How lie detectors exacerbate border problems. By Matthew Yglesias

www.slowboring.com

How lie detectors exacerbate border problems

Matthew Yglesias

13 - 16 minutes

Probably the least important thing happening in the current immigration policy debate is that H.R. 2 — the border security bill that House Republicans wrote before they decided they didn’t want to work on any immigration reform — contains a provision requiring the Department of Homeland Security to hire additional polygraph examiners. On first read, that provision of the legislation tripped me up. I thought maybe they were concerned that asylum-claimants were lying and we could use polygraphs to smoke them out.

But that’s not it.

The issue, instead, is that before CBP can hire you for a law enforcement position you need to take a polygraph test. And, per this CBP recruiting video, if the results of your polygraph are inconclusive, you need to take a second test to try to pass.

The inclusion of this provision is an important little keyhole into the larger immigration debate, in part because it reveals the big lie at the center of the Republican Party’s positioning on the border. They would like to blame Joe Biden for problems at the southern border. And they think — or rather, Donald Trump thinks — that passing a bipartisan border security bill will make that harder to pull off. So now, on Trump’s orders, they need to pretend to believe that border security could be achieved by magic if only Joe Biden were sufficiently magical. The point about the polygraph provision in H.R. 2 is that this isn’t true. Border security involves actual material resources like border agents who need to be hired and trained. And there are various bottlenecks to hiring and training new border agents, including apparently a shortage of polygraph examiners.

Long story short, I think that Biden was too slow to come around on the need for substantive changes to asylum standards (I wrote about this in March 2021, in September 2022, and in July 2023), but it’s also the case that actual resources are needed. I’m glad that Senate negotiators brought the White House around to the correct conclusion here, and it’s infuriating that Trump and House Republicans now won’t take yes for an answer, but I do think this mostly serves to politically backfire on them as long as people in the media are clear about the clumsy political game they are playing.

But there’s also a specific question here. Why is the government relying so heavily on lie detector tests that most people seem to think don’t work?

As most people know, polygraph results are not admissible in most legal contexts. That’s because they are not considered scientific enough to constitute reliable evidence. Those of us who’ve seen “Anatomy of a Fall” know that in the French legal system you can apparently just introduce whatever kind of hearsay or gossip or passages from someone’s novel that you want (French lawyers confirm this is true) but in the United States that doesn’t fly.

You also might think that a reliable lie detector test would have enormous private sector value.

CBP wants to know whether or not its job candidates are lying in interviews, but don’t a lot of employers want this information? It turns out there is actually a federal law — the Employee Polygraph Protection Act — which bans most private sector employers from requiring polygraphs as a condition of employment. And yet this thing that’s banned in the private sector and considered unreliable by the judiciary, is mandatory at many public sector agencies including the CBP.

This is particularly a huge issue at the border, because as Elliott Spagat reported for the AP in 2017, about two-thirds of applicants seem to flunk the CBP polygraph screen.

The Border Patrol Union, which has become an influential force in right-wing politics, has been saying for years that this is a problem. So it’s mostly conservatives who’ve taken up the cause of polygraph reform. Representative Dan Crenshaw is a particular leader on this and he has a bill that “would remove the polygraph requirement for CBP applicants who have prior law enforcement experience or military security clearances.” This would expand some of the limited polygraph waivers the CBP currently offers, which are only specifically available for veterans with top-secret clearances.

And it really does seem odd that you have people flunking the CBP polygraph who’ve been cleared at other agencies. One could imagine a story whereby they just have an unusually low-quality applicant pool, but Spagat’s reporting definitely turned up stories of people who seem awfully legit but are getting bounced by the polygraph:

    David Kirk was a career Marine pilot with a top-secret security clearance and a record of flying classified missions. He was in the cockpit when President George W. Bush and Vice Presidents Dick Cheney and Joe Biden traveled around the nation’s capital by helicopter.

    With credentials like that, Kirk was stunned to fail a lie detector when he applied for a pilot’s job with U.S. Customs and Border Protection, which guards 6,000 miles of border with Mexico and Canada. After two contentious polygraph sessions that lasted a combined eight hours, Kirk said, he drove home “with my tail between my legs,” wondering how things had gone so wrong.

In the coding of partisan politics, it’s right-wingers and the border agent union that complain about the polygraphs while it’s progressives who worry about low standards and corruption at CBP. But the weird thing about this is that Crenshaw’s position only really makes sense if polygraph testing isn’t reliable (which is sort of what he says), and this turns out to be exactly the liberal position about polygraphs.

Crenshaw in his press release says “The polygraph is an interrogation tool. It’s useful for me as an intelligence officer when I want to vet an asset. It is not useful legally. It is thrown out in court. It is not useful for hiring purposes, and this bill aims to fix that.”

To the extent that liberals disagree with Crenshaw, it’s about whether or not the polygraph is useful as an interrogation tool. But to be clear, a big part of the premise here is that a lie detector test can be a useful interrogation tool even if it’s totally fake. There’s a famous scene in The Wire where a detective uses a clearly fraudulent lie detector on a suspect who has reason to believe is guilty. The point of the fake machine is to trick the suspect into believing that the Baltimore Police Department’s evidence is stronger than it really is, so that he confesses.

You can see why investigators might find this kind of tool useful, but as you’d expect people on the left tend to worry that this kind of thing is abusive. The ACLU, for example, says it’s an engine for racial bias. As a criminal justice issue, this sort of comes down to how much confidence you have in the police. In the hands of our heroes from our favorite television show, it’s a great investigative technique. But if we think the cops are bad, then we worry about abuses.

The politics flip when it comes to CBP’s internal use, and the conservative viewpoint is that abusive investigators are unduly sweating earnest border cops. Again, Spagat’s reporting is relevant here:

    But others, including lawmakers, union leaders and polygraph experts, contend that the use of lie detectors has gone awry and that many applicants are being subjected to unusually long and hostile interrogations, which some say can make people look deceptive even when they are telling the truth.

    Republican Sen. Jeff Flake of Arizona said he suspects CBP examiners fail applicants to justify their own jobs. He said he worries applicants are being wrongly branded with a “scarlet letter” in the eyes of other potential government employers.

The whole discourse in this space is redolent with ironies. But the bulk of expert opinion seems to be that polygraphs are highly unreliable and this has been the consensus for some time. A big National Research Council study in 2002 concluded that the use of polygraphs in routine security screening is bad. Steven Aftergood of the Federation of American Scientists notes that infamous spies like Aldrich Ames had no problem passing polygraph tests.

Perhaps more tellingly, to Aftergood’s point, the polygraph itself really doesn’t even say whether or not a given candidate is lying. It’s just lines on a page that the polygraph examiner then interprets — with different qualified experts giving different interpretations of the same lines:

    A DOE security contractor who administered a 1998 polygraph test to Los Alamos scientist Wen Ho Lee found him to be innocent of committing espionage, providing classified information to an unauthorized person, or intending to harm the United States. Unusually strong readings indicated an absence of deception. The test was reviewed by two additional polygraphers, who concurred with the finding. But then the Federal Bureau of Investigation (FBI) examined the very same data and concluded that Lee had failed the test.

The American Psychological Association is against polygraphs, here’s a Vox article about why polygraphs are bad, and there’s even a whole website — AntiPolygraph.org collecting information about why polygraphs are bad.

Given all the bad press and often vehement language with which polygraphs are discussed, you might be surprised to learn that studies show them to be accurate as much as 85 or even 90 percent of the time. But as even bullish takes on the efficacy of polygraphs, like this 2003 article in the MIT Technology Review, concedes that this means that using them as a random screening tool is going to uncover a lot of fake spies for Bayes’ Law reasons. Now in practice, the fairly high level of discretion that the polygraph process gives to polygraph examiners leads to results that differ a bit from what a crude statistical analysis would expect. But the fact that different examiners reach different conclusions is part of the problem here. The intelligence community starting using polygraphs in the 1950s to reassure people that it wasn’t being penetrated by spies, and that’s mostly what it shows. But CBP started doing polygraphs much more recently, as a result of a 2010 law that passed in response to corruption concerns. So they seem to be much more active at finding people to flunk.

On the narrow question at hand here, I think that rather than hiring more polygraph examiners to address the hiring bottleneck, immigration hawks and civil libertarians should join hands to admit that polygraphs are kinda fake and we don’t need to be making it harder to hire border agents.

That being said, the corruption concerns that motivated the polygraph testing are not fake.

There is a lot of money to be made by smuggling people or contraband (usually drugs) into the United States of America. People are willing to run significant risks to their personal safety to reap the rewards of illegal cross-border traffic, and they are also willing to invest time and resources in things like digging tunnels and cutting through segments of border wall. It’s natural that under the circumstances there is also an interest in bribing border agents. Unfortunately, a lot of anti-corruption efforts — certainly including the emphasis on polygraph screening — seem to be motivated by the very specific concern that Mexican drug cartels will recruit people to go become border agents who then work on their behalf.

That kind of scenario, basically a southwestern version of “The Departed,” is certainly something that could happen. But in CBP’s own review of 120 major corruption cases, they only found five instances of infiltration. The most common scenario, in 56 cases, was for an employee to “self-initiate” corrupt behavior — offering illicit services to criminals rather than being recruited by criminals.

One issue the review notes is that corruption seems particularly common in “hometown” assignments where an agent is working close to where he grew up, in a community where he has wide-ranging contacts.

If you sent all border agents to far-off postings, that could be a potentially very effective anti-corruption measure. But realistically, if you tried to do that without making any other changes, recruiting would collapse. Lots of people from South Texas who join CBP wouldn’t want the job if it meant a posting to Buffalo. And if you had to try to exclusively recruit people from out of the region to go move to the Southwest border region, you’d be drawing on a much thinner pool of people — especially if you were trying to get a critical mass of people with Spanish-language skills.

Now of course you could pay a lot more. But not only would that cost money, you might end up in effect just draining people away from other law enforcement agencies. Is it better for America if a given person decides to be a border agent rather than work for the FBI or as a homicide cop in Chicago? That’s not obvious to me, though mostly I’m not sure I’ve ever seen anyone even try to estimate the relative value of different kinds of law enforcement roles.

So now we’re back where we started. There’s a rhetorical construct around “the border” whereby presidents could just secure it with willpower. But even if you go to a really low-crime state like Maine, there are still people getting murdered every year. There was a bank robbery in South Portland last week. Trying to get people to not rob banks is a worthwhile (and I would say fairly uncontroversial) policy objective. But it’s also a non-trivial problem that involves material resources and competing priorities. The border is at least in part like that. It would be nice to have more border agents and a higher standard for recruiting and conduct. Those are certainly achievable goals. But it would take money and it would take leadership and various other things. There isn’t a magic force field that a president can just turn on or off.