QuantNet
Staff member
- Joined
- 10/23/03
- Messages
- 38
- Points
- 28
By STEVE SHREVE
In his article The flawed math of financial models, Financial Times, November 29, Pablo Triana seeks to fix a large portion of blame for the world-wide financial crisis on "quants'' in the finance industry and the programs that educate them. Mr. Pablo recommends radical reform in such programs. Others, carrying these ideas farther, call for a diminished role for quants in finance.
Any discussion of quants in finance must begin with the recognition that the global integration of economies and the associated complexity of our financial system has made the use of mathematical models an indispensable tool. Rules-of-thumb and intuition will not suffice when multi-national firms face exchange rate risk, funding risk and commodity price risk, when insurance companies and pension funds face longevity risk, when financial institutions are called upon to mediate these risks, and when regulators are charged to oversee these institutions. This was recognized in the recent U.S. financial reform legislation, which authorized a government Office of Financial Research whose task in 2008 would have been to alert policy makers to the ridiculously large naked position in credit default swaps held by AIG and to predict the effect of the failure of Lehman Brothers. Such an office must necessarily be populated by quants, people who can build models into which information about financial institutions is fed.
What then is the appropriate training for quants? I believe we should focus on three aspects.
Most importantly, a quant must be competent in the technical disciplines of mathematics, statistics and computer programming, and she must be knowledgeable about financial markets. Achieving competence across this broad spectrum is a tall order. But it must be done because a well-intentioned incompetent quant is as dangerous to the financial system as a well-intentioned incompetent doctor is to personal health. The primary focus of the educational programs at Carnegie Mellon will remain the creation of competent graduates. This is what we do best.
But a good quant also needs good judgment. A wise quant takes to heart Albert Einstein's words, "As far as the laws of mathematics refer to reality, they are not certain; and as far as they are certain, they do not refer to reality.'' All models are wrong. Judgment is needed to know when an admittedly wrong model can be helpful and when it is dangerous. This kind of judgment is acquired primarily through experience, but we can begin teaching it in the classroom. Since the financial crisis, we have invited participants in the crisis to speak in detail to our students about deals that went bad, describing how the deal was analyzed, why it was approved, and what was overlooked.
Finally, we need people with integrity managing our financial systems. Teaching ethics is difficult, and guaranteeing that listeners will implement those teachings is impossible. It is not easy for a quant to sound the alarm that his models are being stretched beyond their limits, knowing that if he is taken seriously it will result in the loss of business to competing firms and may result in the loss of his job. We cannot instill in sixteen short months behavior that properly requires years of nurturing and mentoring. We do what we can, leading by example, penalizing students for academic dishonesty, setting and enforcing rules for ethical conduct when interacting with potential employers, posing ethical dilemmas for classroom discussion, and encouraging our graduates to consult with fellow graduates when facing tough ethical decisions.
A lesson that can be learned from the present crisis is that if everyone implements the same good idea, their collective action can invalidate the assumptions that made the idea good. If everyone assumes that U.S. housing prices cannot decline and makes large bets based on that assumption, their collective action will ultimately bring about a decline in housing prices. This is not a new lesson; it is the lesson of every bubble. A feature of the most recent bubble is that quantitative analysis contributed to a false sense of security that encouraged firms to scale up risks. In some cases senior managers and even quants themselves did not appreciate the limitations in the models on which they based their risk analysis. Our students do not begin their careers at the level where the disastrous decisions were taken, and only a handful of them will ever reach those positions of power. Nonetheless, in the short time they are in our care, we seek to the extent possible to make them competent quants who exercise sound ethical judgment.
About the Author
Steven Shreve's books Stochastic Calculus for Finance I: The Binomial Asset Pricing Model and Stochastic Calculus for Finance II: Continuous-Time Models are the required textbooks for many MFE programs' Stochastic Calculus courses. He is a professor at Carnegie Mellon University and one of the co-founders of the M.S. in Computational Finance at Carnegie Mellon.
Editor's note: Following Prof. Shreve's article, we received a response from Mr. Triana on Jan 10 which we have published in full. It can be seen directly after Prof. Shreve's article.
The following is a response by Mr. Triana sent to Quantnet on Jan 10, 2010.
By PABLO TRIANA
Let me first say that I deeply admire Professor Shreve. Though my mathematical background does not empower me to fully appreciate his scientific prowess (not that his unparalleled global reputation would ever necessitate my feedback as further support), I am aware that in replying to his analysis of my recent FT article on quant education I am addressing most possibly the world´s leading light when it comes to stochastic calculus and mathematical finance. And far from an aloof researcher, Professor Shreve is also a very successful and ingenius academic entrepreneur, having taken a leading role in the development and management of one of the most exciting and path-breaking university graduate programs ever devised. To top it all, I can personally attest to his human generosity and kindness, getting misty-eyed as I recall the time when Professor Shreve, now about a decade ago, kindly accepted my invitation (as President of NYU Stern´s Financial Engineering Association) to regale us with a wonderful lecture and an even more pleasant follow-up dinner at a fancy Soho restaurant. I vividly recall him being impressed by my thorough knowledge of and interest in his legendary Computational Finance program at Carnegie Mellon, to the point of asking me why I had chosen NYU instead (I didn´t even try to apply to terrifyingly intimidating Carnegie Mellon, acutely aware of my negligible chances at getting in; I ain´t no rocket scientist, folks!).
In sum, it is not only my responsibility but also my pleasure to try to address Professor Shreve´s rebuttal as respectfully as possible, given the caliber of the counterparty. I hope I manage to succeed at this, if not so much at triumphing in the debate.
Some initial clarifications are in order. I don´t really blame quants and quant programs for the crisis. I blame the use of certain models for the crisis. I don´t really care if those using, peddling, and imposing the deleterious models were quants, traders, salesmen, or fast food caterers. My goal is not to target specific groups of people, my goal is to target specific analytical concoctions. Having said that, it is true that a lot of quants vouch for those models both inside and outside the financial industry and, much more critically, vouch fanatically for the quantification of finance in general. As long as such belief is held and enthusiastically pushed, we can get in trouble because the potential for bad models to infiltrate the markets would be made that much larger. We need to create much more restrictive filters when it comes to welcoming mathematical finance wizardry into the realm of practice. Quants and quant programs could and should have been much less permissive and much more critical. Roadblocks to dangerous models should have been forcefully built by those who best understand the mechanics. So, yes, quants and quant programs could in the end be subjected to one type of accusation: neglect.
Everything stated in Professor Shreve´s response makes a ton of sense, and one can´t help but wholeheartedly agree. But, like other famed quants too graced with the ability to muse gracefully and the valour to challenge flawed quanty practices, Professor Shreve does not go far enough. Just like Emanuel Derman, Paul Wilmot, or Ricardo Rebonato, Professor Shreve needs to get closer to Nassim Taleb (and, maybe, my very humble self) and take things a step or two further and engage in a healthy dose of loud name-calling and unabashed denunciation. It is not enough to state that quantitative analysis played a role in the crisis by encouraging misplaced confidence or that many misunderstood the maths. It is imperative to endlessly fingerpoint the main culprits, essentially VaR and Gaussian Copula (to Professor Shreve´s credit, he went after the latter in a recent piece), and to make sure that such utterly failed tools are never again given the keys to the risk kingdom. Demonstrably flawed and lethal models should be banned from the land, and the real reasons for their original embracement intrusively inquired. VaR can no longer be part of banking regulations. These things can´t continue being taught, unless they are presented as the bad that can emerge from the quant lab. More pressing still, those failures must serve as catalyst to force everyone to revisit whether finance can and should indeed by mathematized. Are VaR, Gaussian Copula, Black-Scholes, Portfolio Theory, or Financial Econometrics isolated cases of failure, or rather inescapable proof that financial theory is bound to be at best useless and at worst crisis-igniting? We urgently need a Mathematical Finance Council of Nicaea, so that these pressing questions are answered once and for all. I wrote my Lecturing Birds On Flying in a naively idealistic attempt to help kick-start such process. Will the best that the discipline has to offer, like Professor Shreve, pick up the gauntlet?
This is no time for mincing words, it´s time to act. Back in 1994, Carnegie Mellon showed untold innovativeness and courage by correctly embracing the forcefully emerging field of financial engineering. It became the indisputable world beater. Now, with the discipline in tatters and accused of horrible crimes, the same institution should once more display one-of-a-kindness and lead the second quant finance revolution, the one that ought to make sure that models and financial stability can coexist side by side and the one not afraid to terminally castigate those naughty analytical concoctions that wreak havoc.
About the Author
Pablo Triana is the author of Lecturing Birds on Flying: Can Mathematical Theories Destroy the Financial Markets?
In his article The flawed math of financial models, Financial Times, November 29, Pablo Triana seeks to fix a large portion of blame for the world-wide financial crisis on "quants'' in the finance industry and the programs that educate them. Mr. Pablo recommends radical reform in such programs. Others, carrying these ideas farther, call for a diminished role for quants in finance.
Any discussion of quants in finance must begin with the recognition that the global integration of economies and the associated complexity of our financial system has made the use of mathematical models an indispensable tool. Rules-of-thumb and intuition will not suffice when multi-national firms face exchange rate risk, funding risk and commodity price risk, when insurance companies and pension funds face longevity risk, when financial institutions are called upon to mediate these risks, and when regulators are charged to oversee these institutions. This was recognized in the recent U.S. financial reform legislation, which authorized a government Office of Financial Research whose task in 2008 would have been to alert policy makers to the ridiculously large naked position in credit default swaps held by AIG and to predict the effect of the failure of Lehman Brothers. Such an office must necessarily be populated by quants, people who can build models into which information about financial institutions is fed.
What then is the appropriate training for quants? I believe we should focus on three aspects.
Most importantly, a quant must be competent in the technical disciplines of mathematics, statistics and computer programming, and she must be knowledgeable about financial markets. Achieving competence across this broad spectrum is a tall order. But it must be done because a well-intentioned incompetent quant is as dangerous to the financial system as a well-intentioned incompetent doctor is to personal health. The primary focus of the educational programs at Carnegie Mellon will remain the creation of competent graduates. This is what we do best.
But a good quant also needs good judgment. A wise quant takes to heart Albert Einstein's words, "As far as the laws of mathematics refer to reality, they are not certain; and as far as they are certain, they do not refer to reality.'' All models are wrong. Judgment is needed to know when an admittedly wrong model can be helpful and when it is dangerous. This kind of judgment is acquired primarily through experience, but we can begin teaching it in the classroom. Since the financial crisis, we have invited participants in the crisis to speak in detail to our students about deals that went bad, describing how the deal was analyzed, why it was approved, and what was overlooked.
Finally, we need people with integrity managing our financial systems. Teaching ethics is difficult, and guaranteeing that listeners will implement those teachings is impossible. It is not easy for a quant to sound the alarm that his models are being stretched beyond their limits, knowing that if he is taken seriously it will result in the loss of business to competing firms and may result in the loss of his job. We cannot instill in sixteen short months behavior that properly requires years of nurturing and mentoring. We do what we can, leading by example, penalizing students for academic dishonesty, setting and enforcing rules for ethical conduct when interacting with potential employers, posing ethical dilemmas for classroom discussion, and encouraging our graduates to consult with fellow graduates when facing tough ethical decisions.
A lesson that can be learned from the present crisis is that if everyone implements the same good idea, their collective action can invalidate the assumptions that made the idea good. If everyone assumes that U.S. housing prices cannot decline and makes large bets based on that assumption, their collective action will ultimately bring about a decline in housing prices. This is not a new lesson; it is the lesson of every bubble. A feature of the most recent bubble is that quantitative analysis contributed to a false sense of security that encouraged firms to scale up risks. In some cases senior managers and even quants themselves did not appreciate the limitations in the models on which they based their risk analysis. Our students do not begin their careers at the level where the disastrous decisions were taken, and only a handful of them will ever reach those positions of power. Nonetheless, in the short time they are in our care, we seek to the extent possible to make them competent quants who exercise sound ethical judgment.
About the Author
Editor's note: Following Prof. Shreve's article, we received a response from Mr. Triana on Jan 10 which we have published in full. It can be seen directly after Prof. Shreve's article.
The following is a response by Mr. Triana sent to Quantnet on Jan 10, 2010.
By PABLO TRIANA
Let me first say that I deeply admire Professor Shreve. Though my mathematical background does not empower me to fully appreciate his scientific prowess (not that his unparalleled global reputation would ever necessitate my feedback as further support), I am aware that in replying to his analysis of my recent FT article on quant education I am addressing most possibly the world´s leading light when it comes to stochastic calculus and mathematical finance. And far from an aloof researcher, Professor Shreve is also a very successful and ingenius academic entrepreneur, having taken a leading role in the development and management of one of the most exciting and path-breaking university graduate programs ever devised. To top it all, I can personally attest to his human generosity and kindness, getting misty-eyed as I recall the time when Professor Shreve, now about a decade ago, kindly accepted my invitation (as President of NYU Stern´s Financial Engineering Association) to regale us with a wonderful lecture and an even more pleasant follow-up dinner at a fancy Soho restaurant. I vividly recall him being impressed by my thorough knowledge of and interest in his legendary Computational Finance program at Carnegie Mellon, to the point of asking me why I had chosen NYU instead (I didn´t even try to apply to terrifyingly intimidating Carnegie Mellon, acutely aware of my negligible chances at getting in; I ain´t no rocket scientist, folks!).
In sum, it is not only my responsibility but also my pleasure to try to address Professor Shreve´s rebuttal as respectfully as possible, given the caliber of the counterparty. I hope I manage to succeed at this, if not so much at triumphing in the debate.
Some initial clarifications are in order. I don´t really blame quants and quant programs for the crisis. I blame the use of certain models for the crisis. I don´t really care if those using, peddling, and imposing the deleterious models were quants, traders, salesmen, or fast food caterers. My goal is not to target specific groups of people, my goal is to target specific analytical concoctions. Having said that, it is true that a lot of quants vouch for those models both inside and outside the financial industry and, much more critically, vouch fanatically for the quantification of finance in general. As long as such belief is held and enthusiastically pushed, we can get in trouble because the potential for bad models to infiltrate the markets would be made that much larger. We need to create much more restrictive filters when it comes to welcoming mathematical finance wizardry into the realm of practice. Quants and quant programs could and should have been much less permissive and much more critical. Roadblocks to dangerous models should have been forcefully built by those who best understand the mechanics. So, yes, quants and quant programs could in the end be subjected to one type of accusation: neglect.
Everything stated in Professor Shreve´s response makes a ton of sense, and one can´t help but wholeheartedly agree. But, like other famed quants too graced with the ability to muse gracefully and the valour to challenge flawed quanty practices, Professor Shreve does not go far enough. Just like Emanuel Derman, Paul Wilmot, or Ricardo Rebonato, Professor Shreve needs to get closer to Nassim Taleb (and, maybe, my very humble self) and take things a step or two further and engage in a healthy dose of loud name-calling and unabashed denunciation. It is not enough to state that quantitative analysis played a role in the crisis by encouraging misplaced confidence or that many misunderstood the maths. It is imperative to endlessly fingerpoint the main culprits, essentially VaR and Gaussian Copula (to Professor Shreve´s credit, he went after the latter in a recent piece), and to make sure that such utterly failed tools are never again given the keys to the risk kingdom. Demonstrably flawed and lethal models should be banned from the land, and the real reasons for their original embracement intrusively inquired. VaR can no longer be part of banking regulations. These things can´t continue being taught, unless they are presented as the bad that can emerge from the quant lab. More pressing still, those failures must serve as catalyst to force everyone to revisit whether finance can and should indeed by mathematized. Are VaR, Gaussian Copula, Black-Scholes, Portfolio Theory, or Financial Econometrics isolated cases of failure, or rather inescapable proof that financial theory is bound to be at best useless and at worst crisis-igniting? We urgently need a Mathematical Finance Council of Nicaea, so that these pressing questions are answered once and for all. I wrote my Lecturing Birds On Flying in a naively idealistic attempt to help kick-start such process. Will the best that the discipline has to offer, like Professor Shreve, pick up the gauntlet?
This is no time for mincing words, it´s time to act. Back in 1994, Carnegie Mellon showed untold innovativeness and courage by correctly embracing the forcefully emerging field of financial engineering. It became the indisputable world beater. Now, with the discipline in tatters and accused of horrible crimes, the same institution should once more display one-of-a-kindness and lead the second quant finance revolution, the one that ought to make sure that models and financial stability can coexist side by side and the one not afraid to terminally castigate those naughty analytical concoctions that wreak havoc.
About the Author
Pablo Triana is the author of Lecturing Birds on Flying: Can Mathematical Theories Destroy the Financial Markets?
