Thursday, May 13, 2010

Illegal immigration, grand theft auto, and racial profiling

Here is a comment from one of the comment box, which is something one typically hears in the 1070 debate.

Racial profiling is a red herring; if 99% of the illegal aliens in a state are hispanic, then hispanics are going to be targeted in any enforcement of immigration law.

I don't see that this follows. Suppose 70% of car stolen in the Phoenix area are stolen by Hispanics. It doesn't follow that the police, in tracking down car theives, need to consider ethnic status in their procedures for finding car theives. 70% percent of the car theives they catch them will be Hispanic. That doesn't mean that profiling or anything like profiling was involved.

10 comments:

Anonymous said...

That doesn't mean that profiling or anything like profiling was involved.

Which is precisely the point of the comment. Racial profiling is a red herring brought up by liberals who just don't want the laws enforced, period. You seem to agree with the sensible position that disparate impact doesn't imply disparate application, but I'm sad to say that you are in the minority among liberals on that one.

Blue Devil Knight said...

Anon is wrong to say that profiling is a red herring used by those that don't want the law enforced. I am fine with the law being enforced, but think racial profiling (in most cases) is not a very American thing to do. America likes its laws to err on the side of liberty, not intrusion/enforcement. So there is a libertarian case to be made against racial profiling that has nothing to do with the immigration topic.

My original thought was that the focus should be on
P(illegal|hispanic)
rather than
P(hispanic|illegal)
as relevent for determining the proper course of action when the legal status is unknown.

The first is quite low, the second is presumably quite high. But can't we say that only the first, if high, would provide a justification for using racial criteria?

That was my first thought, but then of course the two probabilities are connected.

From probability theory, we know:
P(I,H)=P(I|H)xP(H)=P(H|I)xP(I)

Which shows the two probabilities from above are connected:
P(I|H)=[P(H|I)xP(I)]/P(H)

So I guess the question is, what is our threshold probability that would justify profiling? And note this assumes we should always enforce the law, doesnt' incorporate relative values of different laws being enforced. I'd rather have them catching murderers and rapists than illegal immigrants: what is the worth of catching the immigrants?

That wrinkle, which is crucial, makes things very complicated. Why aren't people complaining that cops aren't putting more effort into catching women wearing pants (which is still technically illegal in Tuscon, I believe).

I'm not saying illegal immigration is that trivial, of course, but just pointing out that getting all hot and bothered about this law implicitly assumes it is an extremely important law to enforce, that it should be a higher priority than other laws.

Clearly, it should be higher priority than some laws (e.g., the women-pants law), but things are not so cut and dried.

At any rate, it would be cool to see a fully worked out analysis of all the probabilities involved, and the relative costs/benefits. From a purely fiscal perspective, even, it would be interesting. I know that illegal immigrants are said to drain the health care/education systems, for instance.

Victor Reppert said...

A complete cost-benefit analysis, but of the process of illegals entering the country now, and the presence of illegal immigrants in the country now (and these are two different questions, surely), is not available to me, but I do know that the economic benefits of illegal immigrants is does exist. The movie "A Day Without a Mexican" underscored this, and in 9500 Liberty (required viewing for anybody who wants to think clearly about this issue), Prince William County suffered a foreclosure rate far in excess of neighboring counties, because of its 1070-like law.

This issue seems to me as complex as most problems in metaphysics.

DL said...

There has to be some level of justification for the police to be able to investigate a suspect. If you're allowed to collect evidence only when you already know a person is guilty, there could be no law enforcement. I wonder, is everyone who's so worried about "profiling" also against the police using sketch artists?

Blue Devil Knight said...

DL: they aren't the same thing. One is a description of a specific suspect, the other is general suspicion of people that look/speak a certain way. Police don't stop people randomly, they must have reasonable cause (for thinking the person is guilty). Someone smelling Mexican should not be reasonable cause. Police can't just come into your house if they have a "hunch" that you are a murderer.

I started out sympathetic to the law, but the arguments I'm seeing for it are actually pretty unamerican seeming to me. When hunches replace reasonable cause, we have hurt America. When skin color or language becomes evidence of guilt, something is fucked up.

Note, this is different from when at location X, a 'black male aged xyz, height q' was seen running from a murder. Again, that's a specific person associated with a specific crime. With this law, there is no specific person.

And I'm not convinced this issue is as important as other issues the state could put resources into. E.g., going after the companies that hire illegal immigrants. Does the cost-benefit analysis really justify all this hubbub?

Blue Devil Knight said...

On the other hand, I'm not saying police shouldn't be allowed to enforce this. What if someone comes to a cop and is like "I'm here illegally but you can't do anything". I'd hope the cop would have the authority to be like "I'm gonna send you back to Mehico, beaatch!"

I really think Victor's (and others') idea is best: make everyone show the id in the same situations. Don't leave it to police discretion. (Obviously it will be up to their discretion anyway, but at least this way the law is moral).

Blue Devil Knight said...

On the third hand, I do find it frustrating when little old white grandma in a wheelchair gets the full search at an airport. So I'm probably not being consistent. Or perhaps I am, but the cost-benefit analysis for terrorism is so much more important (by my lights), and we literally do not search everyone but a small random subset, that perhaps some profiling is justified.


That's it, I'm not inconsistent, just my utility function is different in the two cases. :O

DL said...

BDK: they aren't the same thing.

But aren't they — at least from a philosophical viewpoint? Facial profiling might be more specific (though any police sketch I've seen is a long way from a photograph), but surely that's just a matter of degree, not of kind. You still risk bothering innocent people because they might happen to "look like" a criminal. Surely the ethical problem with profiling is that you shouldn't harass innocent people, but since that is a necessary evil if police are to do anything, then anything we can do to cut down on that number is a moral advantage. In other words, to whatever extent profiling is bad, not profiling must be even worse.

Police can't just come into your house if they have a "hunch" that you are a murderer.

Yes, certainly; there have to be limits on what police can do that's in proportion to how justified their suspicions are. (Such as the considerations a judge might weigh in deciding on a warrant, etc.) But I don't think there's any a priori reason why race, creed, color, or anything else should be excluded. If it is deemed not worthwhile or reasonable in a particular case, that's fine, of course.

Again, that's a specific person associated with a specific crime. With this law, there is no specific person.

Hm, that's a good distinction. I'm not sure it's sufficient, though: when cops stop random cars to check for drunk drivers, there's no specific crime, but that doesn't seem so objectionable. (Of course, that's another form of profiling, esp. when done on certain dates, like New Year's Eve!) Again, the moral problems seem to come down to pragmatics or effectiveness rather than the profiling itself.

Does the cost-benefit analysis really justify all this hubbub?

No idea! (I doubt anyone's worked it out, though on the other hand I don't know how reliable any such calculation could be....) There may be a hundred other reasons why it's a bad or pointless law. I'm just curious what philosophical reason there could be against profiling (beyond it's got "race" in the name so it must be politically incorrect!). There may even be practical reasons not to use profiling (for example, because your police force is full of bigots, or because the mere appearance would cause a societal backlash, etc.), but those still aren't a problem with the actual profiling itself (and we would hope that the root problem gets dealt with, not just the symptoms). (In this particular case, yes, having everyone show ID seems rather obvious....)

I do find it frustrating when little old white grandma in a wheelchair gets the full search at an airport.

Yeah, exactly. If profiling were inherently immoral, then no ends would justify it, but since I don't see that it is, it just needs to be factored in accordingly. It may be less specific than an eyewitness sketch, or whatever, but that means it should be weighed accordingly, just as a partial description will carry less weight than a full one, and so on. Not that I have much trust in airport screening to begin with, but if profiling actually did help catch terrorists, then that would surely be a good thing.

Victor Reppert said...

I think you have to recognize a couple of different levels of profiling.

It is also important to assess the priority level of immigration enforcement? Enforcement of any law requires some collateral damage, but you have to ask what collateral damage we can inflict to get some particular law enforced. Jaywalking and murder are both illegal, but given a choice, where should law enforcement's emphasis go?

Consider high-speed chases. Here we are talking about pursuing serious criminals. Surely, the cops have the right to exceed the speed limit in order to catch someone who has committed a crime. But high-speed chases put innocent people at risk, so we have to question how far law enforcement should go in engaging a criminal on a high-speed chase in order to cathc and offender. You can't just ask "Well, don't you want the law enforced," and settle the issue that way.

I keep emphasizing the movie 9500 Liberty, but nobody seems to have bothered to respond to it, or even watch it. In Prince William County, Virginia, it looked pretty clear to me that the costs outweighed the benefits. On the face of things, I don't see why Arizona would be any different.

DL said...

VR: you have to ask what collateral damage we can inflict to get some particular law enforced.

Right, when it comes to practical application, it has to be proportional, prudent, effective, etc. Nobody's going to deny that.

In Prince William County, Virginia, it looked pretty clear to me that the costs outweighed the benefits.

Ah, so your argument really is about the costs outweighing the benefits. OK, that's always a possibility....