#statistics  #travel  #airplanes  
Issue 29
June 2, 2019
___ _ _ ___ _ _ __ __ _ ___ _ _ / __|| |__ ___ __ __ __ ___ __| | | \ __ _ | |_ __ _ (_)\ \ / /| |_ _ _ | __|| | _ _ (_) _ _ __ _ \__ \| / // -_)\ V V // -_)/ _` | | |) |/ _` || _|/ _` | _ \ \/\/ / | ' \| || | | _| | || || || || ' \ / _` | |___/|_\_\\___| \_/\_/ \___|\__,_| |___/ \__,_| \__|\__,_|(_) \_/\_/ |_||_|\_, | |_| |_| \_, ||_||_||_|\__, | __ __ _ _ _ _ _ _ ___ _ ___ |__/__ |__/ |___/ | \/ |(_) __ _ | |_ | |_ | \| | ___ | |_ | _ ) ___ /_\ ___ / __| __ _ / _| ___ | |\/| || |/ _` || ' \| _| | .` |/ _ \| _| | _ \/ -_) / _ \ (_-< \__ \/ _` || _|/ -_) |_| |_||_|\__, ||_||_|\__| |_|\_|\___/ \__| |___/\___| /_/ \_\/__/ |___/\__,_||_| \___| |___/ _ __ __ _____ _ _ _ /_\ ___ \ \ / /___ _ _ |_ _|| |_ (_) _ _ | |__ / _ \ (_-< \ V // _ \| || | | | | ' \ | || ' \ | / / /_/ \_\/__/ |_| \___/ \_,_| |_| |_||_||_||_||_||_\_\

At The Quintessential Centrist, we are ardent fans of social economist Steve Levitt. Mr. Levitt, who currently plies his trade at the University of Chicago, possesses an uncanny ability to seamlessly explain complicated statistical information using fun, fascinating, real-life examples that keep his students and readers alike engaged and entertained. We recommend his enjoyable books, Freakonomics and Super Freakonomics.

Like most aviation experts, airplane manufactures and operators often make the case that flying is the safest form of travel. In Super Freakonomics, Mr. Levitt makes a similarly compelling case supported by exhaustive data: that driving is safer than flying. It probably is. However, in our view, some of the analytical data comparing the safety of those two respective modes of transportation – while not flawed per se - are materially skewed in favor of driving.

The most commonly used metric to measure safety in the driving vs flying debate is the morbid statistic of fatality rate per mile traveled. Sure enough, on page 65 of Super Freakonomics, Levitt argues, "per mile, driving is much more dangerous than flying." While factually correct, a more granular analysis suggests that this is a misleading statement with limited practical application when considering which mode of transport, air or auto, is truly safer.

Why?

When comparing the safety of air travel vs that of an automobile on a per mile basis, one fails to control for an important variable in the equation, the total number of hours of exposure. Consider the following: a commercial jetliner travels ~500 miles per hour (mph) at cruising altitude. A passenger vehicle goes ~70 mph on an open, interstate highway. As Mr. Levitt so eloquently does, let us use a real-world example to help frame our argument. A lady drives for one hour, she travels 70 miles. A week later, she decides to fly for one hour; she goes 500 miles. However, in order to travel a distance of 500 miles in a car, our lady friend would have to drive approximately 7 and 15 minutes, or 7.25 times further than a comparable trip in a plane, leaving a lot more opportunity (time) for bad things to happen. Hence, using per mile metrics when comparing the safety of various means of travel is deceptive, because that form of measurement does not control for the total amount of time spent driving a car vs flying in a plane. If our subject drove and flew for the same amount of time, the data would tell a very different story and airplanes would appear less safe. So if we are going to use fatalities per passenger mile as a default metric upon which to base our analysis, we must also control for the amount of exposure or time spent in each mode of transport; but we don’t.

Pilot Peter Jones noted in his commentary to the American Scientist magazine in 2003,"...while the article's conclusions are acceptable in the broad sense, they are highly skewed by the common definition of safety in terms of passenger-miles per fatality…as a human (and pilot) with a finite life span, I should point out that the numbers on gravestones represent time, not miles traveled..."

Another data point where the metrics tilt in favor of driving involve how airlines report total miles flown. Let’s make the reasonable assumption that 200 people are aboard a plane from New York City bound for Los Angeles. NYC & LA are ~2,750 miles apart. According to standard reporting metrics, that flight would count as 2,750 (miles) x 200 (passengers), or 550,000 miles (flown). However, this is grossly misleading. Regardless of whether a plane is empty, half-full, or full, it is still ~2,750 miles from NYC to LA, not 550,000 miles.

This is imperative to appreciate because when examining data on a per mile flown basis, the appropriate way to manipulate the data is to not only control for the amount of time spent driving vs flying, but also for the total distance traveled, not the total distance traveled multiplied by everybody in a vehicle or on a plane. Motorcars typically carry one or two people, and rarely more than four passengers. Commercial jets on the other hand can easily accommodate more than 200 passengers. Therefore, multiplying the number of passengers on an airplane by the total distance between destinations grossly inflates the total number of miles flown, which in turn gives a false sense of security when measuring the accident or fatality rate per mile flown. This distorts the data in favor of driving and makes air travel appear safer than it probably is.

Commercial airplane accidents that involve mass casualties are exceedingly rare in America. The last one occurred on February 12th, 2009 when 49 people perished on Colgan Air Flight 3407 from Newark, New Jersey, to Buffalo, New York. The plane stalled out and crashed in inclement weather. By contrast, ~35,000 people are killed per year in automobile accidents. At first glance it would appear that travel by plane is exponentially safer than by car.

But is it?

While commercial airplane crashes are highly atypical events, private plane accidents are quite common. In 2013, there were ~1,300 general aviation crashes which resulted in ~400 deaths. That was a “safe” year. In 2012 there were over 1,500 private plane crashes. On the surface, that pales when compared to the ~35,000 people killed in ~6,000,000 auto accidents each year. However, this information should not be interpreted in a vacuum. We must consider that ~220 million people hold drivers’ licenses in the United States. By comparison, only ~600,000 people hold pilots’ licenses; of those, only 1/3rd or ~200,000 are private licenses. Furthermore, ~30% of those casualties in auto accidents involve a drunk driver. Very, very few people fly while intoxicated. Finally, if there is an aviation accident, it is much more likely to result in death than that of an auto accident. Hence, when controlling for the number of licenses issued, alcohol, and certainty of death in the event of an accident, air travel becomes less appealing.

Let us be clear: we are by no means arguing that driving is safer than flying – we do not have data to support that claim - nor that flying is inherently unsafe. Our position is similar to that of Mr. Levitt and the general consensus. Flying is probably safer than driving, though it is not as safe as most people perceive it to be, particularly outside of the parameters encapsulated by major American commercial airliners.