Another Round of Millman and Manzi
Prompted by another post about the limits of what economists actually know.
Jim’s post is really two posts, making two very different arguments, so I want to tackle them separately – among other things, because I don’t really agree with the first, but find the second really interesting.
Jim’s first post returns to his earlier argument about how scientific economics actually is. Last time around I responded by comparing economics with medicine, and asked: is if economics isn’t a science, is medicine a science? Now I’ll try another tack: is geology a science?
Geologists can’t run controlled experiments, after all. We’ve only got one planet, and there are real limits to how much we can mess around with it. Moreover, much of the planet can’t be observed directly – and indirect observation is heavily theory-dependent; we infer what is going on below the surface based on what we can observe from the surface, plus our current theory about how the interior of the planet works. If the theory is materially wrong, the same observations could different things are happening.
And, as a result, geology gets all sorts of predictions wrong. For one example, geologists cannot agree on how much petroleum there is in the ground – and these disagreements have really material policy implications. For another, Japan has just suffered a catastrophic earthquake and subsequent tsunami – and that earthquake was not predicted. (Obviously, any geologist could tell you that Japan is earthquake-prone – but that’s very different from predicting the timing and severity of an actual earthquake.)
So: is geology a science?
I hope we can all agree that the answer is “yes.” But if that’s the case, why does economics fail the test of being a science, given that it has some of the same deficiencies?
Part of the problem, I think, is that with any science we take settled matters for granted and focus on the unsettled areas, but with the human sciences we somehow forget that any settled questions were, in fact, settled scientifically. Economists can predict all kinds of boring things with great accuracy. Even things we’re interested in, they do well-enough at that people with money on the line – investors, business owners; not just politicians – rely on economic forecasts all the time. They aren’t perfect, but they are better than any available alternative. No, they can’t predict the things we’re really interested in – such as when the next recession will hit. But the geologists can’t predict the next earthquake either.
Another part of the problem, I think, is that economic policy is value-laden, so there’s an emotional charge associated with it that doesn’t obtain with, say, decisions that are based on geology. So geologists may disagree violently with each other about, say, the likelihood that we have already reached “peak oil” production, and we recognize this for what it is: a debate between scientists on a question where there is no consensus yet. But when two economists disagree on whether a stimulus plan will work or not, we throw up our hands and say – these are the people who are supposed to tell us how to run our economy, and they can’t agree on something this basic? But it isn’t basic at all, and there isn’t anything “weird” about the fact of that disagreement.
Where I always end up with these debates with Jim is: how does his skepticism actually impact our decisionmaking? When Jim says economists are “more like historians than biologists” – what does that mean when the question on the table is “what should the Fed Funds rate be right now?” That’s a policy question that the Fed has to answer every time it meets. If we don’t make it primarily by weighing the views of trained economists, leavened with the views of market professionals with a kind of practical, under-the-fingernails appreciation for how markets will react to this or that decision – then how do we make it? And if he agrees that this is how we should make such decisions, then what is the practical impact of his skepticism, given that this is basically how these decisions are made now?
The second part of Jim’s post is more interesting to me. Jim makes a variety of points about the limitations of “economic man” as a model for human behavior even with respect to economics, and cites these as marks against economics as a science as well. Humans have irrational attachments. They are not always motivated by material self-interest. They do not always understand the utility of the institutions and habits that they themselves depend on. And so forth.
This argument is made frequently in various forms, and I’m curious whether actual economists generally dispute any of these limitations. Perhaps they do. But one hot field these days in economics is “behavioral” economics which explicitly attempts to see how limitations on rational cognition affects human economic decisions. This is, in fact, something you can run experiments on, and economists are running those experiments – and then disputing what the results mean.
Ultimately, this point is a complication to economics rather than an argument against it. I’ll take an example. New York City has a medallion system for taxis. The supply of taxis is artificially restricted by law. The argument in favor of such a system relates, basically, to quality control. Most economists would argue that this is self-serving – that, really, this is just a law to protect incumbents, and not a law that promotes “quality” at all, and that anyway maybe consumers would prefer lower prices to the higher “quality” taxi ride produced by the law. Jim’s caution, I think, would be to say: don’t assume you know how much of Manhattan’s commercial culture depends on the reliable quality of yellow cabs. Before you tinker with the medallion law on the basis of an economic theory that opening up competition will improve the consumer experience, try to determine the impact of such a change on the unique culture of Manhattan.
But either that’s something you can study, and try to measure, or it isn’t. If it is, then economists would seem to be especially well-placed to measure it. If it isn’t, though, then how do you know whether the argument in favor of the medallion system is, in fact, a good argument rooted in the importance of taxis for a specific culture, or whether it’s just special pleading by incumbents? Again, we return to this question of decision-procedure: once we’ve been properly cautioned that change may be dangerous, how do we actually make policy decisions?
Jim quotes Douglass North at the end of his piece, saying that neo-classical economics can’t tell us how to promote “development” because it’s all about allocative efficiency rather than adaptive efficiency. I will note, first of all, that “development” and “growth” are not the same thing, and that there are other factors that bear on human welfare – including material welfare – than development. If neo-classical economics does know something about how to achieve full employment, but very little about how to achieve long-term productivity growth, that doesn’t mean neo-classical economics is useless. Unemployment is very bad for the unemployed.
The insight becomes interesting if there is a material trade-off between allocative efficiency and adaptive efficiency – if, say, it could be demonstrated that using monetary and fiscal policy to try to achieve full employment actually resulted in lower long-term productivity growth. But, again, who is going to demonstrate this, if not economists? In the absence of a good scientific theory of what drives innovation – which, in turn, is the driver of productivity growth – we have no way of distinguishing a strong argument from special pleading.
It seems to me that Jim shouldn’t be arguing from a position of skepticism – that economics is insufficiently scientific. Rather, he should be arguing for the development of a more scientific understanding of precisely those factors that he thinks (but can’t prove) are relevant to development and that he thinks economics scants. If they are scientists, economists should be enthusiastic about any project of that sort – provided it is undertaken scientifically. And I suspect they would be.
UPDATE: See also here. It’s kind of funny that I’m defending the economics profession in debate with Jim Manzi, and here’s someone defending the economics profession in a debate with . . . climate scientists.
So: is geology a science? I hope we can all agree that the answer is “yes.” But if that’s the case, why does economics fail the test of being a science, given that it has some of the same deficiencies?
This seems to be rather circular reasoning. Why is geology a science? You don’t give any reason, just say that we should all agree it is. Then the next move is to say that since economics is like geology, it should be considered a science too.
The implication is that anything that is theory-dependent and relies on indirect observation could be a science even if it gets all sorts of predictions wrong. But I’m sure you there are many topics that you would want to exclude despite their fitting that criteria (e.g., young-earth creationism).
I’ve seen this type of argument twice today (We consider X a science, economics is similar to X, therefore economics is a science) and I’m surprised that people think it is sufficient. I’m agnostic on whether economics is a science or not (thought I tend to agree with Manzi that it probably isn’t). But once you declare economics to be a science, it will be allowed to be substituted into future arguments.
Instead of saying “X is like geology, therefore we should consider it a science” they can simply use the lower criteria of “X is like economics, therefore we should consider it a science.” The problem, as I noted above, is that there are too many topics that we would not want to be declared sciences that would fit this standard.
If you want to declare economics to be a science, you first must clearly define what you mean by science and just the field against that criteria. Arguments from analogy aren’t going to be too helpful in this situation.
— Joe Carter · Mar 17, 07:11 PM · #
All these discussions pretty much amount to the same thing. They are rhetorically nice, but neither Millman nor Manzi have the philosophical chops to make an argument about what science is/isn’t that doesn’t gut itself. All Manzi’s arguments apply to all forms of inductive reasoning. Whether it’s geology, economics, weather forecasting, psychology, etc. You can’t simply stop short at the sciences you think are useful. The skeptics path is much longer then that. Theories of causality… epistemology in general, these things are much harder to think about than simply siding with sciences that give nice and clear (if equally wrong) answers.
I mean, evolution is one of the most important concepts ever made in science. But Darwin didn’t get there by breeding animals for millions of years. Observational science is still science.
— Console · Mar 17, 07:52 PM · #
I think the bigger issue—and the reason so many people care about whether economics is classified as a “science“—is because the term carries with it unwarranted epistemic weight in our culture. If a claim can be considered “scientific” then lay people who disagree are excluded from the discussion. Science can only be discussed by scientists, so if economics is a science, then the rest of us should sit quietly and leave the discussion to our technocratic overlords.
I should clarify that I don’t think this is what Mr. Millman is doing. But I think it is, generally speaking, the reason why the debate is more than a mere semantic detail about how a field should be classified.
— Joe Carter · Mar 17, 08:22 PM · #
_So: is geology a science?
I hope we can all agree that the answer is “yes.”_
Perhaps we can agree, but some don’t, along with astronomy and paleontology.
— Freddie · Mar 17, 09:13 PM · #
Jim is a test marketer, so he brings a test marketer’s perspective, which is a useful thing. Still, as a former test marketer myself, I’m struck by how seldom we discuss: Is test marketing a science? Jim seems to think it is. Me, I dunno … I haven’t thought about it much.
— Steve Sailer · Mar 18, 03:41 AM · #