I'm having a little trouble finding time to keep up the conversation, but Scott Greenfield at Simple Justice has responded to my speculative post on why clients might prefer to hire a lawyer who refuses to represent snitches. He says some nice things about me, but his response makes it clear that I get way too much of my information about the criminal justice system from television shows.

My basic theory was that one if the reasons a client might hire a non-snitching criminal lawyer is to signal to a third party that he is not planning to snitch. The third party has to be someone who (a) doesn't want the client to snitch, and (b) is in a position to harm the client if he suspects snitching.

In casting about for that third party, I settled on fellow gang members who might want to kill him to prevent him from snitching on them. By hiring a lawyer who doesn't represent snitches, the client would be insuring his own safety by making it clear that he's not taking that route to freedom.

Greenfield doesn't dismiss my made-for-TV theory outright, but he points out a third party that is somewhat less dramatic, but who should have been much more obvious: Co-defendents. They clearly don't want the client to snitch, and they can harm the client simply by snitching on him first.

This is an example of what economists---borrowing from game theory---call defection. In the standard scenario, if a bunch of soldiers are in a battle, any individual soldier can improve his chances of surviving by running away. But if all the soldiers run away, they will be routed by the enemy and lose badly.

A more prosaic example is littering along the highway. If I throw my empty fast-food containers out the window of my car, I will have a clean car, and the small amount of unsightly litter along all those miles of highway will cause me little grief. So, on balance, I can improve my life by littering. But if everybody litters, we're all a lot worse off.

The co-defendent situation is a little different, however, in that a single defector can do a lot more damage. One fleeing soldier won't lose the war, and a single litterbug won't destroy the environment, but a single talkative co-defendent can really hurt everyone else. This sort of situation has a rather appropriate name in game theory: It's called the prisoner's dilemma.

The defining characteristic of the prisoner's dilemma is that everyone is worse off when defections occur (because the prosecutor will have an easier time getting convictions) but that the best choice for any individual person is to defect regardless of what everyone else does (because the prosecutor will always go easier on someone who flips than on someone he has to convict at trial).

This particular situation differs from the classic prisoner's dilemma in one important way that makes it especially harrowing for the people caught up in it: The order of defection matters. The first person to snitch can break the case wide open for the prosecutor by giving up everybody else. The second snitch has less to give up (the first snitch already being in the bag) and can only provide corroborating testimony. It gets worse for each successive snitch, until the last one, who has nobody else to give up.

This makes the situation devastatingly brittle for the defendents. If any defendent suspects any other defendent is about to snitch, it's to his advantage to snitch first. This could cause a defense strategy to collapse all at once as everybody races to defect. In fact, while I was writing my previous post on this subject, Greenfield posted a good description of exactly this dynamic in a group of 26 defendents.

I don't know how criminal cases work when there are multiple defendents, but I'm guessing that non-snitching lawyers are only one of the safeguards against what Greenfield calls "the foot-race to cooperate."

Greenfield also discusses some other ways non-snitching lawyers affect the case, and he also addresses my prediction that non-snitching lawyers will appeal more to gangsters than white-collar criminals because gangsters have more to fear. He thinks I'm right, but for the wrong reason. It's worth reading his post to find out why. 

On a totally different subject, Mark Bennett asks in a comment:

What does economics have to say about people who choose a lawyer who reflects their own values or their own self-image?

I'm no expert, but I think that's a subject outside the bounds of economics.  There might be curiousity about the magnitude of the client's preference---as indicated by the trade-offs he's willing to make---but unless the client is trying to advertise his values to someone else ("I don't turn on my friends when the going gets tough"), this is a matter of the client's taste in lawyers, and economists tend to treat taste as a given.