etoposido

What matters is how you are at the research stage.

The stars in my department are the result of really good and innovative ideas combined with a great execution. Our research will move the frontier by epsilon, the research of a star by 2*epsilon. Also, the execution of their research involves a lot of data work (a new dataset is a plus) and the ability to build economic models to rationalize what you find in the data.

This info is misleading as you are not applying to all those places. They include jobs in countries all around the world, and all available fields. The number of posts increase, but you will be applying to 150-200. That includes places hiring in your field (if you are a macro person don't expect consulting firms to be calling you), places where you actually have a chance (where your advisors feel ok about sending lors to) and places where you would actually go (I did not apply to countries I did not even know the language). Fill in 200 applications is already a hard task. It is physically impossible to do more than that. The number of fields has been expanding so it does the interest in those new fields. In those 200 applications, you are competing against new PhDs, Post-docs and current assistant professors, and that is not a minor issue. (Many place where I had flyouts ended up hiring Post-docs) And finally, there is an issue with the total number of jobs available because not all those posts are actually filled. Some places cancel their search because of lack of funding, some simply couldn't find a match (In general this occurs, when they aim too high and do not a have a back up plan). Some others are just bluffing, they interview people even though they are not planning to hire anyone (there is a reputation thing about not going on the market each year).

Hi, as I mentioned before, I was lucky. But several people in my cohort were not. Some had to adapt and get jobs in the private sector or Post-docs, something they were not planning to do. In fact, most of international students returned to their countries.

Ok guys, I was in your position 7 years ago. This year one of my students is applying to PhD programs and I am super excited for him, so I am visiting this forum again to see how the market is moving and to offer you my two cents. I applied in a complicated period (2007-2009 crisis) and I was rejected by every top 20 that I applied to. I had a couple of top 30s and accepted the one with funding. If I were in that situation today, I would think it twice. Let me tell you that the job market is becoming more brutal every year, and if you are in a top 30 and you are not a star, you will struggle a lot. For instance, this year there is a top university with almost 40 job market candidates. At the end, pedigree matters a lot. I had a really innovative job market paper, but I was not the star in my cohort. The number of interviews and flyouts I had was just a third of any below average candidate in a top 15. I got a job in a top 50 program, but I was very very lucky (it involves many people rejecting that offer, I was like their fifth alternative). And this is, because my program is focused on the median placement; it pushes almost everyone at the same rate on the market, perhaps the star a little bit more. We don't have a wonderful placement but we are all at least placed. Most of the programs at top 30s are more concerned about the top placement and you will see many candidates struggling a lot to find a job. You will see this strategic behavior also in a couple of top 20s. As basic as it sounds and as alarmist as it can be, if you are thinking about which program to choose, choose the one that is ranked the highest. Then you will maximize the chance of getting the job you want. I don't want to discourage any of you, at the end I have a pretty decent job, but this is how the job market process is going to be. I will probably check this forum a couple of times more, so if you have any question for me, feel free. No PM, let's try to make public all the info.

I am just going to give you my point of view: Berkeley dominates the other two in international economics, in particular International finance. However, topics like the ones you mentioned are closely related to what macroeconomists in Northwestern do (Although, people in Upenn have made a lot of progress in the area of the zero lower bound plus some guys from Minn. FED). In Columbia, you will also focus more on international economics with Uribe and SG and probably in price-settings (e.g rational inattention) since they have Woodford.

I think Duke is doing a great job trying to improve its ranking. Maryland used to be a good place for international macroeconomics, but since Reinhart, Mendoza and probably Vegh left the program, there would be only some junior faculty in the field. I don't know anything about BU, but it should be third in my personal ranking.

To my experience if you do macro and if you're interested in model of heterogeneous agents Fortran is a must. First, because it is in general faster than Matlab (we are talking about routines that may take weeks to achieve convergence) and Second, because you are gonna be part of the "Fortran community", so you don't necessarily need to begin from scratch your codes but take advantage of the routines in the web. I invite you to see codes from papers published in AER or AE Journal of Macroeconomics where those are publicly available. You will be really surprised to see that most of the people publishing in those journal are still using Fortran.

You forgot to mention that Maryland also has on international finance: Korinek and Sebnem Kalemli-Ozcan who will be joining the program this summer (according to what they said during the open house)

Georgetown.

I think the software you choose depends on your needs. Unless you want to do all your codes, you can take advantage of different software for different purposes (If somene else already coded and posted, you minimize the likelihood of making mistakes). Here my two cents about software and uses: If you want to do Dynamic Macro you should devote time to improve your skills in Matlab or Gauss, but better in Fortran since is faster than all those, so it'll save you time at some point. For metrics I also believe depends on, if your research focus on cross-section/panel data models Stata is a friendly package, but for time series there are some other better, like E-views in case you need to filter series it provides you with several important tools or like RATS and its cointegration module named RATS and MALCOLM for SVAR or SVEC. I also know the code for VAR Switching models is in Ox language. S-Plus has a lot of routines for Financial Econometrics really really useful. However, I also discover every day in R the most useful tool, it seems like for everything that I have to do (with the few exceptions cited above) there exist a package (I'm not gonna list, 'cause they're many). I'd just add SAS to deal with huge datasets, but it's a really, really expensive software. Obviously Latex in any version (Lyx, Winedt or TexnikCenter) is a must for any phd student........

I heard Duggan wouldn't be the only one leaving UMD.....that makes me wonder, Is UMD also in decline????.... :hmm:

It depends on, the most important part is what economics tells you about the relation among those variables....If Y is also I(1) and you think there is a long-run relation among them, a cointegration panel approach must be addressed (there are several methodologies PDOLS, PSUR or Groenn-Kleibergen), either be the case you can always include an I(0) variable in cointegration, just be careful....because then you already have a trivial contegrating vector (0 1 0 0). However if there is no long run relation you can differentiate your variables and use some of the PVAR techniques available, which are better known than their counterparts on cointegration. I hope this helps.....

I wonder, if the economic crisis has been reflected in the size of the class for 2010....I have little info to make any inference about this...The class size in Berkeley is 21 (Average), UCLA is 36 (Above Average) and UMD 21 (Average, too). Let´s share info and opinions...

You should verify first the existence of panel unit roots....after you do that, I can give you more accurate references of papers......