NBZ

If you are going to write that kind of letter, it is really better just to refuse to write one in the first place. Perhaps that professor may have thought, "But the point of the recommendation system is to provide honest and unbiased assessments." Maybe, but the vast majority of such assessments (I suspect) are truncated below at the level of "This is a competent candidate who is diligent and has shown good aptitude for coursework." So if you're going to go below that, it's going to send a massive negative signal. The adcoms are not just going to think that the candidate "will likely not succeed" (that is after all true for many people), the adcoms will think "this candidate has absolutely no chance of succeeding."

Not once he gets in, certainly. But will it help him bump up from T15 to T10 (say)? For theory I'd imagine it's even more important to go to a higher-ranked place. (Although I must say it'd be ironic if a potential theorist needs to do an empirical RA job to get into an econ program from which he can place into a theory job).

Hmm the limit of 3 letters is strange. All the sunk cost is in writing the letter, and once they actually write the letter it is not a big deal to send it to 10 schools vs. 3. I hope for your sake that agreeing to send only to 3 schools does not also mean they don't put much effort into writing the letter. On your questions: 1. I would spread A and B over your top few choices. Basically you want your best set of letters for the highest-ranked school for which you have a decent shot. 2. I would still go ahead and apply (but then again it's because I believe it's best to err on the side of applying to more places than less). But can't you get any one of A, B and C to send 4 letters and then you have solved your problem? 3. Don't send a 4th letter of recommendation! This is the one question of yours which I think has a clear unequivocal answer. The consensus on the forum is to stick with 3 letters, unless you have a strong reason to use a 4th (i.e. they can speak about a dimension of your abilities that the other letter writers cannot). It's especially true in your case since the quality of the 4th letter is in question.

I agree with the advice about applying widely and to a lot of places. This would be a really strong profile if not for the undergrad GPA, so you should expect a lot of variance in your outcomes. I'm not sure whether it's worth it to apply to any in the top 10, but I'd definitely throw in some 10-20 applications, and apply all over the 20-50 range. Maybe others can chime in on this, as I think this is a difficult profile in terms of choosing where to apply and which range to target most intensively.

This obviously will vary across schools, but although my school has the same official requirement (B- in core courses), there have been cases where the student gets a C+ in a core course, passes the quals and the department turns a blind eye to the GPA requirement. If you fail the quals, on the other hand, there is almost no coming back.

If you are from a top 2 masters program in India, I imagine your program has a fair bit of experience sending candidates to PhD programs. So if you haven't done so already, talk to your professors about this. It will be hard for posters on this forum (except those familiar with Indian schools) to give a good answer to your questions. (Anecdotally I can think of people who had middling grades in their early math courses but compensated by doing well in more advanced math courses; but that seems different from your case).

It also depends on the type of profile. At the risk of over-generalizing, here's what I think: -If you have a solid undergrad record that's competitive for top 30-40, I'm not sure what difference a masters might make. -If you have a patchy undergrad record, getting good grades in a masters program could improve your chances at the margin. -But if you are missing key math courses, it'd be better to focus on remedying that directly rather than jump into a masters. -If you have a good undergrad record but from an unknown school, doing well at a better-known masters program should reduce the uncertainty adcoms have about your profile (which should mean less variance in your admission outcomes). -If you have a strong undergrad record but little research experience, then getting an RA position will help more than doing a bunch of additional courses at a masters program (you've already shown you can handle coursework).

Most schools will want to know your undergraduate GPA. Top 10% of the year is nothing to sneeze at - what were your math grades like? Assuming they are decent, you should keep the applications in the 10-30 range but add a few more 30-50 programs.

OP, I was in NUS too and can corroborate what cheateauheart is saying. Feel free to take a look at my profile (http://www.www.urch.com/forums/phd-economics/149607-schools-should-i-apply.html). As you can see I had nearly all B's in my math courses, and fully expected to be shut out of T30, but ended up with a T20 admit: so don't be too pessimistic (much of the regular advice on this forum is geared towards applicants from US schools where the grading policies are different). I think in your case you still do need a strong masters performance because there is a lot of variance in your grades, but you still have 3 years (final year at NUS + presumably 2 years of Masters) to put together a consistent string of good grades in upper-level courses. Some specific points: 1. You should take the graduate mathematical economics class and try to score well in it (it was comparable in difficulty to the Analysis I course when I took it). Taking other graduate courses at the PhD level might not be a bad idea either. 2. I'd only suggest taking higher-level math courses if you're confident of getting A- or better in them. Despite all the qualifiers about the rigor of NUS and whatnot, a B in let's say Real Analysis III still hurts you at the margin. 3. I was told by an NUS professor that the coursework masters at NUS is not at all useful in preparing for a PhD. It may be different with the the research masters - but talk with your professors and find out for sure before jumping into the program. 4. I reckon a European/Canadian masters is the best option for you at the moment. I don't know enough about these programs to advise you on where to apply, but in your position a) it would be nice to have the flexibility of taking and acing a couple of math courses at that university just so that the adcoms don't view your current math grades as some kind of red flag b) I notice that you are a little short on research experience, and that seems to indicate going to a program where you are required to write a masters thesis. 5. When asking your NUS professors for LORs, consider asking them to indicate your rough rank within the cohort (like top X%). There will be adcoms unfamiliar with NUS and this information may help them put your grades into context. 6. Once you do end up applying to PhD programs, apply widely. With a background like yours, you should expect a lot of variance in admission results (not all of it bad).

This is very sound advice - but I did not really believe it when I heard it last year. "Everyone says first year is hard," I thought, "So why not make it easier by preparing in advance?" So I made a half-hearted attempt to study math during the time between applying and starting math camp. In retrospect, whatever time I spent studying then (which thankfully was not that much) was probably completely wasted. The truth is it's much easier and much more efficient to learn once you're actually in the program, surrounded by your peers and professors and with every motivation to master the material. I'll say first year isn't really structured to favor students who "learnt stuff in advance" - there will be so much that is new that the comparative advantage from doing so is rather small. It seems much more important to come in fresh (and not burnt out from a summer of studying) and ready to get into the flow of things. If you really feel like you need to spend some time preparing, then it might be a good idea to read up on research and work on developing one or two of your own ideas (you will find very little time to do that in your first year, and I feel it's nice to develop the habit in advance).

I went to the Maryland ARE visit day last year. My impression was actually that the two fields the department was most focused on bolstering were development and environmental. Many grad students were doing development and there are supposed to be good opportunities for doing field work and working with the many institutes in DC. The downside (for me anyway) was that most students placed into DC institutes and there have been relatively few academic placements - though it may be that students who are less interested in academic placements self-select into the program. I'd suggest talking to someone from the department (e.g. Kenneth Leonard) if you want to find out more about development econ at UMD ARE. Good luck with your choices!

letemps, you might want to check out last year's Application Sweat Thread, I remember some people posted about what their UVA interview experiences were like.

Well of course I know the practice is common. But this is the first time I have heard a convincing argument in favor of manipulating your econometrics to drive home an agenda. I am going to have to rethink my entire research philosophy now.

I want my research to have social impact. I guess I'll just round up my R-squared to 1....

I should add that most of the "inferences" I made based on what I saw on gradcafe (e.g. someone got admitted to University of X 5 days ago! That must mean I haven't been accepted!) turned out to be wrong. It's so very tempting to feverishly check gradcafe during early February but for me it was basically a waste of time.