Next week marks the fifth anniversary of the April 27th, 2011 tornado outbreak, which claimed 319 lives across five states, including 238 in Alabama. The anniversary allows us to reflect on the lives lost and our vulnerability to future tornadoes.

I still lived in Texas on April 27th but had just published a book with Kevin Simmons of Austin College on the societal impacts of tornadoes. We documented the life-saving effects of tornado warnings and the National Weather Service’s nationwide network of Doppler weather radars. Kevin and I wrote a follow up book on the tornadoes of 2011 titled Deadly Season.

Meteorologist James Spann on Alabama Weather Blog always assures readers that potential upcoming severe weather outbreaks will not be as bad as April 27th, simply because this was a once in a lifetime outbreak. The number of reported tornadoes surpassed the Super Tornado Outbreak of April 3-4, 1974, and included four EF-5 tornadoes (the highest rating category) and 11 EF-4 tornadoes. For perspective, the U.S. has had just five other EF-5 tornadoes in the last 16 years, and not state had experienced 100 or more tornado deaths in a year since 1971.

My research with Kevin Simmons established that tornadoes in the Southeast are more deadly than elsewhere. One reason is a confluence of risk factors: tornadoes that strike mobile homes or during the night or fall and winter months are more deadly. The Southeast disproportionately faces all of these factors. The April 27th outbreak, however, featured long track, violent, spring, and (mostly) daytime tornadoes.

Tornadoes in the Southeast are still more deadly even when controlling for strength, timing, mobile homes, and so on; indeed about 30% more deadly on average. Researchers do not know what drives this extra deadliness. But the April 27th death toll might have been 70 lower if tornadoes here were only as deadly as in the rest of the nation.

Can we make future tornado outbreaks less deadly? The ongoing VORTEX-SE research effort will undoubtedly help. Beyond this, we must also think carefully about reducing tornado vulnerability. Building underground tornado shelters or above ground saferooms into all of our homes, businesses and schools would definitely help. And building shelters everywhere is a good idea if your goal is to minimize tornado fatalities. The cost, however, would run into the billions of dollars, just for Alabama. Shelters everywhere are not as good of an idea when we recognize what we would have to give up.

Protecting one’s self and family is perhaps the most basic human drive, so we could let people decide how to protect themselves against tornadoes and life’s other risks. Generally I favor this, but the evidence shows that people sometimes make predictable and repeated mistakes. Particularly we sometimes ignore small risks like tornadoes, and put off doing things like installing a shelter.

Our goal, I think, should be ensuring that people make informed decisions about risk. Research can help here. Economists studying risk calculate a cost per life saved for various protective measures. Research finds that Americans typically invest in safety when the cost per life saved is $1 million or less. (Smoke detectors or optional features on cars typically have modest costs and help reduce a small risk of death). If few people took an action where the cost per life saved was say $10,000, this may indicate poor or ill-informed decisions.

Kevin Simmons and I have calculated the cost per life saved for tornado shelters for a dizzying array of assumptions, and this value typically exceeds $40 million for single family homes, even in Alabama. Perfectly informed Alabamians will frequently and deliberately choose not to purchase shelters. Of course, each of us must decide how much we will spend to protect against life’s many different risks, so Alabamians who buy shelters are making the right decision for them.

Our nation has made great strides in reducing the lethality of tornadoes, nature’s most violent storms. Tornadoes today are perhaps about one-tenth as deadly as at the start of the 20th Century. If we approach the remaining vulnerability with reason and deliberation, we can continue to make tornadoes less deadly.

Daniel Sutter is the Charles G. Koch Professor of Economics with the Manuel H. Johnson Center for Political Economy at Troy University and host of Econversations on TrojanVision. Respond to him at dsutter@troy.edu and like the Johnson Center on Facebook.