2020 MSc/PhD in Statistics, Econ background

[redburned]

Hi everyone,
My goal is to apply for a good PhD program in Statistics (not applied) in EU zone or US
Profile:
Bachelor in Economics, cum laude, in a small Italian university GPA 3.8/4
Master's degree in Economics and Finance, ongoing, current GPA 4/4, in a famous Italian university + Erasmus exchange in a famous German Uni
Exams (both degrees): General Mathematics (Calculus I, II) B+, Intro Statistics A-, Financial Mathematics A, Statistics Theory A, Econometrics A+, Time Series A+,  Optimization A-, Game Theory A+, Panel Data A+, Probability Theory, Numerical Methods and Machine Learning (next year)
(and I have A+ in different microeconomics-based exams but I think they're not useful for my application, I don't know)
I won some merit-based scholarships
I'm planning to take the GRE in October
I have a ref from a stats professor and a ref from a strong finance professor (but I don't know if it is useful!)

Do I have any chance to get into a PhD program in Statistics at UCL, Imperial College London, Yale, Duke, Upenn or other Universities of this tier in EU/US?
Or I should do an MSc in Statistics first? If so, what do you think about the Statistics MSc (2 years) organized by Humboldt University of Berlin + Technical University of Berlin + Free University of Berlin? Would it help me?

Thank you for any help!

[redburned]

I just scored 161 V (88%) and 164 Q  (86%) on GRE, what should I do? Should I retake it? I know I can do better on quant part but I hate studying for GRE, I don't learn anything...
I update my list for US PhD: UC Davis, UCLA, UCSB, Penn State, Yale, Duke, UC Irvine, UPenn.
 Should I add/remove some uni? Do I have chances in getting in one of these?

 

[Stat Assistant Professor]

You don't need to retake the GRE.

For the UK and European schools, their PhD programs do not contain any coursework. And based on my familiarity with them and the statistics papers I've read from academics there, their research focus tends to be extremely theoretical (these schools' Business and Economics departments publish the more "applied" statistics papers in applied statistics journals, but their statistics depts seem to be very heavily focused on theory). So for the top schools there like UCL, Imperial, etc., i imagine that they might prefer to enroll students who have more extensive mathematics backgrounds -- especially since their doctoral programs don't have any coursework to bring you up to speed, and your background may fall short. To have a shot at these, it may be advisable to do an MS in statistics or mathematics.

I think you have a shot at some U.S. Statistics programs. Yale and UPenn might be difficult since these programs admit very few students a year and have a preference for those with very strong math backgrounds. But the others on your list seem plausible with your background (Duke might be a bit of a reach but I wouldn't say it is impossible for you to break either).

Edited September 22, 2018 by Stat PhD Now Postdoc

[bayessays]

Have you taken linear algebra and real analysis? The bar is high for international applicants, so your math background will keep you out of those US schools.

[redburned]

Thanks a lot for your replies ?
@ Stat PhD Now Postdoc: I see, then I save my money from applying to UK universities. Many thanks. Could you suggest me some other good schools in which I could have a shot beside the ones I mentioned? 
@ bayessays: In the General Math for Economics exam there was also Linear Algebra (15h) in which I studied stuff like vectors, norms, matrices, determinants, ranks, matrix inversion, linear eq systems, cramer and rouchècapelli theorems, symmetrical matrices and spectral theorem, orthogonality, eigenvalues, eigenvectors, diagonalization, quadratic forms, symmetric bilinear forms. Do you think it's fine? Also, the Econometrics exam I've taken was very linear algebra intensive and based on Hayashi's textbook, very theoretical. For what concern real analysis I will study a glance of it in the Probability Theory exam (convex sets, hyperplanes, separation theorems, fixed point theorems).
Actually I have also studied differential equations, system of differential equations and some calculus of variations in the Optimization exam, maybe it is an useful information? 

Edit: Also, the Statistics Theory exam was actually a Mathematical Statistics exam (casella-berger as textbook)

Edited September 23, 2018 by redburned

[Stat Assistant Professor]

34 minutes ago, redburned said:

Thanks a lot for your replies ?
@ Stat PhD Now Postdoc: I see, then I save my money from applying to UK universities. Many thanks. Could you suggest me some other good schools in which I could have a shot beside the ones I mentioned? 
@ bayessays: In the General Math for Economics exam there was also Linear Algebra (15h) in which I studied stuff like vectors, norms, matrices, determinants, ranks, matrix inversion, linear eq systems, cramer and rouchècapelli theorems, symmetrical matrices and spectral theorem, orthogonality, eigenvalues, eigenvectors, diagonalization, quadratic forms, symmetric bilinear forms. Do you think it's fine? Also, the Econometrics exam I've taken was very linear algebra intensive and based on Hayashi's textbook, very theoretical. For what concern real analysis I will study a glance of it in the Probability Theory exam (convex sets, hyperplanes, separation theorems, fixed point theorems).
Actually I have also studied differential equations, system of differential equations and some calculus of variations in the Optimization exam, maybe it is an useful information? 

Edit: Also, the Statistics Theory exam was actually a Mathematical Statistics exam (casella-berger as textbook)

You may want to ask your letter of recommendation letter writers to explicitly point out your real analysis and linear algebra backgrounds in their letters (which seems more than sufficient for doctoral study in Stats), so that there is no doubt in your application. Your LOR writers should also point out that the statistics class you took was based on Casella and Berger and that you took advanced probability theory. You can also mention this in your statement of purpose or provide a description of your coursework in a supplemental document, but the letters of recommendation will be read more closely so this should definitely be mentioned in there.

It is true that there is a lot of competition for international applicants, but I think that mainly applies to Chinese and Indian applicants because of the sheer volume of qualified applicants from these countries. I do not think there are many applicants from Italy, and plus, you go to a famous university. I think adcoms would also expect an Economics degree from a top school in Europe to be pretty technically rigorous. Statistics schools in the range of 10-30 (for statistics only) and 20-60 (in aggregate USNWR rankings) are a good range to target, but you can also apply to one or two "reach" programs like Duke or University of Washington.

[redburned]

Ok then I will try to get these things written in my letters. Thanks a lot for all your helpful advices ?
@Stat PhD Now Postdoc