Interior Economics

Imagine you asked what fraction of whipped cream should be air, and someone responded "it is impossible to have whipped cream without air."  Or imagine you asked how much discretion judges should have, relative to following clear legal rules, and someone responded "it is impossible to eliminate all discretion from legal decisions."  You’d think they were avoiding the question.  Similarly, I asked:

For good policy advice, what is the best weight to place on economic theory, versus (individual or cultural) intuitive judgment?

And offered a tentative answer:

My guess is over 75% weight, so I try to mostly just straightforwardly apply economic theory, adding little personal or cultural judgment.

Tyler Cowen "answers":

There is no such thing as "straightforwardly applying economic theory." … Theories are always applied and interpreted through our personal and cultural filters and there is no other way it can be.  Robin believes in an Archimedean point for using theory, I do not. … Robin’s post is the clearest example I have seen of what I call Robin’s logical atomism.

Free Exchange chimes in:

There is no simple fact of the matter about the content of "economic theory". There are many bodies of theory that give conflicting answers to the same questions. … While I greatly admire Mr Hanson’s ambitions to overcome bias, sometimes his approach reminds me of tying one’s hands to the bedframe to avoid self-molestation.

Commenters at Marginal Revolution pile on: 

I wonder why some people are so determined to keep trying to conduct life as though the point-of-view conundrum can be willed away. … People have used "logic" and "rationality" to "prove" all sorts of things throughout history. … There is no such thing as presuppositionless exegesis. …

Geez. One should be able to ask where one wants to sit on a spectrum without being accused of saying the extremes are feasible or desirable.  Yes of course economic theory isn’t an autonomous computer program, able to answer any policy question posed.  And yes there can be other legitimate influences on policy. 

Nevertheless, the process of becoming an economist, including studying economics models, does influence our policy opinions.  So we can reasonably ask how strong we want that influence to be.  We can ask how much subjective weight on average we want to give to what we perceive to be "economic theory" versus other internal influences.  Or we can ask what fraction of our policy opinion variance we want to be explained by exposure to common economic theory (including frameworks, default assumptions, and standard data sources). 

But instead of engaging such hard questions, Tyler prefers to accuse me of thinking I see things from some perfect vantage-point, or that I know the "atoms" of thought.

Added: Tyler responds in the comments, and I reply.

GD Star Rating
loading...
Tagged as:
Trackback URL:
  • http://entitledtoanopinion.wordpress.com/ TGGP

    < ?xml version="1.0" standalone="yes"?>
    < !DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">

    I guess even economists think that economists are bar haids.


  • http://michaelkenny.blogspot.com Mike Kenny

    Robin writes:

    ‘We can ask how much subjective weight on average we want to give to what we perceive to be “economic theory” versus other internal influences.’

    I think the most important word is ‘average’. It might not have been clear to everybody that this is what you meant initially. Sure there are a lot of economic theories, but there’s a mainstream, which I assume would be more or less the ‘average’ of what all economists think. That would be the 75 percent. “Straightforwardly applying economic theory” would then mean “going with the mainstream of economics” and cultural and personal considerations would be the presumably necessary interpretative process Tyler points to, and which you don’t discount. Does that make sense?

  • http://hanson.gmu.edu Robin Hanson

    TGGP, I can’t find a version of your link that works.

    Mike, I had in mind averaging over policy topics.

  • http://profile.typekey.com/briarandbramble/ briarandbramble

    The problem is that taking a question seriously and answering it carefully would give a lot of credence to the questioner, and this would distract attention away from the answerer. A better approach is for the answerer to wait some period of time, reformulate the question as his own question, and then answer it.

  • Jef Allbright

    Tyler wrote:
    “Robin believes in an Archimedean point for using theory, I do not. … Robin’s post is the clearest example I have seen of what I call Robin’s logical atomism.”

    This highlights precisely the jarring feeling I quite often get while reading Robin’s thoughts. It’s as if Robin assumes not only a consistent reality, but some kind of objective access to it. It doesn’t mesh with my internalization of subjective Bayesian epistemology and thus the ineluctable importance of context.

    I find it fascinating that I get nearly the same impression from Michael Vassar (a frequent commenter and professed Bayesian), whom I also appreciate as a thoughtful, rigorous, but distinctly atomistic thinker, whereas with Eliezer I used to see it, but hardly anymore.

    This is profound, I think.

  • rukidding

    With whipped cream, you can separate cream from air.
    You can not separate economic theory from personal or cultural judgment.

  • Mark Stoehr

    On Tyler’s comment: “There is no such thing as ‘straightforwardly applying economic theory.’ … Theories are always applied and interpreted through our personal and cultural filters”, I strongly sympathize with Robin’s implied point that Tyler is avoiding the question given that, by my best reasoning, Tyler’s comment seems to have very little to do with what I presume Robin to be talking about. Robin’s post seems to me to be a follow-up on the question of why there aren’t many economists who support taxing the tall despite (what he claims to be) straightforward reasons for why economists should support it given standard economic theory. As I understand it, straightforwardly applying economic theory, according to Robin, is taxing the tall because that would be able to increase income for the government without hurting incentives: whereas going by cultural influence would be not taxing the tall because most people would consider it a very strange tax.

    My best guess on Tyler’s point is that he is commenting on how economic theory reflects the cultural influences of economists: that is that in developing any economic theory the economist is influenced by the society around her or him. I don’t think Robin has denied this at any point. I think that it would be easier to understand this presumed charge if somebody pointed out exactly to what influences Robin was succumbing, or, better yet, what influences has he unsuccessful avoided. In naming the influence be sure to name the source and the way that it operates: indeed this could be a new bias to post about on the blog.

    But, if nobody can name such an influence then the critique of “atomism” might be something akin to the Fully General Counterargument pointed out by Eliezer Yudkowsky http://www.overcomingbias.com/2007/04/knowing_about_b.html, since it’s so general that we could use it critique almost anything.

  • Tom Breton

    I think briarandbramble nailed the interpersonal dynamic going on.

    That said, one problem I’d have answering the question is that I don’t know what giving it (say) a 75% weight would mean; it might mean different things to different people. One might interpret that as spending 75% of their time on every policy question trying to understand it economically, another might aim for 75% on average, another might want to consult 3 economists for every one (whoever is supposed to stand for the other pole), another might list the weights of various pros and cons and give 75% of the total weight to economic factors. Another might do the same, but categorize the relevant factors differently and get a different answer. Another might multiply all the economics factors’ weights by 3.

  • Alan Gunn

    I don’t much like the original question because it seems to me to assume that all “economic theories” are the same. When the theory in question involves which way demand curves slope, as when the issue is whether raising the minimum wage is likely to affect unemployment or whether rent control will lead to queues of potential tenants, the only sensible answer is a very high percentage. But when the theory involves value judgments, as theories of what kinds of taxes are best typically do, a much lower percentage seems to be in order. So a good answer would necessitate estimating what fraction of economic theories are well established, don’t involve ideas of what’s good, etc. I don’t know enough about economics to guess at that.

    By way of example, when I studied geology, long ago, continental drift was quite controversial (it isn’t any more, I’m told); other geologic theories were much more widely accepted. If you asked a geologist how much weight he’d put on “geologic theory” when faced with an apparent anomaly, the answer would depend more on what kind of theories the person had in mind than on how conventional or committed a scientist he was. Same here, I think.

  • Tyler Cowen

    I think Robin’s defense would make perfect sense if he had originally written: “I define “theory” and “intuitions” in terms of descriptive sociological categories, but of course what we call theory is full of intuitions, cultural biases, personal judgments, etc. Given that division, I say 70-30.” Robin instead wrote: “I try to mostly just straightforwardly apply economic theory, adding little personal or cultural judgment.” If we interpret the question as Robin now does, I’ll say 30-70, to his 75-25. So there are two differences between us. The first is that Robin could have ever phrased things the first way he did. Even if this is me and others misunderstanding Robin, it’s not the kind of misunderstanding that my framing of the problem would give rise to (I’d more likely be misunderstood as a relativist, for instance). The second is 75-25 vs. 30-70.

  • http://hanson.gmu.edu Robin Hanson

    Tyler my original post referred to “economic theory (including frameworks, default assumptions, and standard data sources).” You apparently think that unless I explicitly said “but of course what we call theory is full of intuitions, cultural biases, personal judgments” that it was reasonable for you “accuse me of thinking I see things from some perfect vantage-point, or that I know the `atoms’ of thought.” This seems to me an excessive inferential leap of yours. To merely mention “theory” is not at all to make extreme claims about its immaculate conception, independent of any personal or cultural contamination.

  • Nominull

    If our economic theory is not an autonomous computer program, able to answer any policy question posed, that’s just a sign that we need better economic theory.

  • James Bach

    I don’t interpret Tyler that way. I interpret him as saying that your answer is unintelligible. Which it was, at least for me.

    75%? Come on, man. You know better than that. Do I really have to ask you 75% of what? You don’t mean 75%, you mean “lots”. And by lots you mean… I have no idea what you mean. I think Tyler doesn’t know either. You might as well grunt rhythmically.

    In my business I am often asked to give mathematical sounding answers to poorly formed questions. I won’t do it. Neither should you.

  • http://hanson.gmu.edu Robin Hanson

    Jef, I agree we access reality from a context, but I don’t see why that isn’t “objective” access, nor why that implies I think I know the “atoms” of thought.

    Tom, Alan, and James, it is hard to talk about big social questions without invoking averages or weights. To address “are men happier than women”, for example, one must average over people and moments of their lives, even though these may differ greatly, and weigh the degree of happiness of each. At lunch Tyler, I and the gang are willing to try to estimate answers such questions. Yes, that can produce misunderstandings due to our implicitly using different averages or weights. But if that is the price of talking about the big questions, it seems well worth it.

  • Caledonian

    What percentage of water should be hydrogen?

  • Jef Allbright

    Robin wrote:
    “Jef, I agree we access reality from a context, but I don’t see why that isn’t “objective” access, nor why that implies I think I know the “atoms” of thought.”

    Robin,

    [No definite resolution below, but perhaps some relevant observations.]

    There’s no assertion that you know the “atoms” of thought, but that your statements often convey the assumption that there are or could be such, playing an active part in any model of the world under discussion. I detected (what appears to me) the same epistemological gap when several weeks ago you expressed that you were “stunned” by a particular paper by Meltzer.

    I’m a natural-born scientist, all my life fascinated and driven by a desire to understand the nature of things. Up until a certain point in my life it was clear to me that I was in a process of discovering pieces of objective Truth, and I was mightily irritated by those who professed pragmatism, relativism, or post-modernism and thereby stood in the way of true progress. These days, I see relativism as morally deplorable, and post-modernism as partly sad and partly laughable — but I see pragmatic truth (no longer Truth) as epistemologically unavoidable.

    Note that I don’t deny absolute Truth, but claim there is wisdom in recognizing that it can never be known, and it plays no part in any knowable **model** of reality.

    ******
    For those who read the above statement and immediately wrote me off as [vague|mystical|mush-head|relativist|post-modernist] I can only suggest there’s something subtle here that may warrant your attention. And I’d be offended by those labels if I thought they applied.
    ******

    When I was a teenager in the 70s I encountered the idea of induction, and I remember walking for hours in a mild fugue pondering how to integrate it with my belief in “objective” Truth. Soon after, I took a course in statistics and was fascinated that they were teaching “randomness” when all I could see was “uncertainty.” Likewise in physics class I was introduced to entropy and intrigued by its obvious subjectivity. The teachers and the textbooks, however, hadn’t a clue about my concerns. Twenty-some years later I was introduced to Bayes, and it was like we were already old friends. Such was a part of my path, and it may be pertinent.

    Please excuse me for being blunt, but I would guess that for you Bayes theorem is objectively applicable, but doesn’t ring with a broad and beautiful resonance. I would also venture to guess that in the Meyers-Briggs scheme (notwithstanding its weaknesses) you measure closer to “S” than to “N.”

    Precision and accuracy are orthogonal to meaning, although they derive from the same reality. Like “overcoming bias”, where it’s about overcoming bias in our tools (inc. rationality), but it would be incoherent to think of overcoming oneself, despite the knowledge that one’s values-complex is necessarily an example of bias. Lacking an “Archimedian point”, our models don’t contain Truth, but we must believe they are refined by reality.

    Sorry if this leaves you thinking WTF, but it’s subtle, and my intent here is that I might only plant a seed.

  • J Thomas

    Without looking at much context, to me this:

    “My guess is over 75% weight, so I try to mostly just straightforwardly apply economic theory, adding little personal or cultural judgment.”

    looks ambiguous. On the one hand you say over 75% weight, which would imply something close to 25% individual judgement. On the other hand, you say you’d add “little” judgement, and to me close to 25% is not little. That phrase would put you over on one fringe, in the area where people would reasonably disagree and say that it isn’t possible to avoid a fair amount of judgement even in figuring out how to apply theory to complex real-world situations.

    Your original wording was so unclear that you shouldn’t be surprised it got misunderstood, and further discussion should have cleared up the original issue by now, leaving just the residue of people arguing about who was stupid to make which mistake.

  • http://hanson.gmu.edu Robin Hanson

    Jef, your guesses about me are wrong.

    J, to my reading “less than 25%” is consistent with “little” but does not imply “none at all.”

  • J Thomas

    Robin, I’m saying that when you use the language that way, some people misunderstand your meaning. I tend to. To me around 25% is not little. 3% is little, maybe 5%. Close to “none at all”. I might have supposed you were close to the “none at all” range if you hadn’t also said 25%.

    It looks to me like a wording issue, nothing more.

  • Jef Allbright

    Robin wrote:
    “Jef, your guesses about me are wrong.”

    The “guess” I expressed on the MBTI felt wrong even as I hit Send. You come across as INTP (almost INTJ) by nature but ISTJ possibly by training, and it’s the apparent conflict (and it’s potential resolution in the bigger picture) that I find interesting. I wrote a fairly long paragraph expressing the basis of my guess and uncertainty and then discarded it in favor of brevity. In retrospect I should have left it out altogether.

    As for your subjective appreciation of Bayes, distinct from its obvious utility, I hope for the opportunity to someday compare thoughts and observations with you in person.

  • ScentOfViolets

    I’ve never really held with these types of classifications myself. I will note that time and again I see Robin trying to pass off his personal opinion as the ‘standard consensual agreement’ on a given subject. This particular posting was just one instance demonstrating one type of strategy in doing so. Are there any Myers-Briggs categories which are more accomodating to this type of behavior than others? I’ll him the benefit of the doubt and assume he’s sincere.

  • Jef Allbright

    SofV wrote:
    “I’ve never really held with these types of classifications myself. I will note that time and again I see Robin trying to pass off his personal opinion as the ‘standard consensual agreement’ on a given subject.”

    What fascinates me, because it so fundamental to progress but so transparent and overlooked like the air we must breath, is how we lack an effective popular lexicon and grammar for everyday discussion of topics clustering around cooperation, agreement, synergistic advantage…the whole field of interactive epistemology applicable to intentional action of groups.

    Myers Briggs is useful but flawed; the “Big Five” classification scheme is statistically better, but still flawed because the model is statistical rather than generative, therefore increasingly wrong with increasing deviation from the mean. [Significant to many of us deviants hereabout.]

    I erred in bringing temperament into the discussion, providing a facile target detracting from my main point, which has not been acknowledged. (Indeed, a very strong “N” by temperament might be expected to over-compensate within academia by stressing the “S” of “objective” (and tacitly authoritative) facts.)

    But more significant seems your statement regarding a tendency to “pass off” personal opinion as the “standard consensual agreement.” I don’t see Robin doing that at all, indeed I think he’s typically very clear about distinguishing between consensus and the (capital T) Truth to which he seems to refer.

    That this topic is so resistant to elucidation is, in my opinion, why it’s interesting.