Home > data science, finance > Emanuel Derman’s Apologia Pro Vita Sua

Emanuel Derman’s Apologia Pro Vita Sua

September 17, 2012

Why, if I’m so aware of the powers and dangers of modeling, do I still earn my living doing mathematical modeling? How am I to explain myself?

It’s not an easy question, and I’m happy to see that my friend Emanuel Derman has addressed this a couple of weeks ago in an essay published by the Journal of Derivatives, of all places (h/t Chris Wiggins). Its title is Apologia Pro Vita Sua, which means “in defense of one’s life.” Please read it – as usual, Derman has a beautiful way with words.

Before going into the details of his reasoning, I’d like to say that any honest attempt at trying to answer this question by someone intrigues and attracts me to them – what is more threatening and interesting that examining your life for its flaws? Never mind publishing it for all to see and to critique.

Emanuel starts his essay by listing off the current flaws in finance better than any Occupier I’ve ever met:

After giving some background about himself and setting up the above question of justifying oneself as a modeler, Derman reveals himself to be a Blakean, by which he means that “part of our job on earth is to perceptively reveal the way the world really works”.

And how does the world work? According to Norman Mailer, anyway, it’s an enormous ego contest – we humans struggle to compete and to be seen as writers, scientists, and, evidently, financial engineers.

It’s not completely spelled out but I understand his drift to be that the corruption and crony capitalism we are seeing around us in the financial system is understandable from that perspective – possibly even obvious. As an individual player inside this system, I naturally compete in various ways with the people around me, to try to win, however I define that word.

On the one hand you can think of the above argument as weak, along the lines of “because everybody else is doing it.” On the other hand, you could also frame it as understanding the inevitable consequences of having a system which allows for corruption, which has built-in bad incentives.

From this perspective you can’t simply ask people not to be assholes or not to use lobbyists to get laws passed for their benefit. You need to actually change the incentive system itself.

Derman’s second line of defense is that the current system isn’t ideal but he uses his experience to carefully explain the dangers of modeling to his students, thereby training a generation not to trust too deeply in the idea of financial engineering as a science:

Unfortunately, no matter what academics, economists, or banks tell you, there is no truly reliable financial science beneath financial engineering. By using variables such as volatility and liquidity that are crude but quantitative proxies for complex human behaviors, financial models attempt to describe the ripples on a vast and ill-understood sea of ephemeral human passions. Such models are roughly reliable only as long as the sea stays calm. When it does not, when crowds panic, anything can happen.

Finally, he quotes the Modelers’ Hippocratic Oath, which I have blogged about multiple times and I still love:

Although I agree that people are by nature tunnel visioned when it comes to success and that we need to set up good systems with appropriate incentives, I personally justify my career more along the lines of Derman’s second argument.

Namely, I want there to be someone present in the world of mathematical modeling that can represent the skeptic, that can be on-hand to remind people that it’s important to consider the repercussions of how we set up a given model and how we use its results, especially if it touches a massive number of people and has a large effect on their lives.

If everyone like me leaves, because they don’t want to get their hands dirty worrying about how the models are wielded, then all we’d have left are people who don’t think about these things or don’t care about these things.

Plus I’m a huge nerd and I like technical challenges and problem solving. That’s along the lines of “I do it because it’s fun and it pays the rent,” probably not philosophically convincing but in reality pretty important.

A few days ago I was interviewed by a Japanese newspaper about my work with Occupy. One of the questions they asked me is if I’d ever work in finance again. My answer was, I don’t know. It depends on what my job would be and how my work would be used.

After all, I don’t think finance should go away entirely, I just want it to be set up well so it works, it acts as a service for people in the world; I’d like to see finance add value rather than extract. I could imagine working in finance (although I can’t imagine anyone hiring me) if my job were to model value to people struggling to save for their retirement, for example.

This vision is very much in line with Derman’s Postscript where he describes what he wants to see:

Finance, or at least the core of it, is regarded as an essential service, like the police, the courts, and
the firemen, and is regulated and compensated appropriately. Corporations, whose purpose is relatively straightforward, should be more constrained than individuals, who are mysterious with possibility.

People should be treated as adults, free to take risks and bound to suffer the consequent benefits and disadvantages. As the late Anna Schwartz wrote in a 2008 interview about the Fed, “Everything works much better when wrong decisions are punished and good decisions make you rich.”

No one should have golden parachutes, but everyone should have tin ones.

Categories: data science, finance
  1. September 17, 2012 at 11:57 am

    I don’t use mathematical models as much as I use boundary conditions to deal with my investo-speculation.

    I would certainly enjoy working in finance, being that I love economic history. However, my love of it tends to be in terms of mania/panics/crashes.

    Once the set up is there, you can easily pick the right side of the trade.

    I *understood* 2007-2009. That is to say, it was in line with my model of reality. I was watching it and making money off of it. It was one of those situations whose outcome was predictable and my model worked.

    Right now, the model at issue is the new-issue QE model. I *do not* have a good model for this one. It’s very liquidity driven and we are now way over the top in terms of normal historical outcomes.

    Will Bernake ignite commodity inflation, ultimately tanking the economy and the stock market with the latest round of Infinite QE?

    I don’t have a good post-crash QE model. That is to say, I can pretty much tell you where the market is extremely likely to be in 7 years in terms of real valuation. However, I can’t tell you the path to get there, which is what I need to do if I want to get more than 0%.

    I don’t have a defense for my own life. I don’t know whether it *is* justifiable. Am I doing things incorrectly and am I underachieving? Absolutely. Do I consider myself a failure? Absolutely.

    I’m not sure what is wrong with admitting that your life has no defense. I have no means of justifying my actions.


    • September 18, 2012 at 7:58 am

      jonlaw2 :
      I don’t use mathematical models as much as I use boundary conditions to deal with my investo-speculation.

      Hey jonlaw2, you don’t happen to be a former UCLA professor?


      • JP
        September 18, 2012 at 10:19 am

        Nope. I did fill in for an IP professor once on the other side of the country from UCLA. I’m a practicing attorney.

        I just have a chemical engineering background, which is why I use boundary conditions, having been force fed them in undergrad.

        Who do I resemble from UCLA?


  2. Bindicap
    September 17, 2012 at 9:19 pm

    I thought it was weak. For a while now, Derman seems to have lost his way. You dropped out the punch line.

    “I’ve tried, and I believe that the models I’ve helped to create have been used, for the most part, wisely and without harm.”

    So it suffers from the same thing that bugs me about a lot of essays along similar lines. I want to hear where the author conned others, or cheated, or was dishonest. Or watched others do that around him and couldn’t prevail upon them to do the right thing, maybe got pushed out. And then the apology for that wrong or for that failure.

    But the apologies are never like that. By now it’s a drag to read a rant about how corrupt everyone in the system is coming from the one pure exception. It’s more like the author doesn’t really have a good handle on things.

    This one is so unclear you also struggled to find his point: “It’s not completely spelled out but I understand his drift to be…”

    I feel a bit mean laying it out like this, because I like Derman, and I think he did some neat stuff. Oh well…


  3. September 17, 2012 at 11:08 pm

    “Why, if I’m so aware of the powers and dangers of modeling, do I still earn my living doing mathematical modeling? How am I to explain myself?”

    Am I way off base, or am I correct in assuming that you and Mr. Derman already know the answer: Because models, in and of themselves, cannot truly be evil. An unwillingness to recognize their limits has the potential for evil. A willingness to knowingly understate their limits is evil to the core.

    We knowledgeable skeptics (not the cynics) had better figure out a way to maximize the information tech at our disposal for our skeptical purposes. Finance isn’t the last sector that will have a meltdown brought on by value-free Manhattan Project too-clever-by-half technicians.

    We’ll miss the point if we blame the “art and science”, and not the lack of ethics with which they have too often been applied.


  4. September 22, 2012 at 7:49 pm

    Much of modern financial modeling disasters reminds me of the situation in fluid mechanics of knowing rather well the dynamic equilibrium states on either side of a shock wave, but not having a clue as to the dynamics of getting from one side to the other. The rationalized-linearized assumptions and limitations of the equilibrias will tell you nothing about the nonlinear, approximationless zone about to hit you, and only the vicious experience of ever-unfolding reality with ever inform you, never gaussian-statistical nonsensical hand-waving.


  1. September 22, 2012 at 12:01 pm
Comments are closed.
%d bloggers like this: