How to find a dissertation topic?

[injak82]

Those who have been through the dissertation defence or those who are familiar with the process, would you mind answering a few of these questions?

1)First, how do you come up with a topic? How do you go from a broad topic to a specific one? Is it after taking a course, or reading a couple of papers?

2)What is the role of a dissertation advisor? Does he/she tell you exactly which topic to investigate? Does your advisor even propose a title for your dissertation? Or is it completely up to the student? Ie is the student responsible for finding a topic on his own and the advisor only role is to make sure that the topic makes any sense?

3)Finally how do you go from reading the theoretical models explained in textbooks such as game theory, information economics and other micro theory topics to writing your own theoretical model? It seems me that doing a dissertation on a purely theoretical topic is extremely difficult

[asquare]

injak82, there is a lot of information on these topics available online. I think the best of it is Hal Varian's paper "How to Build an Economic Model in Your Spare Time."

Your first question is extremely broad and there is no clear answer. The research process is different for everyone, and it changes as you become a more experienced economist. Remember that no one sits down one day to start writing a dissertation. The dissertation (which is typically three separate papers that may or may not be on related subjects) grows from ideas or papers generated earlier in the PhD process. People find ideas in different ways. Reading seminal articles in journals is a good way to build understanding of your topic but not such a good way to find ideas. Attending conferences and seminars, and reading recent working papers, gives you a better sense of active areas of research. Reading general-interest media (NYT, The Economist, other news sources) may point out phenomena that don't make sense or policy changes that create research opportunities. In general, you should write down any idea you have, any observation that makes you say, "huh, that doesn't seem like what they taught us in class." Most of these ideas won't pan out, but if a few do, you are well on your way. In general, I find it easiest to develop a research question by contrasting observed, real-world behavior with the predictions of economic models, and then trying to understand why the real world doesn't look like the model. That's just me, though. Other people do things differently.

The relationship between student and advisor is also quite heterogeneous. However, the advisor does not provide the topic (or title!). The advisor helps steer the student, helps refine the question, provides feedback on research strategies, results, econometric techniques, and (if you are lucky!) drafts of the eventual papers. Your advisor can point you towards related literature, and help you develop your interests by suggesting papers to read or seminars to attend. You may be able to spin off a dissertation project from RA work you do for your advisor. However, unlike hard sciences, economics PhD students don't join labs with ongoing research agendas they adopt as their own. Students are responsible for coming up with their own research ideas. It's one of the bigger challenges of the PhD process!

I'm not a theorist, so I can't really help you out with your last question.

I have to ask, are you a current graduate student or a prospective student? If you are a prospective student, I wouldn't spend too much time worrying about a dissertation topic yet. Virtually no one starts an econ PhD with a topic in mind, and even those who think they know what they want to study often wind up changing their minds or refining their question substantially. The research process is extremely challenging but it does become somewhat less mysterious as you slowly immerse yourself in it.

[asianeconomist]

One can find topics from reading the "Areas of Further Research" section in a journal article.

[Golden Rule]

Originally Posted by injak82

3)Finally how do you go from reading the theoretical models explained in textbooks such as game theory, information economics and other micro theory topics to writing your own theoretical model? It seems me that doing a dissertation on a purely theoretical topic is extremely difficult

I'll bite on this last one. One of my advisors likes to say, "if you want to do empirics, read theory. If you want to do theory, read empirics."

What that means is a good empirical paper might test distinct theoretical models. Similarly, a good theoretical paper might explain stylized facts or empirical observations better than what came previously.

So for an example of the latter (though this might be more "applied theory" than "pure theory"), the Melitz model in trade was successful because it provided a foundation for recent empirical work suggesting that firms that export have higher measured productivity, that exporting involves high sunk costs, and that market share shifts from lower to higher productivity firms after trade liberalizations and that lower productivity firms exit. It was better than competing theoretical models from Eaton & Kortum & others because it didn't rely on strong distributional assumptions and it allowed the number of varieties produced to be endogenous.

And I'll add, your question seemed to be asking how does reading pure theory inspire you to do more pure theory? My impression is that it just doesn't. Again, I think you need some real world problem or observation to inspire the development of that new theory. Otherwise you're just doing pure math.

[Canuckonomist]

Originally Posted by asquare

injak82, there is a lot of information on these topics available online. I think the best of it is Hal Varian's paper "How to Build an Economic Model in Your Spare Time."

Great suggestion, asquare. I love this paper. Hal is my hero.

[italos]

I would say choose your field first and then try reading some papers to get more into details.Afterwards, try to find the ''hall'' in the literature to fill with.I mean you must have an idea what do you like(i.e. micro,IO, Monetary,Macro,econometrics etc).So go read some papers ,make a idea what they are talking about and try to come up with a research inspiration.For example a very standard approach is to read a paper and try to add something more.If the author is doing an empirical work estimating xy factors you should add also z in the model and soon you have just came up with some new!

[italos]

TO THE FORUM:

Why when you have a great idea on a research topic almost always you discover somebody else had it before you????

[jeeves0923]

To italos:

If that weren't the case, then everyone interested in economics would be a famous economist. The field has been around long enough that many of the most intuitive ideas have already been "taken."

In that case, I comfort myself by saying that I wanted to know more about this subject, and now I can know more about it because someone already researched it.

[italos]

Originally Posted by jeeves0923

To italos:

If that weren't the case, then everyone interested in economics would be a famous economist. The field has been around long enough that many of the most intuitive ideas have already been "taken."

In that case, I comfort myself by saying that I wanted to know more about this subject, and now I can know more about it because someone already researched it.

It was more an exclamation in form of question rather a question!

[GymShorts]

Here's what you do to find a dissertation topic:

1. Take a saying from a fortune cookie, say, "You will find happiness in generosity"
2. Add "at the margin" to the end of it.
3. Replace "You" with "Rational economic agents"
4. Then test the hypothesis "Rational economic agents will find happiness in generosity at the margin."
5. When you are finished, remember to cite GymShort's dissertation method.