昨日紹介したAdam Ozimekのエントリでは、Russ Robertsが1月に書いたブログエントリを議論の出発点にしていた*1。しかし、そのRobertsのエントリは、ScienceBlogsの「Mike the Mad Biologist」によってかなり手厳しく批判されている(H/T Econospeak)。以下ではそれを紹介してみる。

I have often said that economics, to the extent it is a science, is like biology rather than physics. Let me try to make that clearer. By biology, I do not mean the study of the human cell, which we have made a great deal of progress understanding though there is more to learn. I am thinking of biology in the sense of an ecosystem where competition and emergent order create a complex interaction of organisms and their environment. That sounds a lot like economics and of course it is. But we would never ask of biologists what the public and media ask of economists. We do not expect a biologist to forecast how many squirrels will be alive in ten years if we increase the number of trees in the United States by 20%. A biologist would laugh at you. But that is what people ask of economists all the time. Economists should be honest and say that the tasks they are often asked to do are outside the scope of economics as we know it and perhaps outside the scope of economics as it will ever be known.

Erm, no. Actually, ecologists not only answer those types of questions for regulatory purposes all the time (e.g., forestry biologists, fisheries biologists, pest control, crop management), but they routinely answer much more difficult questions, such as how does a temperature increase combined with less rainfall affect pine populations due to a change in bark beetle populations. Biologists have been running simulations (which are different from models) for a very long time.




Is economics a science because it is like Darwinian biology? Darwinian biology is very different from the physical sciences. Like economics it is a very useful way to organize your thinking about complex phenomena. But it is not a predictive or very precise science or whatever you want to call it. Before seeing any direct fossil evidence, no biologist can tell you how long the giraffe's neck was ten million years ago. They cannot make accurate backcasts of any precision such as the year that the forerunner of the giraffe began to lengthen its neck through natural selection. It cannot model why the giraffe's neck isn't longer. Darwinism, like much of economics, exploits tautological reasoning. If the fossil record is incomplete or shows no change or vast periods or the pace of change is inconsistent with the fossil record, the theory is not discarded but modified with the concept of punctuated equilibrium. Is punctuated equilibrium true? There is no real way of knowing. It is our best hypothesis given very limited data. Is it a science? Sure. But it is a science that is unlike physics. That's OK. It is still a very useful way of organizing one's thinking about evolution. And the "imperfection" of biology is fine unless you really want to know when the elephant got his trunk. Then you are in unscientific territory. It doesn't matter whether our understanding of natural selection is imperfect or that we simply don't have enough fossil data. Biologists understand the limits of their field.

Whenever you hear the term 'Darwinian' from anyone other than historians of science, assume the crash position; it's going to get real ugly. There's a lot here to correct (but we like helping!). First, evolutionary biologists do predict past states: whenever we reconstruct evolutionary histories (phylogenies), we reconstruct the ancestral past states. And if we have molecular data, we can often attach a rough estimate of time to those states. We certainly can get the order in which events occurred estimated reliably.

The part about tautology is an egregious misunderstanding of the theorem of natural selection. The theorem of natural selection is, well, a theorem because it is repeatedly supported by multiple, independent lines of evidence. The punctuated equilibrium hypothesis isn't a theorem. It is simply a testable prediction of how the outcome of natural selection should be reflected in the fossil record. Sometimes 'punk eek' does fit the observed data, and sometimes it doesn't. This is what one expect of a discipline that deals with historical contigency--as does economics. As a dissertation committee member once told me, "It all comes down to those stupid fucking natural history facts."


経済学が科学であるのは、ダーウィン的生物学に近いためであろうか? ダーウィン的生物学は、物理科学とはまったく異なる。経済学と同様、それは複雑な事象に関して考えをまとめるのに非常に有益な方法である。しかし、どう呼ぼうとも、予測科学ではないし、精密科学でもない。実際の化石の証拠を目にしない限り、1000万年前のキリンの首の長さが分かる生物学者は一人もいない。キリンの祖先が自然淘汰によって首が長くなり始めた年を少しでも正確に述べるなどということは生物学者にはできない。また彼らは、キリンの首がなぜもっと長くないかをモデル化することはできない。ダーウィ二ズムは、多くの経済学と同様、同語反復的な推論を行う。もし化石の記録が不完全、もしくは長い期間に亘って変化が無い、もしくは変化の速度が化石の記録と不整合な場合には、理論は棄却されるのではなく、断続平衡説という概念を用いて改変される。断絶平衡説は正しいのか? それを知る術は無い。極めて限られたデータのもとでは、それが我々の考え得る最善の仮説なのだ。それは科学なのか? それは確かにそうだ。ただし、物理とは異なった科学である。そしてそのことは問題では無い。進化に関しての考えを整理する上で非常に有益な方法には変わりないからだ。そのような生物学の「不完全性」は、象の鼻が今のようになった時期をどうしても知りたいなどと思わない限り、問題とはならないのだ。そうしたことを知ろうとすると、非科学的な領域に足を踏み入れることになる。我々の自然淘汰に関する知識が不完全であることや、十分な化石のデータが手元に無いということは、さして問題では無い。生物学者は自分たちの分野の限界をわきまえている。




But the failure to predict the housing crisis isn't equivalent to getting the specific adaptive change involved in a particular case of resistance wrong; it's the equivalent of saying, "Don't worry, be happy" because resistance won't evolve. I'll forgive missing the specific, but the general outlines should be correct.


  • ここで言う予測は、一般的な予測ではなく、個別具体的な予測のことである。気象学は、毎年12月〜2月のNY市の気温が低いということは高い確率で予測できるが、明日コートが必要か雨傘が必要かについてはほとんど当てにならない。同様に、環境学者は、棲息領域の縮小といった環境変数の変化によって絶滅確率が上昇するとは言えるが、いつ、どの種が絶滅するかを予測はできない。
  • 株価予測でも、景気の良い時にこういう株が上がる、悪い時にはこちらの株が上がる、といったことを言うモデルはあるが、明日どの株を注文すれば良いか決めるには役立たない。




経済学者vs物理学者・続き - himaginaryの日記


*1:そのRobertsのエントリは、Cafe Hayekというブログ名に違わず、科学としての経済学へのハイエク懐疑論(cf. 本ブログの以前のエントリ)から話を始めている。