Dissertation Topics

I’ll collect good dissertation topics here. They might be for economics, or sociology, or political science, or law. Anybody can use them freely. Probably you won’t, because most students can’t spot a good topic even if you thwack them across the forehead with it. They all want to come up with a theory of business cycles or a new equilibrium concept or apply a fancy technique to an old tired question using the old tired data. That’s rational, actually. Paul Graham said on Twitter:

The median investor’s first question is “who else is investing?”

The reason really good investors don’t ask this is not because they’re more disciplined, but because averaging their opinion of you with those of other investors would decrease its accuracy.

This pattern holds through many different fields.

The better you are at something, the more dangerous it is to care about conventional opinion.

Also, see the comments there:

“Obviously you don’t make money if you’re wrong. What most people don’t realize is that you don’t make money if you’re right in consensus. Returns get arbitraged away. The only way you make money is by being right in non consensus. Which is really hard.”

– @arachleff

Vincent Daranyi replied:

Don’t be interested in the opinion of others. Be interested in the reasons for their opinion. Then use that to form a stronger opinion of yours.

Somebody else noted, however, that this advice only applies to unusually smart people. If you’re average, then your opinion is probably stupid and you should just go along with what everyone else is doing. That will work okay in most contexts— business, choosing a dissertation topic, choosing a husband— but will fail in finance, because there, profits get arbitraged away by the smart people, and the ordinary people are left with losses. Thus, if you aren’t especially smart, avoid finance. As far as PhD topics go, if you are especially smart, you can come up with a good one, with effort, but you will also recognize that my list is a good one and it is much less effort to choose one of them. If you are not especially smart, you can fool people into thinking you are by choosing one of my topics even though you think they’re all crazy because you can still be wise, even if not intelligent or creative, and realize that I am an old scholar who probably knows a good topic better than you do. But if you are both dull and foolish, you’ll do okay. You’ll choose a mediocre, safe, topic and do average on the job market even though you’ll probably never get it published and if it does get published it will probably never be cited and even if it does get cited it will probably never have been read, even by whoever is citing it. So you’ll do okay.

That said, here is the list.

1. Bob Ellickson famously studied customary versus court law in Shasta County, California regarding whether you have to fence your cattle in (you pay the neighbor if your cattle eat his plants) or fence them out (you have no recourse. I note that I have heard different assertions from people who hunt as to what the law is in Bloomington, Indiana regarding bowhunting for deer in your backyard. It is agreed that in season, at least, you can do it and it’s fine if the deer then dies in your yard. What happens if it runs over to the neighbor’s yard and dies there? Does it depend on whether you have it in sight when it dies? (hot pursuit) Who owns the meat, and are you in criminal trouble? What is the actual law? What are the beliefs about the law? How are disputes resolved? Do the answers differ across cities? Harvard Law’s J. Mark Ramseyer and I were talking about this today.

2. What has covid-19 done to small businesses in Bloomington, or any small city you care to examine? What do left-wing, progressive owners whose businesses have died think of the lockdown policies?

3. How does the small-seller Persian rug business operate, especially with respect to finance and pricing? How often is a rug sold? Does the owner reduce prices if he needs cash to go on vacation? Who provides capital? Is the inventory easily resold at wholesale in case of foreclosure?

4. Regulations were imposed to require toilet tanks to be small to save water. This is unjustifiable from economic efficiency unless consumers are stupid about simple things in their daily lives. What would be interesting to see if is whether it even saves water. Smaller tanks mean more blockages, which results in repeated flushings. Thus, the regulations might well have resulted in the threefold badness of water wastage, lower quality, and greater expense. Showerhead regulation effects should also be analyzed.

5. How did the regulations get put into place that forbid practisting law iwthout a license? This is a case of reuglatory capture by an intrest group: it is for attorneys’ profit, not consumer protection. It no doubt sprang out of justified regulation: the requirement of being a member of the bar to argue before the court. That is jsutified because you cn be disbarred for misconduct, and you will not waste court time because of your inexpertise. In England, most lawyers were not members of the bar— they were solicitors, not barristers– for this reason. But that is no reason not to allow non-lawyers to offer legal advice outside of court.

6. Look at the hearing aid industry. It is a racket, like optometry used to be. Deregulation is coming on graduatelly.

7. Look at misinformation on oil changes, in particular how often you have to do it. Oil change outfits have clear incentive to lie. So do car manufacturers, since they do not pay for the oil, but they want their cars to have a reputation for lasting a long time. Many people do not think about this. Even Consumer Reports seems oblivious to the obvious conflict of interest when it comes to manufacturers.

Note that many good topics involve going out and talking to people and collecting your own data and applying simple theory. Fancy theory and fancy statistical techniques are optional. You can always find a legitimate reason to be fancy and show off your grad school training, which is legitimate showing off when you are on the job market. So don’t worry about that. But these are no-risk topics because the descriptive statistics are bound to be interesting, whatever they are, and will sell the paper and make people actually want to read it. The reasons students don’t do topics like this is because (a) they’re unusual, and (b) they seem like a lot of work, and (c) they’re shy, being scholarly people, and (d) they are homely topics, not like telling the Fed what its policy ought to be. I was going to write, “they are homely topics of no interest to the World Bank or the IMF”, but that is false: the World Bank and IMF are very interested in topics like Persian rug business finance and the effect of covid-19 on small business, and they are totally uninterested in what a grad student has to say about Fed policy. And the first topic, hunting law, is exactly the kind of scholarly work that is going to interest judges and get cited in court opinions. The general rule is: the more homely the topic, the more likely you are to have important policy influence, especially as a no-name scholar.

I should mention that on reason (b), “they seem like a lot of work”, the key word is “seem”. They are a lot of work, of course, but any successful topic will involve a lot of work— it’s just that most of them seem easy. Data collection topics actually are the easiest, probably even in terms of hours spent. Spending 8 hours a day for 2 months would do it. It looks easier to take existing data and apply a new statistical technique, but that’s because you are inexperienced and don’t realize how much time coding takes, and fixing your mistakes, and starting from scratch again when you realize you made a fundamental mistake, and explaining the technique so your readers aren’t suspicious, and figuring out whether all the assumptions need to apply it are valid (they never are, quite, so you have to figure out what is close enough). Same goes for pure theory work. It looks easy to state a theorem and then prove it. But read Lakatos on that. The end product theorem is never the one you start with, which is always false. And your proof will be flawed too. And when you find your flawed proof is flawed and your theorem is false, all the theorems that follow will have to be redone too. And probably you will throw away the first idea altogether and start over. And once you have a true theorem, your proof will still be flawed, and you’ll have switched a minus to a plus somewhere in the middle, or in two places that cancel out, etc. etc. So it’s actually safer and less work to pick a straightforward topic: you can see the path all the way up to the top of the mountain, and you know there isn’t a giant chasm near the topic, blocking your route.

8. Unicode has lots of strange characters. Can a websearch be done to see how often the strange characters are used, and how the meaning evolves over time?