Quant programs MUST CHANGE - Pablo Triana

[Andy Nguyen]

Pablo Triana has written articles on quantitative finance education in the past. He also penned about MBA education.

In this article on FT, he gives a list of things that schools offering MFE degree should consider

● Quant finance programmes must start focusing on why and how financial modelling can be useless and dangerous.
● The demonstrably broken and troubling models must be presented as failures of the discipline, not, as has often been the case, symbols of triumph. When teaching derivatives or risk management courses, ensure that the models’ flaws are emphasised and highlight the potentially negative real consequences of such defects.
● Hire faculty that is not analytically dogmatic and openly sceptical to the value to be derived from financial modelling and quantitative finance. At the very least, students need to be exposed to both sides of the debate.
● Question the wisdom of having key (and deeply influential) aspects of financial regulation based on theoretical models.
● Encourage critical thinking among your students; intellectualise the programmes by emphasising history and big-picture ideas.
● Investigate how things were done in the markets before quantification became prevalent and analyse without bias whether the models truly represented valuable improvements and whether market crashes became more likely than before.
● Look at how models are used as alibis by otherwise no-nonsense operators in order to engage in reckless actions.
● Consider whether so-called quant funds (among the most profitable investment managers yet) truly employ financial theory when making their decisions. If they do not, ask why.
● Run mandatory courses on ethics for students; make them take an oath agreeing not to promote the use of knowingly erroneous and lethal mathematical concoctions that can lead to social unrest.
● Overall, the notion that finance can be quantified needs to be questioned at every turn. If that results in uncomfortable answers, perhaps the kind that make administrators wonder about the need for the degree, so be it.

FT.com / Business education / Soapbox - The flawed maths of financial models

 

[bob]

Those who know me would, I hope, not hesitate to attest to my credentials as an amused skeptic when it comes to many aspects of this field we all know and love. Still, I guess I believe in the project if not, maybe, its particulars, and evidently Triana (Can I just call him Mini-Taleb?) is in favor of chucking bathwater, baby, and maybe the bath and bather all out the window as soon as possible.

So what I'm most curious to know is...if we decide not to quantify finance, then what do we do with it instead?

 

[euroazn]

Run mandatory courses on ethics for students; make them take an oath agreeing not to promote the use of knowingly erroneous and lethal mathematical concoctions that can lead to social unrest.

I LOL'ed.

 

[Andy Nguyen]

While I agree with Triana's premise that quant programs must change to adapt to the new reality, I have problems with the practicality of many of his suggestions.

And his suggestion about an oath may be just a joke from his part. He should already have read about the MBA Oath initiative and probably believe that will save future of business.

To put this discussion in more context, here are few things that you should know about Pablo Triana

He is the author of a book called Lecturing Birds on Flying: Can Mathematical Theories Destroy the Financial Markets? (Taleb foreworded the book for him)

new quant fin critique - Lecturing Birds on Flying by Pablo Triana - Forum | Quant Network

He wrote a piece on quant finance vs general finance vs MBA programs. The PDF file is fittingly hosted on the website of the Princeton MFin program
http://www.princeton.edu/bcf/newsevents/in-the-news/links/20070900-GARP-metamorphosis.pdf

And he wrote an article on HuffPost last year arguing why Business Schools will not change even after the crisis.
Pablo Triana: Why B-Schools Don't Change

 

[bigbadwolf]

So what I'm most curious to know is...if we decide not to quantify finance, then what do we do with it instead?

Does finance as an academic discipline exist without quantification? Nope. Does finance as an activity exist without quantification? Nope. At the same time, do quantitative models in finance work? Again, nope.

Triana's gripe seems to be that models known to be flops are still being taught in MFE programs. The rejoinder to the call for "better" models should be there are none. Models in finance and economics do not have the explanatory and predictive power of models in physics.

Triana is going after bad models in quant finance. In economics, equally worthless macro models -- with no real explanatory or predictive power -- have been taught for generations. Something has to be taught -- even if it's utterly meretricious -- thus justifying the existence of faculty.

As for courses in ethics -- *snort of derision* -- does anyone think an employee, who's been chosen out of a large pool of hungry and jobless quants, is going to raise moral questions about what he's asked to do? Likewise, does anyone think a student who's forked out serious cash for tuition fees is going to ask unsettling questions (thus possibly jeopardising both his grades and his references)? No way, Jose. Quant students will conform. There's too much at stake.

 

[woody]

Simply adding a history course would go a long way. For example, a history of financial crises like MIT's OCW course. I'm not aware of any such course being offered as part of an MFE.

 

[lalala]

I think I did post similar questions in one of my older threads, questioning models.

Just my 2 cents worth, I think the real key thing is lack of ability or perhaps motivation to really understand/interpret what the models/equations represent. Just like physics, we know where math modelling comes from, laws, principles... But it seems to me that in finance, many people are just too concerned with $/profit. When you are optimizing a function, you need to know what exactly it represents, or in short, reasoning. It is happening in my field too, its getting very math, but I really doubt if those things are necessary. Same goes to logistics/OR.

I think model does work to certain extent, as Connor said before. The theory is there, the main problem is that these models are misused or abused, because of greed.

Regarding Bob's question, I think both extremes (quali/quanti-tative) are not the direction we want, perhaps experience+quantifying is. Finance can never be = physics.

I do agree with wolf that cost is another thing that is stopping people from pointing that out. But I would expect senior people up there to work on it, they are financially independent, especially faculty, they are the main driver for innovation.

And lastly, it seems to me that the big $ still goes by relationship. So if you are in the circle/lucky enough to catch the news, you make it and sorry if you are not. Sometimes I doubt if current quant fin system really works.

Just my personal opinions, do feel free to correct me if I am wrong. I am reading Martin Wolf's Fixing Global Finance, nice book.

 

[Andy Nguyen]

Steve Shreve just sent me a rebuttal to Pablo's article on FT.
Steve Shreve on Pablo Trianas The Flaw Math of Financial Models

 

[Stephie]

I have a choice between an economic history of the 20th century course and a beginning principles of banking and finance course. Knowing I can think for myself, I'm taking Banking and Finance to later prepare me for Corporate Finance. Why? I'll be the first to tell you the system needs to change, but I think there should be more quants, more math, and more encouragement from society to those in this field.

 

[Andy Nguyen]

Steve Shreve just sent me a rebuttal to Pablo's article on FT. I'll post them on Quantnet later.

Pablo just emailed me a rebuttal to Shreve's rebuttal of his article. What an interesting world.
The article is updated https://www.quantnet.com/threads/st...’s-the-flawed-math-of-financial-models.10045/

 

[DominiConnor]

I teach some ethics on the CQF, and Paul Wilmott integrates much of those ideas into his teaching, but...

One big issue is advocacy.
It's all very well someone thinking "we ought not to be doing this", but for it to be useful, you have to be able to get others to share your position. Otherwise all you are is some guy who pointlessly makes trouble.

An important issue in ethics which are expressed in law is that of competence.
For example
Goldmans has got a lot of hassle over selling stuff that seemed guaranteed to go bad.
However, as I understand the case, if you really understood what they were selling, you could have worked that out for yourself.

Let's de-personalise that to the situation where someone is prepared to pay much more for something than your calculations say it is worth.
Case 1: They are a professional working for a big firm

Case 2: They are an occasional participant in the market, whose math/finance education doesn't go much beyond (1 + r ) ^n level compound interest.

Do you decline to sell it to them ?
How much "competence" is required for you to say "it's their problem" ? An MFE ? An MFE at a school that's as good as yours ? A PhD + ten years experience ?

How about selling an instrument that on average pays out less than it is bought for, and this loss is by design ? This instrument to be sold only to consumers, few if any who can understand the maths involved ?
Bad ?

Maybe, but it's called insurance, which is regarded as OK by most people.

Does the market exist ?
Do you have a duty to this entity, if you accept that it is a 'thing' not just a model.

 

[Andy Nguyen]

Pablo Triana's response to Prof. Shreve's article can be seen in the second part of this article
Steve Shreve on Pablo Triana’s The Flawed Math of Financial Models

 

[bigbadwolf]

Triana talks about bad and dangerous models but if one has a financial and economic system that is fundamentally irrational at root, and one whose ground rules keep changing, how is one ever going to have "good" models? Model-building in finance and economics seems to be about providing an intellectual fig leaf for the status quo. Everything that one has in real science is absent in finance: the theories cannot be tested in a controlled way (i.e., holding some variables fixed); they cannot be repeatedly tested; they are not falsifiable. Predictive power isn't there. It seems that tools have been unthinkingly imported from the hard sciences to provide a patina of intellectual legitimacy and respectability. Maybe all those equations will fool someone ....

 

[euroazn]

Triana talks about bad and dangerous models but if one has a financial and economic system that is fundamentally irrational at root, and one whose ground rules keep changing, how is one ever going to have "good" models? Model-building in finance and economics seems to be about providing an intellectual fig leaf for the status quo. Everything that one has in real science is absent in finance: the theories cannot be tested in a controlled way (i.e., holding some variables fixed); they cannot be repeatedly tested; they are not falsifiable. Predictive power isn't there. It seems that tools have been unthinkingly imported from the hard sciences to provide a patina of intellectual legitimacy and respectability. Maybe all those equations will fool someone ....

1. Invest in treasury bonds
2. Wait infinity years
3. ?????????????????
4. PROFIT!

 

[Tsotne]

Quant finance programmes must start focusing on why and how financial modelling can be useless and dangerous.
● The demonstrably broken and troubling models must be presented as failures of the discipline, not, as has often been the case, symbols of triumph. When teaching derivatives or risk management courses, ensure that the models’ flaws are emphasised and highlight the potentially negative real consequences of such defects.

Alan Greenspan: How much dependence should be placed on financial modeling, which for all its sophistication can get too far ahead of human judgment?! That'd be another type of quant profession monitoring the danger arising from financial models and economic assumptions.

 

[DominiConnor]

It's very easy to say "we must teach X" where X is something we don't currently teach.
It's so easy to say that a bit of googling on my name will prove how easy it really is.

But an MFE is a year...

Maybe you could add a week, maybe even a month to the taught part of the course, is this enough to turn out people who are ready for the financial world of 2012, and the world of 2020 when some of them will be in positions of considerable responsibility ?

If you want to lecture bankers ethics (as I do on Paul Wilmott's CQF), how much is enough to turn out 'good people' ?
My ambition is to get people to think a bit more clearly about ethical issues, I do not kid myself that I am some sort of divine moral leader whose words are followed reverently by cults of bankers.

If you want MFE grads to have scepticism of models, does that mean you teach them less about models ?
That sounds bizarre, but teaching the failure modes of any system or model is necessarily much more work than the system itself. So with a finite budget of time, you get fewer models or less informed scepticism.
I'm not making a judgement on that trade off, just telling you there is one.

Maybe MFEs should be two years, perhaps ?
Aside from the extreme unpopularity of such a move, is it enough ?

Ethics, scepticism, banking history and a far better grasp of basic economics than currently possessed by most MFE grads. Other people will add other skills to that list.

Also, as a professional interviewer of bankers, it's clear when you talk to people, that some people "get it", and others don't.
So we need peer interviews, ie you can't be any sort of financial engineer unless and until other FEs question you on your ability and suitability.

On top of that...
My wife is a scary lawyer (beautiful, but scary).
She's got a degree from Oxford, and undertook a two year course in lawyering from London's top school for that purpose, so good was she at lawyering that a big law firm decided to pay her fees in the hope that she might work for them.

So did that make her a lawyer ?
No.
She then had to do a training program, at a firm whose training program has to pass tests, else the law firm is not allowed to take on newbie lawyers.

Imagine if a bank was only allowed to hire newbie quants if it had to train them properly.
If it had to make sure that real grown up quants, traders etc spent time with them explaining how all this crap worked.
Even then, not all lawyers are up to the job.

It's not so different for accountants or doctors or dentists, all have longer courses and mandatory post-qualification mentoring.

Most here on Quantnet are studying an MFE, or have done one recently...
So on Monday look at the student sitting next to you, and imagine he was one exam away from being given a host of knives, interesting chemicals, and of course electric drills to stick in your mouth.
Would you be happy with that ?