How to think about uni-disciplinary advice

Let’s say its 1990, and you are proposing an ambitious privatization plan to an Eastern bloc county, and your plan assumes that the enacting government is able to stay on a non-corrupt path the entire time.

While your plan probably is better than communism, it probably is not a very good plan.  A better plan would take sustainability and political realities into account, and indeed many societies did come up with better plans, for instance the Poland plan was better than the Russia plan.

It would not do to announce “I am just an economist, I do not do politics.”  In fact that attitude is fine, but if you hold it you should not be presenting plans to the central government or discussing your plan on TV.  There are plenty of other useful things for you to do.  Or the uni-disciplinary approach still might be a useful academic contribution, but still displaced and to be kept away from the hands of decision-makers.

Nor would it do to claim “I am just an economist.  The politicians have to figure the rest out.”  They cannot figure the rest out in most cases.  Either stand by your proposed plan or don’t do it.  It is indeed a proposal of some sort, even if you package it with some phony distancing language.

Instead, you should try to blend together the needed disciplines as best you can, consulting others when necessary, an offer the best plan you can, namely the best plan all things considered.

That might fill you with horror, but please recall from Tetlock that usually the generalists are the best predictors.

Ignoring other disciplines may be fine when there is no interaction. When estimating the effects of monetary policy, you probably can do that without calculating how many people that year will die of air pollution.  But you probably should not ignore the effects of a major trade war, a budgetary crisis (“but I do monetary policy, not fiscal policy!”), or an asteroid hurtling toward the earth.

If that is too hard, it is fine to announce your final opinion as agnostic (and explain how you got there).  You will note that when it comes to blending economics and epidemiology, my most fundamental opinion is an agnostic one.

This is all well-known, and it has been largely accepted for some time now.

If a public health person presents what is “only an estimate of public health and public health alone” to policymakers, I view it as like the economist in 1990 who won’t consider politics.  Someone else should have the job.  Right now public health, politics, and economics all interact to a significant extent.

And if you present only one of those disciplines to a policymaker, you will likely confuse and mislead that policymaker, because he/she cannot do the required backward unthreading of the advice into its uni-dimensional component.  You have simply served up a biased model, and rather than trying to identify and explain the bias you are simply saying “ask someone else about the bias.”

If an economist claims he is only doing macroeconomics, and not epidemiology (as Paul Krugman has said a few times on Twitter), that is flat out wrong.  All current macro models have epidemiology embedded in them, if only because the size of the negative productivity and negative demand shock depends all too critically on the course of the disease.

It is fine to be agnostic, preferably with structure to the opinion.  It is wrong to hide behind the arbitrary division of a discipline or a field.

We need the best estimates possible, and presented to policymakers as such, and embodying the best of synthetic human knowledge.  Of course that is hard.  That is why we need the very best people to do it.

Addendum: You might try to defend a uni-disciplinary approach by arguing a decision-maker will mainly be fed other, biased uni-disciplinary approaches, and you have to get your discipline into the mix to avoid obliteration of its viewpoint.  But let’s be clear what is going on here: you are deliberately manipulating with a deliberately non-truthy approach (I intend those words as a description, not a condemnation).  If that’s what it is, I wish to describe it that way!  I’ll also note I’ve never done that deliberately myself, and that is along many years of advising at a variety of levels.  I’d rather give the best truthful account as I see it.


Comments for this post are closed