准版主moontory和majia20112011吵架的神贴～技术or直觉？

BusFin

I really appreciate your posts. No doubt many nice theoretical papers in corporate finance have learned and
borrowed those "showoff" tools which are first employed in asset pricing (see, e.g. zhiguo he's papers).
Technical stuffs are kindof necessary when one needs to express his/her views in a rigorous manner.
Although it is good to put one's (economic, physical, psycological, whatever) intuition into one or two
simple sentences, many many times our common sense or intuition turns out simply wrong.

三人行，必有我师焉

"I agree with your basic point but I may not holding such a strong attitude as you. Start with a great idea and then find a way to model it is a good way to do research, but there are many other possible way to contribute.
There are indeed some very technical papers at their era but without too many fresh economics insight, but they provide some very useful tool for following researches(a not so good example might be theoretical econometrics, they mainly try to solve some problems in applied field, or just theoretical interest, but the solution itself often lack of any economic insight).
There are also many other great works with some intuitions difficult to express in plain sentences, like Arrow's impossibility theorem, in those situations, technical work often provides some results out of people's intuitive thoughts, and it is those technical results which push people to rethink some so called "common sense", and change our view of the world. There are many examples in Micro theories generating those "counter intuitive" results, and many of them are not easy to explain in a very few sentences.
And I think you know that there are many models and famous papers analyzing the same issue but generate totally different results with their own great ideas(like Macro, since you mentioned Lucas and Prescott), and the first step for us to compare them is to understand the technical details.
And even though technical level itself has no direct relationship with the quality of the paper, I would say it is still a not so good indicator of quality of research in different field, the history of economic or finance research in a field is a history of relaxing assumption or finding problems in previous methods, both often requires a more technical method to deal with.


I'm not saying you are wrong, and no offense. I just say technical issue is not a secondary factor as you mentioned, it is a very important part of the research and I would be very cautious to make those statements.

-- by 会员 moontroy (2011/3/29 10:58:18)"

majia20112011

My point is simple: with only technical skills you can never write a great paper, but if you have economic ingenuity you can always learn technical skills later. Also great economic insight can and should be expressed in plain English. Any Nobel Prize winners' major contribution can be summarized in a very short paragraph outlining the finding as well as intuition. Technical skills are necessary but of a second order relative to economic ingenuity.  Bob Lucas majored in political science when he was in his undergrad so he did not know of a lot of technical stuff when he went to do a PhD in Chicago. As a result, he got quite a few Cs during his course study. However, Lucas learnt the necessary technique by himself and because of his economic talent later wrote great papers and went on to win Nobel Prize. Nobel Prize winners do not win Nobel Prize because of their technical skills but because of the deep economic insight in their papers, so economic ingenuity is of the first order.

I really appreciate your posts. No doubt many nice theoretical papers in corporate finance have learned and
borrowed those "showoff" tools which are first employed in asset pricing (see, e.g. zhiguo he's papers).
Technical stuffs are kindof necessary when one needs to express his/her views in a rigorous manner.
Although it is good to put one's (economic, physical, psycological, whatever) intuition into one or two
simple sentences, many many times our common sense or intuition turns out simply wrong.

三人行，必有我师焉

"I agree with your basic point but I may not holding such a strong attitude as you. Start with a great idea and then find a way to model it is a good way to do research, but there are many other possible way to contribute.
There are indeed some very technical papers at their era but without too many fresh economics insight, but they provide some very useful tool for following researches(a not so good example might be theoretical econometrics, they mainly try to solve some problems in applied field, or just theoretical interest, but the solution itself often lack of any economic insight).
There are also many other great works with some intuitions difficult to express in plain sentences, like Arrow's impossibility theorem, in those situations, technical work often provides some results out of people's intuitive thoughts, and it is those technical results which push people to rethink some so called "common sense", and change our view of the world. There are many examples in Micro theories generating those "counter intuitive" results, and many of them are not easy to explain in a very few sentences.
And I think you know that there are many models and famous papers analyzing the same issue but generate totally different results with their own great ideas(like Macro, since you mentioned Lucas and Prescott), and the first step for us to compare them is to understand the technical details.
And even though technical level itself has no direct relationship with the quality of the paper, I would say it is still a not so good indicator of quality of research in different field, the history of economic or finance research in a field is a history of relaxing assumption or finding problems in previous methods, both often requires a more technical method to deal with.


I'm not saying you are wrong, and no offense. I just say technical issue is not a secondary factor as you mentioned, it is a very important part of the research and I would be very cautious to make those statements.

-- by 会员 moontroy (2011/3/29 10:58:18)"

-- by 会员 BusFin (2011/3/29 12:02:47)

moontroy

I would be very curious about how to present the insight of Arrow's impossibility theorem in a few PLAIN sentences, if you find this one too difficult, then try Aumann's Agree to disagree. I am very cautious to say "any, all", since I have faced many false intuitive claims in technical courses.

majia20112011

Good models usually do generate counterintuitive points. But the intuition of those seemingly counterintuitive points can still be explained by economic intuition. For example, the famous Kydland-Prescott paper which won the Nobel prize is counterintuitive but the driving force can still be summarized in a few words: ex-ante and ex-post preferences differ because of agents' rational expectations which fails the optimal control theory. I was actually trained ( and I myself continued doing that now) to make this kind of summarization to sharpen my understanding. Arrow's theorem can be done in a similar fashion. Since you are in Stanford, probably you can ask Arrow about it. Technical issue is important but still secondary compared to economic ingenuity. You just notice those Nobel Prize winning papers. They won prize because of the deep economic insight, regardless of what kind of techniques you use. Theoretical econometrics papers are ,strictly speaking, not economics because their purpose is to provide econometric tools for economists to use. So, by definition, they are not supposed to bring deep economic insight but serve as useful technical references wherever economists want to use them. However, they are very important. I am merely referring to non-econometrics papers.

[bgcolor=#ffffff]My point is very simple: there are quite a few papers nowadays which are purely a showoff of techniques without any economic insight. The rest is what you have said. Papers with good economic insight is because their ideas are good and intuitive and their conclusion can be summarized in at most a couple of plain English sentences.  If those idea needs to be illustrated using whatever technique as necessary, then use it. But do not use technique solely for the purpose of showing off something unless you have some real interesting economic story to tell. The art of economic modelling is to illustrate your point using the simplest possible model, although sometimes the simplest possible model might be quite technical. At least this is what I have learnt from Bob Lucas and Ed Prescott.  They write technical papers, but they use technique for good reasons.


I have no idea if the technique in B-S is basic to physicists, but their idea is simple and intuitive and can be expressed in plain English. If that needs to be done in such technical way,  that is fine. You seems to be indicating finance people do not have as much technical skills as physicists. But, rather than derivative pricing, I do not see any physicists making real big contributions to the finance literature (exclude John Cochrane or Steve Ross who has a undergraduate physics degree because their PhD training are solid econ and finance), in particular the corporate finance area. What's more, as somebody in the board pointed out a long time ago (sorry I cannot remember the ID) in recent years most job market stars (especially Chinese) are from traditional finance and econ background and that does contribute to the point that economic intuition and  institutional knowledge are more important than pure technical skills.

[/bgcolor]

I agree with your basic point but I may not holding such a strong attitude as you. Start with a great idea and then find a way to model it is a good way to do research, but there are many other possible way to contribute.
There are indeed some very technical papers at their era but without too many fresh economics insight, but they provide some very useful tool for following researches(a not so good example might be theoretical econometrics, they mainly try to solve some problems in applied field, or just theoretical interest, but the solution itself often lack of any economic insight).
There are also many other great works with some intuitions difficult to express in plain sentences, like Arrow's impossibility theorem, in those situations, technical work often provides some results out of people's intuitive thoughts, and it is those technical results which push people to rethink some so called "common sense", and change our view of the world. There are many examples in Micro theories generating those "counter intuitive" results, and many of them are not easy to explain in a very few sentences.
And I think you know that there are many models and famous papers analyzing the same issue but generate totally different results with their own great ideas(like Macro, since you mentioned Lucas and Prescott), and the first step for us to compare them is to understand the technical details.
And even though technical level itself has no direct relationship with the quality of the paper, I would say it is still a not so good indicator of quality of research in different field, the history of economic or finance research in a field is a history of relaxing assumption or finding problems in previous methods, both often requires a more technical method to deal with.


I'm not saying you are wrong, and no offense. I just say technical issue is not a secondary factor as you mentioned, it is a very important part of the research and I would be very cautious to make those statements.

-- by 会员 moontroy (2011/3/29 10:58:18)

angelabsc

这个帖子变学术讨论帖了。。。

majia20112011

Here is my understanding which might not be correct: transitivity of ordinal preferences can be violated whenever the number of alternatives is no less than three.

Please tell us the insight of Arrow's impossibility theorem in a few PLAIN sentences.

-- by 会员 moontroy (2011/3/29 12:48:53)

moontroy

[quote]
ex-ante and ex-post preferences differ because of agents' rational expectations which fails the optimal control theory.
[quote]

Sorry, although it maybe a very plain sentence in your mind, I could not understand it. Maybe I am wrong, but not every layman of economics know optimal control theory and ex-ante, ex-post preferences. I agree that SOME great work can be summaries into several sentences, but your example is not of the cases.


Kreps uses around 5 pages trying to give some intuitive explanation of Arrow's impossibility theorem to students, maybe you could do a far better job than him. I think many big names use less than 5 pages to explain Agree to disagree, but non of them would say it is very intuitive, and in fact there are tons of new papers trying to explain why it is so counter-intuitive.


Do not kick out econometrics so carelessly, talk to your classmates doing theoretical econometrics and try to tell the difference between econometrics and statistics.

moontroy

Here is my understanding which might not be correct: transitivity of ordinal preferences can be violated whenever the number of alternatives is no less than three.

Please tell us the insight of Arrow's impossibility theorem in a few PLAIN sentences.

-- by 会员 moontroy (2011/3/29 12:48:53)

-- by 会员 majia20112011 (2011/3/29 12:56:00)

Tell me the economic insight in this sentence:transitivity of ordinal preferences can be violated whenever the number of alternatives is no less than three.

majia20112011

Does every layman in economics in late 1970s or early 1980s know optimal control theory?  Rational expectations models originate from optimal control, but I guess maybe you assume every layman in economics now know about rational expectations models. The original purpose of Kydland and Prescott is to illustrate the universality of optimal control theory in solving dynamic optimization problem but they failed because economic agents are rational, unlike machines.  I hope this is not what every layman understands. In fact every layman nowadays understand that earth rotates around the sun, but in the Medieval Century this is certainly not the case.

[quote]
[quote]
ex-ante and ex-post preferences differ because of agents' rational expectations which fails the optimal control theory.

Sorry, although it maybe a very plain sentence in your mind, I could not understand it. Maybe I am wrong, but not every layman of economics know optimal control theory and ex-ante, ex-post preferences. I agree that SOME great work can be summaries into several sentences, but your example is not of the cases.


Kreps uses around 5 pages trying to give some intuitive explanation of Arrow's impossibility theorem to students, maybe you could do a far better job than him. I think many big names use less than 5 pages to explain Agree to disagree, but non of them would say it is very intuitive, and in fact there are tons of new papers trying to explain why it is so counter-intuitive.


Do not kick out econometrics so carelessly, talk to your classmates doing theoretical econometrics and try to tell the difference between econometrics and statistics.

-- by 会员 moontroy (2011/3/29 13:08:53)

majia20112011

Kreps uses 5 pages to give some intuitive explanation because he is teaching a class and he has to make sure a complete picture is taught to the students. If you are asking me to teach this in a class, I probably will use more than 5 pages, but that doesn't mean that Kreps or Arrow himself cannot summarize in a plain way. It is just that if you want to explain in detail,  you have to expand and that certainly would be more than a few paragraphs. In fact, if you are going to an AEA/AFA interview saying that I need 5 pages to explain the intuition in my paper, I promise you are not going to get any fly outs.

Econometrics is different from statistics because econometrics serves for economists while statistics is way more general. But relative to economics, econometrics is still a technical field and the main purpose is still to develop better tools for economics research.
[quote]
[quote]
ex-ante and ex-post preferences differ because of agents' rational expectations which fails the optimal control theory.

Sorry, although it maybe a very plain sentence in your mind, I could not understand it. Maybe I am wrong, but not every layman of economics know optimal control theory and ex-ante, ex-post preferences. I agree that SOME great work can be summaries into several sentences, but your example is not of the cases.


Kreps uses around 5 pages trying to give some intuitive explanation of Arrow's impossibility theorem to students, maybe you could do a far better job than him. I think many big names use less than 5 pages to explain Agree to disagree, but non of them would say it is very intuitive, and in fact there are tons of new papers trying to explain why it is so counter-intuitive.


Do not kick out econometrics so carelessly, talk to your classmates doing theoretical econometrics and try to tell the difference between econometrics and statistics.

-- by 会员 moontroy (2011/3/29 13:08:53)