we do. Danger and risk are not synonymous.

Danger exists as a condition; risk measures the probability that you will encounter a danger, and the probability that the danger you encounter will have a bad outcome.

Loose nukes and volcanoes are like that ravenous tiger: inherently dangerous. Our risk of being consumed by hot ash or a ball of radioactive fallout will wax and wane depending on what path we choose, but the built-in dangers never go away. We ourselves define the risk by what we choose to do. A hiker climbing the side of an active volcano is taking greater risk than a trekker taking in the sites on a Bahamian beach, but the danger represented by the volcano remains constant.

Managing Does Not Mean Eliminating

With the distinction of risk and danger clarified, we need to tackle another serious misconception. Risk management does not mean in the real world eliminating risk although the two are commonly confused. Rather the idea is to recognize risk, reduce potential danger to the greatest extent practical, and mitigate that which cannot be avoided. Note that "greatest extent practical" does not mean risk-free.

Since we cannot eliminate risk, and no form of risk management is perfect, we come to an odd conclusion, but one important to public policy. An accident does not automatically mean someone did something wrong. An accident can simply be the manifestation of inherent risk, even if everybody did everything right. The very existence of risk means that we can see a bad outcome even under the best circumstances; if not true, there would be no risk. Do not misunderstand this: most accidents are avoidable, and an accident usually is in fact due to mistakes, failures, negligence or specific intent to do somebody harm, but not necessarily so.

You do not believe me? Consider a scenario from our favorite source of examples, aviation. A large bird strikes and severely damages the propeller of a single-engine airplane flying between Greenland and Iceland. The high level of training, meticulous aircraft maintenance, detailed planning, beautiful weather, and well-honed pilot skill that brought the aircraft to this moment matters not, and will not change the inevitable fatal outcome of that strike. "Stuff" happens. Not every accident is an indictment nor necessarily reflective of poor skill or judgment.

Giant Asteroids and Stubbed Toes: Impact versus Probability

A giant asteroid hitting the earth would be catastrophic, but fortunately has a low probability of happening during your lifetime. In contrast, you have a high probability of stubbing your toe during your lifetime, but with little consequence. So nerds worried about risk assessment and management need to consider together not only the impact of an event but the chance that the event will happen in a defined timeframe (your lifetime, next decade, in 10,000 years). We have four extremes that define our event boundaries:

• Small consequence and not likely to happen: Butterflies invading your garden and damaging your tomatoes.

• Small consequence and very likely to happen: Unexpected rainstorm ruining your weekend at the beach.

• Huge consequence and not likely to happen: Sarah Palin being elected president. Or an asteroid impact.

• Huge consequence and very likely to happen: Oil rig explosion leading to massive oil spill. Or terrorists exploding a dirty bomb in a major city somewhere in the world.

Making Sausage and Public Policy

We can now apply these ideas to the important task of governing. Unfortunately, many of our laws and regulations are currently based on emotional instinct (choosing 30% chance of success over 70% chance of failure) rather than on refined risk management strategies, sometimes with tragic and costly consequences. We need to do better. Let's see how, using just two major public policy arenas as examples.

Public Health

Heart disease (all forms) and cancer combined cause more than half of all deaths in the United States (28.5% and 22.8% respectively). Of the deaths from cardiovascular disease, about 30% can be attributed to smoking. Of the cancer deaths, 30.9% are due to lung cancer. About 90% of lung cancer deaths are due to smoking. The bottom line is that tobacco kills about 400,000 people every year in the United States. Excluding smokers, obesity kills an additional 300,000 people per year.

Tobacco is that tiger in the jungle, an inherent danger. Our risk of being harmed by tobacco is entirely a function of our own actions (excluding for the sake of argument second-hand smoke). The tiger lurks, but we can choose to avoid the beast entirely. This is a perfect opportunity for public policies to reduce risk by influencing our behavior. So what have we done? We have provided tobacco farmers with nearly $1 billion in subsidies since 1995. At the same time we devote nearly $1 billion per year in biomedical research to combat lung cancer.

Obesity is the lion in the Serengeti, a predator we encounter only by choice. Here too is a golden opportunity for public policy to have a huge impact. So what do we do? NIH just recently funded a paltry $37 million research program on human behavior to develop more effective interventions to reduce obesity. NIH has an annual budget greater than $30 billion. Health insurance policies rarely pay for wellness visits.

We as a nation can assess accurately the risks of smoking and obesity. But we fail to manage those risks. With reasonably simple changes in lifestyle we could every year cut the number of deaths nearly in half. Pause and consider for a moment the mind-boggling implications of that reality. But in spite of this astronomical potential for good, we focus our resources and energy elsewhere, victims of emotional instinct. Knowing what we do, the vast majority of all biomedical research funds should clearly be directed to smoking prevention and promoting healthy eating habits and related lifestyle changes. Our tax laws should make eating healthy foods more attractive than chowing down another fry. But we cannot muster the courage to institute even a small tax on sugared beverages, which have become a significant factor contributing to obesity. The average American now drinks 54 gallons of soft drinks annually, five times as much soda as Europeans. Our policies fail to manage a known risk.

Homeland Security

The U.S. Department of Homeland Security has an annual budget of about $50 billion. Probably a great deal more is spent on classified programs. Their mandate falls into four major categories: guarding against terrorism, securing our borders, enforcing immigration laws and preparing for natural disasters. Perhaps all $50 billion is well spent in ways we cannot know since thwarting terrorists must largely be done in secret. But we have evidence that DHS efforts may ignore the basic tenets of risk assessment and management that we learned earlier, falling victim, like us, to emotional instinct.

In 2007 Congress mandated 100% screening of ships bound for the United States at more than 600 ports of origin. The problem is that the technology for scanning about 15 million containers every year simply does not exist. The Bush White House opposed the mandate stating that the scanning requirement was "neither executable nor feasible." Remember impact versus probability? Imagine the impact of a nuclear bomb exploding in a major U.S. city smuggled through one of our ports. Has this extraordinary threat been a major focus for DHS? Apparently not, for we appear to be no more protected in 2010 than we were in 2007. We are not devoting significant resources to reducing the probability of what would be a catastrophic event.

We cannot live without water, yet our water supply is extremely vulnerable to terrorist attack. The nation's water infrastructure is comprised of three main elements: supply source from lakes, rivers, and wells; treatment facilities; and the distribution system. Some money has been spent to protect supply sources, but they remain vulnerable, and nearly no resources have been devoted to protecting the distribution system. Consider this: less than one-twentieth of a quart of ricin (or anthrax) can render one million gallons of water toxic. That volume of water fills more than 100 miles of six-inch diameter pipe, anywhere along which the toxin can be injected. Impact versus probability? Can we imagine a consequence much worse than contaminating the water supply to millions of people? But we seem to be spending little on reducing the probability.

A shoe bomber attempted to bring down an airplane, so we now all take off our shoes at airport security. Why then after the underwear bomber failed are we not all required to take off our skivvies? Thankfully we have not yet seen a suppository bomber, although a failed bra bomber might have better consequences for some of us. A would-be bomber tried to create a bomb by combining two reactive fluids, so now we are all restricted to that ridiculous quart bag and 3 ounces per fluid container. What would prevent four or five conspirators from combining their 3 ounces once on the airplane?

We focus our attention and resources inappropriately like this because we ignore the basics of good risk assessment and management. We need to understand and design our programs and policies to accommodate inherent versus operational risk, danger versus risk, the difference between managing and eliminating risk, and the differential consequences of impact versus probability. If we did we would be protecting our ports and water supply instead of (or in addition to) disrobing at the airport. We would do everything possible to reduce smoking. We could begin by stopping our tax payer subsidies to grow tobacco. We would tackle the epidemic of obesity instead of cowering before the soft drink and junk food industries.

We have assessed the risks. Now let us gather the courage to massively redirect our resources to those that are high probability and high impact. Walking barefoot through airport security while five TSA officers look on makes little sense if a nuclear bomb is at the same time entering one of our ports unobserved and undetected.

Jeff Schweitzer is a scientist, former White House senior policy analyst and author of Beyond Cosmic Dice: Moral Life in a Random World (Jacquie Jordan, Inc)( Follow Jeff Schweitzer on Twitter ( and on Facebook.