COMPARE Columbia MFE vs Columbia Math Finance

[chrisd]

I am wondering if anyone can tell me about the obvious differences between Columbia's two programs: MS in Financial Engineering in the IEOR dept, and the MA in Financial Math in the Math Dept. In terms of coursework, how do they differ? Relative strengths/weaknesses? Which one would you say gives better prep for a quant career? And finally, how do recruiters view the two degrees?

Thanks.

 

[Yuriy]

Side by Side comparison of the two Columbia programs

 

[Yuriy]

I will check the individual course contents and report later

It might be a good idea to do this kind of comparison of the top N programs.

It is interesting to observe that Baruch has a couple of unique [required] courses not featured among the requirements at Columbia

 

[Vic_Siqiao]

I will check the individual course contents and report later

It might be a good idea to do this kind of comparison of the top N programs.

It is interesting to observe that Baruch has a couple of unique [required] courses not featured among the requirements at Columbia

is Columbia's MFE a one-yr program?

 

[Yuriy]

is Columbia's MFE a one-yr program?

Yes, Columbia's MFE is a one-year program. The other program can also be completed in one year.

 

[Andy Nguyen]

1 year because it's a summer-summer program with no internship in between.

 

[Yuriy]

I have collected course descriptions for the two programs above. But before I post my comparison, maybe we can also compare other programs as well. By "comparison" I mean description of the areas that programs span. For example, both programs have stochastic modeling component, but other components are covered in different depth.

What other programs do you think are interesting to compare? I can only think about CMU, NYU, Chicago, Berkeley, who else?

 

[DailyVaR]

Isn't Derman in the IEOR department? Nuff said.

 

[alain]

Isn't Derman in the IEOR department? Nuff said.

That means absolutely nothing.

 

[Andy Nguyen]

Saw a post by someone who claimed that he is a student in the MA program

I completed a course at CU last fall (part-time) offered through the math dept in quantitative finance. I interacted with many students from the MFE and I can speak to item #2 on your list.

The main difference between the MA and the MFE lies in the structure of the each program. The MFE seems more structured and organized. There are many more course offerings per semester in financial engineering, quantitative methods, etc... There are profs in the FE program who have proven track records in research... The MA program seems to offer many fewer classes and the profs are typically adjuncts from industry. IMHO, the MA program was hastily thrown together to make some money.

 

[EricFleming]

Derman is and yes, enough said.

that being said, those with Maths degrees are still marketable, assuming they have some programming ability or the hiring managers like the way they think.

 

[Yuriy]

Side by side comparison:

FM program is 30 credits, FE program is 36 credits.
FE program is 1 year

Similarities:

1. Both programs offer standard introduction to mathematical finance: Black-Scholes, lattice models.
2. Both programs offer stochastic processes courses which include "discrete-time Markov chains, poisson processes, renewal processes" - that some schools don't teach in FE/FM curriculum - among other things.
3. Both programs have statistical component in the form of inference, time series, regression; two courses in each program.

Differences

1. FM program has a separate course on Stochastic Methods in Finance, which I assume is another name for Stochastic Calculus. I do not see a course entirely devoted to Stochastic Calculus in FE program (they teach part of it in Stochastic Processes).
2. FM program requires numerical methods, FE program has it as an elective.
3. FM does not have any programming component, FE has programming as an elective.
4. FM does not have portfolio management, monte carlo simulation, optimization.
5. FM program does not have as many electives as FE program. One useful elective FM has is on Risk.
6. FM does not have Structured Finance, Fixed Income, Credit Derivatives, Foreign Exchange, Implied Volatility Smile and other electives as opposed to FE that has these and many others.

Overall, FE seems to be more practical and applied, where as FM goes deeper into theory and has less financial applications (electives are in categorical data analysis, inference, multilevel models).

 

[alla]

I graduted from the MA program. In terms of the electives - the above comparison is incorrect. 1) There is a good class "Statistical methods infinance" offered, where you can can learn about portfolio management and optimization. And they obviously teach you Monte Carlo simulations there.
2) "Stochastic Methods in Finance" is not another name for "Stochastic Calculus". They are 2 different classes and both are required for the digree. You take the "St. Calculus" first and learn the math behind it. And then in "St. Meth. in Fin." you learn how to use stochastic calculus for pricing and modeling in financial industry.
3) It's useless to compare coursework in my opinion, since you can always take any class for the FE program as elective being the MF student. You can also take up to 1 class per semester in business school. The same about programming - we had many students taking programming classes.
4) One problem with the MF is that it is only 2 semesters vs, FE 3 semesters - you miss an important for your future career summer internship. Although if you're lucky enough to get a good job without an internship on your resume, you end up saving money for tuition and half a year of your time.

You said there is Derman in FE? We have Karatzas there, not bad either.
In general, the major difference is definitely not the coursework, since as Columbia student you can pretty much take any classes from many schools - SIPA, CS, Statistics, Engeneering. It is supposed to be only 1 year - you have only that much time. However if you think about it, you can stretch your program in 3 semesters, take more electives in whatever field you want and these 2 programs will be pretty much the same. And you can even get the internship in summer - employers don't really distinguish which program you're from - MF or FE. The major difference in the kind of people admitted to the program. In FE you will mostly find fresh graduates in their early 20s - it is a full-time only program. In MF there are many working students - all core courses are offered at night and can be taken part - time, they are older and don't remember any math (to be precise there is about 50%-50% mix of school graduates and experienced people).
I heard NYU has the best program of it's kind. (And the difference mostly comes form the career center). Remember, whatever program you choose, much more depends on you personally, rather that the program you're in.

 

[menovive]

@ alla : you are right. In their website they mention that 3 elective courses can be taken and If you dont want Linear Regression then a total of Four electives can be taken from Eng or B School. I think MAFN is more flexible to attain the necessary skill for a student.

"5. Elective. Can be Business School course with Business School approval, can be Economics department or SIPA course 4000 and higher, and any Math, Stat, Applied Math, Engineering course with course number 4000 and higher. Research seminars can not count as electives."

 

[Andy Nguyen]

Someone just left a comment on our review of the MAFN program
http://www.quantnet.com/review-columbia-mafn-program/

 

[Andy Nguyen]

More comment from the above link

Columbia MSFE isn’t too good on career support either. It all depends on the state of the financial markets at the time you are seeking employment. You are referred to general career services which supports undergrads and others. No one at Columbia is there to help you in the job search at all. They think the Columbia brand is enough and it is all on you to sell yourself. Getting interviews is a real challenge- whether you’re MAFN, MSFE, MSOR (finance). On-campus recruiting has two streams- undergrad (analyst) and MBA (associate). Quant masters doesn’t fit into either stream so there really isn’t a structured recruitment process.

Maybe my perspective is biased because I went to Columbia from 08-09 which was one of the worst times in history to be an MFE or MBA student with S&T/asset management interests. Hopefully the job market has recovered to some extent. I agree with the sentiment that among most Wall Street recruiters, Columbia has a bigger brand than CMU or NYU, just because of the overall prestige of the university. That said, unless you are French with prior internship experience, or American and already work at a bank/fund, you’re going to need some career help. CMU MSCF and NYU MathFin have much better career support. Columbia is like Chicago where staff doesn’t provide enough placement support to masters students.

On a bright note, most Columbia grads eventually get good jobs because they are smart and ambitious. Top MAFN grads are partners and executives at AQR, Bridgewater, Goldman Sachs, etc. (MSFE is a newer program so most alums haven’t reached the highest levels yet).

 

[KPC]

I graduated from the MAFN program, thought i'd shed some light on this topic. Comparison of the courses
1. Stochastic Processes in MFE is no match for Stochastic Processes which is taught in MAFN(This is the most essential course). Students from the MAFN program are generally very good at stochastic
2. Data Analysis in MFE sucks( course is very superficial, they are taught ARMA in 1 day, ARIMA in another day ARCH and GARCH few more days and this course also covers Linear Regression, In a nutshell MFE students dont learn anything) , Time series in MAFN is a very good course. The course covers material from time series theory and methods. I did the course under professor richard davis. He is amazing and he covered all the stuff in great detail.
3. Linear regression is covered in full detail in MAFN while its covered superficially in MFE in the data analysis class.
4. Intro to MAFN is not a great course. I'm pretty sure that course is better taught in MFE, this course is a very rudimentary course and covers the material from HULL(first 17 chapters). Most students entering the program are already familiar with the content in the book, even if the students havenot seen the material before they should be able to tacke it without any hassles.
5.Continuous Time Models in MFE and Stochastic Methods in Finance in MAFN are pretty much identical.
6. Barring the above courses there are absolutely no good courses in the MAFN program. At this juncture a student can take good courses in MFE and learn more than an identical student from MFE or take fraudulent seminars in the MAFN department and learn nothing. This is probably the best part(could be the worst part too) about MAFN, You can take all the good courses.. do 12 courses and pay 40k usd than paying close to 65k usd in MFE.
7. I took Implied Vol Smile and Monte Carlo methods in the MFE. Implied Vol Smile is probably the most practical course and the best course that i did at Columbia apart from Stochastic. Professor Derman is Fantastic and he provides great intuition. This course is a "MUST TAKE". Monte Carlo Methods is a good course but i think the material covered in the course is pretty straightforward and one can learn it in less than a week.
8. I also audited QRM and personally i dont like risk management, its a very boring topic and i regret not doing term structure models.
BottomLine : If you are good and if you have prior programming background then opt for MAFN. You will definitely learn more if you take the right courses. If you are'nt smart then both MAFN and MFE will be hell but MAFN will cost you less :P

 

[falcon]

Thanks, KPC!
I will most likely be taking MAFN for the upcoming Fall, so your post has been very helpful and encouraging.
Could you also comment on the career prospects at MAFN, provided that you have programming background and take the right courses like you have done? I'm coming from electrical engineering background, and thus pretty confident with MATLAB, C, C++, Python, Java. How tough is it to get that first interviews?

 

[KPC]

There are lot of people who tell me that MFE is more applied ..more practical more finance related. I say its non sense. If one is aspiring to be a quant one needs to be very very good at Math. More the math the better the program. (I'm assuming that the student has an adequate programming background which is a MUST).....In addition to the math and programming all you need is Professor derman's Implied Vol Smile. Professor derman will give you all the finance knowledge that you need. BottomLine: Take MAFN take Vol Smile , term structure models in MFE and learn monte carlo or audit the course ! and there is also a lot of controversy surrounding whether professor Karatzas teaches at columbia or not. He is very much with columbia but he only teaches PHD and Advanced PHD courses. I'm actually considering doing Probability theory under him in the coming fall !!

 

[Andy Nguyen]

Thanks a lot KPC
I hope incoming or students interested in the Columbia MAFN will take good notes of your suggestion.
Can you talk a bit about career services there? Is there anything the program (not the university) providing to help their students with career preparation? I have a quick question here and some good info is coming in.
Is the 9 month duration of the course a big problem with students trying to find internships and jobs?

I hear a lot of people mention that IEOR dept is a lot richer than the Math dept so they can afford to hire dedicated career staff, practitioners, etc while the Math dept seems to be run like a traditional math dept. What is your take on this?