Tuesday, April 23, 2013

A Bad Case of Magical Thinking



A FB friend shared something from an outfit called "Americans Against the Tea Party."

It was something that was written/prepared in response to the disaster at the fertilizer plant in Texas recently. But there is always something. I have to presume that this post is aimed at a federal initiative to improve safety inspectors, that it isn't aimed at the state government in Texas. After all, the state government can't print money, and may be limited by scruples against running deficits and borrowing money from the Chinese.

But I'm going to ignore a lot here in order to sharpen my focus. I'll ignore federal/state questions, philosophical taxation-as-theft arguments, arguments about the folly of printing money, etc. Sometimes you have to pick your targets. My target at the moment is a certain magical two-goals-at-once thinking that goes with many such appeals.

Here is a bit of what the FB "share" said,' "Don't support deregulation. Hiring inspectors in numbers enough to make a difference would put a small dent in unemployment. And we'd be safer...."

Wow. Those are presented as two separate arguments, and the first one is a prelude to the second: put a dent in unemployment and be safer.

How much of a dent in unemployment is desired from the creation of new safety inspector positions?

However many inspectors you deem the right number, "enough to make a difference" in safety, ask yourself: why not twice that many? Why not three times? Wouldn't that make twice as big a dent? Three times as big a dent? I suspect the phrase  "enough to make a difference" is meant to indicate some level of modesty in the plan. Not too many, just enough for a specific purpose. or even only enough to make a "dent" in a specific purpose.

But if you -- and I address the actual author of this silliness -- if you really believe that hiring inspectors is an unmitigated good thing, why be modest about it? Why not abolish unemployment overnight by declaring every unemployed adult an inspector of fertilizer plants so you can start sending out the checks?

Because you sense a practicality issue with that?  Ah ... we're getting somewhere then.

There is always a tension between two different reasons for doing X, when it comes to the specifics of X. If you want to hire more safety inspectors for a certain industrial process because (a) you want to make that process more safe and (b) you want to put a dent in unemployment, there will come times when those goals pull in opposite directions. What if a heavily automated process turns out to be the best way of making the process safer? Are we supposed to favor a labor-intensive process anyway, because of our high-employment goal?

Here's a not-at-all-farfetched scenario: suppose the best way to make fertilizing plants safer involves the hiring of a small group of human inspectors, assisted by robots, and involves giving advice to the management of those plants that leads to the plants' themselves becoming more automated and less labor intensive than they are.

Suppose that after a month or two of such inspections and re-structuring, there is more unemployment than when we started, because the loss of jobs in the newly re-designed plants exceeds the initial "dent" made by hiring the inspectors. Yet (by hypothesis) the neighborhoods are all safer.

Good plan ... or bad plan?

There is a certain sort of social welfare policy that consists chiefly in refusing to ask one's self such questions. But among those who do ask those questions, safety-as-jobs-policy doesn't find may defenders. Safety for safety is one thing, jobs policy is another, and only a bad argument confuses the two.

Perhaps safety can be best served by a vigorous tort law system. If businesses see the bottom-line costs of safety lapses through an effective litigation system, those costs are internalized, and they'll hire the optimal number of inspectors themselves. But in no case should we think of that as a jobs policy: even then, they might well decide the best way to lower their liability costs is to automate, and should they not be free to make that call?



1 comment:

Steve Finnell said...

you are invited to follow my blog