A good strike

Back in September, I warned against a three-strikes law that eliminated marginal deterrence. It looks like ACT and National have come up with a compromise that sounds pretty sensible. There's a list of serious "strikeable" offenses. On the first strike, the offender is warned that it's his first strike, with sentencing as normal. On the second strike, "truth in sentencing": no parole period applies, so the offender serves the full sentence. On the third strike, the judge must assign the maximum sentence, again with no parole. Since the only penalty for murder is life in prison, murder as a second strike gets the same treatment as murder on a third strike.

This makes much more sense than ACT's initially proposed "twenty five years for a third strike" rule, which of course eliminated marginal deterrence across third strike offences. There may still be a bit of a severity shift within categories: offences within the same category can be more or less serious, but the losses from this kind of shift are pretty small relative to the losses from a blanket "25 years" policy. They also ought to be small relative to the deterrent effect achieved.

Recall that California's three-strikes law deters even first offences on the strike list. The NZ law seems to avoid the biggest problems of the California legislation: there won't be any serious severity shift the way it's written, and the increase in prison costs will be rather smaller than that experienced in California where any third strike draws a fixed and long sentence rather than one proportionate to the offence committed.

Note that strikes only count going forward; this also is pretty reasonable.

Good job ACT and National. I'll be interested in seeing the full list of strikeable offences and the rest of the details, but this gets a big provisional thumbs up.

Three strikes law for NZ?

The ACT party put forward a bill proposing a three-strikes rule that, in its initial form, eliminated marginal deterrence at the third strike: 25 year sentence on third strike regardless of the offense. Of course, we know this can induce a severity shift among offenses on the strike list and provides little inducement for criminals to leave witnesses hanging about.

Kiwiblog reports that negotiations between ACT and National has led to a bill that National would support at second reading which would require judges to impose the maximum penalty on a third strike. Where Farrar emphasizes the fairness aspects of not having relatively minor offenses draw a 25-year sentence, I'm far more interested in the maintenance of marginal deterrence. Having the third strike draw the maximum sentence for that offense does a pretty good job: there's no reason to expect a severity shift among offenses since they all still draw differential penalties. Perhaps we could expect such a shift within a particular offense, but that can't have too large an effect.

Joanna Shepherd's work shows rather convincingly that strike-type legislation in California was remarkably successful in deterring even first strikes. The approach here proposed would avoid some of the worse potential consequences of the California law (severity shift, excessive prison costs) while keeping a decent chunk of the benefits.

We'll have to wait for the final bill that comes through. There are lots of ways of maintaining marginal deterrence:

• Second strike: maximum penalty; Third strike: some multiple of the maximum
• Second strike: minimum penalty is no less than 75th percentile of penalties awarded on first strike for that offense; Third strike: maximum penalty
• If you were paroled after 1/3 of your sentence on the prior strike, adding 2/3 to your next-strike sentence

David Friedman's work shows reasonably convincingly that you need to have a higher penalty for a second offence than for a first one just to maintain deterrence; maintaining marginal deterrence then also requires that there aren't flat portions of the expected punishment curve across severity of offence. Lots of ways of achieving that, but a mandatory 25-year sentence for any third-strike isn't one of them.

Strategic incompetence

The ever-wonderful Scott McLemee today provides a wide-ranging review of Gambetta's work on signalling and crime (and academia). For the mafia, signalling incompetence at running a business credibly shows the subject of the protection racket that the mafia just wants to keep extracting money that way rather than take over the business fully. In academia, at least in Italy, something similar happens:

"Being incompetent and displaying it," he writes, "conveys the message I will not run away, for I have no strong legs to run anywhere else. In a corrupt academic market, being good at and interested in one's own research, by contrast, signal a potential for a career independent of corrupt reciprocity.... In the Italian academic world, the kakistrocrats are those who best assure others by displaying, through lack of competence and lack of interest in research, that they will comply with the pacts."

Kakistocracy: government by the worst. Love that word.

Gambetta's book has now moved onto my "must read" list.