PDA

View Full Version : How Much Math do PhD Economists Know?



GymShorts
11-25-2007, 01:15 AM
There are hundreds of threads on this forum about how much math is needed to get into a PhD program. But once you completely finish an economics PhD, how much math will you have acquired?

zsla
11-25-2007, 01:31 AM
There are hundreds of threads on this forum about how much math is needed to get into a PhD program. But once you completely finish an economics PhD, how much math will you have acquired?

It depends. The most important thing is to accumulate the potential to be able to study and understand everything (even after your PhD) when you read it. Other than that, people generally prefer to gain field-specific skills, which should be the optimal strategy I think.

kartelite
11-25-2007, 02:23 AM
It's a different sort of math than you would expect to see in a math grad program. A lot of economists are probably better than mathematics phd's with optimization and calculus, just because abstract math doesn't involve that stuff. Once you have an econ phd, you should be able to teach an intro calc, stat, or linear algebra course quite comfortably I would imagine. As zsla said, economists typically won't have the breadth of knowledge (econ doesn't make a lot of use of the Sylow or Ramsey's theorem and you don't usually see econ students debating the axiom of choice as nerdy math kids like to) but they will probably be very strong with several areas.

YoungEconomist
11-25-2007, 05:31 AM
A lot of economists are probably better than mathematics phd's with optimization and calculus

Economists are strong at math, but I find the above statement hard to believe.

Bayern
11-25-2007, 05:56 AM
Is it true that current average econ phd graduates are better than their predecessors. My Berkeley prof says that he had no idea about differential equations when he went to grad school. If that is true, how does the older profs keep up-to-date with the younger generation's highly mathematical econ papers?

Elly
11-25-2007, 05:57 AM
Economists are strong at math, but I find the above statement hard to believe.

This reminds me of a story my abstract algebra professor told once, where a family member asked her if being a mathematician meant she added up really big sums all day long.

Advanced math is all about proving complicated things on classes of abstract spaces- whether or not you remember anything about integration in R or how to row reduce a matrix doesn't really matter anymore.

Karina 07
11-25-2007, 06:16 AM
Is it true that current average econ phd graduates are better than their predecessors. My Berkeley prof he had no idea about differential equations when he went to grad school. If that is true, how does the older profs keep up-to-date with the younger generation highly mathematical econ papers?

I don't know, but just because something's published doesn't mean it has no errors. We had a funny story here the other day, we were supposed to read a paper by Summers and we come in and the prof asks how many people actually read it -- not very many. "Good," he says, "when I looked at it again it was rubbish, half the equations were wrong." (And he proceeds to insult it through the rest of the class, clearly amazed it was ever published like that.)

Ironic, given the hubbub surrounding Summers, that apparently he had a ton of math mistakes in a published paper, himself.

YoungEconomist
11-25-2007, 06:42 AM
Is it true that current average econ phd graduates are better than their predecessors. My Berkeley prof he had no idea about differential equations when he went to grad school. If that is true, how does the older profs keep up-to-date with the younger generation highly mathematical econ papers?

I imagine that they picked up math throughout the years. I've had many older profs tell me the same thing (that they didn't have much math when they entered grad school). I imagine they learned some math in grad school and I figure as econ got more mathematical they picked up some stuff.

IntEcon80
11-25-2007, 04:34 PM
IMHO, a lot of econ students tend to have some sort of arrogance. Many of them really believe that their Math skills are just absolutely amazing! If you read many posts on this forum, you get a feeling that some people really think that they can easily succeed in a Math Phd program.

However, like kartelite said, econ Phd students are definitely good with optimization theory.

kartelite
11-25-2007, 04:48 PM
I dunno. I actually did the bulk of coursework required for a math phd, and if I had stayed on I don't think I would have spent much time working actual calc/linear algebra problems except for as a TA. I took courses like probability, financial math, numerical analysis, and grad LA that involved a little computation but honestly that's not what math phd's do for a living, while economists are getting their hands dirty with probability distributions, systems of equations all the time. I remember trying to talk to my undergrad addvisor on some pretty basic topic (either DifEq or RA related I think), and he said "sorry I'm a graph theorist, I haven't seen that stuff in 30 years."

I can guarantee that the bulk of my former math professors would stare blankly at a moderately tough optimization problem for a good bit before figuring out how to proceed...

pevdoki1
11-25-2007, 05:06 PM
Is any theoretical linear algebra used in econ. PhD programs? You know, stuff relating to bases, eigenvalues etc? Or do you only have to know how to row reduce a matrix/find determinants etc?

kartelite
11-25-2007, 05:43 PM
I would imagine econometrics makes a lot of use out of theoretical LA (using different decompositions and canonical forms).

Epictetus
11-25-2007, 06:09 PM
My casual experience has been that most arguments in econometrics use fairly low-tech canonical forms and decompositions... rarely does the Jordan form come into the picture, for instance. But it's much more comfortable to have a good base of knowledge all around so that you never feel like you are missing the key argument due to a gap in your knowledge.

Terd Ferguson
11-26-2007, 02:39 PM
Is any theoretical linear algebra used in econ. PhD programs? You know, stuff relating to bases, eigenvalues etc? Or do you only have to know how to row reduce a matrix/find determinants etc?

Probably the opposite. Besides for math camp, I've never had to reduce any matrix. On the other hand, we have used eigenvalues a lot in macro and math econ. It looks like to me that linear algebra is the most applicable math course for most subjects. I would also highly recommend taking a numerical analysis course.

bscout
11-26-2007, 05:32 PM
This is a very interesting paper that addresses, among others, this issue: Economic Science: A Search Through the Hyperspace of Assumptions? (D.N. McCloskey) (http://www.econmethodology.org/methodus/pdf/v3n1/v3n1p6.pdf)

snappythecrab
11-26-2007, 11:19 PM
Is any theoretical linear algebra used in econ. PhD programs? You know, stuff relating to bases, eigenvalues etc? Or do you only have to know how to row reduce a matrix/find determinants etc?

No. Reducing a matrix, finding determinants, etc is for monkeys. Having a good handle on, say, linear space theory, is much more helpful than any sort of computational BS.

octavio
11-26-2007, 11:38 PM
No. Reducing a matrix, finding determinants, etc is for monkeys. Having a good handle on, say, linear space theory, is much more helpful than any sort of computational BS.


It's posts like this that make me worry that I'll be in way over my head next year! :doh:

constrainedoptimizer
11-27-2007, 04:03 AM
No. Reducing a matrix, finding determinants, etc is for monkeys.

That is hilarious. I'm seeing monkeys banging on an NxN with monkey wrenches like in the car mechanic commercials.