Let Markets Help Criminal Defendants
posted by Dave Hoffman
Dan’s interesting post on plea bargaining made explicit the informational inequalities faced by criminal defendants and their lawyers. Indeed, one of the advantages public defenders have over private defense counsel is that they can more easily share information internally about the informal norms that “really” govern the system (judge sentencing practices; which cops tell what stories; which labs are sloppy; etc.) But even so, the instruments of law and order almost always will know more about the facts and the law than the defense, at least until the eve of trial and probably throughout the process.
That there are exceptions (Enron; OJ) proves the rule that informational asymmetry is a significant part of the prosecutor’s arsenal – indeed, this asymmetry justifies constitutional attempts to remedy the problem through mandatory discovery procedures. But I’m skeptical that legal rules alone are a panacea to structural problems. Why not try markets?
To be more concrete, the major decision that criminal defendants face is whether or not to plead guilty. The decision depends on a prediction about what will happen at trial. Assuming that defendants are risk averse, they will take pleas when rational actors would not, but generally will go to trial when the expected time served post-trial is less (by some margin) than the actual time proposed in the plea agreement. The problem is that (1) defendants are unsophisticated; (2) defendants’ lawyers are incented to push pleas; and (3) neither defendants nor their lawyers have as much information as prosecutors about likely verdicts.
If I were running a public defender service, I’d consider setting up an online prediction market for the conviction of my clients. Prediction markets did a fantastic job in the Enron trial. At the beginning of the trial, the odds of conviction were about 50% for each defendant; by the end, the odds were significantly higher. Now, I can understand why neither defendant would have pled facing a coin-flip’s chance at conviction. As I argued at the beginning of the trial:
I’d guess that the reason Skilling and Lay have not pled and Fastow has is demographics. Fastow is a young(ish) man, who can serve significant time and still emerge with earning power. Lay and Skilling don’t have the years left to do the time that the government (apparently) would find appropriate.
But for most criminal defendants, 50% odds would translate into a pretty hefty expected sentence that might make a plea more attractive. And, assuming that such markets would be sufficiently liquid, the predictions generated by traders ought to be both more accurate and less prone to bias than defense counsel’s odds. I imagine that the result would be a net decrease in pleas, and in the long term, as prosecutors reacted, less net jail time. That is, the current system is biased by risk aversion and agency problems – as others have observed – toward more jail. This effect may serve the forces of law and order, but it doesn’t necessarily serve the search for truth. Why not try something different?
Obvious objections: (1) the idea is “”utterly repugnant to a civilized society“; (2) thin markets are prone to manipulation; (3) incentives would increase to violate the attorney-client privilege; (4) it would look like public defenders are selling out their clients. Of these objections, I’d be most worried about #3.
Incidentally, if you are interested in thinking more about criminal law and the Enron trial, the Conglomerate is hosting what promises to be a great forum on the topic for the next two days. Check it out!