# Quantitative roles in finance?

1. Oct 7, 2012

### 11thHeaven

What are the different kinds of quantitative/math heavy roles in finance? I'm doing an undergraduate maths degree in the UK and beginning to think about my eventual career options. I know that finance as a field is one where a lot of maths/science graduates end up working in, but I really don't know much about the industry/area.

I know this might be common knowledge around here so I apologize for my ignorance! I just don't know where to start. I type similar searches into google and only seem to come up with specific job ads and nothing in the way of general information.

I realize this is a very broad question so I'm not asking people to answer me in entirety - just pointing me in the right direction to somewhere I could learn more about this stuff would be extremely helpful. Thanks :)

2. Oct 7, 2012

### meanrev

If you want more detail, something like Kuznetsov.

I'm not the best reference for this, since I'm technically not employed; I went on an entirely different path and quit college to start a quantitative trading firm. (So far, so good.) However, comparing the work I'm doing to some of my college/grad school peers' experiences in the field, I'll say that I'm not doing anything particularly different from other prop trading firms (e.g. Jane Street, GETCO) or quantitative hedge funds (e.g. D.E. Shaw, AQR). Most of the mathematics is not cutting-edge - if that's the sort of challenge you're seeking, I suppose you want to work on the pricing side.

If you're not taking a drastic path as I did, I'd recommend you keep your options open as the job climate is changing rapidly, due to (1) regulations and (2) tighter competition. (My college buddies commented that the number of recruiters for financial firms have doubled this year, with names such as Blackedge, Chopper Trading, and guys developing iPad apps to simulate order book dynamics and practice for bid-ask spread scalping etc. Feels like a bubble environment, if you asked me.)

I've also noticed that >90% of the opinions you can find on the internet fall short of what you'd find out from actually working in an internship. It will be harder to land one with a bulge bracket investment bank if you are doing an undergraduate math degree (for some reason, they explicitly prefer econs/business majors), but easier to land one with a trading firm. Now is the time to start applying. Just be careful to avoid scams; some call themselves "prop trading firms" but are really what we call "trading arcades". Also, be very good at classic probability challenges (https://www.amazon.com/Challenging-Problems-Probability-Solutions-Mathematics/dp/0486653552), useful to get you through the interview.

More importantly, get lots of practice in programming.

Hope this helps.

Last edited by a moderator: May 6, 2017
3. Oct 8, 2012

### twofish-quant

A good snapshot circa 2007 is

https://www.amazon.com/Complete-Quantitative-Professionals-McGraw-Hill-Investment/dp/0071468293

It's a good book that describes the situation right before the world blew up. As time passes it's going more and more out of date, and if you are an undergraduate now, I really couldn't tell you what jobs will be available in seven years.

4. Oct 8, 2012

### twofish-quant

One problem is that anyone that is working on anything that is cutting edge isn't in a position to talk about it in public.

But regulations can be a good thing for jobs. When a regulator wants numbers to convince them that they shouldn't shut you down, you give it to them, and if it turns out getting those numbers involves hiring lots of people to babysit supercomputers, then so be it.

One difference between today's world and the description of Kuznetsov is that today you can't breathe without asking for permission from some government official.

Getting a physics/math/statistics Ph.D. is another way in. Also, getting a non-financial programming job and then switching to finance as a lateral career move is another path in.

One other thing about interviews is that they are psychology tests as much as skill tests. If you answer a question, the interviewer is going to give you a harder one. And the skill level is going to increase until you are at your breaking point. It's a psychology test to see how you react when you are pushed to your limits.

5. Oct 8, 2012

### meanrev

True, I wouldn't disagree.

It's widely known that one of the most passionate (and talented) mathematicians of our generation abruptly left the PhD program at Harvard in his last year to join Renaissance Technologies, if this gives the rest of us hope that there is holy grail material in algorithmic trading.

6. Oct 8, 2012

### 11thHeaven

Thanks for the replies guys, they've given me plenty of food for thought.

I've been told that securities and risk are areas of finance/banking which can be quite mathematical - would you guys say (in your experience) that there's much merit to this statement?

I've just started my first year and still settling into uni life etc. so I'm not so keen to start applying to internships just yet. I know getting into a spring week/summer placement during the first year is a great advantage and so on but I just don't have the energy for it yet(I don't even have a proper CV, much less anything impressive to put on it).

7. Oct 8, 2012

8. Oct 9, 2012

### twofish-quant

One thing about those slides was that they were written in 2005. Obviously there is something fundamentally wrong in the those slides or else the world wouldn't have fallen apart. A lot of what people are doing right now is figuring out what broke, and to figure out how to keep things from breaking again.

One of the fundamental problems with pre-crash risk management is that they assumed that assets were fungible. You have $1000 worth of gold, you can convert that into$1000 worth of Treasuries or $1000 worth of cash. So when you do your equations, you just assume that you have$1000 worth of "stuff".

The problem is that you may have $1000 worth of gold, but if people are demanding$1000 worth of cash, and the mechanism to convert one to the other have broken, then you have a big problem.

There's also the problem of "time". What ended up happening was that the Federal Reserve ended up saying "give us your stuff and we'll convert it to cash." The trouble is that this may take a few days, and if you have a crisis then you need to make sure that you have enough reserves to get you through a few days while people are scrambling to figure out what is going on.

Finally there is a problem with feedback. If you look at those models, you are in the market, stuff happens in the market to you. The trouble is that in the real world, you will react to the market, and then your reaction will cause the market to change.

Just to give you an idea of what physics Ph.d.'s do in investment banks. Someone hands you 200 pages of equations, and your job is to figure out what's the problem with those equations. Gaussian copula models misused were one of the key things that caused the world to blow up.

9. Oct 9, 2012

### twofish-quant

There are parts which are mathematical, there are parts that aren't. There are rather few jobs that are extremely mathematical, but there are even fewer people that are interested in doing math.

One reason I like the field is that it's not "pure math." There are elements of physics and even elements of politics all mixed up. For example, on those slides, you have statements like "assume that you have stocks that are correlated in X way, then crunch the equations and ...." At that point, the physicist in me says "whoa there, how do we know that those stocks are correlated in that way?" You could reply "we'll historically stocks behave like that" and then I'd ask so what reason so we have to think that stocks will behave in this way and what happens if they don't?

Then the politics comes in. If you figure out that these equations are just fudge factors to give people the answers they want so that they can pay themselves bonuses, then you have a problem. At that point if you are lucky, you work somewhere that thinks this is a bad thing, and you do something about it. If you work somewhere where you get into trouble for pointing this out, then ethically, the only thing you really can do is to resign and want for the world to blow up.

After the world blows up, you'll have a ton of government officials trying to put into place processes to make sure that people just don't make up random fudge factors. This means that if you write an equation, you will have to be prepared to justify every factor and assumption that goes into it. Conversely, there are lots of jobs for sadists whose job it is to take any equation anyone proposes and bunch holes in it.

10. Oct 9, 2012

### twofish-quant

One thing about some hedge funds is that if you want to work for them, don't get a job in finance. It's common among some hedge funds that they hate hiring people with any finance experience, and having anything remotely financial in your background is the kiss of death. If you want to work for some hedge funds, skip Wall Street, get a job at Google or Facebook.

Also, proprietary algorithmic trading is inherently zero sum. There's only a limited amount of money you can make, and as more people do it, the markets become faster and more efficient until "something bad happens" and lots of people think that we are near or past that point. What will happen is that people will put in rules..... And then people will figure out ways of making money with those rules..... And some of those rules are informal. I got into a conversation once with someone that was doing algo trading, and the topic ended up being on when brokers eat lunch and how that impacts market behavior (and he showed some slides showing that yes that matters.....)

Here is a good readable explanation of what is going on...

http://www.lrb.co.uk/v33/n10/donald-mackenzie/how-to-make-money-in-microseconds

The other thing is that a lot of this stuff is behind the scenes, and it's fun to pull away the covers. You go to your favorite online broker, buy or sell a mutual fund, and it all works. However what happens is that your broker batches up the orders and then goes to their broker, and their broker uses an algorithm to do the sale.

Last edited: Oct 9, 2012
11. Oct 10, 2012

### twofish-quant

One other thing. People make the mistake that the only job for a physics Ph.D. is a physics professor. Don't assume that every job in finance is the anything like working for Rentec.