Common Misconceptions of Economists
Tyler Cowen did a matched pair of posts on what he believes to be the common mistakes of left-wing and right-wing economists. What seems so striking to me is not the difference between mistakes made by different kinds of economists, but rather what I believe to be the set of misconceptions that are endemic to the profession.
To begin, let me provide two caveats. First, everything that follows is a generalization about what I take to be the dominant tendencies of professional economists (and especially American economists working in academia and government). It is possible to cite counterexamples to each statement; in fact, every criticism I will make has been anticipated by canonical economists including Hayek, Coase, Knight, Schumpeter, North and Smith (Adam and Vernon), among others. Second, I am observing the economics profession from the outside as an entrepreneur and business executive engaged in the economy directly. I think that the last formal economics training I had was Evsey Domar’s Comparative Economic Systems seminar at MIT in the 1980s.
I’ve grouped these observations into two broad themes to provide some structure; but in order to get beyond very general abstractions, I’ve also tried to give some of what I think are the most important examples of each theme. At the conclusion of each theme, I’ve highlighted what I see as the negative result of these specific problems, from the point of view of a consumer of the outputs of professional economics.
1. Strategic elision between economics as predictive science, and economics as informed advocacy. Economists will sometimes make explicit claims that “the economic science says X,” and will more frequently make implicit claims for scientific knowledge by flatly asserting the known truth of some predictive assertion. This is normally a statement made around some specific policy question – we should (or should not) execute the following stimulus program; we should (or should not) raise the minimum wage right now, etc.
When pushed to provide the scientific evidence, they will normally reference some combination of empirical analysis of naturally occurring phenomena and mathematical models derived from axiomatic statements about human decision-making. In scientific terms, this is all sophisticated theory-building. What’s lacking is dispositive evidence of the accuracy of the predictive rule that allows the statement about this specific case to be an example of a more general rule that has scientific provenance. Otherwise, all we have is an informed opinion of the type we might have from an expert historian rendering an opinion about something the likelihood that Libya would revert to an authoritarian government within ten years if it overthrew Gaddafi
Among the most important manifestations of this problem are:
a. Lack of focus on controlled experiments as falsification trials. Theory and experiment are to science as inhalation and exhalation are to breathing. Even in scientific fields in which experiments are infeasible, our knowledge of causal relationships is underwritten by traditional controlled experiments. Astrophysics, for example, relies in part on physical laws verified through terrestrial and near-Earth experiments. Economics has traditionally been a consciously non-experimental science (though this is slowly starting to change). This creates a very weak feedback loop to weed out false belief. One can argue that controlled experiments cannot be done for many important economics questions. Fair enough, but then the claim to scientific status for these beliefs is hard to sustain, and leads to the next problem…
b. Ad hoc retreat to non-falsifiable “all else equal” arguments when confronted with apparently disconfirming evidence. The attempt to use predictions about future non-experimental events as falsification tests for beliefs tends to founder on what I have called the “causal density” of society. There is always some plausible excuse for why the prediction was wrong, but the theory is still right. The lack of control in most so-called “natural experiments” is deadly, but is often obscured by the next problem…
c. Hiding behind pseudo-technical jargon. All scientific fields have jargon, but not all fields with jargon create scientific knowledge. Normally, jargon must exist so that statements can be made precise enough to be falsifiable. As one layer of insight builds upon another in a given scientific specialty, jargon comes to include the role of explaining an ever-widening scope of phenomena with reference to prior insights. Jargon is a bug, not a feature: it is a necessary evil for specialists to make progress within a paradigm, but has the disadvantage of preventing non-experts from contributing meaningfully to the discussion. Without the discipline of experimental verification, however, this becomes more like philosophy than science. For economists, this can be a feature, not a bug, if it can be used to intimidate non-experts (generally those who are less comfortable with mathematics).
Result: The lack of a body of useful, reliable and non-obvious rules to predict the impact of proposed government interventions. As somebody who sits outside the profession, debates among economists are a means to an end. All I want is output: tell me the value-creating rules to predict the results of potential courses of action on the major issues of the day that your collective enterprise has produced that I would not have in the absence of your work.
Greg Mankiw, an economics professor at Harvard, is the author of one of the most widely-used economics textbooks in the world. In a chapter specifically devoted to arguing for the scientific nature of economics, he presented “a table of propositions to which most economists subscribe.” I found this pretty underwhelming as an argument for economics as predictive science: about 10% – 20% of economists apparently disagree with the central results of the field; half of the propositions are value statements concerning the way the world should be run (literally using words like “should”), rather than the kind of predictive rules produced by science; and even the other half, which are theoretically-falsifiable predictive rules, are mostly neither practically testable, nor specific enough to guide rational action. That doesn’t sound like any scientific field that I know about.
I’m not arguing that economics has produced nothing of value, but rather that its most useful outputs are more like those of historians than those of biologists. Draping the cloak of “science” over its findings can often be a rhetorical strategy designed to increase the leverage of economists in policy debates.
2. Use of a model for human mind, and by extension human society, that is simplified to the point of caricature. All rational disciplines, of course, must use abstractions that ignore some of the complexity of the real world. The question as a consumer of the work of the discipline is whether this abstraction supports or precludes the development of practically-useful guidance. The point of the first half of this post is that economics mostly hasn’t done this. It is my view that an over-simplified view of human mind and society is a key reason why not.
a. Ignoring the “irrational” psychological importance of group affiliation, and therefore under-emphasis on the role of institutions in promoting self-image rather than merely material self-interest. At the timescale of biological evolution, an extended commercial republic is brand new invention. While we have more just-so stories than legitimate scientific knowledge about the role of evolution in shaping human nature, one should expect that the faster pace of social change than of biological evolution is likely to create profound conflicts.
Any sustainably great collective – IBM, the Berlin Philharmonic Orchestra, the Pittsburgh Steelers, the U.S. Marine Corps, the University of Cambridge, or the United States of America – appeals to the rational self-interest of its members, but also creates a sense of irrational identification with the enterprise. Individuals within each will, to some extent and in some circumstances, sacrifice narrowly-construed perceived self-interest for the good of the whole. This kind of motivation is far more central to the lives of most real people than it is to most economic theories.
b. Many other examples include the mercurial nature of “utility,” the difficulty of forming commitment bonds across kin lines, and the central role of culture in creating economic outcomes. As is sometimes said in business “the soft stuff is the hard stuff.” Though not amendable to analysis, and especially not to quantification, the weird crevices of the human mind manifest themselves powerfully in our daily lives. In combination with the prior example, this tends to scale up to the dizzying nature of the institutions (in the broad sense of the formal and informal “rules of the game”) that determine the economic success and failure of societies. This is not ignored by economics, just radically under-emphasized.
c. Treating uncertainty as if it were risk. Under-emphasizing the complexities of the kind highlighted in the first two examples tends to lead to the problem of excessive belief in that previously-observed patterns are reliable predictors of future behavior. Repeated coin flips are complicated in the sense that we can’t normally predict heads vs. tails on a specific flip; but the series is still subject to probabilistic regularities, such as fair coin tosses should come up heads almost exactly half of the time. Human society is yet more complex, and patterns that seem reliable can suddenly change. This is frustrating to analysts, and therefore often ignored, or given only lip service, en route to making recommendations that rely on the assumption that these patterns will persist.
d. Ignoring the resulting complexities of the evolution of institutions over historical time. Our institutions are often mechanisms for organizing human behavior in light of human complexities, and for making decisions in light of true uncertainty. In an environment of true uncertainty, they have often evolved though trial-and-error, and therefore resist analysis. This is also frustrating to analysts.
Result: Excessive focus on allocative efficiency, at the expense of adaptive efficiency. I really can’t say this any better than Douglass North in his 1993 Nobel lecture: “Neo-classical theory is simply an inappropriate tool to analyze and prescribe policies that will induce development. …. It is adaptive rather than allocative efficiency which is the key to long run growth. Successful political/economic systems have evolved flexible institutional structures that can survive the shocks and changes that are a part of successful evolution.”
In sum, academic and government economists routinely overstate their actual degree of reliable, non-obvious knowledge about the answer to the practical question “What will happen if we execute policy X,” because it serves the class interest of economists to do so. Economists follow incentives like everybody else, rather than somehow sitting outside and above the process of buying and selling. Buyer beware.
So, it’s just economists who work for academia and government who are arrogant and unreliable.
Who is over-simplifying, again?
— Deb · Mar 17, 03:02 PM · #
I found this post excellent.
On your identity formation point, I was just thinking about that this morning with regards to (usually libertarian) theories of rational political ignorance and voting. What they missed, I thought, was the connection most people feel to their group. Most folks take political action as a part of – and in reliance upon the other members of – their group. It may (or may not) be worth their time to vote. But, their decision is not guided solely by their calculation of the value of their time versus the likelihood of their single vote turning the election. Rather, they usually have some belief that this is the type of thing “we” (meaning good citizens) do, or that if “we” (meaning whatever tribal ideology they subscribe to) don’t do it, then the other side will win. It’s the thought that while their vote may not make the difference, if everyone in their group thought that way then it would fall apart, so they act as part of that group, and in confident reliance on its other members.
Anyway, that’s a bit off topic but I thought relevant in the sense that it’s a repetition of the mistake commonly made by economists that you discuss. I found the post great and had nothing else to add directly on point. Bravo.
— Vermando · Mar 17, 03:17 PM · #
Typical whining from non-economist — economists don’t do this thing I think they should do. Guess what? We do all of these things, when they are appropriate and change the answers. Go back to your whatever and stop speaking of things you know little about.
— Oversimplifier · Mar 18, 03:30 PM · #
If there is one thing wrong with this post, it is the preposition in the title: this is a list of common misconceptions about economists. Your earlier post about Mankiw’s list betrays your ignorance: for instance, the only problem with the statement “Cash payments increase the welfare of recipients to a greater degree than do transfers-in-kind of equal cash value” is that it is a tautology, given the technical definition of welfare. The fact that purportedly smart people like you think that “welfare” is a “pretty broad term” may be why some economists answered “no” to this question, for fear of their technical assertion being mistaken for a policy recommendation.
I suggest you look up a paper called “Regulation for Conservatives: The Case for Asymmetric Paternalism.” It was written by a group of academic economists (and one lawyer) and published in Penn Law Review in 2003. The authors give examples of economic research which avoids all of the “misconceptions” you list, but then show (regretfully) that the resulting policy recommendations aren’t necessarily any better.
— Aspirant · Mar 18, 04:07 PM · #
Vermando,
Thanks for the kind words. I don’t think your observation is off-topic at all.
Oversimplifier:
Thanks for the charming, though evidence-free, response. Based on your moniker, you seem to be very self-aware.
Aspirant:
So let me make sure I’ve got this straight. In research to understand whether economists actually have achieved consensus on important issues, the researchers decided to ask them if they agree with a tautology. Then, 15% of economists disagreed with a tautology because they are afraid that it will be misinterpreted. Have I got that right? I guess that’s one theory. You seem to have taken lessons at the same charm school as Oversimplifier.
— Jim Manzi · Mar 18, 05:33 PM · #
great post, I’m not being sarcastic. Though there is little that can be done for an alternative methodology. Many of the criticisms in the first half about the nonscientific nature of economics are valid, widely applicable to almost all of the social sciences.
The comments in the second half are a bit tired. Caricature nature of economic agents often are designed to make models tractable interpretable. The results aren’t as robust as economists presuppose its hard to say they don’t contain meaning.In terms of methodology you’ll be hard pressed in the social sciences to find a discipline which lies a clearer or more coherent methodology for doing research, and as abject to criticism in the first place.
I do agree about your comments that economists do not pay attention to institutional details, however that also is not something that is easily incorporated within that framework. I do think the discipline should do more case studies with empirical analysis. However, I think one needs to ponder that aren’t other branches of social sciences paying attention and placing emphasis institutional details such as research done in business schools, political science (I.R. Comparitive Politics, Public Poicy), sociologists. One might argue that issue with economists is their prominence and clout in affecting policy is the issue. However, I do think sincerely that the discipline provides more trust worthy results as flawed as they are than other disciplines doing research in the same subject areas.
— economics grad student · Mar 18, 10:41 PM · #
economics grad student,
Thanks for the comments.
I’ve written a lot about these same issues as applicable to social sciences more generally, for example: http://www.city-journal.org/2010/20_3_social-science.html
— Jim Manzi · Mar 18, 11:46 PM · #
Another reason economics is less than a science: its most fundamental (neoclassical) constructs — notably the derivation of demand and supply curves, and the pricing theory of perfectly competitive markets — have been demonstrated by its own leading practitioners and its own methods to be self-contradictory. (Absent some truly heroic — read: ridiculous — assumptions/conditions that have been adopted in response to those demonstrations.)
http://www.amazon.com/Debunking-Economics-Emperor-Social-Sciences/dp/1856499928
Absent those assumptions, neoclassical theory fails by its own terms.
Economists enthusiastically adopt notions that conform to those assumptions, such as Hicks’ IS-LM model (he called it a “gadget”), even when those notions are later disclaimed by the very people who dreamt them up — i.e., Hicks’ acknowledgment that since IS-LM assumes certainty (IOW, fails to include the central concept of Keynes’ General Theory), it is fatally flawed. But they just keep teaching it because it fits with the rest of the hermetic* body of theory.
The adoption of those assumptions in a rather desperate rear-guard action (multiple, actually), and the failure to explain those assumptions until graduate school when indoctrinees have invested a great deal in the world view, is perhaps the best evidence of economics’ failure as a science.
This is not to say that it could not be a science, or at least more of one. But it’s in need of a Copernican-scale revolution.
— Steve Roth · Mar 19, 05:25 PM · #
Milton Friedman’s answer to Mr Roth’s point was that the assumptions in the model don’t have to be true. If the model allows out-of-sample prediction, then the differences between the assumptions of the model and reality just don’t matter much to what you are trying to explain.
That doesn’t address whether economics is generating out-of-sample prediction.
— Pithlord · Mar 19, 08:02 PM · #
Jim,
Great essay!
I believe that a simple economic analysis can show that Economics itself is not a science. Tell me if I’m off here…
Science is dependant upon the Scientific Method, a fundamental mechanism of scientific development. The purpose of the Scientific Method is to confirm theories that accurately predict outcomes, reject theories that fail to predict outcomes, and suggest alternate theories and alternate experiments.
For the most part, economic theories are not suitable for controlled repeatable scientific experiments. So there is little or no application of the Scientific Method, and thus no mechanism to confirm or reject economic theories based on their ability to predict outcomes.
So that’s the parting with Science, right there.
Now, it seems to me that there is effectively a market for scientific theories. Theories that can predict outcomes have a high value in this market, and conversely, theories that fail to predict outcomes have very low value. Experimental results effectively set the market values of scientific theories.
After that, the market values of proven scientific theories are also increased if they propose, directly or indirectly, more practical applications with high demand (like building transistors or curing diseases) or promote other theories to advance the field further.
Since Economics has no mechanism to confirm or reject economic theories based on their ability to predict outcomes, the values of economic theories are then mostly determined by the needs of their customers. I guess that would primarily be politicians, academics and business.
So an economic theory will then be far more valuable if it appeals directly to a particular political/academic/financial goal, as opposed to whether it can accurately predict an outcome. So there is an economic incentive for an economic theory to not be scientific.
The value of an economic theory is greater still if it can increase its appeal by appearing or claiming to be scientific, and thus more authentic. So there is an incentive for Economics to claim the imprimatur of Science without actually being Science.
I think I’m saying that Economics tells us that Science has an economic incentive to grow and progress forward, and that Economics has an economic incentive not to be Science, but to claim that it is.
(Wow, is that right?)
— Don Tillman · Mar 20, 04:40 AM · #