Jim Manzi and Ryan Avent continue their erudite food fight over Cap and Trade.

Manzi quoting Avent and then replying

Second, the cost overestimates have nothing to do with any underlying issue bias; as Brad Plumer notes, those favoring and opposing regulations both historically overestimate costs.

Presumably the same awareness of the track record of asserted prior under-estimation of environmental costs was available to both the EPA and CBO as they prepared their cost estimates. Unless we wish to assert that they are biased or simply irrational, why would we assume they failed to incorporate this information into their (very similar) forecasts of costs by 2020?

Avent back again

One thing that recurs in Manzi’s writing on climate change issues is an extreme devotion to the infallibility of models. There is a belief, seemingly, that the moment something becomes known, it is seamlessly and perfectly modeled. In practice, this almost never happens.

Here I have to side with Ryan. In my limited experience building and evaluating models for government there are two types of modelers. I won’t tell you which I am.

First, there are the purists. To the purist, nothing matters more than consistency. If we did it that way when the first model came out in 1946 then unless there is a darn good reason then we should be doing it that way now, the purist argues.

The reason is simple and compelling. No model is perfect, but people calibrate their expectations to the model. If the model says that proposal X will raise the baseline deficit by 1% of GDP then people have an idea what that means. That idea might or might not have anything to do with how much 1% of GDP actually is, but that’s not important. What’s important is that people can get a sense of the consequences from reading the results. This can only happen if consistency is paramount. Furthermore, the purist is afraid of the moralist.

The moralist has a different approach. Sometimes consciously, sometimes unconsciously the moralist shaves the model one way or another. Before everyone draws in a collected breath, ready to exclaim “I knew it was all bunk” allow me to explain.

Ninety-eight times out of a hundred the moralist’s goal is to constrain the excesses of policymakers. This means estimating higher deficits, larger costs, or smaller benefits. The moralist knows that policymakers will cherry pick the scenarios to find the most favorable and so he or she shaves the favorable ones. The moralist also knows that every model contains an “um, let’s just say four” factor or constant within it and he or she doesn’t want to be responsible for the institution or government taking on obligations it cannot meet. A cynic might also say that the moralist realizes that no one gets yelled at when the deficits turn out lower than expected.

Now, I have never worked with the CBO and I can’t in any possible way claim to know the culture there, but I can tell you that if anyone came to me and said, “Our models have been consistently conservative, maybe we should goose the costs downward” I would laugh out loud. Unless of course that person was a lobbyist in which case I would launch in to a long and to my colleagues well-known monologue that begins with “Well, you see . . .” and ends when the lobbyist has been worn down by technical jargon.

On the other hand I do think Manzi is right on the basic logic of the precautionary principle. But, more on that later.