Derivatives trader looking to jump to academia/quant PM role

[bumaye]

Hi all - please excuse the erudite nature of this post. Have done a good deal of scouring, and had a few idiosyncratic circumstances I'd love to glean some perspective on.

I'm currently an interest rate derivatives trader at large investment bank, four years out of undergrad (econ/social sciences dbl major). I largely trade a proprietary/discretion-based book (i.e, make PnL based on fundamental/technical theses/trade ideas). This style of trading rankles me for two reasons: a) it's incredibly stressful, and b) I believe the world of proprietary trading is moving away from irrational human/discretionary traders sitting in individual silos and ultimately toward a quant/algo structure. What I love the most about my role is the macroeconomic nature of the markets I follow, and being able to formulate large scope theses and allocate assets/risk based on this. Still, I want to be in a role where I'm looking at the big picture more quantitatively and systematically. Moreover, I have a very strong/nuanced understanding of the markets, pricing mechanics, etc. The caveat however: my math acumen is limited to calc (1-3), linear algebra and ordinary differential equations; i.e. my college-level classes, with very limited statistics and almost no programming experience. Hence, I'm considering a quantitative M.A at Columbia to fill in these gaps. With this in mind, I'm trying to jump out of the world of discretionary prop trading to do either of the following:

a.) Become a pseudo discretionary/systematic/quantitative portfolio manager/asset allocator: I want to be able to merge my macro strength/understanding of the markets/global economy with a more nuanced quantitative and systematic approach. I don't think I would be the Physics PhD-type (whatever that implies) given the amount of my career (4 years) I've devoted to trading the markets and developing idea expression through asset allocation, but I'm deeply intrigued by the process of creating/developing the fundamental inputs, technical signals, and data points into a robust algorithm that can then be tweaked and implemented. Perhaps I wouldn't be the one creating the algo from scratch, but I'm interested in understanding an algo's ins and outs, having some sort of say into the flags and inputs that drive it, and ultimately finding the ideal asset classes/trades to express the algo's findings. My programming is limited to creating pricing models in VBA, but I'm thoroughly interested in programming and developing a strong skillset therein (whether it means learning Python, C++, etc.) Does such a role exist in a quant fund? A birds eye, top-down macro-focused synthesis of a trader/portfolio manager and a quantitatively oriented investor? Given my interest and end-goal (a mouthful), and my relative lack of advanced math, what type of quantitative M.A would you recommend (if any?) Statistics, mathematics of finance, MFE? I would basically be doing the M.A while working (so 2-3 years hopefully no more).

b.) Jump into finance/macroeconomic academia: as stated above, I have a strong passion for financial economics/macroeconomics and political economy. In a different route from the above, I would ideally pursue a career in finance academia as an end, with the means being a PhD in a top 10 b-school/finance program. With this in mind, which type of M.A program do you think would give me the math/stats coursework needed to be a competitive applicant to a top finance PhD program? Stats/MAFN/MFE?

Apologies for the lengthy nature of this post - just most of my career anxiety condensed for the first time into written form.

Thanks!

 

[Yike Lu]

(a) and (b) could not be more different. (a) does exist.

It's hard to point to a specific degree that gets you everything you need, as none of them in my experience do. The closest for (a) is probably MFE and for (b) stats, but there's going to be both (1) a significant amount of extra self study (2) a significant amount of coursework that isn't applicable in either case.

 

[bumaye]

(a) and (b) could not be more different. (a) does exist.

It's hard to point to a specific degree that gets you everything you need, as none of them in my experience do. The closest for (a) is probably MFE and for (b) stats, but there's going to be both (1) a significant amount of extra self study (2) a significant amount of coursework that isn't applicable in either case.

Thanks so much @Yike Lu - as far as significant coursework for (a), anything in particular you would suggest? Also, any reason you would recommend MFE over MAFN? I'm erring towards MAFN just because it can be taken part-time in the evenings, while MFE needs to be full-time. thanks again -

 

[tfors]

My background is not too dissimilar from your's, allthough your's seems better both academically and professionally. I decided to start all over from undergrad-level (I attended liberal arts college, hardly anything quantitative), doing CS now . Really loving everything I am learning, still not sure whether I will be going for a PhD or just for a masters. To be a competitive applicant for a PhD in general, better focus on learning theoretical subjects in depth as opposed to a more practical MFE, imo.

 

[Yike Lu]

The extra stuff you want to learn is along the lines of applied math like statistics, optimization, software engineering, ie the data science toolkit. But it's not very useful without trading / finance / quant knowledge. The stuff that isn't so useful from an MFE (to me at least) is the exotic pricing stuff, deep risk management.

MAFN vs MFE -- it looks like you are set on Columbia? To me MAFN and MFE are same category for most schools. For Columbia, MFE seems to be the more prestigious of the two programs, but I don't have that much information on either. I personally recommend Baruch MFE (where I did my MFE), which has both PT and FT options.

 

[mhy]

Thanks so much @Yike Lu - as far as significant coursework for (a), anything in particular you would suggest? Also, any reason you would recommend MFE over MAFN? I'm erring towards MAFN just because it can be taken part-time in the evenings, while MFE needs to be full-time. thanks again -

CMU and NYU also offer part-time programs. You should check them out.

 

[TBG]

Thanks so much @Yike Lu - as far as significant coursework for (a), anything in particular you would suggest? Also, any reason you would recommend MFE over MAFN? I'm erring towards MAFN just because it can be taken part-time in the evenings, while MFE needs to be full-time. thanks again -

I totally get the whole "prefer part-time" thing because that was what I ended up doing myself. When I was considering the different programs and speaking to a bunch of colleagues and friends who were alumni of CMU/Columbia/NYU programs etc.. The overall takeaway for me (focusing just on PT offerings in NYC) was - CMU PT would be the most rigorous and most prestigious.. the quality of the quants coming from the program is top notch. But it's also going to be a solid 3 years of work without much flexibility in the program. It's also a lot of being on the NYC campus and watching videos streamed from Pittsburg.. Though the professors travel to NYC here and there to teach in person but it's just not the same as in person learning.

Moving on to Courant PT - also very reputable and solid program.. Courant is one of the top, if not THE top applied math department in the states.. But having gone to stern for undergraduate I just didn't want to go back to NYU again / wanted more diversity in my education/resume..

Baruch PT - the class size is tiny.. Which is why admission stats are low and very competitive. I hear great things about the program in terms of the stuff you learn and the kind of focused attention from the professors. But I'm a sucker for brands and I just couldn't / didn't want to go to baruch having come from Stern..

Which left me with Columbia MAFN. The program is good, I'm learning a lot. Though unless you scout out for programming focused electives, you wouldn't necessarily graduate with a solid programming skill set. But it's possible to take cross electives in Computer Science... in Statistics... etc. And it's the flexibility that I picked this over everything else. Also, it's possible to graduate in 1.5 years while doing it PT Doesn't have to be a 2-3 year drawn out process