Thursday, June 17, 2010

Economics as Science

In the comment section Xenophon raises this point of economics as a science (he asserts it isn't).

Why do people think it isn't a science? I've never understood the rationale. It clearly makes use of the scientific method. Some people talk about "relationships aren't constant between variables", but (1.) that depends on our ability to observe all relevant factors and conditions, which is harder in economics than, say, physics, and (2.) that certainly doesn't disqualify other sciences. Others are concerned about the fact that empirical economics isn't experimental. Again (1.) yes, there is experimental empirical economics, (2.) since when has non-experimental empiricism been a disqualifier for being a science - lot's of sciences are non-experimental?

It's not that this is even a tough question - none of the counter-arguments even come close to making sense. It's just this cultural sense that if you're studying human beings you can't be objective or scientific. I suppose that's the source of the antipathy - I don't know.

For those of you out there who don't think it is a science - why not? I never understood this meme.

42 comments:

  1. Let me repeat what I wrote earlier...

    I think economics can be a science, but so far it is at such a juvenile level of development that it isn't one.

    ReplyDelete
  2. Or let's put it this way; maybe one day economics will be a real boy, but it isn't right now.

    ReplyDelete
  3. Ya - I saw that after I posted this.

    The "but it isn't one now" point still suggests you think there is a reason it isn't at the moment.

    Again - why not?

    ReplyDelete
  4. Modern evolutionary biology, after all, was inspired by Malthusian empirical work. And empirical economics has certainly advanced since Malthus. What is so juvenile about a science that inspired the theory of evolution, exactly?

    ReplyDelete
  5. Well, for one thing, economics isn't terribly empirical, it isn't involved in on the ground field work much, and it is riddled more with ideology and premise defending than anything else. That's my view as an outsider.

    Not that economics is entirely useless; but what I reminds me of is medicine as it was practiced in Pharonic Egypt - it has insights, practical suggestions, etc. but that is not the result of it being scientific as we understand the term.

    ReplyDelete
  6. Hate to say it, but I agree with Xenophon here... I have no trouble with calling it a "science", but I don't think that it is to the same extent as some other fields that have claimed the name, and I think the scientific aspects of economics and other social sciences have a good deal more development available to them.

    I say this as someone working in theology, of course, so you shouldn't take this to mean that economics isn't a knowledge-producing critical inquiry. It's also worth saying that "science" is a rather bland and broad term, and one could offer a plausible conception of it that includes both economics and theology. That's my $.02, at least.

    ReplyDelete
  7. Why do you hate to say that you agree with me?

    ReplyDelete
  8. I hate to say I disagree with Daniel. Don't read too much into it. "I hate to say it" is a relatively unspectacular phrase.

    ReplyDelete
  9. Could someone explain why they think it is underdeveloped?

    I guess I'm having a hard time understanding exactly what you think is underdeveloped or juvenile, as it's repeatedly left unstated by both of you.

    I suppose Xenophon mentioned two things: field work and ideology.

    First, some field work does occur (I'm involved in two projects right now with major field work components). But it's certainly not emphasized. So? Why is "field work" necessary for science? It seems to me this depends on the science (just like the importance of experimental or non-experimental work depends on the subject you're studying).

    As for ideology - all of these have implications. Nuclear scientists have had ideological imperatives throughout the decades. Richard Dawkins has an ideological agenda. I'm not quite sure why having an ideology means you're not doing science - UNLESS your ideology determines your answers. That would of course not be scientific.

    Is it just field work and ideology? Evan - do you have any reason for your assertion? I'm still left wondering exactly why you're thinking it's underdeveloped.

    ReplyDelete
  10. I think the issue is simply that social behavior of any sort, and especially human social behavior, is a complex thing. While other sciences have been well on their way for a number of centuries, I think that for the social sciences things were slower until probability and statistics became a larger part of scientific method. In something like physics this allowed for a significant development of the science, while in the social sciences I think it was required simply to get them off the ground, and simply because their object of study operates on a more complex level of emergence.

    I'm not sure about Xenophon's mention of medical practice, although I do think the analogy could be put to use. As I see it, medical practice today, much less in ancient Egypt, is much less scientific than it is usually thought to be. The difference I see with medicine and economics, however, is that medicine is more of a practical craft- it is more akin to engineering, or law, or (if we want to try to relate it back to economic science) to entrepreneurship. The end-game of medicine is practical rather than theoretical.

    ReplyDelete
  11. Evan,

    Ahh, brothers in arms.

    Daniel,

    "I suppose Xenophon mentioned two things: field work and ideology."

    That's a strange turn of phrase. You suppose I did?

    "So? Why is "field work" necessary for science?"

    I can't think of a single scientific discipline where the object of study isn't directly studied. Economics just doesn't seem to do this very much at all.

    "I'm not quite sure why having an ideology means you're not doing science..."

    Oh here we go again. I didn't say that it did, I said this:

    "...and it is riddled more with ideology and premise defending than anything else."

    This doesn't say that the sciences don't have ideological components, it does say that they don't outweigh everything else.

    ReplyDelete
  12. I agree on medicine. Entrepreneurship is a good example - but also simply economic policymaking. In many ways it's the difference between applied and basic research - or even just research and engineering. Greg Mankiw had a good essay on the economist as scientist vs. the economist as engineer a while back. I think both are valid - you just have to be cognizant of what you're doing, and if you are engaging in engineering you have to remember Hayek's warning about what we think we can design.

    I think you'd be surprised about how statistical practice by economists compares to statistical practice by physicists or others. The mathematics of their theory is more sophisticated than economic theory, but their use of statistics substantially trails economics. The only discipline that may come close is biostats/epidemeology, but I'm not even sure about that. And the reason is simple - economists have much harder data to work with and much thornier (causally) research questions to answer so they have done a lot of work in advancing our knowledge of statistics.

    ReplyDelete
  13. RE: "I can't think of a single scientific discipline where the object of study isn't directly studied. Economics just doesn't seem to do this very much at all."

    Oh you just mean direct study? I thought you meant qualitative observational study. Who says we don't directly study the objects of our research?

    RE: "This doesn't say that the sciences don't have ideological components, it does say that they don't outweigh everything else."

    Right - and I agreed with that. When economists make an ideological position or take a normative stance they're not doing science. But that's a small part of what economists as individuals do.

    ReplyDelete
  14. "I think you'd be surprised about how statistical practice by economists compares to statistical practice by physicists or others."

    And yet that statistical practice is largely barren in its results.

    ReplyDelete
  15. I can't think of a single scientific discipline where the object of study isn't directly studied. Economics just doesn't seem to do this very much at all.

    This strikes me as flatly wrong. Physicists don't engage in any direct interaction with a huge chunk of what they study... a lot of it is working off of measurable effects and positing back to the object itself. Working on the cellular, atomic, or sub-atomic level employs the same sort of indirect interaction. This has been the whole point of major critiques in empiricist philosophy over the past 150 years, and it has great implications for our understanding of how science is done.

    ReplyDelete
  16. RE: "And yet that statistical practice is largely barren in its results."

    Again - could you explain? You regularly assert you're an outsider of the discipline. On what do you base your claim that the results are barren? What was the last empirical economics paper that you even read? I'm not saying you're obligated to read these things - I am saying claims like that should be based on a familiarity with what those results you're commenting on are.

    What was the last work of empirical economics that you (1.) read, or (2.) conducted/wrote yourself.

    I'm seeing strong confirmation of credit rationing (a la Stiglitz and Weiss) in empirical work I'm running right now - and the results are robust across five different alternative financial services. That's just this week of course.

    What do you base your particular claim on?

    ReplyDelete
  17. Daniel, on statistics in various disciplines, I wouldn't deny it. In fact, these methods are likely more sophisticated and heavily used in the social sciences precisely because these fields required such methods in order to approach something that we might call "science"... which was largely my point.

    ReplyDelete
  18. Daniel,

    "Who says we don't directly study the objects of our research?"

    I am obviously. Economists generally study people through multiple layers and abstractions. It would do your discipline a fair amount of good to study people far more closely; and that would seem to include a lot of field work. Having had a number of economists tell me that field work is just not that common in economics I have to say I was surprised every time I heard it.

    "But that's a small part of what economists as individuals do."

    I disagree; which is why you have these warring ideological camps amongst economists who argue over the fundamentals of the discipline. I have a biology background and that just looks, well, bizarre to me.

    ReplyDelete
  19. "Physicists don't engage in any direct interaction with a huge chunk of what they study... a lot of it is working off of measurable effects and positing back to the object itself."

    Actually, they do; which is why they build super-colliders and why fields like string theory are so controversial. Math really only gets you so far in physics; without the hands on research physicists quickly turn on merely math based notions of the physical world (which is of course why it took twenty to thirty years for the physics community to be convinced of Einstein's paper debunking of "the aether" - well respected physicists were still working on experiments regarding "the aether" well into the 1930s).

    ReplyDelete
  20. 1. Once again - insofar as we battle over ideology my point is that isn't science. Almost all the fighting over at a place like Cafe Hayek, for example, is over normative questions. It's akin to physicists arguing over what to do about the atom bomb. The fact that we have opinions about the implications of the science says nothing about the science itself.

    2. What "fundamentals of the discipline" do you have in mind? Some of the biggest clashes: the real balances debate, the size of the multiplier, the question of labor supply and demand elasticities (ie - the impact of the minimum wage, etc.) are all empirical questions. These aren't differences on fundamentals. Even something like Ricardian Equivalence is an empirical question. The only "fundamentals" argument I can really think of is the question of something like Say's Law. And it's notable that the people who dissent on that one explicitly reject the scientific method anyway! So among economists pursuing economics scientifically, I think you see much less disagreement on fundamentals, reasonable disagreement on empirical questions (which is fine for scientists), and certainly a variety of normative positions.

    ReplyDelete
  21. Daniel,

    Where do I get it from? From what I have seen of the work of Ed Leamer. Maybe I am making the wrong conclusion based on that. If so, explain.

    ReplyDelete
  22. Xenophon, I think the same point applies to superconductors. What they're working with isn't so much the particles, as the numbers that spit out of the monitors of the collisions that they're orchestrating.

    Various theories aren't controversial because they're so weak on experimentation, but rather because they are so significantly different than currently utilized theories that have been accepted in the scientific community as adequately making sense of things.

    ReplyDelete
  23. Daniel,

    Physicists don't argue over atomic theory and biologists don't argue over evolution the way that economists argue over the fundamentals of their profession.

    Quick question ... outside of some courses you may have taken as an undergrad have you ever work in a field other than economics? If not, then trust me, scientists find the sort of squabbling that economists do (and political scientists and sociologists, etc.) just bizarre.

    ReplyDelete
  24. Evan:

    "Various theories aren't controversial because they're so weak on experimentation..."

    Au contraire, that is exactly why they are controversial; or as one of my professors used to say, if you cannot experimentally verify it, then all it is religion or philosophy.

    ReplyDelete
  25. Ed Leamer's cautions are good, but what I've always found odd about Leamer is that he seems to have no concept of the fact that people are deeply concerned and cognizant of precisely the issues he raises.

    Economists deal with complex subjects that are not amenable to experimentation. That makes empirical work different, and empirical economists simply need to be cognizant of the concerns that Leamer and others raise before affirming a result. I reread the conclusion to Leamer's 1983 article, in response to your point and everything we're doing on the alternative financial services work is consistent with the steps that he suggests for making econometrics more plausible.

    So I guess I'm not sure what you're getting at. Of course there are pitfalls that aren't there in experimental work. So?

    I do find it a little disconcerting that you're commenting on an empirical literature that you don't seem to be familiar with yourself.

    ReplyDelete
  26. Xenophon -
    I have not worked in a field outside of economics, and I suspect that non-economist scientists are not as familiar with the discipline as they think they are.

    How many times has the fundamental paradigm of physics been overturned since 1870? Economists have maintained essentially the same fundamental understanding of their discipline since about 1870. Innovations have been made. Empirical results have been disputed (as we've established, non-experimental empirics is more work than experimental empirics). Ideologies springing from normative interpretations have come and gone. But the fundamentals of economics have remained the same for about 140 years. We seem to have gone through at least two or three "oops - we were wrong on the fundamentals" moments in physics.

    ReplyDelete
  27. Well, right... experimental verification comes into it. But I think I've affirmed as much in explaining why they are controversial.

    Any theory whatsoever is going to be distinct from empirical givens. My point is that this distance from the object of study isn't what makes a theory controversial. Rather, the theory's inability to make good sense of the object of study makes it less adequate, and the less familiar a new theory is to current scientific practice and sense-making, the more controversial it will be until scientific practice can more comfortably incorporate it and perform appropriate experimentation in order to determine its usefulness. But it's not controversial because it's non-empirical. No theory is empirical.

    ReplyDelete
  28. Nice talk guys...

    I have to go do some empirical science now (I don't feel guilty this morning because I was up until 2:30 last night finishing this crap as well).

    ReplyDelete
  29. *One more note -

    One thing that does bother me about economics is the insularity of a lot of people.

    I personally think real business cycle theory, Austrian business cycle theory, Keynesian depression economics, the inventory cycle, Minsky's work on financial fragility, and the accelerator-oscillator model ALL make sense as explanations of the business cycle. It's like explaining current in a stream... there are going to be lots of things influencing the flow of the water. For the economy, sometimes there will be real shocks (a la real business cycle theory), sometimes there will be liquidation of miscalculated investments. Sometimes demand will fall. I think it's ridiculous that people think of these as distinct or opposing theories.

    In that sense, I do think that economists could and should take a more "necessary but not sufficient" approach to their work. I think New Keynesianism is probably the best at this because it is a very flexible framework that can incorporate a lot of these insights.

    Explaining the business cycle is like explaining human evolution and why we are what we are today. We have lots of different theories about where we came from and when, why certain features were evolved, what niches they could have responded to, and we have disagreements over how all of this plays out. Some of these insights inevitably turn out to be wrong, but it's not like just a single one is right. They all explain different forces acting on the evolution of the human being. Business cycle theory is the same way. But just as biologists really agree on the fundamentals of the evolutionary mechanism, you'll very rarely find disagreement on the fundamentals of economics.

    Often the disagreements that do exist are just semantic anyway (ie - the whole "bounded rationality"/"rationally irrational"/"irrational" debate in which they all basically are talking about the same thing).

    ReplyDelete
  30. "I do find it a little disconcerting that you're commenting on an empirical literature that you don't seem to be familiar with yourself."

    Well, I am not an economist; you'll just have to take that into account if you want to discuss something like this with me.

    "We seem to have gone through at least two or three "oops - we were wrong on the fundamentals" moments in physics."

    Sort of, and sort of not. But the changes that have occurred are to the credit of physics. Anyway, you say that the fundamentals of economists have not changed since 1870; of course I think it is safe for me to say that a significant number of economists would disagree with that position.

    Evan,

    "...and the less familiar a new theory is to current scientific practice and sense-making, the more controversial it will be until scientific practice can more comfortably incorporate it and perform appropriate experimentation in order to determine its usefulness."

    Seems like it always comes down to empiricism then. Thus one of the knocks against "string theory" is that it is not amenable to empirical study; same with "intelligent design" for that matter. For something to be science it has to be empirical, and so in a significant sense - that doesn't mean that non-empirical work is useless, irrational, etc., but it does mean that it isn't science (and it also means that science does have some fairly significant limitations).

    ReplyDelete
  31. Daniel & Evan,

    Well, you guys have good days then.

    ReplyDelete
  32. Xenophon -
    RE: "Well, I am not an economist; you'll just have to take that into account if you want to discuss something like this with me."

    I understand that - and note I said that there is no obligation to.

    I would just hope that comments would be commensurate with knowledge of a subject.

    "The way economists approach X" is something you could probably speak to reasonably well - at least enough to have a discussion. "Econometric findings are X" is something that you really need to know the econometric findings to say!

    ReplyDelete
  33. Seems like it always comes down to empiricism then.

    That's fine. The origin of this back-and-forth was your insistence upon direct study of an object, however. My point is simply that theories indirectly tied to such empirical bases remain perfectly scientific so long as they are acceptable as useful methods for scientific work committed to empirical grounds. As non-empirical work, however, I don't see how theorizing isn't "scientific". If you're defining science simply as the observation itself, why even use the term science? The whole point of scientific study is that there is a larger methodological complex around empirical observation that leads to knowledge production through theorizing, categorizing, measuring, and other such non-empirical activities.

    ReplyDelete
  34. "The whole point of scientific study is that there is a larger methodological complex around empirical observation that leads to knowledge production through theorizing, categorizing, measuring, and other such non-empirical activities."

    Sure, but it is bit like trying to bake a cake without a leavening agent. While it isn't a sufficient condition, it is a necessary one.

    ReplyDelete
  35. Okay... I'm still not sure where this gets us. No one's disagreeing about that.

    ReplyDelete
  36. "I would just hope that comments would be commensurate with knowledge of a subject."

    That's what is great about blogging; you can take positions and try them out and be quite vocal and forthright about it.

    Leamer's position so far seems more convincing.

    ReplyDelete
  37. Of course economics is a science, and the axioms and implications of interest, capital, credit, and business cycle theory have been known for 80 years.

    Human Action, anyone? :P

    Edit: Also, above, in regards to the claim that science must study the object it describes: What about geometry? Perfect lines and angles never exist in reality, but that doesn't stop scientists from making conclusions and positing results from deductive claims.

    ReplyDelete
  38. Mattheus,

    Whether mathematics is a science has been highly controversial for a couple of centuries.

    ReplyDelete
  39. What would you call mathematics if not a scientific discipline?

    ReplyDelete
  40. Or simply a logic or a philosophy. Empirical verification is at the heart of science.

    Actually, empiricism and simulations are playing a bigger role in math than they used to - so I guess now we may be doing some "empirical derivation of logical principles". I'm still not sure it amounts to a science, though - I'd agree with Xenophon on that.

    Is it an art? I think I could buy that. One of the most challenging and interesting math courses I've had was Real Analysis. I have to say - many of the proofs in that course came from "ah ha!" moments and a little creative speculation as much as anything else.

    ReplyDelete
  41. An art, Xenophon? Really?

    Mathematics is not open to subjective claims of truth or value like art is. Mathematics is the greatest example of deduction we know. It is the single greatest topic our minds have been able to understand.

    Mathematics is logic, sometimes philosophy; I would still call it science. I don't believe science requires empirical derivation. The science of cardinal numbers, and how they function, is the study of mathematics.

    ReplyDelete

All anonymous comments will be deleted. Consistent pseudonyms are fine.