Counterterrorism Blog

My latest article in the Daily Standard, "Spare No Resource," was in large part a response to a San Francisco Chronicle op-ed by Mike German.  German's op-ed argued that racial profiling is an ineffective tool in anti-terror policing, and used statistical analysis to try to prove this point.  In response, I critiqued German's use of statistics and pointed out that law enforcement can use a number of factors (not just race but also age, gender, dress and behavior) to identify the most likely terrorists.  I made the intuitive argument that "[i]f our last line of defense is searching bags before riders enter the subway, our searches should target the passengers who are most likely to be terrorists.  Only through intelligently targeted searches can we have a reasonable chance of disrupting terrorist plots."

I'm on friendly terms with German, and told him that if he wanted to write a response, I'd post it on the Counterterrorism Blog.  Here's German's response to my article, along with my reaction:

The sub-title of Daveed Gartenstein-Ross�s Weekly Standard article �Spare No Resources� promises much:  �Terrorist profiling is the most efficient, and effective, method of anti-terrorism policing.�  Unfortunately, Gartenstein-Ross never makes any attempt to support this statement with facts, and instead only re-states his thesis and declares it �intuitive and hard to refute.�  If only that were true.

The difficulty with terrorism is that much of it is counter-intuitive.  In reading Gartenstein-Ross�s article I was reminded of a scene from the Steve Martin comedy The Jerk, in which an assassin tries to shoot Martin with a high-powered rifle.  The initial shots miss Martin and hit some oil cans stacked behind him.  �These cans,� Martin yells, �he hates these cans!� �The subways,� Gartenstein-Ross yells, �jihadists hate the subways!�  Just like Martin, Gartenstein-Ross mistakes the result of the enemy�s action for the objective.  The objective of a terrorist campaign is not to commit an act of terrorism, but rather to force the government they attack into adopting a defensive position that is both expensive (because it is inefficient) and demoralizing (because it is ineffective).

Gartenstein-Ross argues, �The question�is who's targeting New York's subway system right now.  The answer is, of course, jihadist groups.�  As it turns out, of course, nobody was targeting the New York subway.  It was a hoax.  Just like the apparent hoax that shut down the Baltimore Harbor Tunnel this week.  But the response was real� and really expensive, both in terms of actual dollars spent to temporarily increase security and in terms of the public�s loss of confidence in the government.  This misspent effort is the terrorist�s real objective.

Gartenstein-Ross spends most of his article criticizing a piece I wrote for the San Francisco Chronicle entitled, �Racial Profiling no tool in thwarting terrorism.� Gartenstein-Ross misstates my argument though, and while that makes it much easier to refute, it doesn�t really add to the discourse.  I did not argue that �random bag checks are just as effective as profiling,� as Gartenstein-Ross states, but rather that profiling is just as ineffective as random bag checks.  Protective security in the subway is no more effective as a counterterrorism strategy than covering the oil cans would be for Steve Martin.  It�s not about the cans, and it�s not about the subway.  Would New Yorkers feel any better if a terrorist blew up people in the street instead of in the subway?

The government uses protective security procedures like random checks and profiling to give the public the perception that they are being protected, but the truth is these methods are neither effective nor efficient.  A patient terrorist can easily find a way around security procedures, much as the 9/11 hijackers used small knives to circumvent the airline security procedures in place at the time.  Start profiling men and they�ll use women (don�t worry I�m not giving them ideas- watch The Battle of Algiers sometime).

An impatient terrorist could simply detonate his explosive at the checkpoint, accomplishing his mission by killing the police officers and anyone else waiting in line.  Or, just as likely, he�d go find a softer target.  Terrorists don�t attack subways because they don�t like underground transit systems.  They attack subways because they�re difficult (and expensive) to protect.  Spend the money to protect the subways and they�ll attack the buses.  Protect the buses and they�ll attack shopping malls.  Carlos Marighella laid out the strategy of a terrorist campaign decades ago in his Mini-manual of the Urban Guerrilla:  �The government has no alternative except to intensify its repression.  The police networks, house searches, the arrest of suspects and innocent persons, and the closing off of streets make life in the city unbearable�In spite of all this, the police systematically fail.�  This is the game terrorists want us to play, and unfortunately, it�s the game we are playing.

Remember after 9/11 when everyone thought we needed to have a domestic intelligence agency like the United Kingdom�s MI-5?  Remember how security experts marveled at London�s vast network of security cameras set up to let those intelligence agents see what�s going on anywhere in the City?  Well neither MI-5 nor their high-tech security cameras prevented the 7/7 subway attacks, and, even more tellingly, they failed to stop an almost identical attack two weeks later.  Yet even after the London attacks New York City pursued a $200 million dollar contract to outfit the New York subway with security cameras.

Gartenstein-Ross suggests we have a wealth of data on how terrorist look and how they act and we can use this data to create a terrorist profile that will help us identify the terrorists before they act.  A lot of people have been spending a lot of money trying to find this magic bullet.  I hope they succeed, but it�s unlikely.  The problem is that someone comparing every detail about the diverse lives of Mohammed Atta, Germaine Lindsay, and Kevin James will find more differences than similarities.  James, in case the name is not familiar, is the convicted felon who�s accused of hatching a home-grown terror plot from inside a California State prison.  The alleged plot, arguably the most serious jihadist plot on our homeland since 9/11, wasn�t uncovered by our multi-billion dollar counter-terrorism intelligence program, or by security systems that must be ubiquitous in a prison, but rather by good police work from the Torrance California Police Department, who stumbled into the plot during a routine robbery investigation.

Rather than spend money searching for a terrorist profile, we should spend money searching for terrorists.  Instead of spending $200 million on cameras let�s give it to the police officers investigating real threats to their communities.  MI-5 came under criticism when it was learned that 7/7 bomber Mohammad Sidique Khan had been �indirectly linked� to a previous terrorist plot uncovered the year before.  Khan was not put under surveillance because, according to an unnamed British government official quoted in the Sunday Times, MI-5 was constrained by �finite resources.�   It turns out high-tech security cameras are even more expensive than we thought.

Unfortunately, this response highlights one of German's bad habits as a writer, which is to embark on tangent after tangent, often losing sight of the main point.  My article argued that German's statistical analysis did not in fact prove the inefficacy of profiling.  Rather than defending that analysis, German builds a whole new case that includes the argument that jihadists are not actually targeting the New York subways, that mobilizing police in response to terrorist threats is expensive, that terrorists will shift their attack strategy in order to circumvent security procedures, that the U.S. does not need a domestic intelligence agency like Britain's MI-5 . . . you get the point.

Moving to German's argument, he offers a red herring when he argues that the last subway terror scare was a hoax.  I never argued that it was genuine.  Instead, I pointed out that "the subway terror alert highlighted two basic facts.  The first is that the terrorists would like to strike our mass transit system; the second, is that this system is still highly vulnerable."  The plain meaning of my statement that these facts were "highlighted" is that attention was drawn to them -- not that the terror threat was real in that particular instance.  The fact is that the terrorists do want to hit our mass transit system.  Since 9/11 there have been major Islamist attacks on London's underground, Madrid's passenger trains and Moscow's subways.  Not only is there no reason to believe that the U.S. is an exception, but there's also hard evidence suggesting that we're not:  Two men were arrested last summer in an alleged plot to blow up a New York subway station.  Lesson:  If you want to make a tangential point, at least make sure that you're right.

(Incidentally, while German bizarrely accuses me of fixating on the subways, he's far guiltier of this -- for example, by arguing that if we protect the subways, the terrorists can hit our buses and shopping malls.  My article discusses the subways because we know terrorists have targeted them and because the amount of civilian traffic makes them hard to protect, but the solution I offer is by no means narrowly circumscribed to subways.  The same problem exists for buses, shopping malls, and other soft targets.  Terrorist profiling can protect these other targets as well.)

German's point about the cost of responding to terrorist alerts is a non-sequitur.  To the extent we disagree here (and I'm not sure we do), it's not relevant to whether we should have a system of terrorist profiling.

What does that leave German with?  Two arguments:  that searching for terrorists is a better use of our money than policing soft targets and that terrorists can adapt to our anti-terror policing measures.  The first of these arguments is also a non-sequitur, as it's only a question of resource allocation.  In a previous article on terrorist profiling I acknowledged that terrorism investigations should take precedence over policing soft targets:  "Solid intelligence, arrests and prosecutions are a cornerstone of protecting U.S. citizens."  But we should also police soft targets, since we're highly unlikely to be able to unravel every terrorist plot before it gets underway.  Even if we devote more resources to intelligence-gathering and arrests, we should still have a last line of defense

That leaves German with the argument that terrorists will adapt to our policing measures.  There are two fatal flaws with this argument.  First, German falsely suggests that we're up against an omnipotent foe capable of evading every defense we erect:  "Start profiling men and they'll use women . . . .  Spend the money to protect the subways and they'll attack the buses.  Protect the buses and they'll attack shopping malls."  The fact is that the terrorists are far from omnipotent.  Just as our resources are finite, so too are the terrorists':  They have limited financial resources and a limited number of operatives.  If we arrest terrorists who are trying to hit the subway, it won't be easy for them to come back and hit the malls the next day.  And if our profiling effectively targets some of their operatives, it won't be easy for them simply to switch over to women.  If some of the most well-trained and effective terrorists cannot do their job because of our profiling, that makes a successful terrorist attack that much harder.  The new operatives they switch to may be less well trained, and hence less deadly.

The second fatal flaw with German's argument is that he's correct that terrorists adapt to anti-terror policing measures.  But if we don't use profiling, we're left with static policing against an adaptive opponent.  Random bag searches, by nature, aren't adaptive since they don't distinguish between Al Gore and Mohamed Atta.  Terrorist profiling, on the other hand, is highly adaptive.  A profile can be adjusted based on our latest intelligence.  If we know that the terrorists have recruited more women, we can search more women.  If they've moved from the subways to the buses, we can shift our officers there.  While we can't certify the subways or any other terrorist target 100% safe, profiling can move us in the right direction by concentrating policing resources on the likeliest terrorists.

The bottom line is that in addition to arrests and prosecutions, we'll need to police soft targets.  Thus, we should police as effectively as possible.  Our anti-terror policing can improve if we don't remain wedded to the ridiculous system of purely random searches.