Discover more from Overcoming Bias
Weak tall-tax support among economists suggests to me this question:
How much does economic theory influence economists’ policy recommendations?
Or more precisely:
Considering all policy judgments of all economists (made as economists, thinking economic theory relevant), what fraction of judgement variance is explained by the fact that they know a common economic theory (including frameworks, default assumptions, and standard data sources)?
I asked (a version of) this question at lunch yesterday. Tyler C. suggested about 15%, while Bryan C. thought the number could be much higher for some average of economist judgments. (Alex T. and I withheld judgment.) Low estimates suggest economic theory usually says little, or that we are usually unwilling to give it much weight relative to our considerations (legitimate or illegitimate).
I know of two relevant studies. In one study, economist (and physician) policy choices had no relation to their beliefs on health economics facts. Similar results were found for labor policy. This raises two more questions:
Which economists rely more on economic theory, relative to other considerations, for their judgments?
For example, are libertarians better able to judge economic theory implications because libertarians explicitly reject such implications as a basis for policy? Also:
For good policy advice, what is the best weight to place on economic theory, versus (individual or cultural) intuitive judgment?
My guess is over 75% weight, so I try to mostly just straightforwardly apply economic theory, adding little personal or cultural judgment. Which is why I endorsed the tall tax. What weights are other economists shooting for?
Added: That lunch seems to have also inspired Alex’s post today.
Economist Judgment
This is my comment from Cowen's response to this post:
I'm not entirely sure what Robin is arguing.
For good policy advice, what is the best weight to place on economic theory, versus (individual or cultural) intuitive judgment?
My guess is over 75% weight, so I try to mostly just straightforwardly apply economic theory, adding little personal or cultural judgment.
Good policy advice is presumably a function of two things: values and facts. If we lack the proper values, it doesn’t matter how good our facts are. And if we lack the proper facts, it doesn’t matter how good our values are.
Economic theory attempts to answer the question “How do people respond to incentives?” That’s a factual matter. If this is all Robin intends by economic theory, then I’m inclined to agree we should give little weight to our personal or cultural judgment. But if that’s the case, I assume that what Robin means by personal judgment is our own anecdotal experiences, and I distrust those because people tend to overweigh the anecdote.
For example, I think this is flimsy: “I know people who would stop working if the income tax rate was raised even one percentage point, therefore the entire economy would stall if taxes were raised.”
Whereas this is stronger: “I’ve read many studies on large samples of the population on the disincentive effects of the income tax, and conclude raising taxes has minimal impact on taxpayers’ amount of labor.”
(Note neither statement says whether or not we should alter tax rates.)
Not only do I think the latter statement is stronger, I think the latter is properly described as being more unbiased than the former.
If this interpretation of Robin’s stance is proper, than Tyler’s argument is either wrong or uninteresting. If Tyler in saying “Theories are always applied and interpreted through our personal and cultural filters and there is no other way it can be” means that no matter how we try, we will always have some residual bias leftover, then that’s true, but trivial. We will always have bias, yes, but we can both do things to reduce and increase that bias—we can move towards or away from an Archimedean point, even if we can never completely reach it.
On the other hand, if Tyler means that there is no such thing as objectivity or no way to lessen our bias, then his point seems intuitively wrong. A postmodernist might accept that assertion, but if any other wants to, he’ll have to also accept much of the more disconcerting aspects of the postmodernist paradigm and its alarming stress on subjectivity.
But, perhaps what Robin is speaking of is not so much the empirics of economics but rather the value of economic efficiency. Perhaps he means efficiency should weigh 75% and other values, presumably personal or cultural, should get 25%.
If that’s so, there’s no reason I’m aware of for that weighting. And if that’s so, perhaps Tyler’s criticism is simply that all values are either personal or cultural—that there is no objective point to stand on and judge morality. I disagree—I think—but the issue is very contentious and I can understand Tyler’s position.
James:
You say 'It's not a question of intuition vs. reason', but your argument against the tall tax is pretty much just an appeal to intuition or common sense (which you seem to be metaphorically equating with oxygen).
If you're just making an argument against *advocating* the tall tax, this is a point that Robin made in the original tax-the-tall post. Experts may feel the need to bias their findings towards 'common sense' so as not to look 'silly' in front of others, especially those who don't have time to take in the arguments behind the 'absurd' idea (which in the case of economic proposals would include a whole course on basic economic theory). But we should be aware that this is a bias, unless you consider common sense opinions on a statement to be overwhelming evidence of its truth or otherwise. At best, you could regard it as the experts telling white lies in order to steer people closer to an optimal position, because they estimate that this will work better than telling people straight out where they think the optimum lies. At worst, the experts actually believe their common-sense-adjusted position and are unaware of the bias.
(Incidentally, there are two senses of 'absurd'. One is a logical absurdity, in which one concludes both a statement and its negation. If you can show a theory leads to this kind of absurdity, then you can conclude the theory is false. A 'fuzzy logic' version of this would be to find that conditioning on statement A would cause statement B, a statement which has been previously assigned very low probability, to jump to a much higher probability, something which counts as evidence against statement A, though a Bayesian might then say statement A is 'implausible' rather than absurd. The other is common-sense absurdity, which means something most people would intuitively reject out of hand, or a priori assign effectively zero probability to. In the second sense, much of modern physics is absurd, for instance, at least from the perspective of someone who isn't familiar with it. It's the ambiguity between these meanings that gives the word 'absurd' its rhetorical strength.)