Commenting on a paper that looked at clusters of citations in economics research and found evidence of ideological tribalism, Tyler Cowen writes,
Berkeley and MIT have the saltiest taste, while Minnesota and Rochester are the freshest of the fresh. Chicago has a more neutral set of citation practices than many economists (not I) might think. Chicago cites saltwater school papers at a higher rate than the general average, nonetheless Chicago ends up strongly in the freshwater camp because it is cited so much by other freshwater schools, and not so much by the saltwater schools. A cynic might wonder if the Chicago economists are more open-minded than their critics, and I must confess that is consistent with my own anecdotal experience.
Robert Nozick once wrote of
Normative sociology, the study of what the causes of problems ought to be
In my new book, without using the term normative sociology, I give the example of different views of the cause of lower average wages for women. Using Nozick’s formulation, sociologists study gender bias and power, which they believe should cause the problem. Economists study human capital and lifestyle choice, which they believe should cause the problem.
In the book, I claim that economics is not a science. The usual narrative for scientific progress is that someone observes a phenomenon, comes up with an insight to try to explain it, turns that insight into a testable hypothesis, and tests the hypothesis. The problem in the disciplines that study social phenomena is that hypothesis tests never seem to be definitive, for familiar reasons, some of which are stated in the book.
In the book, I say that economists create interpretive frameworks, and that we have a hard time choosing from among different frameworks. Thinking about it further, I would say that in the absence of definitive empirical testing, economists are tempted to champion frameworks that focus on what they think the cause of a problem ought to be. If you’re pro-market, you think that the cause of the crisis of 2008 ought to be Fed misbehavior or housing policy. If you’re pro-government, you think that the cause of the crisis ought to be deregulation.
Of course, the best thing to do would be to come up with something better than normative sociology. Meanwhile, however, I think it would be better if we were to admit that is what we are doing. I we did, then I think we would be better off than we are now, when we think that we are applying scientific standards. Believing that science is possible leads to a mindset that thinks, “I’m doing science. Those guys are just investigating what they think ought to be the causes of the problem.”
Would admitting one doesn’t do science increase of decrease the status of economists?
Don’t people want status? Don’t people want the power and money that comes with status? The social adulation? The feeling of righteousness?
Imagine being highly intelligent, going through years of schooling, turning down lucrative private sector gigs, only to find out what you’ve been doing your whole life is probably just useless guessing most of the time. Who would admit that? Not just to the outside world but to oneself.
The change will come when employers of economists start looking for the skills that Arnold cherishes.
I, for one, am building a business I hope will have lots of economists. I know they will need ‘standard’ training, but I will be looking for people with Arnold’s general worldview.
Isn’t the main employment of economists as propagandists? Why would a business want an honest propagandist? Defeats the whole purpose.
You don’t want the propaganda pointed in, only out.
Is economics is more science-y than sociology?
Some parts of Economics are better than others. Micro is typically better than Macro, though not always.
Maybe a better way to frame it (heh) is to say that for any particular question, there are a lot of potential avenues of approach and degrees of thoroughness of study, and these can be ranked according to theoretical and practical reliability. So maybe a ‘science-ish’ spectrum.
There is ‘gold standard science’ including but not limited to double blind controlled experiments with low causal density, very large numbers of observations, well-defined and precisely measurable quantitative metrics, a high number of sigmas of confidence in the results of statistical analysis, and very high replication in other similar experiments.
That kind of result can form a good basis for reductionist approaches of rigorous theories in which mathematical rules of symbolic manipulation can be expected to yield useful results.
However, investigating questions at this high standard of inquiry is not always possible, practicable, acceptable, and/or feasible. And that’s fine, except, the further away one gets, the less justifiable it is to try and mimic the harder sciences to develop theoretical regimes under the pretense that they are of the same type, status, and reliability and that they can just as reasonably provide a basis for policy advice as in the gold standard cases.
I would say that in every major School, Economics of course included, there are questions and approaches to study that rank very differently in the science-ish spectrum. “Economics is not a science” is I think a short-hand way of saying that on average, and especially in Macroeconomics (and most especially Monetary Macroeconomics), apparently rigorous theories that dominate the high-status mainstream of the field at present purport to have high science-ish characters where the ‘science is settled’ on major, important questions, whereas, in fact, they have pretty low science-ish scores and confidence in advice and prognostications regarding major political action is totally unwarranted.
So, another way to say it is that Macroeconomics is a very weak, soft social science that is masquerading as – and inappropriately going through the motions of – the strong, hard natural sciences.
Nozick’s witticism prompts me to a passing philosophical observation:
The phrase ‘ought to be’ or ‘ought to exist’ is problematic, since the notion of moral obligation applies in the first instance only to actions. Still, it is not impossible to make sense of this phrase, even though in it the notion expressed by ‘ought’ is being applied to a *thing* or *state of affairs* (for brevity I will henceforth omit the word ‘thing’). For sometimes the existence of a certain state of affairs is fully dependent on the prior action of a certain agent, in which case the state of affairs may, as it were, “inherit” the moral status of the action. Such dependence will exist provided, given the state of the world at time t0, at which time the agent (call him/her “G”) had a choice between actions A0 (presumably *inaction*, which is always an option) and A1, the existence or not of state of affairs S at later time t1 depends on which choice the agent made. (Of course, I am taking the laws of nature as fixed.) Suppose the agent, as a matter of morality, *ought* to have chosen A1, in which case S *would* exist at t1 (if A0 had been chosen S would *not* exist). Then we can say that S ought to exist at t1 *relative to agent G and time t0*.
But note that this relativizing clause is necessary; if we pick a different agent or a different time we may get a different answer to the question, “Ought S to exist at t1?” Now, most people who use the phrase ‘ought to exist’ seem to have in mind an absolute notion, not relativized in this way. It seems to me that the best way to make sense of *this* notion of *ought to exist* is to imagine a theistic world, in which everything is *brought about quite directly by the action of God*, so that there is only one agent (or, at least, only one *primary* agent), eliminating the need for relativization to an agent; and either God is outside of time or His actions are all compressed, as it were, into the first moment of time, eliminating the need for time-relativization. In such a world, what “ought to be” (absolutely) is simply what God, through His action(s), ought to have created or caused to exist.
Of course, if one accepted a traditional conception of God, according to which He is morally perfect, we wouldn’t get much mileage out of the concept *ought to exist*, for what ought to exist would be always and everywhere precisely what *does* exist. But if we think of God as morally flawed, perhaps simply as amoral (though still subject to moral judgment), then we might well find discrepancies between what exists and what ought to exist.
So the absolute concept is quite understandable. But as an atheist I have no use for it (and a traditional theist would have virtually no use for it), and I suspect that its widespread use is due simply to confused thinking. Perhaps people are not clearly distinguishing *value* from *morality*: they want to impart moral force directly to their judgments about what is *better* or *worse*. If a certain state of affairs seems to them *the best possible* (out of some subset of all possible states of affairs–which subset in particular probably being somewhat unclear in their minds) they want to express this in a way that has *moral force*–hence their resort to ‘ought-to’ language. But this had better be avoided, for clarity’s sake.
This reminds me of the witticism:
“Economists study how people make choices. “Sociologists study how people don’t have any choices to make.”
It’s not a perfect description of economics, but it’s a first order approximation–combined with the contrast to the stereotype of sociology tending toward a study of individuals helpless before the forces of stratification and structures limiting their agency.
“Normative sociology, the study of what the causes of problems ought to be.” Is way above average. Nozick at his best was brilliant.
Is there a modern day Ambrose Bierce collecting such gems for a modern day _Devil’s Dictionary_?
Maybe I am more charitable but I believe most are trying to do science, it is just very difficult to do science and very easy to let our prior beliefs guide or deceive us. Yes. there are those whose priors are so strong no evidence is or will be sufficient, but that is true of all people, and these are often the people that scoff at the idea of science as that might force them to admit they are wrong. Those who predicted hyperinflation and skyrocketing interest rates and haven’t changed their views would be among them, but the predictions these would not happen is still positive evidence at some level. We should know more than we do, but that shouldn’t lead us to knowledge is impossible.
The point (probably) is that those who approach a topic with a “normative viewpoint” have hypothesized or (more likely) conjectured the existence of norms.
I read the book and quite enjoyed it, especially the reframing of economics, the contrast of command vs. decentralized markets, and the explanation of the recent financial crisis.
Let me push back though on your “interpretive framework” framing. I see much value in it, but taken to extreme it devolves into complete relativism — you frame it as such and we frame it as this. I think this is unfair. Why not include mystical framings? Why not include framings based upon the labor theory of value? You say potatO we say potAto.
I agree economics isn’t a full fledged science, but there are still superior framings of issues. The point is that once economists look at the choices women actually make, there is little or nothing left for sociologists to explain, other than perhaps why it is they make these choices in first place.
But note that isn’t what sociologists are doing. They are taking on an economics subject (wages), with little or no grasp of the economists knowledge, and making recommendations based upon whatever topic they wanted to recommend before starting the study. It isn’t an honest exercise in the slightest. It is a rhetorical dance to promote the interests of those they favor and demote those they don’t favor.
In the end, I am convinced that politics and related fields (Evonomics, climate change, regulatory issues) is such that people will convince themselves and others of some seriously wacky frameworks. This of course implies WE need to be constantly on guard for doing this ourselves. Hence the demand for skepticism. But frameworks are still necessary and valuable. In the end, the frameworks will compete in the worlds we create. Most will fail, but some will do better than others. What this calls for is actually more competition between places and institutions with competing frameworks.
Otherwise the best rhetoric wins. See Aristotelian Scholasticism to see how that turns out.
The real failure of macro as a science is the almost total inability to replicate experiments. There’s also the pesky fact that, despite every “socialism” that’s been tried and has resulted in bad economies, when people want to believe that somehow socialism / free stuff for them really work, they want to believe so many find a way to believe.
How often does socialism have to fail before those who profess it are considered crackpots?
Yet, on the issue of “intellectual property”, which is a gov’t granted monopoly, I’d be happier in a more sharing, arguably socialist, economy. And I think it would work better than what we have now for distributing the benefits, tho I’m not certain that the long term research effectiveness would be as good.
There’s no good way to have experiments on such important issues.
tiny typo, “I” should be “If”: >>are doing. I we did,<<
More interventionist IP is more socialist.
I don’t even want to use words like “stronger” IP. Stronger IP would be IP that didn’t require as much intervention to keep it standing on its own.