Schneier on Security
A blog covering security and security technology.
« David Brooks on Resilience in the Face of Security Imperfection |
| Gift Cards and Employee Retail Theft »
January 6, 2010
Nate Silver on the Risks of Airplane Terrorism
Over at fivethirtyeight.com, Nate Silver crunches the numbers and concludes that, at least as far as terrorism is concerned, air travel is safer than it's ever been:
In the 2000s, a total of 469 passengers (including crew and terrorists) were killed worldwide as the result of Violent Passenger Incidents, 265 of which were on 9/11 itself. No fatal incidents have occurred since nearly simultaneous bombings of two Russian aircraft on 8/24/2004; this makes for the longest streak without a fatal incident since World War II. The overall death toll during the 2000s is about the same as it was during the 1960s, and substantially less than in the 1970s and 1980s, when violent incidents peaked. The worst individual years were 1985, 1988 and 1989, in that order; 2001 ranks fourth.
Of course, there is a lot more air travel now than there was a couple of decades ago. Although worldwide data is difficult to obtain, U.S. air travel generally expanded at rates of 10-15% per year from the 1930s through 9/11. If we assume that U.S. air traffic represents about a third of the worldwide total (the U.S. share of global GDP, which is probably a reasonable proxy, has fairly consistently been between 26-28% during this period), we can estimate the number of deaths from Violent Passenger Incidents per one billion passenger boardings. By this measure, the 2000s tied the 1990s for being the safest on record, each of which were about six times safer than any previous decade. About 22 passengers per one billion enplanements were killed as the result of VPIs during the 2000s; this compares with a rate of about 191 deaths per billion enplanements during the 1960s.
Why? Because over the past decade, the risk of airplane terrorism has been very low:
Over the past decade, according to BTS, there have been 99,320,309 commercial airline departures that either originated or landed within the United States. Dividing by six, we get one terrorist incident per 16,553,385 departures.
These departures flew a collective 69,415,786,000 miles. That means there has been one terrorist incident per 11,569,297,667 mles flown. This distance is equivalent to 1,459,664 trips around the diameter of the Earth, 24,218 round trips to the Moon, or two round trips to Neptune.
Assuming an average airborne speed of 425 miles per hour, these airplanes were aloft for a total of 163,331,261 hours. Therefore, there has been one terrorist incident per 27,221,877 hours airborne. This can also be expressed as one incident per 1,134,245 days airborne, or one incident per 3,105 years airborne.
There were a total of 674 passengers, not counting crew or the terrorists themselves, on the flights on which these incidents occurred. By contrast, there have been 7,015,630,000 passenger enplanements over the past decade. Therefore, the odds of being on given departure which is the subject of a terrorist incident have been 1 in 10,408,947 over the past decade. By contrast, the odds of being struck by lightning in a given year are about 1 in 500,000. This means that you could board 20 flights per year and still be less likely to be the subject of an attempted terrorist attack than to be struck by lightning.
In 2008, 37,000 people died in automobile accidents -- the lowest number since 1961. Even so, that's more than a 9/11 worth of fatalities every month, month after month, year after year.
There are all sorts of psychological biases that cause us to both misjudge risk and overreact to rare risks, but we can do better than that if we stop and think rationally.
Posted on January 6, 2010 at 2:59 PM
• 34 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
Main article: Economic effects arising from the September 11 attacks
A satellite view of Manhattan shows a large smoke plume a day after the attacks.
A New York City fireman calls for 10 more rescue workers to make their way into the rubble of the World Trade CenterThe attacks had a significant economic impact on the United States and world markets. The New York Stock Exchange (NYSE), the American Stock Exchange (AMEX), and NASDAQ did not open on September 11 and remained closed until September 17. When the stock markets reopened, the Dow Jones Industrial Average (DJIA) stock market index fell 684 points, or 7.1%, to 8921, a record-setting one-day point decline.
By the end of the week, the DJIA had fallen 1,369.7 points (14.3%), its then-largest one-week point drop in history, though later surpassed in 2008 during the global financial crisis. U.S. stocks lost $1.4 trillion in value for the week. This is equivalent to $1.69 trillion in present day terms.
Economist and Crisis Consultant Randall Bell was retained by the city and state of New York to compute the economic damages to The World Trade Center site. He writes in his book, Strategy 360, "The World Trade Center damages, estimated by the New York City Mayor's Office, were staggering: Clean up and stabilization of the WTC site - $9.0 billion; Repairing and replacing damaged infrastructure - $9.0 billion; Rebuilding the World Trade Center as smaller buildings - $6.7 billion; Repairing and restoring other damaged buildings - $5.3 billion; Lost rent of the destroyed buildings - $1.75 billion."
"There are all sorts of psychological biases that cause us to both misjudge risk and overreact to rare risks, but we can do better than that if we stop and think rationally."
Stop AND think rationally? If we were prone to either behaviour accurate risk assessment would never have been a problem. I submit the best we can hope for is to find work arounds for our hard wiring that can be implemented as needed and which will be effective from time to time.
I think comparing terrorism numbers to car accidents is not a good comparison, since the airline numbers didn't include fatal crashes that were not the result of terrorism.
However, I think it would be a good to compare airline terrorism numbers to something like deadly carjackings.
When talking about public policy and the best way to allocate limited resources, it's perfectly fair to compare car accidents to terrorism. The point is not whether a fatality is intentional. It's whether a fatality is likely enough to be worth the money you've spent trying to prevent it.
If we took the piles of money we've spent on anti-terrorism programs which have limited likelihood of success, and put those funds towards drunk driving checkpoints and treating alcohol addiction, we'd save far more lives. But we wouldn't get re-elected, because we'd be "soft on terror".
I don't consider it obvious that airplanes are more secure than cars. 37,000 people dying in automobile accidents is not that much considering that the number of hours spent on the roads probably is orders of magnitudes larger than the number of hours flown.
Just what I keep saying.
And if call correctly, people switching from planes to cars after 9/11 caused more additional traffic death than 9/11 itself.
Petter has a fair point; the article calculates terrorist incidents vs. distance, or vs. time. That might be a better comparison, to calculate car deaths vs. those two factors.
I'm always a little skeptical about arguments that include mind-boggling but completely irrelevant comparisons -- "This distance is equivalent to ... two round trips to Neptune." So?
@Rob & @Petter
You're pulling arguments from the "Which is safer (planes vs. cars)" debate which don't apply here. There is no reason to believe that more air traffic would result in more terrorist attacks, which is the only way that number of hours would be relevant.
IIRC, for it's year, 9/11 didn't even contributed to any the top 10 (or was it 20?) cause of death.
It's revelent to compare the risk of dying to a terrorist in an airline flight to the risk of dying in a car accident on the way to the airport.
If you're 10 times likely to die on the way to the airport, what possible excuse could there be for the ridiculous loss of freedom we're putting up with for post-9/11 airport "secutity"? I suspect the trip to the airport is, in fact, 10 times as dangerous as the terrorists.
Hmm I'm a little suspicious of some of the statments.
Take this for instance,
"Therefore, the odds of being on given departure which is the subject of a terrorist incident have been 1 in 10,408,947 over the past decade. By contrast, the odds of being struck by lightning in a given year are about 1 in 500,000. This means that you could board 20 flights per year and still be less likely to be the subject of an attempted terrorist attack than to be struck by lightning."
Does anyone else spot a problem with "you could board 20 flights per year"?
Let me make it more obvious,
~10,000,000 in 10 years -v- 500,000 in one year.
Now you can work either figure to normalise to either 1 year or 10 years.
However 0.5M/year becomes 0.05M/decade.
Thus 10M/0.05M = 200
That is the risk is about par at 200 flights per year not 20 flights per year which I'm sure "0.25M airmiles/year" Bruce feels more comfortable with 8)
Another "figures problem" with 9/11 is,
That might have been the number on the day but how many have died since?
That is due to medical complications due to the dust etc 9/11 caused at ground zero and the movment of it in the "clean up".
Or self inflicted harm either directly or indirectly by stress related bad habits (alcohol / cigeretts / pills / caffine / etc).
As far as I can tell the meter is still running on 9/11 and will do for the next 50 years or so...
@Clive: we can't count the deaths over the next 50 years, because doing so would be admitting just how absurdly successful bin Laden was! If there's one thing we're good at, it is not admitting things we don'like
"These departures flew a collective 69,415,786,000 miles."
I haven't been able to determine, is that plane miles or passenger miles? If it's passenger miles, then it also has to be divided by the number of passengers directly affected by a terrorist incident, and the 11bn number would be much lower.
@Clive - if you add those figures to terrorist attacks, you must also add them to all other causes of death - it is not only terrorism which leaves such scars, you get the same problems from road traffic accidents (RTAs), etc.
Better to ignore them for all deaths, as virtually every death has an emotional effect on the survivors.
The US has the highest car fatality rate of developed countries thanks mainly to the refusal to submit to the oppressive seatbelt
Risks of Airline security are being taken very seriously now, and perhaps for a good reason, if they reasons come from high up trusted people.
Group think gets me nervous, the old, something is going to prove the group think wrong, changer is coming.
But overall, airplanes are very safe. Terrorists surprisingly have not figured out the real fears in the USA, which are better left unsaid, unwritten, and keep the terrorists, working some bs.
PackagedBlue: isn't that security by obscurity?
Comparing car travel to air travel, air is about 3 times more dangerous per journey, car 4 times more dangerous per hour, car 60 times more dangerous per kilometre. (From which we can deduce that air journeys are on average 12 times longer (time) than car journeys, 15 times faster and 180 times the distance.)
Data allegedly from an article in 2000 in a railways magazine. I don't know what country/countries these data are from, nor whether the air statistics include general aviation or just airlines.
For amusement, see also
http://www.airtravelsurvival.com/pages/... which hypes up the 'per journey' comparison (they say 12 times more dangerous by air) to scare you into buying their book. (If air travel is 12 times more dangerous per journey, and I make 200 times more car journeys than air, how many 'car safety' books should I buy before I bother with their air safety one?)
So we find that there is a coin-toss chance of fatal terrorist hijacking on every return flight to Neptune. That's scary. Hope NASA and the DHS are doing something about it!
If you include non commercial flights, then even per trip flying is safer that cars when i last ran the numbers. There is reason for it. Even a minor incident is investigated and taken seriously. Recommendations made and changes made if needed.
If you are a pilot, they don't tolerate loose fliers like the movies suggest. We had a near miss when some guy joined the circuit to fast and too low, we were taking off. His license was revoked permanently.
If there is a car accident, unless someone is killed the cops only write it up enough for the insurance companies to take over.
@ Nathan Tuggy 10:37pm
As for "cheap" terrorists trying their efforts at airplanes, and if that is security by obscurity, ask the philosophers of security.
It might be simpler to call it security through misdirection, or active misdirection, or meaconing.
Either way, security theater works and is part of the game.
I've always speculated that the reason we don't react the same to auto fatality risk as we do to air travel fatality risk is the illusion of control. Deep down, every person believes that he, personally, is a better driver than the poor schmuck who just ended up under that 18-wheeler and would have been able to evade and escape.
In an airplane, however, you're just a passenger clearly at the mercy of the crew.
I'd be curious to see what the total cost of damages and collateral damages is between 1 terrorist attack on a plane where a death is involved versus the average car accident where a death is involved..
Wouldn't that be a better estimation for where money is better spent?
I have big disagreements with the way some security experts seem to toss off airline security by comparing the numbers to something more mundane or less extravagant that happens way more often (ie. car accidents).
Looking at terrorism caused airline incidents (or even airline incidents as a whole - a much better comparison) vs something else by quantifying numbers of fatalities is really short sighted. Some have already mentioned the economic effect of airline incidents - this definately needs to be considered. But I also have another perspective on the fatality statistics. Using the example of airline incidents vs car accidents, the issue for me (and as I see it, most other people) is not so much "is an incident going to happen"...the issue is "what is the result/impact of the incident occurring?" To explain further, consider this:
1) We can not ever be 100% secure (either in a plane or driving a car).
2) Knowing that, we recognise that somewhere, someday, an incident will happen (a plane incident or a car incident).
3) What is the typical result of the incident occurring? I have been in several automobile accidents in my lifetime so far, none of which resulted in any fatalities. Many other people have as well. How many people are out there that have survived their first aircraft incident? How about their second? Third?
My point is that there is a legitimate reason to be concerned with a terrorist (or other incident involving aircraft) more than an auto accident even though one is very common and the other is not. When the rare incident occurs (not if, when) - the passengers, people on the ground, etc - are really really screwed. Your chances are much better in a car, as there are tons of car accidents that do not result in fatalities every single day.
If you want to use stats, compare the number of fatalities as a result of an airplane incident vs the number of fatalities as a result of an automobile incident.
People wouldn't have such a problem with airline incidents if for example 100 airplane crashes occurred every day and the worst thing that happened in 90 of them is that everyone got a case of whiplash and a fight with their crappy insurance provider. When a commercial airliner gets blown up or crashes, the 200-something people on it are almost certainly going to die.
The last statement in this entry concerning the effect of "psychological bias" and "irrational thinking" is also a problem.
We are human beings and humans are irrational by nature. Part of being human is the need to satisfy at least some of those irrational tendencies. While I agree that "security theater" provides minimal protection (if any), some of it is necessary to calm the irrational human tendencies. Call it "securing the human psyche" or "emotional security" or something like that, but there is a place and need for it in combination with the things that will really provide effective security.
One thing that gets me is the fact that most of the fatalities were on the ground for Sept. 11. Should we be comparing time spent in buildings or always attach it to air travel? No one associates being in a building with a risk because there is no issue that can be addressed, but buildings- not airplanes were the targets.
We should work to reduce or eliminate the "no fly"list. It is well known that the "no fly" list has been used as a tool of political harassment. To avoid hurting the human rights and civil liberties of millions of people, to maintain a "no fly" list that is not politically oriented and does not hurt innocent people would require a vast amount of intelligence to determine who is a real threat to airplanes and who is not.
Airline security needs to become more concerned with WHAT is on an airplane, and away from WHO is on an airplane. Airline security needs to be about preventing opportunities for harm to airplanes from things that can be brought on board an airplane.
The example I like to use is that we can have 50 hard-core al-Qaeda terrorists on board ANY airplane, say an A330, with 200 other passangers, and the plane is safe if the plane has been made secure from opportunities for attack by terrorists on-board.
It's not WHO is on the plane, but WHAT is on the plane.
"we can't count the deaths over the next 50 years, because doing so would be admitting just how absurdly successful bin Laden was! If there's one thing we're good at, it is not admitting things we don'like"
@ Steve Parker,
"Better to ignore them for all deaths, as virtually every death has an emotional effect on the survivors."
Sadly you have both hit on "socio-political" points that make life easier for some (our political lords and masters and the lucky) and a lot harder for others (the rest who "get unlucky").
In reality we need to know how many people and what the cost/loss to society is to correctly assess risk and thus make sensible economic choices.
But the answers are as you two and others have observed supposadly "not to our liking" one way or another.
I guess the unlucky become another part of societies "hidden waste" much like, physicaly or mentaly disabled, those disabled as part of service to society such as war veterans/peace officers/fire fighters/paramedics, then the abused, runaways, homless, down to drug addicts, prostitutes etc.
Rather than square up to the issues we would rather ignore them or worse blaim them for their misfortune.
I guess I'm odd because I do care not just about the "waste" and "cost to society" but also because they are or were just like other people, now they are seen as the modern equivalent of leppers to be shut away or forgoton, and as has been said "there but for fortune go you or I"...
It has been pointed out that if you die "before average" you are either "stupid" (overweight, smoke, drink to much etc etc) or are "unlucky". But we are talking over 50% of the population dying before their time and over half of those through no real fault of their own the just got "unlucky" somehow.
I guess it say's something about our moral values as a modern society that soon the only real freedom we will have is how we will die before our time.
"... is the illusion of control. Deep down, every person believes that he, personally, is a better driver... ...In an airplane, however, you're just a passenger clearly at the mercy of the crew."
A thought normalise against "car driver fatalitie and pilot fatalities".
Then I think the results would be a bit of an eye opener...
@ Clive Robinson
"...how many have died since...due to medical complications due to the dust...or self inflicted harm either directly or indirectly by stress related bad habits...?
As far as I can tell the meter is still running on 9/11 and will do for the next 50 years or so..."
Clive, you’re obviously a compassionate and thoughtful man, but no way am I letting OBL take credit for an entire generation's inability to cope with life stress. To respond to terror with perpetual fear is a recipe for depression. Yeah, I watched his goons kill 3000 people live on CNN, but my eating habits, alcohol consumption, coffee intake, exercise regimen, and life skills were always and will always be my personal responsibility. The proper solution to risk is a measured response, not perfect prevention. I wear a seatbelt every time I'm in a car. I make sure the boat I’m in has enough lifejackets for everyone. I change the batteries in the smoke detectors every October. But I do not cower in the basement whimpering in fear of death by lightning strike every time I see a thunderstorm. If I'm ever "unlucky" enough to be on a plane with a man trying to detonate a wad of PETN in his boxers I hope I'm "lucky" enough to be seated next to him. The appropriate reaction to outrageous behaviour is righteous anger, not fear, or Valium.
@ Cerberus: "Deep down, every person believes that he, personally, is a better driver ..."
Quite a few studies confirm this. See http://preview.tinyurl.com/ycbtgxl [ScienceDirect .com]. Most people think they're above average drivers.
I don't see how flight time has any relevance to terrorist attacks. As far as distance and passenger numbers, they can have relevance for public policy, but they aren't meaningful to the average person. When someone gets on a plane, they're worried about something going wrong on that flight, not how many other passengers are present, how many miles they're flying, nor how long they'll be in the air.
One reason for this is that people think of plane crashes in which everyone dies. In fact, according to the NTSB, in cases of "accidents" (i.e., anything involving serious injury to a person or plane, including terrorism) in the past 25 years, only 10% involved fatalities, and 96% of the people aboard survived. ( http://www.ntsb.gov/aviation/Table5.htm ) Even counting just the flights that did have fatalities, about half had survivors. ( http://www.ntsb.gov/aviation/Paxfatal.htm )
Of course, only a tiny fraction of those were cases of terrorism, but fear outstrips logic every time.
Comparing odds with that of being struck by lightning as always annoyed me; Llike most everything, it depends.
The odds of getting hit by lightning, say, while golfing in Florida, or hiking down from the summit of a Colorado 14er while a storm rolls in, are actually quite good.. (1 in ~3000, i've read)
Averaged out over the whole US population, the chances are small.
I suppose such anecdotes, though, are really referring to the freak occurances, where someone gets hit by a strike out of the proverbial blue sky (from some storm 12-20 miles away ...), but even then, I would suppose those freak occurances number 1-2 orders of magnitude below the hikers, and golfers who get hit)
Foreigners killing Americans - ZOMG PANIC LOCK THEM ALL UP AND STRIP SEARCH THEM!
Americans killing Americans - eh. That's life.
Schneier.com is a personal website. Opinions expressed are not necessarily those of BT.