The News

↑ Grab this Headline Animator

14 January 2010

Another Perspective on the Risk of Airline Bombings

Since 2001, terror events, including attempted bombings, have been a serious concern for passengers and governments around the world. While high profile terror events of all kinds have been a concern, the probability of a bombing event involving a US airliner or a flight departing or arriving from the US hasn't seen any major changes since 1960.

Terror Events in the Previous Decade
Nate Silver, better know as the person behind the political polling site, ran an article that used data from the Bureau of Transportation Statistics (BTS) to evaluate the risk of a terrorist attack on airliners that either departed from the US or had a US destination.

Looking over the ten year period from 2000-2009, Silver identified six attempted terror-related events on board these airline flights, including the four planes hijacked on 9/11, the December 2001 attempted bombing of an American Airlines 767 by convicted shoe bomber Richard Reid, and the January 2009 attempted bombing of a Northwest Airlines A330 by alleged underwear bomber Umar Farouk Abdulmutallab.

Odds of Being a Terror Victim
Silver reviewed BTS records and estimated that there were 99.3 million commercial airline departures that either originated or landed within the United States, or one terrorist incident per 16.6 million departures.

He further estimated that excluding crew members, bombers and hijackers, there were 674 passengers on those six flights on which these incidents occurred, compared with seven million estimated passenger enplanements during that decade, giving the probability of being on a given departure which is the subject of a terrorist incident of about 1 in 10.4 million. He contrasts this risk with the 1 in 500,000 chance of being struck by lightning.

Odds of Being a Bombing Victim
If you only look at the bombing events in there previous decade, there were two, including last month's attempt on Northwest Airlines and the 2001 show bombing event. Using Nate Silver's numbers, the rate of bombing events on US airliners from 2000 to 2009 was 1 in 31.2 million departures.

Previous US Bombing Events
Nate Sliver only looked at events during the previous decade. has tracked US bombing attempts going back to the 1950s, and found that in the four decades from 1960 to 1999, there were eight events on US airliners that involved a bombing attempt or a successful bombing, with seven involving passenger fatalities. This averaged two per decade, compared with the six in the decades of the 1990s. However, the average number of flights was lower during this 40-year period. According to BTS data, there were about 311.6 million flights during this four decade period, which means the rate of bombing attempts from 1960-1999 was 1 per 39 million flights, compared to 1 per 31.2 million flights for 2000-2009.

How These Two Periods Compare
One of the things that stand out with the events from 1960-1999 was that only six of the bombing events involved a deliberate attempt by individuals or groups that could be associated with terrorism. If only those six are counted, then the rate for the 1960-1999 period drops to 1 in 59.1 million flights, roughly half the risk of the 2000-2009 period.

Does this mean that the risk of bomb related terrorism has doubled in the previous decade? Maybe. However, if the alleged underwear bomber had waited another week and attempted a bombing on January 1st, the 2000-2009 rate would have been 1 in 64.4 million flights, even less than the terror-related bombing rate for 1960-1999.

1 comment:

  1. Thanks for the number crunching; I will definitely re-use the example/numbers.

    The (aviation) security world is driven by the so-called risk-paradigm. Unfortunately, there is a tendency to use 'standard' risk techniques without an assessment of their applicability. E.g. over the recent years more or less all AVSEC players work with a typical risk-chart plotting impact vs likelihood (probability).

    A key problem is that humans are bad in estimating risk-likelihoods. There is plenty of literature on biases etc. Typically, common (security) risks are underestimated, whereas 'extreme events' - such as 9/11 - are overestimated. The article further demonstrates that making use of historical data adds statistical questions and challenges the usage or added value of statistical analysis of historical events.

    I personally favour an asset-centric view. This helps me to articulate risk as a combination of criticality, threat (conditions) and vulnerabilities. This way I am able to overcome the likelihood-related shortcomings of the classical - typically attacker-centric - perspective of aviation security risks.