Does Quant MS/PhD Make Sense?

[Aj Kappor]

"Beware quants bearing algorithms. A key factor to the expansion of esoteric financial products on Wall Street has been the hiring of "quants," often doctorates in quantitative sciences such as math and physics, who built intricate formulas to profit from minute market "inefficiencies." The algorithms became increasingly divorced from financial reality, masked by complexity and greed." (Poular Press, US)

"Modern banking is about pure and unmitigated greed. What's new is that it is mathematical greed engineered by some of the finest minds of their generation, the so-called rocket scientists' and quants', who dedicated their intellects to inventing ever more obscure and complex financial instruments the better to promote personal enrichment. Trouble is, they were too clever by half as the banks have discovered." (Popular Press, UK)

"In fact, most Wall Street computer models radically underestimated the risk of the complex mortgage securities, they said. That is partly because the level of financial distress is “the equivalent of the 100-year flood,” in the words of Leslie Rahl, the president of Capital Market Risk Advisors, a consulting firm. The people who ran the financial firms chose to program their risk-management systems with overly optimistic assumptions and to feed them oversimplified data. This kept them from sounding the alarm early enough. There was a willful designing of the systems to measure the risks in a certain way that would not necessarily pick up all the right risks."(NYT)

"Is it possible that the quants were so buried under leveraged layers of derivatives and other exotic instruments that they didn't see the coming storm? This seems like a pretty big movement to miss. After all, if you're sitting on top of a few billion in debt that is on the razor's edge of liquidity, wouldn't you spend some time looking at it more closely, especially with such broad macroeconomic factors staring you in the face?" (Trade Press)

Any suggestions, tips, for those experienced technical graduate (non quant MS/PhD) professionals contemplating full-time quant MS/PhD in computational, mathematical, and quantitative finance?

Bigbadwolf...?

Possible answers... (multiple responses may apply)

1. Don't pay attention to all the negative press about quants...

2. Quant and computing skills gained can be used across other industries...

3. Financial services need such skills much more beyond quant roles...

4. This is as bad as it gets, things should turn around in 2 years (or 4)...

5. Stop thinking about further education in this area,...

6. There are other ways of doing hara kiri...

7. Anything else...

 

[bigbadwolf]

I've already made my sentiments clear. But also pay attention to more mature people like Andy who says it's difficult to look even one year ahead. With all humility, I feel that pursuing a degree in quant finance is too committal today: it's a sucker bet, with the odds stacked against one. But pursue a degree in PDEs/probability/numerical analysis, and then see which way the cookie crumbles. The idea is to retain flexibility so that if one can't become a quant, one can still teach calculus to morons in some community college.

 

[Navestax]

Maybe its time to look at the big picture, as they imply, instead on focusing solely on math.. Not underestimating other disciplines as economics, dont know...

just MHO.

 

[IlyaKEightSix]

People that think that numbers can ever replace reality are not only ineffective--they are downright dangerous. As Nicholas Nassim Taleb has said with an aeronautical analogy:

"Give a pilot a broken altimeter and he'll crash the plane into a mountain. Give him no altimeter and he'll look out the window."

This is why if I was running a quantitative hedge fund, I'd never hire some poor-English immigrant from an overseas technical school who doesn't know left from right or up from down in terms of rational (or the lack thereof) human behavior. I don't care how many derivatives you stack up on top of one another, in the end, that tower of derivatives is built upon the foundation of some real security, and if that goes south, then your tower of complexity built upon a thousand lines of code will crash down all the same. And not only did you get paid for X hours you spent coding BS, but you also just lost the firm X millions of dollars.

It's also why I subscribe to NNT's beliefs that "take a street-smart kid with a thick Brooklyn accent and teach him the technicals and he'll be a strong statistical arbitrageur. But take someone with a phD--especially in the physical sciences--and try to have them do stat arbitrage and they become reckless speculators..." or something along those lines.

I honestly think that the age of the "idiot savant" quant, as NNT describes them, as these people with no real world experience but amazing textbook skills, is over. I think that quants will still very much be in demand. With all of the raw computing power and all of the numbers whirling around Wall Street, those who can play the numbers and program the algorithms will still be EXTREMELY important. What I believe is on its way OUT is the idea that a phD replaces 5 years of in-the-trenches experience.

In fact, this is what I am banking on. I'd rather learn on the job and replace the phD with E-X-P.

Clearly, there are exceptions to the rule, such as Renaissance, but even King Simons's RIEF is lagging the S&P by 3%. Hopefully the old man isn't losing his touch...because we need someone to continue to be a crusader for MfA, and I don't have a seat on that board yet!

So no, I don't think that quants are going out of style. I just think that the job role will change drastically, requiring far more real-world smarts to actually formulate the foundations of the models, rather than "black magic" which, if anybody plays any sorts of role playing games, completely backfires if handled recklessly.

 

[Aj Kappor]

Albeit it is rare, some use a technical PhD to do a research fellowship, write up a doctoral thesis, teach at a university, launch a globally renowned technology company while making fundamental contributions to theory and practice in key areas of global practice. Again that may depend upon individual circumstances and happenstance and may not be feasible in all cases. The only reason for thinking about doing (first or second) PhD - in contrast to just a Masters - is to do serious work with the potential of making major long term impact. Also, after having spent some years in full time "work experience" one may realize that there is too much that keeps one occupied on a day-to-day basis to make any serious contributions to big things that may really matter. The current global devastation that has put the world in a tailspin and has caught everyone's attention has occurred despite long known 'common knowledge' that the fundamental paradigm about how risk is modeled or managed in financial markets had fundamental flaws. Therefore, if one can afford the opportunity cost of getting away from the day-to-day hassles to focus on serious and fundamental improvements - like rethinking risk management - (first or second) PhD may be one possible avenue to pursue. On the other hand, for those that operate in the manner stated above, opportunity costs could be significant - say to the tune of something between a half- and a million-dollars. Therefore, this pursuit is not for those interested in quick returns or returns that may be ever recouped. Hence, capacity for taking such losses also needs to be there besides primary interest in pursuit of what could contribute to the future foundation of global financial systems.

 

[IlyaKEightSix]

Well I'd think that you could learn on the job. Companies do have research teams (at least the good ones do) that do this day in and day out, and if you have an idea, I'm sure something can get worked out.

But for me, the opportunity cost of a PhD is just staggering.

 

[DominiConnor]

Almost all medical errors are caused by qualified doctors.

IlyaKEightSix cites NNt's pilots, but again almost all plane crashes have a fully qualified pilot at the controls.

Is the answer to plane crashes to put unqualified people at the controls.

Or for medicine to hire nice , but stupid people ?

The second is as we all know coming to pass. Many people are choosing homoeopathy, herbs, crystals and faith over medical science. They "feel more comfortable" with these bogus cures, and homoeopathic "remedies" proudly boast that they have no side effects, so they're safer right ?

Even if it costs 500 billion to fix this mess, the taxpayer will still come out ahead.
The growth caused by allocating capital more efficiently swamps the 1-2% of GDP that is being cited. Credit derivatives can easily be shown to have helped this growth.

Even if I'm hopelessly optimistic and the whole 700 gigabucks just vanishes, the US economy will still have made a net profit.

Retreating to "faith based finance" is not the solution.

 

[IlyaKEightSix]

You're absolutely right, Sir Connor, but that begs the question...

What defines a "qualified" quant? A qualified pilot learned how to fly a plane. A qualified doctor went through medical school. However, how do you define someone who's "qualified" to do well in the markets?

Doesn't that beg the question that if there *were* a formulaic way to be qualified to be assigned these high-risk high-reward responsibilities that wouldn't everyone be making a profit who could get their hands on these "qualified" individuals?

Which also leads me to another question...if the model of the "pure" quant otherwise becomes a professor in a highly quantitative field, why is it that the big firms go and specifically HIRE PhDs to do research? Why can't they outsource their research to respectable quant finance colleges or individuals? What if the big firms would put quant professors in their pocket and have their research be done by the professors as their research field? I know here at Lehigh, we have one "technical" quant (mix of genius/programming), who happens to be my favorite professor, and one "pure" quant--the professor that teaches stochastic calculus (which I didn't get to sit in on! Rargh!). What keeps places like the big quant funds from saying "we'd like you to investigate into such and such a field"?

 

[doug reich]

What keeps places like the big quant funds from saying "we'd like you to investigate into such and such a field"?

Secrecy?

 

[IlyaKEightSix]

Non-disclosure forms? With our senior project, that's what everyone has to fill out when they visit a company to do consulting. Except mine is intro to quantitative research XD

 

[doug reich]

I would venture a guess that the average person working as an academic is not only not interested in the questions wall street has to ask, they don't even want to deal with the culture. Those that are interested already have the connections you say, come to think of it.

 

[Aj Kappor]

Most of the research in finance and in other business functions such as consulting has already been outsourced even in what used to be "bulge bracket" banks. Many big firms across industries including those in NY City have been and are 'off-shoring' research and analysis to 'cheaper' countries (say India, Hong Kong, Philippines). Those countries in turn are now further off-shoring to other countries and continents (say Africa, South America). In case of the financial sector, this has been exacerbated after the incorporation of the "Chinese Wall" in US financial firms that separates analysis and investment banking functions: there is not much money left to be made in research.

Real world "experiential" perspective of most well known traders, investors, and businessmen - such as Bloomberg, Buffet, Kovner, Simons, and Soros - can shed some understanding about what is behind the success of many of these as well as others who are ranked in the Forbes 400. The distinguishing characteristic of most such 'success stories' was / is a high level of autonomy and the capacity and capability of risk taking by going against the tide. Most of their biggest bets were based on huge contrarian risks: some panned out, while others tanked, nevertheless, overall score has been net plus. Not many such success stories have come from those who have spent most of the lives being 'cogs' in the wheels of the 'big banks' or other mechanizations in the finance industry. Not many of them are known to have had any formal quant finance credentials. Nor have we seen many from 'traditional' academia [folks such as NNt is known to detest] despite several Nobels being awarded for financial research known to have fundamental gaps - being used in standard CFA curricula - as observed by Mandelbrot among others.

The key question that baffles most who may not be as savvy as most elite quants is: How could those deemed to be the smartest people drawing the largest paychecks and controlling the greatest level of [information, capital, other] resources go down so fast and with such a loud crash while taking the national and global economies down with them. Did any of them really have a clue about what they were doing? If so, how do we explain the "real" outcomes about 'most everything taught, learned, and practiced in the world of global finance and risk management coming to a nought with a thundering crash. Can we trust any of the stuff preached or practiced by the high priests of finance anymore?

As evident from above comments, I tend to agree with Dominic about maintaining healthy skepticism regarding anything faith-based, and specifically finance - particularly when it involves one’s own money. However, the critical question given the 'hard bottom line' results which are there for everyone to see is how can one "trust" anymore that anything being taught or practiced by the high priests of finance is based on anything but faith: Poof! One body blow and it all came down like a house of playing cards. How can one trust anything that has been taught, learned, and practiced in the world of quant finance after seeing the grim outcomes? Does anything get more real than the current reality to underscore that the "emperor has no clothes" and perhaps never had to begin with.

These are key concerns about NOT taking anything based merely on faith but taking a hard look at the bottom line results. In any area of practice, be it heart surgery or piloting planes, it is the success versus failure rate that finally matters. If a doctor kills more patients than (s)he saves or a pilot crashes more planes than (s)he flies, then soon either of them will be out of business, stripped of their license, and perhaps facing criminal and civil lawsuits and jail time.

BTW, Dennis, if you read Taleb's quote that you cite about NNt in its original context (published perhaps a decade ago in Derivatives Strategy?), the gist of his quote is consistent with what you are arguing: not to take anything based upon faith including "autopilot" controls AND always cross-checking machine-based smarts with "real" human smarts. The key question remains: in the world of quant finance, where are the "real" human smarts, and were they there to begin with.

Almost all medical errors are caused by qualified doctors.

IlyaKEightSix cites NNt's pilots, but again almost all plane crashes have a fully qualified pilot at the controls.

Is the answer to plane crashes to put unqualified people at the controls....

Retreating to "faith based finance" is not the solution.

 

[Stefan Zota]

Most of the research in finance and in other business functions such as consulting has already been outsourced even in what used to be "bulge bracket" banks. Many big firms across industries including those in NY City have been and are 'off-shoring' research and analysis to 'cheaper' countries (say India, Hong Kong, Philippines). Those countries in turn are now further off-shoring to other countries and continents (say Africa, South America). In case of the financial sector, this has been exacerbated after the incorporation of the "Chinese Wall" in US financial firms that separates analysis and investment banking functions: there is not much money left to be made in research.

Outsourcing is not something as easy as you may think. It is true that Wall Street companies are global and they have offices in dozens of countries. However, outsourcing is a big challenge including back-office and middle-office technology. There are significant resources spent in training and especially in hardware. Building high-available scalable systems on top of a fragile infrastructure is a limiting factor. For instance, a network grid in Delhi might be more than enough for a domestic bank, but it cannot satisfy requirements of a global bank with dozens of data-centers and high traffic.

If we look only at quantitative side, outsourcing becomes almost impossible. There is a barrier of knowledge, a proximity needed to the traders, to the market. Quantitative strategies will grow with the market so there is a long way to catch up for U.S. and U.K. markets ...

 

[Aj Kappor]

If reports of the U.S. IB "meltdown" can be relied upon, the back-office positions were the first to be chopped en masse. Here is one random report pulled right now to confirm similar observations reported elsewhere earlier in major US media sources. Industry insiders such as top corporate executives may know better. With the recent technology trends such as computing clouds - openly accessible globally from Amazon.com, IBM, Google, among others - the notion of local grid or enterprise data center has been already rendered obsolete. Most significant growth of the U.S. IT companies in terms of major new facilities and headcounts over the past 5-8 years has been outside the U.S. such as on the Asian subcontinent. It is another matter that some hedge funds have been moving their trading transactions intensive servers right in the middle of market exchanges to shave milliseconds in transaction time. However, all the economization of milliseconds in transaction time can't make up for the havoc wreaked by fundamentally flawed risk management models and methods currently in widespread use. Regarding barriers to knowledge about commodity quantitative knowledge, they are already eroding with many top US schools such as Columbia and Chicago already offering parallel quant programs in Asia and elsewhere. Proprietary trading may see severe restrictions on this side of the Atlantic given the current regulatory environment. [This report just in: UBS cuts 2000 in IB and puts prop trading on the chopping block.]

"Outsourcing is not something as easy as you may think. It is true that Wall Street companies are global and they have offices in dozens of countries. However, outsourcing is a big challenge including back-office and middle-office technology. There are significant resources spent in training and especially in hardware. Building high-available scalable systems on top of a fragile infrastructure is a limiting factor. For instance, a network grid in Delhi might be more than enough for a domestic bank, but it cannot satisfy requirements of a global bank with dozens of data-centers and high traffic.
If we look only at quantitative side, outsourcing becomes almost impossible. There is a barrier of knowledge, a proximity needed to the traders, to the market. Quantitative strategies will grow with the market so there is a long way to catch up for U.S. and U.K. markets ..."

 

[alain]

...With the recent technology trends such as computing clouds - openly accessible globally from Amazon.com, IBM, Google, among others - the notion of local grid or enterprise data center has been already rendered obsolete. Most significant growth of the U.S. IT companies in terms of major new facilities and headcounts over the past 5-8 years has been outside the U.S. such as on the Asian subcontinent...

If you think a quant trading firm of any kind will willingly deploy their know-how outside of a datacenter not under their control, you are sadly mistaken. They might outsource their operations business or their back office IT but no pricing engine or risk engine will be send outside of their grasp. There is too much IP knowledge riding on it.

Regarding barriers to knowledge about commodity quantitative knowledge, they are already eroding with many top US schools such as Columbia and Chicago already offering parallel quant programs in Asia and elsewhere.

This is education as a business. Those courses were opened because the demand for this field in Asia is really big (for whatever reason). Hence, if the demand is high, you can sell and charge too (the old supply and demand thing )

 

[Aj Kappor]

Anyone conversant with IT security and controls 101 realizes the issues with control: "physical" control of assets doesn't necessarily imply "real" control. That is why some hedge fund managers don't even use their own "official" e-mail addresses on servers under their control and rather use generic addresses for most sensitive communications as they are less susceptible to be trapped by industrial spies and hackers. BTW, all latest trends in Internet telecommunications and networking including those related to cloud computing allow comparable or greater level of control than ever possible as the notion of control of "physical" information assets is isolated from "logical" information resources. This trend started about 15 years ago and most major companies including billion dollar firms are on record for opting for the best-of-breed virtual access facilities from IT providers who [supposedly] know best about security and control issues. Anyhow, given that a homeless teenager hacker somewhere in South America - among others around the world from countries such as Russia and China - could repeatedly penetrate the most secure financial and defense networks in the U.S. in recent times demonstrates the vulnerability that comes with the assumption that anything is "under control." The only way to be in real control of your networked computer or server is to cut it off from complete network connectivity and then unplug it if it has any wireless capability .

A popular read on this theme which is a classic was written by Kevin Kelly who as editor of Wired had published Out of Control.

If you think a quant trading firm of any kind will willingly deploy their know-how outside of a datacenter not under their control, you are sadly mistaken. They might outsource their operations business or their back office IT but no pricing engine or risk engine will be send outside of their grasp. There is too much IP knowledge riding on it.

 

[Stefan Zota]

Adding to Alain's comments, even back-office infrastructure is only partially out-sourced.
I work in infrastructure, a lot of investments in India by Wall Street. Still the structure is centered around NY. The engineering and direct service comes primarily from NY or LN.
The rest is some part of support/operations.

Runbooks can cover only a small part, experience is essential when it comes to critical applications. If for instance jobs cannot start in an entire division, or monitoring is down, it needs to be resolved within minutes.
To give you an idea, when 30s outage means 1 million lost, you think really hard about the level of service and the cost saved by outsourcing.
Outsourcing can be very successfull for a pure software company immediately, for financial companies it takes tens of years.

 

[DominiConnor]

IlyaKEightSix is right, there is no "formula" for defining a qualified quant, whatever that actually means. As a headhunter, I have stolen staff selection procedures from many domains, and I don't believe quanting is more diverse than law or medicine.

A doctors, lawyers and accountants have formal post-qualification training schemes. They can choose specialities, but one definition of a "professional" is that they are required to keep up to date.

Second is peer review. I teach on the CQF, many here are doing MFEs of various kinds, and education is necessary but not sufficient. You need peer review.

Commercial and military pilots have their performance reviewed by other pilots both at qualification and in service. They are not put in a simulator where they pass if they score enough points

But it's clear that the errors were mostly not tactical, but strategic.

Most of the strategic decision makers had no quant knowledge at all. Look at Paulson. A degree in English, and before his current role was a great man at Goldmans. What do yuo think his mental model of tail events might be ? Do you think he could even calculate NPV ?

This applies right across the industry, most people who decided the "right" exposure to credit derivatives had no detectable knowledge of the bloody bloody things.

If a general thinks in terms of men on horseback, he is not going to run an air force well.

 

[bustweiser]

What defines a "qualified" quant? A qualified pilot learned how to fly a plane. A qualified doctor went through medical school. However, how do you define someone who's "qualified" to do well in the markets?

Doesn't that beg the question that if there *were* a formulaic way to be qualified to be assigned these high-risk high-reward responsibilities that wouldn't everyone be making a profit who could get their hands on these "qualified" individuals?

Comparing a pilot/doctor to a quant is fundamentally flawed way of thinking. That is because the profession of aviation/medicine is not scalable. That of a quant finance (or finance in general), however, is extremely scalable.

A doctor can provide consult to, say at most 20-30, people in a day. Cant do more than that. A pilot can fly no more than one plane at a time. A quant, however, can create a financial instrument for $1M, $10m, $100m or more.

Consequently, there is a hard upper bound to the amount of damage doctors and pilot can do by screwing up. The "risk" is, therefore, bounded by the very nature of their profession. For quants, though, the risk is not bounded by the nature of the profession itself.

Now, a financial firm can devise measures to control risk, and they all do. But by now, we all know how great that works.

Many people obviosly believe that government can step in to cap the risks, but I seriously doubt the efficacy of those measures for financial industry. People will figure out a way to circumvent them.

 

[IlyaKEightSix]

Well of course. But what does any of that have to do with how to tell when a quant is "qualified" to do what he does?