In the wee hours of the morning on Monday, February 15, my house, like many others across the state of Texas, lost power during a winter storm. For the next 48 hours, we wore layers of clothes and huddled under blankets as temperatures indoors dropped to about 40 degrees. Even after the power came on, water supplies were low, and the city of Austin was under a boil-water order.

This cold snap and series of storms were by far the worst I have encountered in the 23 years I have lived in town, but longtime Texans do remember other bad winter storms over the years. They are not utterly unprecedented.

So why was Texas so poorly prepared for a week of snow and freezing temperatures?

A lot of it has to do with how people make decisions about unlikely events. People generally have difficulty with understanding very small probabilities—and, importantly, how those small probabilities affect large populations.

SMALL PROBABILITIES

One key observation that psychologists Daniel Kahneman and Amos Tversky incorporated in their influential Prospect Theory is people have difficulty with small probabilities. Whenever an event is unlikely, people tend to treat it as if it will not happen. In addition, people are not that sensitive to differences among unlikely events, so something that will happen 1 in 10,000 times is not treated as that much different from something that will happen 1 in 10,000,000 times.

That means that most people assume that unlikely events are not going to happen to them. And, of course, most of the time, they’re right. So, people are generally rewarded for ignoring unlikely events, because they haven’t spent any time or energy to avoid something that probably was not going to happen anyhow.

To compound matters, preventive maintenance creates a second problem psychologically, because if you successfully avoid a bad outcome because of what you did do, you are unaware that your actions averted disaster. When you get a flu shot, for example, and you don’t get the flu, it is hard to credit the flu shot with preventing you from getting sick.

Put these together, and you can see how companies would avoid preventive maintenance for unlikely outcomes, such as winterizing Texas power plants for extremely cold weather. In most years, expenditures on low-probability events do not pay off.

SMALL PROBABILITIES AND LARGE NUMBERS

Probabilities themselves are a marvelous invention, because they allow people to compare outcomes despite differences in the size of populations. If you wanted to determine whether the flu is more common in Texas or in Delaware, for example, you might just look at the total number of flu cases. But Texas and Delaware are vastly different in size, so it is better to use a percentage or probability so that you don’t have to worry about the difference in the number of people and can just compare the rate of illness in each state.

Because we get used to looking at percentages and probabilities, though, we often forget about these population sizes. That is a problem, because an outcome that is rare for an individual can still be common across a large population of people. A disease that affects 1 in 10,000 people will still affect 100 people in a population of 1,000,000. One hundred people is a lot of people.

A similar thing happens when you look at the cost of a rare outcome. Early estimates of the damage from the 2021 winter storm in Texas suggest that it could cost $19 billion or more. The decision to engage in preventive maintenance requires thinking not just about the cost of doing that maintenance at the time, but also what will be saved over the long term by avoiding problems that will be costly when they do occur.

WHAT TO DO?

As individuals, it is important to actually do the math when it comes to low probabilities. That is, your gut reaction is not going to give you good information to decide whether to get a vaccine, change your diet for long-term health, or buy insurance. Instead, you need to calculate the costs of actions now, the probability of bad outcomes, and the costs of those outcomes when they occur. Even though we like to feel comfortable with the choices we make, when it comes to low-probability events, we need to listen to the numbers.

When thinking more broadly about society and companies, though, this is also a place where government has a role to play. The free hand of the market is good at finding opportunities that pay off now. Companies invest large sums of money to serve large numbers of customers quickly and efficiently.

Where companies tend not to invest is in small numbers of customers (for instance, developing rural infrastructure) and in low-probability events (such as treating rare diseases or preventive maintenance for unlikely outcomes). When devastating outcomes occur, though, the government often has to step in to cover many costs.

While many individuals and companies bristle at regulation by government, enforcing the purchase of insurance for unlikely events and requiring preventive maintenance are ways to both cover the costs of bad outcomes when they occur and also minimize the chances that bad things will happen.

We cannot afford to pay attention to the need to think about these unlikely events only in the aftermath of a major disaster. We have to live life routinely knowing that unlikely things can happen. The better prepared we are for bad outcomes, the less likely it is that they will have significant consequences when they occur.