Hi everybody. Welcome back.
Last time we answered two frequently asked questions about the model of
criminal liability having to do with efficient defenses and with the
rationality of offenders. Here's a third frequently asked question
that's suggested by the technique that we described earlier as probability scaling.
One of the first economist to write seriously about the application of
economic reasoning to the problem of crime and punishment was the great
American economist, Gary Becker, who teaches at the University of Chicago.
And published in 1968, a wonderful journal article that was the first
serious attempt to apply economic theory and detail to the problem of criminal
punishment. In the course of this article, Professor
Becker makes an argument for probability scaling criminal punishments.
He notes that the operation of a criminal justice system is expensive.
It's not cheap to operate a police force to apprehend offenders.
It's not cheap to operate a court system to adjudicate their guilt and determine
their penalties. And it's especially not cheap to operate
a set of prisons that inflicts punishment upon convicted offenders in the name of
compensating the indirect victims of their crimes.
Where it costs a great deal of money to invoke the criminal liability system
against an individual offender, it would be far better if people simply didn't
commit those crimes. And we didn't have to invoke the criminal
liability system and its attendant expenses quite as often.
And so, Professor Becker prescribes probability scaling.
If punishment is uncertain, he says, lower the probability that any offender
will be captured and punished at all as far as possible.
And then raise the severity of the punishment that's inflicted on that
unlucky few as far as possible. So that when you multiply that low
probability by that very high punishment, you get an expected liability price which
is sufficient to deter people. From committing inefficient offenses and
this raises the third question which is just that.
If apprehension and trials are costly, why not probability scale criminal
punishments and sentence just one or two offenders very, very severely so as to
encourage the others. Notice, just as in our earlier discussion
of probability scaling if we could get the values right, if we could get the
probability right and the severity right, we could produce an expected liability
price for every potential offender. Even those who are not going to be caught
that would just internalize the costs of their crime.
But if things don't work we can present them with precisely the same incentive in
the form of an expected liability price. And this will generate, in principle, the
same number of crimes as would perfect enforcement and perfect matching of price
and cost. So indeed, if we could do probability
scaling it might be a very efficient way to control the level of crime at the
efficient level by not invoking the Criminal Justice System very frequently
at all. The problem, of course, is the norm of
proportionality. Compensatory li-, liability systems can't
do this. They're constrained by strong norms of
justice. In the United Stated they're constrained
by the 8th Amendment, which forbids the imposition of cruel and unusual
punishments. They're prohibited, liability systems are
prohibited from inflicting the kinds of disproportional punishments that
probability scaling would require. American criminal justice, like American
tort liability requires that in each and every case.
The amount of liability that is imposed on a defendant held liable be
proportioned or equal to exactly the amount of cost that this person's
tortuous or criminal act has imposed upon other people.
It's an eye for an eye and a tooth for a tooth.
But it's never an eye for a tooth. Punishments, and liability generally,
must be proportioned to the harm that is done by the particular cost imposer upon
whom the liability price or punishment is imposed.
And so compensatory liability systems simply cannot inflict the kinds of
super-proportional punishments that would be required by any effective system of
probability scale. They're all constrained by the norm of an
eye for an eye and a tooth for a tooth. But not an eye for a tooth, the system
won't tolerate the severe punishment of people beyond the cost that they've
imposed. So as to produce incentives for other
people who are not caught to behave in particular ways.
What this means, interestingly enough Is that when the probability of apprehension
and the payment of compensatory liability price, when that probability is less than
normal. When the probability that somebody will
not pay is greater than zero, then the expected liability prices will be too low
to induce the efficient number of crimes. Again, to use an example I used before.
If I commit a crime that imposes $100 worth of cost, but there's only a one in
five chance that I will be caught. And when I am caught I'm only imposed a
punishment of $100, then the next time I'll face only an expected punishment of
$20 for a crime that actually imposes a cost of $100.
And that's because the norm of proportionality has prevented the court
from inflicting upon me a punishment greater than the $100 that my offense
imposed upon other people despite the fact that only one offender in five has
been caught. In such a circumstance, offenders face an
expected price for their crimes, which is significantly smaller than the costs that
those crimes actually impose. And that will induce inefficienct crimes.
So for example if I got a benefit of $70 from committing crime.
And the expected cost to me of committing the crime were only $20, I'd commit the
crime. And this, despite the fact that when I
commit the crime, I impose a hundred dollars' worth of cost and get only $70
worth of benefit. That's inefficient and it's inefficient
because I face an expected liability price that is too low to internalize all
the costs of my crime. Therefore too many crimes would be
committed for efficiency sake. If uncertainty is present and probability
scaling is not. [BBB] And this leads us to our last
frequently asked question. So, if criminal liability is prevented
from achieving systemic efficiency, as I'm calling it through probability
scaling. In allocating the costs of crimes and the
costs of controlling them through probability scaling, if it's not trying
to produce the efficient number of crimes.
What is it trying to do? And the answer is, it is trying to
internalize the costs of individual crimes one at a time.
By ensuring that, for every case that comes before the liability system.
Liability prices are set as close as possible to the costs of this particular
kind. Liability prices are paid in kind to
indirect victims even though this means that the expected liability prices are
too low, and far more crimes would be committed than would be systemically
efficient. This is a result that economists find it
exceptionally difficult to accept. Economists like Professor Becker, believe
that the goal to be sought by criminal justice ought to be what I'm calling
systemic efficiency. And that is to make sure that only
efficient crimes are committed and only inefficient crimes are deterred and that
this can be achieved through probably scaling.
But because probably scaling is unavailable to the american system of
criminal justice. Punishments can't be set high enough to
internalize costs when there is uncertainty of apprehension and a
requirement that compensation be paid. And thus, many more crimes than are
systematically or systemically efficient will be committed under this regime.
Instead, the system is trying to maximize the number of cases, in which an accurate
liability price is extracted from an individual offender.
And paid to that individual offenders victims.
[BLANK_AUDIO]
Richard AdelsteinProfessor
