The Law and Economics of Tort Accidents, Illustrated

Via Andrew Sullivan comes this nice figure illustrating the effect of a car driver’s care level on pedestrians.  It might have come directly from the chapter on the economics of tort law in Mitch Polinsky’s famous  An Introduction to Law and Economics.  The chart’s authorpresumably unlike most mainstream law and economists, argues that local governments ought to be permitted more freedom to regulate driver speed (as opposed to letting the level vary ex post through a liability regime.)  I think it might be a better example of the importance of regulations requiring sidewalks.



You may also like...

5 Responses

  1. A.J. Sutter says:

    A nice case of misleading labeling and/or sloppy inferences. What does it mean? E.g., among other possibilities:

    (a) If a car is traveling at 30 mph, the probability of pedestrian being struck and killed by it is 45%, whereas if it is traveling at 20 mph, such probability is only 5%?
    (b) If a pedestrian is struck by a car that was traveling at 30 mph, he or she has a 45% chance of not surviving, whereas if the car was traveling at 20 mph during the collision, only 5% of such accidents are fatal?
    (c) If the speed limit is 30 mph, the probability of pedestrian being struck and killed by a car is 45%, whereas if the limit is 20 mph, such probability is only 5%?
    (d) If the speed limit is set at 20 mph, and a pedestrian is struck by an automobile, he or she will have a 95% chance of surviving?

    [To use a kind of Bayesian notation, where Lnn = speed limit is nn, C = collision between car and pedestrian, Snn = speed of car is nn, and K = pedestrian is killed, we can write these as: (a) p(S30 | C∩K) = .45 and p(S20 | C∩K) = .05; (b) p(S30∩C | K) = .45 and p(S20∩C | K) = .05; (c) p(L30 | C∩K) = .45 and p(L20 | C∩K) = .05; (d) p(L20∩C | ~K) = .95]

    My guess is that the data underlying the graph pertains to case (b). But in that case, the graph doesn’t have anything to do with economics — or with law. It’s a matter of physics and biology.

    Case (a), although a reasonable interpretation of this ambiguous graph, is clearly false. Otherwise there would have been many more pedestrian collisions and deaths in the town where I grew up, where there was a 30 mph limit on many streets, including the one I lived on. But anyway, once again this has nothing to to with law or with economics.

    Cases (c) and (d) do involve law (the speed limit), but they don’t have anything to do with economics. One could imagine any city council being persuaded to lower a speed limit if it will mean fewer injuries, merely if its members believed in preserving human life, protecting citizens, etc. Such concerns — and the idea that laws are meant to protect citizens — are much older than L&E, so don’t take credit.

    Unfortunately for these two cases, though, there isn’t anything in the X-axis label to suggest that speed limit is the relevant parameter. Moreover, (c) seems silly for the same reasons as (a). And (d) seems a questionable claim without more: e.g., data showing what percentage of people obey the speed limit.

    OTOH, if you mean to highlight the figure as an example of the shoddy reasoning or manipulative rhetoric of L&E, then of course it’s a good illustration.

  2. A.J. Sutter says:

    Sorry, I inadvertently reversed my probability notation, compared to the usual Bayes theorem notation; I put the known or assumed information before the vertical slash, instead of after it. More conventionally:

    (a) p(C∩K | S30) = .45 and p(C∩K | S20) = .05;
    (b) p(K | S30∩C) = .45 and p(K | S20∩C) = .05;
    (c) p(C∩K | L30 ) = .45 and p(C∩K | L20) = .05;
    (d) p(~K | L20∩C) = .95

  3. anon says:

    I think you’re thinking about this a little too deeply, A.J. Anything with numbers is “law and economics.” Anything to do with incentives is “law and economics.” That’s why most of the stuff that travels under the label “law and economics” is utter garbage, devoid of interesting insight.

  4. Ken Rhodes says:

    I gotta go with anon here. Assuming the intent of the graph is AJ’s interpretation (b), which seems to be indicated by the small print under the big title, then what I see is this:

    If you get hit by a car, you’re likely to get hurt worse if the car is going faster.

    Hmmm … science in action!

  5. A.J. Sutter says:

    Ken & anon, I take your point. Nonetheless, since L&E is a “superior scholarly technology,” as one HLS professor smugly declares in the conclusion of a recent piece, its proponents ought to be able to fend off puny slings and arrows like my comment with their superior scholarship. I’m curious — and still waiting — to see how they do so.