The Precautionary Principle requires that nobody should do anything that could come out very badly. It sounds like a very sound but entirely different principle, which is that we shouldn’t do things whose odds of a very bad outcome are high enough that they aren’t good bets, but it isn’t; it’s fundamentally different. Today’s performance of the PP shows that its real meaning is frequently “don’t let anything happen that I could be blamed for if it doesn’t go well.”
Certainly we can spend a lot of effort assessing “very bad” and “odds” to make risky choices. But there is no escape from the idea of an operational definition, the bedrock scientific rule that a measurement must be reported with (or be understood implicitly with) the protocol by which it was obtained. What are the operational definitions of the italicized words in the previous paragraph, when the PP is invoked? Well, could often means “an ignorant monomaniac with an internet connection said so”, and very badly means “another such has spun out a fact-free fantasy or borrowed it from a movie”. The PP is why ignorant people don’t vaccinate their kids: it endorses believing that a scissors only has one blade.
The superintendent of schools in LA closed 900 schools this morning in the face of an email threatening terrorism, putting a great city into complete chaos as parents missed work and tried to figure out what to do with their kids, not to mention losing a whole day’s learning, for which the district pays about $44m. All in all, a quarter of a day’s worth of the LA basin’s GDP is probably a good guess at the value Cortines and Garcetti put on a bonfire today for no good reason. (In New York City, cooler heads prevailed in the face of the same “threat”.) What would a responsible public official think about this decision?
(1) How does this threat look to a terrorist actually planning to pull off an attack? It looks a lot like intending to minimize the death and destruction on tap, which is inconsistent with the whole idea from the get-go.
(2) How often are real attacks preceded by warnings? …warnings followed by actual attacks? Murders by death threats (domestic violence aside)? Bayes’ theorem, not to mention common sense, makes these questions central to the analysis.
(3) There is no avoiding risk, only choosing the right risks. What can go wrong if we close the schools? Well, in addition to the immediate economic and social costs of the closure, we confirm to every crank and nutcase, and high schooler unprepared for today’s chemistry final, that any of them can close down the schools (courthouses? the Super Bowl?) with an anonymous email. These are pretty bad things to happen, and the LAUSD affirmatively chose to cause them with probability 1.0; not very precautionary, is it?
(4) Economist Michael Spence has given us the very useful concept of a market signal. This is information seeking to induce this or that action whose credibility depends on it being (i) costly to send (ii) less costly when it is true. The classic example is a used car guarantee: the car dealer might have to make good on it, so it’s not like “I am an honest man, and this is a good car! Really!”, and less costly if the car in question really is in good shape than if it isn’t. An emailed threat of mayhem (never mind all the evidence of a low-capacity mind at work that struck the New York leadership about this one) fails both parts of the test.
The LA city fathers, incredibly, failed to order the immediate evacuation of their city this morning even though there is a real, non–zero threat of a great earthquake that will kill thousands, including children, if they don’t get out of town. And after this incredibly irresponsible failure, I am pretty sure they will do it again tomorrow, and the next day! If the safety of kids is everything to us, how are we letting our precious LA children out on the street going to school day after day, when almost one pedestrian a day, including a child every week, dies in LA traffic? Who votes for these heartless, reckless bozos?
Garcetti, save your people!