1

In the opinion of the economists here, what are some of the more important theories in economics that remain untested? By untested I mean theories that, though perhaps evaluated using empirical data, have not been tested in a laboratory setting.

This interests me because of experimentalists track record with fundamentally altering how the field views certain theories. The most obvious example of tested theories are expected utility theory and the convergence of double auction markets with no restrictions.

What theories do you think the field should eventually test and why?

examples:

monetary policy experiments:

https://www.macroeconomics.tu-berlin.de/fileadmin/fg124/heinemann/publications/CHsurvey-2014-08-25.pdf

macroeconomic experiments:

http://www.socsci.uci.edu/~duffy/papers/expmacro2.pdf

123
  • 2,911
  • 1
  • 14
  • 31
  • 3
    Are you defining "test" as laboratory experiment? Then, pretty much everything macro is untested. How do you test RBC, or Taylor Rule, or new trade theory, or endogenous growth theory in a lab? – luchonacho Jul 23 '17 at 07:18
  • Quite easily, depending upon who you test it. For example, there are experiments showing that people naturally adhere to the Taylor Rule whenever stabilizing closed, dynamic systems both with and without macroeconomic context. And yes - by test I mean test using laboratory experiment. As an example - I would love to see a test of the fiscal price theory in a laboratory setting. – 123 Jul 23 '17 at 19:44
  • 1
    But how is that a proof that Central Bank policy-makers, who are not just people but a very high-qualified group of individuals with different views on the economy, act as if the Taylor Rule? Even if the experiment is internally valid, its external validity is very very debatable. Which is actually a problem of every single experiment, and particularly worrisome in Macro, imo. If I were you I would reduce the question to micro only. – luchonacho Jul 24 '17 at 07:32
  • Also, at the moment the question is to be closed as opinion-based. Since the title and your in-text questions are related but different questions, I would stick to the one in the title, which is less prone to opinions (but still). – luchonacho Jul 24 '17 at 07:34
  • @123 "For example, there are experiments showing that people naturally adhere to the Taylor Rule whenever stabilizing closed, dynamic systems both with and without macroeconomic context. " Please edit a link of this into your question. It would greatly clarify the kind of testing you are looking for. (It also sounds really interesting.) – Giskard Jul 24 '17 at 20:05
  • @luchonacho - I never claimed any of those things. Apparently, you did. I made a statement of fact about known experiments. And if one observes human behavior exhibited in multiple laborites and experiments - spanning time, space, and culture - then one has little to question about external validity. Yours is an incredibly lazy argument. If, in your opinion, experimental economics still needs to defend itself ... well, it is possible that you are more dogmatic than you're willing to admit to either yourself or those with whom you interact. – 123 Jul 27 '17 at 09:28
  • Every empirical methodology needs to defend itself, whether it is panel data econometrics, experiments, or machine learning text analysis. I can't see why I'm lazy, nor dogmatic, but anyway. My point is that in my opinion the tools provided by experimental economics do not allow you to test a whole range of macroeconomic things. In the second document you attach (bottom of page 4), you can read it by yourself: "Perhaps the most difficult methodological issue is the external validity of macroeconomic experimental findings. ..." You see, I did not make up this myself. – luchonacho Jul 27 '17 at 14:14
  • @123. Experiments still lack convincing evidence that people's behavior in monitored, highly stylized, low-stakes lab games is closely related to their behavior outside the lab (I know there's some papers on this, but imo this should be the main concern of every single lab experiment). Also, we know from experimental psychology that even if external validity is given, the body of knowledge produced by low n, theory free lab experiments is one hot, p-hacked unreplicable mess. – Tobias Jul 27 '17 at 14:52
  • 1
    @luchonacho - I did not claim a field should not defend itself. I claimed that I am not, in the comment section of an online forum, obligated to defend all of a field myself. Further, experimental economics in general and macroeconomic experiments are two different beasts. My point is that experimental economics no more needs to defend itself than does any other field in economics. Perhaps less so. – 123 Jul 28 '17 at 02:38
  • I think we are derailing a bit here, perhaps seeing bad blood where it is not. The question is interesting, but since you have reduced the testing of theories to experiments only, I think most of macro and other areas like trade are basically untested. To make my point, I will give you an answer from such field (once I have more free time). However, the last question in the post is primarily opinion based. I would instead focus on the title's question, which is more objective. – luchonacho Jul 28 '17 at 05:59
  • This might interest you (just one side of the debate, of course). – luchonacho Aug 04 '17 at 10:39
  • There are far, far fewer macroeconomic experiments than microeconomic experiments. It's not hard to understand why. Consider for example the microeconomic problem of optimal auction design. That's a fairly well defined and constrained problem. In macroeconomics not only do we have a complex system but time is intrinsic to macroeconomics, with all the additional problems of expectations etc.. – Paul Aug 15 '17 at 04:05

0 Answers0