Python7768

A company's profits don't have to be complex to make it useful within economics. I didn't claim that it was used everywhere, I claimed that it was useful in some circumstances particularly within econometrics (more specifically time series econometrics). Here is a very widely cited paper within economics that uses complex numbers (I would claim the majority, if not all, macroeconomists are familiar with this paper): "Measuring Business Cycles: Approximate Band-Pass Filters for Economic Time Series" by Marianne Baxter and Robert King (link http://sws.bu.edu/mbaxter/papers/mbc.pdf). There are others that use very similar math. Complex Analysis is used very often within filtering frameworks. This is a good case against Abstract Algebra. Unless you enjoy it from a recreational standpoint (or are the next John Nash) it will only provide very marginal benefits. Overall, if you have very little programming experience then I think numerical analysis will still be the best course, but if you are lacking in formal math on your profile then complex analysis might be nice to have mostly for signal purposes. Numerical analysis would probably be more useful for passing first year courses.

I agree that numerical analysis is a good one, and maybe even the best of these choices (although I'm biased because I think computational work is fun). IMHO, people can go a little too far overboard on complex analysis and abstract algebra being completely useless... It may not be particularly useful if you are doing mostly applied work, but many of the breakthroughs in economic theory come because there is a problem that no one can solve until someone finds the right tool (often that tool is found within mathematics). Abstract Algebra may not be terribly useful in first year courses, but it may become useful if someone finds the right problem to solve with it. One example where it was used is proving Brouwer's Fixed Point theorem which is used to show existence of NE. Also, Complex Analysis is useful in econometrics.

I know two first year students; one at NYU and one at Colombia who both came straight from China undergrad, so it might just depend on school and the selection of students. They were both quite highly ranked within their department and were at a very good China undergrad.

Yes, in my opinion, the people who are doing the most interesting research in economics are currently developing mathematically rigorous models and then joining that with some kind of data. I think many if not all of those topics are used within economics (some more than others). Probability theory shows up literally all over the place and is a viable contender for having been the most useful undergraduate class I took in preparation for my first year. One example that could use several of the things is a simple macroeconomic model. You will typically have a measure of heterogeneous agents who are maximizing expected utility of their lifetime. In order to solve the majority of the models that look like this, you will need to use at least a little of each of the following: Measure theory(mostly in the context of probability), Probability theory, Some kind of difference or differential equation, Functional analysis (at least to prove the existence of a value function), etc. If you aren't afraid of the math; Stokey, Lucas, with Prescott's Recursive Methods in Economic Dynamics is an excellent book. It is 20 years old, but I think it gives a good introduction of the use of mathematics in macroeconomics. There are lots of other applications in micro and game theory. If you want to see how it is currently being used then you can probably crack open a volume of Econometrica or the American Economic Review. Those should have some good examples.

Agreed. There is no reason for them to suspect that you have the slightest amount of mathematical preparation to handle a masters. I don't know what most masters require, but I would be surprised if basic differentiation and integration (from the high school level) would be sufficient preparation. From what I understand and have experienced, the economics that is typically taught to the undergrads in business majors is not typically at the same level that it is taught to economics majors. i.e. managerial economics \neq (not equal, in case you aren't familiar with latex) economics. If you are set on pursuing this field (and if you are then you should pursue it), I would highly suggest taking at a bare minimum multivariable calculus, linear algebra, and intermediate micro/macro; without this basic understanding, even if you were to be admitted, you would struggle. That's my 2 cents, but take it as what it is: a suggestion from a random person on the internet.

It doesn't look like it is their intro to econometrics. That is econometrics 468 as opposed to the introductory which is 337 (evidenced by google). As far as being similar to an undergrad course in econometrics, it looks pretty similar to what I took at my undergrad for a 2nd course in econometrics.

I guess I must still be a naive first year, but I haven't found my department to be nearly as bad as EJMR. I wonder if we sampled participants of EJMR how their distribution across schools would be... Unfortunately most of them would lie so it isn't possible to get data unless someone hacks their servers and gets the IP addresses, anyone have connections with the NSA? :eager: Is it always a problem or can you get around it by finding a good set of coauthors that you work with?

Hey guys, I don't know if y'all have heard of the stackexchange websites, but basically they are a series of topic specific websites that allow users to ask/answer questions (example topics: general programming, mathematics, physics... Check some of them out to get a feel of what they are like). There is currently a proposal to open an economics specific stackexchange. It would probably have some overlap with this website, but I also think it will be a good place for some broader discussion. I plan on using both. If you are interested in it, check out this site and "follow" the proposal.

For future applicants, these types of problems can be avoided by waiting till you hear back from all of your schools to accept any offers. Accepting March 28th compared to April 14th or 15th won't have any effect on your future. tsonline, I would suggest following the same advice people gave earlier. Call or email the chair and be honest. I doubt they will have any hard feelings since they have invested much in you yet.

Short answer: You need more math. (you'll hear this from at least 6 other people) Take a chance this summer/next fall to take at a bare minimum Calc 3, Linear Algebra, and Real Analysis. Note that these are simply necessary, not sufficient for many schools. I would prioritize Real Analysis and Linear Algebra over Calc 3 because you have probably done some basic multivariate calculus in your other classes. Also, your GRE quant needs to be at the 164+ish range to shoot for places like BU. The lower ranked schools might not be as concerned with how you did, but I'm not sure how their application process works. There is more to be said, but someone else will follow this up.

I think NYU and Michigan both dominate Duke, but I'm not really caught up on the applied micro people at each dept. Don't know more about between NYU and Michigan

If anyone is interested in learning more about NYU Stern, feel free to pm me. It is a biased opinion, but happy to answer any questions that you might have.

Had a classmate that made a very similar decision between a PhD and a hedgefund(not to mention he was a double physics econ major as well). I would set up your schedule to be the PhD prep route, apply to PhD schools (top 5 or whichever you prefer), and apply to jobs. It can't hurt to see what happens and unless the company is very pushy then I don't imagine that they would have a problem with that. He chose PhD based on where he was accepted even though I think his plan is still to end up in the private sector afterwards, but he'll have better tools and skills to do the things that he was planning on doing. And if you decide to go straight to industry then a worst case scenario is that you have gained a few extra skills that you might enjoy having later on down the line if for nothing more than kicks and giggles.

I kinda agree with this, but I won't comment more on it because maybe you have good reasons and I'm not that concerned about it... It sounds like you are doing good research and if you truly love economics enough to survive grad school then it is worth it, but if possible then I wouldn't suggest constraining yourself to top 5/10 schools. There are a lot of students who deserve to be there that get rejected because there is randomness in the application process. Furthermore, there have been a lot of students who have gone to 20-40 ranked grad schools that have placed very well because of their ability to research and it sounds like you can do good research. So take that into consideration, but comes down to your choice. Also, you should talk to your advisor about this. He probably has a better idea of where he can place you than we do.

Don't want to step on anyone's toes, but I just want to go say that I don't agree with this... Stochastic Calculus, PDE, and SDE are all at least pretty useful (I don't know much about the setting of a mathematical finance MA, but the classes themselves are IMO gold). Especially for Macro. I'm a first year and today our professor said that if we took 1 class outside of the economics department that it should be stochastic calculus, he is extremely well-known within economics and I am planning on taking his advice. I agree with everyone else in general though. You'll want more math if you want to do grad school. Not even just for signaling purposes, just to survive. It is a rough 9 months.

The answer is yes, you can get into a top 25-50 PhD program. Make sure that economics is something you want to do though. A PhD in economics is challenging although very rewarding. Can you do intermediate micro/macro during the summer? That might give you a slightly better taste of whether economics is something that you are interested in and could be passionate about (Grad school is much much more advanced than those, but that's when I decided that I wanted a PhD in Econ). One of the students in our program had only done 3-4 classes in economic prior to his enrollment and he has found his niche within economics. I agree with moneyandcredit that some exposure to some field classes is good, but if you don't have time for any extra classes pick up some academic journals and read. See if there is anything you like. After reading a paper, write a 2 paragraph summary of what the paper wanted to communicate. Also, the best thing you can do is go talk to some of the economics professors about going to grad school. See what suggestions they'll have. Ask about their research etc... If I were you, I would apply to a few masters programs and some PhD programs. If you have kind of an idea of what you're interested in or at least know that you want a PhD in economics then it isn't a problem to jump straight to your PhD (not to mention you don't have to pay tuition like you would for a masters). If you do think you need a little more time deciding then a masters isn't a bad thing either and the students with masters have a little more time to prepare than those without and it can show the first year Good luck. Best advice in all of that is to go talk to a professor in your econ department..

Depends on how rigorous your MA is and which classes are you taking. From what I understand (I haven't been in a MA, but talking to classmates) PhD level courses are significantly more difficult than MA courses. So depends on which MA classes and which PhD program etc...

Yeah, it is really hard to eat healthy haha. I've gained like 7-10 lbs since october, which is probably a mix of stress eating and unhealthy eating. I'm doing the same type of stuff now. I'm trying to do all the old homeworks from last semester during this semester so that the material stays fresh in my mind for quals. It is time consuming, but it is fun. Maybe other first years feel this way, I'll be glad when it is over, but jealous of the new first years in the Fall.

A book that covers real analysis relatively well and includes economic applications is Real Analysis with Economic Applications by Efe Ok. Maybe check that out.

It is a really great field to do your PhD and you'll likely have a job afterwards if you do well. Your background is going to be fine to get you into a great school. That being said, make sure that you would enjoy it. If there is any way for you to add/substitute a class then tack on an intermediate economics course next semester. Otherwise, make sure to read some current research papers and make sure it isn't just a random idea because it will eat the next 5-6 years of your life and if you don't like it then it won't be a pleasant experience. Just saw the post above me and I'm just going to go ahead and say that the post above mine is more detailed and gives good direction. Follow his direction.

Yeah, that's why I hesitate to recommend it because it isn't mainstream and they don't have anything posted about their placements (which could be laziness, but could also be a bad sign). Hope you find the right place. Cheers

Just came across this the other day (Home | www.css.gmu.edu). I know VERY little about this program, but I thought it looked kind of neat. Before recommending that you apply, I would seek their placement information (make sure to ask which dept they placed in addition to school). It looks like a cool idea, for everyone who isn't going to open the link, it is a PhD at George Mason University in Computational Social Science. I'm interested in ABM type stuff and mostly applications to Macro, so I chose to go the Economics PhD route, but if you aren't as set then this might be worth checking out. After I finish my time in hell (first year), I'm hoping to get involved with some of the CS professors and bring more collaboration between the two departments. I think there is some neat stuff that can be achieved as a mix between CS and Econ.

I'm looking at the scipy source for the generealized schur decomposition. The sort eigenvalues option is currently disabled only because it doesn't work well on windows. If you aren't running windows and you really need them to be sorted, it wouldn't be particularly difficult to turn back on. Matlab's schur decomposition was faster for small matrices, but scipy performed better for big matrices. I tested dot product and a few other things just now and numpy linalg operations were pretty much consistently faster. As far as the optimization routines go, that is the only place that I think Matlab might still have an advantage, but only in some cases. Worked on a project 6 or 7 months ago that had been written in Matlab. In order to get the optimization to work they had to call the NAG library in Matlab. The scipy optimization routines ended up being able to find the same answer as NAG. I will admit there is room for improvement in the Python optimization packages, but they have come a long way and I would be surprised if the difference were still several orders of magnitude (didn't test though). Also, if you really needed it to be as fast/faster than Matlab then grabbing the routines from Fortran and wrapping them into python via f2py or cython would be something worth learning. I think John Stachurski has the best quote in response to Matlab vs Python "Incidentally, why don’t I use MATLAB like most other economists? MATLAB is a very useful language, but it’s more limited than Python, and the design is starting to show its age. As the marketing people at MathWorks used to say, MATLAB is Fortran for the 1990s." So overall, yes what I said was probably a little of an overstatement, but as a general rule I think that python code can be made faster than Matlab code with vanilla numpy and if it can't be then there are packages such as f2py/numba/cython that can make the speeds closer to C which is certainly faster than Matlab.

Yeah haha, you probably didn't see this coming though; I do my programming/computation in python! Surprise :excited: And yeah, I'm not a big fan of Matlab either. Well, if you want matlab to stop being the norm in economics then you've got to join me :snowhuman: Just got a professor to switch because matlab wanted to charge him $300 to renew license and $200 for the parallel toolbox...

Agree with all the previous posts mentioning python. Python is starting to catch in economics. Tom Sargent and John Stachurski just put together a website Quantitative Economics — quant-econ 1 documentation . It is geared towards doing computational type economics and they chose to use python as their primary language (Matlab won't stay king for long in Macro haha). Everything that can be done in MATLAB can be done in Python (and faster if you are even just a little careful) There are some people working on developing good statistical tools in Python (see statsmodels and you can even call R from python if you wanted) and the package Pandas is fantastic at data handling. If you are going to start learning a programming language, I highly recommend python. Additionally, it helps to get involved with the Python community. If you are trying to do some statistical analysis and the tool hasn't been developed then write some code that will do it and contribute (besides being good for the community it is a great way to learn what is happening under the hood. Don't let things be a "black box").