Derman on Theories, Models, and Science
Mar 12 2012

Emanuel Derman of Columbia University and author of Models. Behaving. Badly talks with EconTalk host Russ Roberts about theories and models, and the elusive nature of truth in the sciences and social sciences. Derman, a former physicist and Goldman Sachs quant [quantitative analyst], contrasts the search for truth in the sciences with the search for truth in finance and economics. He critiques attempts to make finance more scientific and applies those insights to the financial crisis. The conversation closes with a discussion of career advice for those aspiring to work in quantitative finance.

Explore audio highlights, further reading that will help you delve deeper into this week’s episode, and vigorous conversations in the form of our comments section below.


Mar 12 2012 at 11:45am

This talk came at a wonderful time. I’m really hoping to get into an economics program where I can worry about these kinds of methodological worries.

I’ve been digging into these issues in philosophy. Here’s maybe a few good resources for people

John Strong
Mar 12 2012 at 11:51am

What a pleasant surprise to discover that Emmanuel Derman is your most recent guest. This morning my 18 year-old son, who plans to study economics, asked me for some good reading for the barbershop, and I recommended Mr.Derman’s really fun book, My Life as a Quant. When I got the email announcement that Mr. Derman was the latest EconTalk guest, we both got a chuckle.

Mar 12 2012 at 12:47pm

I get a sense that Mr. Derman may have more respect for what the sensible application of financial modeling and financial engineering can bring to the table than what many may surmise from the general flow of this interview. Let’s take options pricing as an example. Anyone who has followed the evolution of options pricing technology from Black-Scholes, Cox and Rubinstein, and through all the various iterations since then, knows that while the insights are brilliant, real life-real time profits can only be earned by retaining premiums received/ insurance bought through “semi-continuous” hedging using real market assets, moving in real market ways, and under all kinds of real liquidity conditions. Contrary to what Mr. Roberts suggested (to paraphrase: Wall Street is the only business where the right price is sort of stated, not discovered) option traders (and their quantitative analysts, who are often the same) are constantly inferring the correctness of their models from a wide range of actual option prices. Thus, reality trumps the models, which are not discarded but modified to recognize that variables have term structure, surfaces, etc. These tweaks are not only elegant, but they are effective ways to adjust simplistic assumptions (e.g., a Brownian motion process producing a measure of volatility) to conform to the realities that go into making a price. Mr. Erman only hinted at that towards the end, when he referred to volatility smiles (part of the volatility surfaces inferred from the pricing of options of a very broad range of strikes, maturities, etc). This is not just playing with numbers, but allowing the risk managers to look into what the market thinks about risk in a very discrete fashion. Discrete enough that in a well functioning market, outlier risks/returns could be priced much more realistically than people think — Black Swans and all. In fact, when markets fail to do this, it is option traders (like Mr. Taleb) who would suspect it first. They are also the ones that for many asset classes would keep cynics (like Mr. Taleb) from continuously making Black Swan profits by systematically buying underpriced, out of the money insurance. We should spend more time exploring what are the systematic biases (regulatory moral hazard, e.g.) that can make some markets, through some periods, immune to pricing an appropriate volatility skew — e.g., housing, developed markets sovereign credit. I don’t think is necessarily the math, and certainly not options pricing, that let us down.

Patrick R. Sullivan
Mar 12 2012 at 4:43pm

About ‘tawdry’; you can say that about any field. I have relatives who are writers in Hollywood and have told me amazing stories about actors, directors and producers. Most of which are not very admirable.

And they were inspired by Shakespeare, Chekhov, Ibsen….

Eugene F. Fama
Mar 13 2012 at 12:31pm

Enjoyed the Derman interview, but a few facts need correcting.

(i) The Efficient Markets model says nothing about the probability distribution of returns. (Two thirds of my Ph.D. thesis, which is taken to be one of the parting shots in the EM literature is about non-normality, specifically fat tails in the distribution of returns.)
(ii) The CAPM is model of market equilibrium (a story about risk and expected return), based on Markowitz’ portfolio model. Like all models of market equilibrium, the CAPM assumes that markets are efficient. But the reverse is not true. Testing the EM hypothesis requires a model of market equilibrium, but not necessarily the CAPM.
(iii) The EM model does say that prices are correct; they are the best one can do based on available information.

Mar 13 2012 at 2:50pm

The most interesting aspect of this interview for me came from the fact that Derman’s background is in physics rather than finance. I came to financial engineering from a background in financial economics (“the branch of economics that deals with risk and time”) and, perhaps because of this, my perspective is very different.

For example, I never met anyone who believed that returns were normally distributed. As Dr. Fama points out, that literature was rich decades ago. The markets didn’t believe it either, as evidenced by the existence of observed volatility skew. Similarly, no one believed that markets were complete and frictionless, which is why option traders both calculate and avoid gamma – it costs money to delta hedge.

Black-Scholes holds that volatility is constant, but no one in the market believes this either. Options, in essence, are trades in volatility rather than price. You might hear an options trader quote an ATM option in vol points (“I’m 25 vol at 26 for Oct”). Likewise (although this didn’t come up during the interview, it often does) I never met anyone who believed that delta or delta-gamma VaR (or historical or expected tail-loss or PCA either, for that matter) worked in any but “normal” markets. Discontinuties and drastic changes in correlations were always understood to be outside the operating specs. This is not to say (as Russ suggests) that they are not valuable. It is very useful to know that your book is not likely to blow-up during normal markets, trust me!

It has already been pointed out, but I will reiterate that financial engineering is all based on observed market prices, not “just” models. You start with building blocks (debt, forwards, futures, swaps, options) and you know that you can turn combinations of these into the others for a cost. As Black and Scholes noted, we can create call option payoffs with combinations of debt and the underlying, and so there should be a predictable relationship between the prices of these two approaches to achieving the same results. If not, we arbitrage until that relationship obtains. This may not exactly happen in the market for apartment buildings, but it certainly happens in physical commodity markets in which people will, for example, blend different grades of crude oil to generate a quality of crude that is fetching a higher market price than the cost of the components.

Finally, try as they might, I don’t think that anyone has really posited a credible connection between the practice of financial economics and the financial crisis, so please ease up on the financial economists! It’s sort of like blaming the Challenger disaster on physicists because somehow they got gravity wrong.

Matt Cooper
Mar 14 2012 at 6:06am

It strikes me that all these financial models assume freely functioning markets, but that assumption is invalid. Over the last 5 years here in the US, we’ve had bailouts, QE I, QE II, Operation Twist and so on. The price and value of money itself is determined or at least greatly influenced by political actors.

John Berg
Mar 14 2012 at 5:12pm

Is a “model” a spread sheet with the relationships between each element of the table carefully explained/documented by mathematical expressions?

John Berg

Sri Hari
Mar 14 2012 at 8:07pm

I enjoyed the the down to earth interview style of Derman. After having experienced lot more financial engineering than most who claim to be experts, Derman confesses how little can be reliably predicted when there is high level turmoil in the markets. Black-Scholes, CAPM are fine to put a value around risk when there is less volatility and when most market participants are playing it fair. As markets are getting increasingly crimogenic (term I learnt from William Black) these models are highly unreliable and almost a childish in essence!!

Mar 14 2012 at 8:08pm

I was going to comment and point out that we’ve known for a long time that stock prices are not normally distributed, after all Fama wrote much of his dissertation on fat tails. But Eugene Fama has already commented on it!!!! What a wonderful world we live in!

I also agree MG that the B-S price has never been the market price. It is reasonably close in many situations, but especially with far out of the money options, the market has always priced in tail risk. Whether it is priced perfectly, who knows, but very smart people are trying to price it correctly and there are large gains to being better than the next guy.

Additional work in Finance is using market prices on assets, options and forward prices to better understand how the market prices risk. It tries to disentangle price of tail risk and variance risk that’s implicit in the equity risk premium. It is very hard to do, but I think there are some large methodological gains being made in the recent literature. I think it is a pretty exciting area. But the larger point is that we know it’s in there, we know it’s being priced, if we can isolate it and study it, perhaps we could learn to price it even better.

Mar 14 2012 at 10:50pm

Dr. Russel Roberts,

You know you are doing something right when Eugene F. Fama comments on your podcast.

Great job.

Jeff Boyd
Mar 15 2012 at 1:48pm

Delightful podcast. Sometimes shake my head in amazement over how something can be as good as econtalk.

Mar 16 2012 at 7:47am

I Agree with MG, Charlie and txslr.
Corrections made by Fama are also very valuable.

Greg G
Mar 16 2012 at 11:43am

Great interview Russ. I especially appreciated you both pointing out that the self-perceived shortness of capital of the investment banks (before they went public) was a feature – not a bug.

I used to be an independent retailer. When I built my own building my dad gave me one of the few pieces of financial advise he ever insisted on. He told me to be sure not to build more space than I needed. Having to operate within that limited space was a discipline crucial to making good buying decisions and achieving the necessary turnover and sales per square foot.

With hindsight we can say that having more limited capital was crucial to investment banks maintaining the discipline to limit themselves to the investments they believed in the most.

Mar 18 2012 at 4:26pm

I enjoyed this podcast more than I expected partly because Derman turns out to be so endearingly modest. His quant skills put him in rarefied company indeed, yet he’s at pains to admit how much he doesn’t know. If only that attitude were more common in all disciplines and for that matter, walks of life.

Emanuel Derman
Mar 18 2012 at 6:25pm

I appreciate Prof Fama’s comments and don’t disagree. Here are a couple of points I would like to accentuate:

  • I have great respect for the efficient market model and for neoclassical finance. It’s comprehensive in scope, and nothing I’ve seen that has come along since then measures up to its ambitions and scope.
  • I don’t think and have never thought that models were responsible for the financial crisis. I have a much more macro view of that issue.
  • I realize that people have known for many years that there are fat tails. The trouble is that there is no systematic way of dealing with them, and most people in practice deal with them by changing the parameters in (say) the BS model. Or in other cases by using stress tests.
  • Prof Fama stresses that efficient markets don’t imply CAPM, and that CAPM is a very particular case of efficient markets. In my book Models.Behaving.Badly, I suspect I do tend to conflate the two as though they were one, which is unfair. On the other hand, though, almost everyone in practice uses CAPM or extensions of it like APT, and I was trying to write something general.
  • From my point of view, the trouble with CAPM is that it assumes all risk is defined by standard deviation. That’s the untrue part of the model, much too limited.

In his article “Noise” Fischer Black wrote

… we might define an efficient market as one in which price is within
a factor of 2 of value, i.e., the price is more than half of value and less than twice
value.1′ The factor of 2 is arbitrary, of course.

Trouble is, how does one know what the value actually is?

Mar 23 2012 at 3:40am

“Trouble is, how does one know what the value actually is?”

It sounds like you’re alluding to the joint hypothesis problem. You know who’s written a lot about that over the last 40 years…. I’ll give you one guess:

“My main contribution to the theory of efficient markets is the 1970 review (Fama 1970). The paper emphasizes the joint hypothesis problem hidden in the sub-martingales of Mandelbrot (1966) and Samuelson (1965). Specifically, market efficiency can only be tested in the context of an asset pricing model that specifies equilibrium expected returns. In other words, to test whether prices fully reflect available information, we must specify how the market is trying to compensate investors when it sets prices. My cleanest statement of the theory of efficient markets is in chapter 5 of Fama (1976b), reiterated in my second review “Efficient Markets II” (Fama 1991a).

The joint hypothesis problem is obvious, but only on hindsight. For example, much of the early work on market efficiency focuses on the autocorrelations of stock returns. It was not recognized that market efficiency implies zero autocorrelation only if the expected returns that investors require to hold stocks are constant through time or at least serially uncorrelated, and both conditions are unlikely.

The joint hypothesis problem is generally acknowledged in work on market efficiency after Fama (1970), and it is understood that, as a result, market efficiency per se is not testable. The flip side of the joint hypothesis problem is less often acknowledged. Specifically, almost all asset pricing models assume asset markets are efficient, so tests of these models are joint tests of the models and market efficiency. Asset pricing and market efficiency are forever joined at the hip.”

Mar 27 2012 at 12:58pm

“So, we did one on the financial crisis last week and read, used as a basis a paper by Metrick–two guys from Yale, Metrick and somebody, weekend’s worth of reading on what happened in the financial crisis and various people’s theories about it.”

This paper is,

Getting up to Speed on the Financial Crisis: A One-Weekend-Reader’s Guide

Gary B. Gorton, Andrew Metrick

Excellent paper. Please get Gary Gorton in for a future podcast, Russ–very interesting economist, used to work at AIGFP, wrote some really insightful papers about the GFC, and is also quite funny.

Mar 27 2012 at 6:44pm

I think the distinction between Theory and Model is a very useful one. The difference intuition makes is quite interesting (and one which seems to mirror the distinction in the Myers-Briggs personality types).

I find the distinction especially fascinating, because the lines in science are blurring again. I’m a graduate student in computer science, working in scientific computing. A lot of science now is moving towards computational modeling. We have massive amounts (petabytes = 1,000,000 gigabytes) of data that are being sifted through by computers. Scientists are trying to derive correlations between thousands or even millions of features. This is not something that can be intuitively grasped. These computer-identified correlations are being assembled into different physical models. I’m not sure how the models translate into theories. I suspect the theories could be derived by inspecting the models, and then using intuitive leaps to derive a theory, which gets tested a la Newton. On the other hand, I think a lot of these models get called “theories”, directly. I would love to hear Derman thinks of this.

Mar 27 2012 at 7:13pm

Russ, I am an ex-physicist quant and want to thank you for your podcast. I have learned so much economics listening to Econtalk, and I find the depth and thoughtfulness of the discussions to be a delight. One thing you said in this podcast caught in my ear, however, and I believe is a misunderstanding of what pricing models are for. You said:

I have to say, the part I find strange about this–not being a finance person–is the focus of Wall Street on getting the right price. We don’t do that anywhere else in the economy. If you hadn’t, if it was a long time ago and we hadn’t had this scientific revolution in mathematical finance and financial engineering, if you had said to me: What’s the right price of the option? I’d have said: Well, sell it. Put it on the market. See what people pay for it. That’s the right price. That’s how we price everything else in the world.

Option pricing models are not useful for determining the market price of the option. They are useful for estimating the future COST of producing that option. The vast majority of products that are bought and sold in the economy are produced before they are sold, and the cost of producing them is known – you don’t need a model. Options are very different, in that they are a contract to provide a product/service in the future where none of the cost associated with writing the option has taken place at the time of the transaction. There are other products like this in the economy outside of finance, and similarly sophisticated mathematical modeling is found in those fields to try and forecast future costs of producing the product. Insurance is an obvious example, where much modeling is done to try and estimate the future cost of the policy to the insurance writer in order to properly “price” the insurance policy. I’m sure that other businesses writing forward contracts for real products they produce also use future cost models in determining what price they are willing to accept and don’t just use a “market price”. Selling future production blindly into a market, without knowing what the cost of that production will be is a sure way to drive any business into bankruptcy. If you have a current inventory of product and the current market price is below your ALREADY SPENT production cost, then you already have a real loss whether you sell the inventory or not (assuming reasonably efficient markets). But to sell future product at a market price below your NOT YET SPENT production costs is to take a loss you could have avoided.

James Oswald
Mar 28 2012 at 3:26pm

The reason why we use mathematical models to determine the value of stock prices and not things like apartments and ice cream, is that the primary value of a financial asset is only the expected monetary value of it. There is no subjective value to account for, like you have in other markets.

C Divakar Dhaveji
Mar 30 2012 at 2:37am

I have just this to say (not being an economist but a lay person with some interest in the subject). I bumped into Econtalk podcasts by accident as I was trying to download podcasts on current affairs and other areas. I have todate heard many of the podcasts in my car on my way back from office in the evening and have immensely enjoyed several of them. Admittedly, not everything I heard, I claim to understand. Russ Roberts does a great job and so do his guests. I am in development sector consultancy and listening to Econtalk podcasts is not only a supplemental learning to my ongoing online management course on the University of People of United Nations but also (more importantly) refreshing.

About this week's guest: About ideas and people mentioned in this podcast:

Podcast Episode Highlights
0:36Intro. [Recording date: February 23, 2012.] Russ: I want to start with your personal story. You began your career as a physicist, and then you went to Wall Street. Tell us about that journey. Guest: I originally grew up in South Africa, and I liked literature and the arts. But, you had to specialize very young when you were in South Africa. It was like a British system. When you were 16 or 17 you went to college; that was it. You did arts or sciences or medicine or law or business. So, I liked physics, and I went into physics. And I got very attracted to it, and I was inspired by all the famous physicists, just like everybody else. You read about Albert Einstein and Erwin Schrodinger. So I did four years of physics in South Africa and then I came to Columbia University, where I teach now; but I teach financial engineering. But I came to do a Ph.D. in physics over there. It took me a long time--7 years. I had the wrong kind of background, a very classical physics background, and Columbia was doing, obviously, very modern physics and quantum mechanics. And then I worked--I did a thesis on weak interactions, on theoretical physics--and I worked as a post-doc and an assistant professor for 7 years, doing research and kind of liking it. But also getting a bit discouraged at times. Russ: And why did you leave physics? And where did you go? Guest: I graduated from Columbia, and it was very difficult to get jobs. Hatfield-McGovern Amendment at the end of the Vietnam War passed [? The bill failed, but the discussions had repercussions.--Econlib Ed.], and a lot of the physics jobs in academic life, even the theoretical ones, were being funded by the Department of Defense. And some of the [?], and I had a 2-year--no permanent jobs. And even temporary jobs you were lucky to get. So, I had a 2-year post-doc at the U. of Pennsylvania; and then I had a 2-year post-doc at Oxford U. in England in theoretical physics; and then I had a 2-year post-doc at Rockefeller University back in New York. And my wife and I were both in academic life and juggling being in different parts of the world. And then I got a tenure-track assistant professor job at Boulder, Colorado. I kind of went there but my wife couldn't get a job there, and we had a two-year-old kid already, and life got complicated. And that was--nominally while I [?] physics, I want to say I was getting a little discouraged. Physics is very difficult and very appealing at the same time, and after a while, unless you are really, really good, you feel like you are no good at all. Russ: That's not uncommon in many fields, but I think it's particularly true in physics. Because you feel like--as a friend of mine once described it--you are just slicing the salami; you are not really extending anything. Guest: Good metaphor. Yes. You run into people whose papers you could maybe understand but you could never do anything like that. Even if you are smart. Russ: It's sobering. So, how did you go from that? You are thinking: I wish this were better. How do you go from that to being on Wall Street? Guest: Well, in two steps. First, I wanted to move back to New York where my wife was still working and where my son and my wife were. And at that time, what Wall Street was for the last 20 years--in 1980 it was telecommunications and energy, because of the energy crisis; and Bell Labs was hiring a lot of ex-physicists and so-called rocket scientists to do business research. And I ran into some people from there; I'd known some other physicists who went there, and I got a job at Murray Hill in a business analysis systems center, using ex-physicists and computer scientists and people to sort of do, I won't even say financial modeling. Maybe financial modeling and then build financial modeling tools for people at AT&T headquarters. And I did that for 5 years. That was sort of a big shock to my system, actually, because I was used to being in academic life where you think you are your own boss and you can do whatever you like; and suddenly I was in this industrial complex where everybody had a supervisor and every supervisor had a department head and every department head had a--I forget what. Russ: But somehow you end up at Goldman Sachs. Guest: I end up at Goldman Sachs because I actually learned a lot of useful stuff at Bell Labs, but maybe to my detriment, I hated it. I took the authority over me very seriously in those days. Russ: You were spoiled. Guest: Yes. I think it's an academic thing. I don't know. I was definitely spoiled. I always wanted to get out of there, although I actually learnt a lot of useful stuff. I kept thinking: Should I go back to physics or should I leap all the way out, and eventually Wall Street came knocking on people's doors and I went all the way out, although it took me 5 years. Well, the first 3 I didn't even think about it; and then it took me like 2 years to decide to go. Russ: So, what year, roughly, did you decide to go to Wall Street? Guest: I went to Wall Street the end of 1985. I actually was offered a job there a year before and turned it down. Russ: And how long did you stay? Guest: I stayed on Wall Street from 1985 to 2002. Russ: Long time. Guest: And I still actually work part time at a fund to funds called Prisma Capital Partners, run by some ex-Goldman people in NY; I usually work there on Fridays.
6:30Russ: So, when you were at Goldman, you actually were there during the transition, I think, from being a partnership to being a publically traded company. Guest: Yes. I came there in 1985; I actually worked at Seligman for one year in 1989 and then I went back to Goldman. I was there till 2002. They went public, I don't know, maybe 1999. Russ: Yeah, late 1990s. I'm just curious, because it's an issue that I'm intrigued by. Did you notice any changes in the company from that? Guest: Yes, very marked. Although some of it had to do with growing in size, and then some of it had to do with going public. But, I always think--maybe it's sad. But Goldman, I think they felt they had the obligation to go public because they thought they were, as a private partnership, up against all the big commercial banks, which could make low interest loans to people in order to get business and not have to mark them to market, whereas Goldman had to mark those things to market. And they always felt a shortness of capital. In 1994, they had a lot of partners leave when the fixed income world had a very bad turn, and I think they got scared. And they went public. I think the firm did change; it got much bigger. It got, I would say, internally more secretive and less congenial. I think it was a very well-run--probably still is a very well-run--firm, because there were partners seated all over the place who suppressed, for lack of a better word, sociopathic behavior. Because their capital is going to be locked up in there for 20 years. Even if they retire they had to leave their capital there. And so the place functioned very well. One thing that changed was that [beforehand--Econlib Ed.] they weren't playing with other people's money--they were playing with their own money; and that's always a big sort of sanity contributor. Russ: That's the thing I'm interested in--when you go from a world--it's true it's nice to be leveraged. It's hard to be leveraged when you are a private partnership and the public doesn't allow them to borrow more than they otherwise would have which made them more risk-taking, I think. Guest: Yes. Although they had episodes before, I think in the mid-1990s when they went public, sort of having internal hedge-fund-like groups. They were always trying to raise money--I forget, the head money fund, there's some Hawaiian school's wealthy endowment; and they had money from them and from some [?] because they were always short on capital. So they solved it by going public. Russ: Seems like a good thing, always being short of capital. The alternative always having plenty is not a healthy situation, maybe. Guest: I agree with you.
9:33Russ: So. In your book, you talk at length and in very interesting ways across science, philosophy, economics, and finance about the distinction between theories and models. Those are two words that we often use interchangeably to mean our attempt to understand the world. But you make a nice distinction between them. Explain the distinction and talk about the differences between theories and models in science and in the social sciences. Guest: Thanks. I came to that--it's my use of the word, somewhat, although I like them; everybody has a slightly different definition of theory and models. And I did used to equate them. I used to, a long time ago, talk about fundamental models and more phenomenological models which just tried to describe the world but aren't fundamental. But eventually I decided I liked the distinction between theory and models. Let me start with theories. To me, theories are really attempts, successful or unsuccessful, to describe the world the way it really is. They are kind of destructive and they look at the world. So, for example, if I were to take Newton's Laws and force = mass times acceleration, and the inverse squared law of gravity--those are descriptions of the world that are theoretically framed in mathematics but they are not an analogy. They are trying to describe the way things actually work. Russ: Yeah; they are not an approximation, except under various situations. Guest: Yeah, they are not an approximation; although they may not be dead right. I think sort of [?] is a theory or, I mention in my book I think Spinoza's theory of emotions is a theory because it doesn't rely on analogies. It doesn't have to be correct, but its aim is to describe things as they are. Russ: And a model? Guest: A model--I sort of came to the conclusion while I was writing my book that it was really much more of a metaphor, an attempt to find an analogy between something you want to understand and something to really do understand, either heuristically or by a theory. So, I have an example in my book--there's a quote by Arthur Schopenhauer where he says, a very pretty metaphor: Sleep is--I'm going to have a hard time remembering it, now. Russ: I'll riff for a minute. If you have the book nearby you can look it up. It's a haunting quote to think about, because what he was dealing there was: Is evil the absence of good or is evil its own entity? Is darkness the absence of light or is there really something called darkness? Is death just the absence of life or is death as vivid and life is the thing that's sort of temporary or otherwise. Those are the things I remember. Guest: I wrote a big section about presence and absence. I'm just trying to find his quote. Here it is. He says:
Sleep is the interest we have to pay on the capital which is called in at death; and the higher the rate of interest and the more regularly it is paid, the further the date of redemption is postponed.
Actually, I have no idea where I came across that any more. But what he's doing is, he is looking at sleep and saying one of the characteristics of sleep is that it happens periodically, every night. And one of the characteristics of bonds is that you owe interest on them and you pay them every month or every six months. And based on the analogy of the similarity between the two periodicities, he then says, since you pay interest because you once borrowed money and you have to pay it back, in the same way you go to sleep at night and subject yourself to darkness because you once borrowed you life from the darkness and you have to give it back at the end. Russ: And that's a metaphor. Guest: So, it's a metaphor. But it's based on a limited analogy, which is the periodicity of coupons and sleep; and then he's extending it or sort of analytically continuing it in a mathematical way to say something about sleep and life by comparison. And I think models are like that. I was going to talk about the efficient market model later. But I think that's similar in that you sort of say: stock prices or the returns on stock prices behave like smoke diffusing. And there's something similar about them, but it's not an accurate description in the way that, say, Newton's Laws attempt to be an accurate description. It's really based on an analogy to something you do understand, which is smoke diffusing, and saying maybe stock prices behave a lot like that. Russ: And it helps you--models are useful. Because they can, if the analogy is accurate enough, help you get at the intuition of what's going on. When I was reading those sections I was thinking about macroeconomics, which you touch on very briefly in the book. You touch on finance, but you mention economic models and I think about them way too much. But obviously, there's no such thing as Aggregate Demand. It's a conceptual idea to help us try to understand our myriad of connected interactions as buyers, sellers, employers, employees, investors, etc. And it may be a useful metaphor; it may not be. But it has a black box quality. In a theory, you are getting at the mechanics of the black box. Guest: Yes, I agree with you. I like what you said about Aggregate Demand, because I read something by Hayek once where he said that in the physical sciences what you really understand are macroscopic concepts like pressure and volume and temperature and tension and stuff like that; and then you deduce from them the existence of atoms, which you never see but which explain those things. And I think he then sort of points out that macroeconomic things like supply and demand and liquidity are really not things that you observe. They are the metaphors. What you really observe are the individual people. There, it's the elements that are real and the other things are more metaphorical. Russ: Yeah, it's true. Guest: People are--he sort of said theoreticians are going in the wrong direction in economics by trying to mimic the direction of physics, going from the large to the small. Russ: That's obviously an ongoing debate that's not settled.
16:21Russ: The other part of this section of your book which I found utterly beautiful and fascinating--and correct me if I'm wrong, I think I read it in your book; I've been reading some other things as well--but it's sort of the order of theory and empirical evidence. You say at one point that Newton saw the way the world worked; and then he did some experiments to confirm it. He didn't do a bunch of experiments, noodle around with the data, and then say there seems to be a relationship here. He actually saw the answer, saw the theory, and then he had to confirm it. It might not turn out to be true, but the flash of insight came first. Is that what you wrote? Guest: Yeah, that is what I wrote. A little bit about Newton and then a lot more about Ampere and Maxwell. I have one chapter about electromagnetism and comparing it to the history of the efficient market model and trying to explain exactly what you said, that there's a succession of physicists who make great descriptive discoveries like Newton's Laws or Ampere's Law of electromagnetism or Maxwell's equations; and each one of them is recursively amazed by the person who came before them could claim to deduce this law from experiments because it looks to them like they must have thought up the law first and then checked that it confirms experiment. Russ: And sometimes that's true because there's a faith, almost, in the underlying mathematics that it implies something that has to be there even though you can't see it or don't have the measurement tools; but eventually it's found; and that's an unbelievable aspect of physics. Guest: Yeah, a lot of things. I'm a big fan--fan is the wrong word--of people who play the role of creative and artistic intuition in all these discoveries--in economics and in physics and in mathematics. In mathematics I think people understand it. I was sort of reading a bit of Kahneman's book on fast and slow thinking, and he points out all the errors in fast thinking. But I'm sort of a fan of equating that roughly with intuition. I'm still sort of a fan of intuition. And I think of observation--at some point there has to be some leap which comes from somewhere or another. Russ: And I think the thing that fascinates me is we don't understand that. And I think even more importantly the people who make the leap don't understand it. There's this wonderful description by Andrew Wiles where he, having proved Fermat's Last Theorem, finds out that his proof is wrong. He spent 7 years proving it; he proves it; it's on the front page of the New York Times, and then someone finds an error in the proof. And it's back to the drawing board, and he spends 18 months. And in those 18 months he must be in total despair most of the time. Guest: Yeah, it sounds like it. Russ: And at one point, I think he said it was Saturday morning, he's sitting at his desk, he's looking at this equation he's been looking probably 8 years; and all of a sudden he says: I just saw it. It's not like, oh, he did all this work, and dug this ditch and then the water flowed in. It's just something inexplicable, miraculous really; he just saw it. He didn't see it before, can't explain how he saw it. It's an amazing thing. Guest: Yeah. And that's what I wrote in my book, about Kepler and Newton and Ampere, Maxell, and Einstein, who all do something like that; and everybody who writes about them--in fact Ampere's paper is called something like: Deduction of the Laws of Electromagnetism from Observation. And Maxwell and Poincare say there's no way he could have done experiments to deduce these laws. He had to know, he had to sort of intuit the law and then confirm it. Russ: There's that story--I've mentioned it before on the program--where Einstein's theory is confirmed by the bending of the light of the star, by the sun during an eclipse; and when asked what he would have done if it had been proved wrong, he said: Well, I wouldn't have believed the experiment; because I know my theory is right. Guest: I just saw today on the Internet that there is now actually a flaw--there are these people at Cern who claim that neutrinos travel faster than light. And there's now some report on some website today saying that their GPS system was connected by faulty wires. Russ: Oh, no! Guest: And there's a 60 millisecond gap there. It proves it, but there's definitely a mistake there. Russ: Truth is elusive. Of course, we are all overly confident in our theories. But sometimes they are right. And Einstein's right more often than the rest of us, more of the time.
21:02Russ: But let's move on to models. So, we have theories, which are testable and are often confirmed. They may not be perfect, they may need tweaking; but they are trying to get at the way the world is. And then we have models, that try to get at--the way I understand your idea, the underlying intuition of what's happening based on an intuition we have of some understanding we have from somewhere else. Guest: Yes; and my claim is it's usually a model that's usually only a partial mapping between the thing you are interested in and the thing you are trying to compare it to. In fact, if I get pedantic, I read some history of Maxwell when I was doing this, and he eventually came up with Maxwell's equations, which are just mathematics and describe electricity and magnetism; but he actually proceeded by first building a bunch of metaphorical models in which he thought of magnetic lines of force as fluid flow. Which they resemble but aren't exactly identical to. He actually says, very categorically: I'm trying to get some intuition and I want to build a model; I don't believe it but I want to see what I can extract from it. Russ: Irving Fisher, economist in the early part of the 20th century, used hydraulics--the flow of water. I don't know which book it was in. He's got a book where he's got these complex hydraulic models, water flowing through pipes and into basins, and it's supposed to capture something about liquidity and interest rates--the level of the water is analogous to the--. And they are really beautiful, really interesting. Everyone knows that financial markets are not bathtubs. They may be like bathtubs in some dimension, but they are not bathtubs. And that can lead you astray. Guest: Jeremy Bernstein, who is a physicist I know, gave me a quote that I use in my book that says--I think it comes from Herbert Simon, but I'm not sure now--he has a drawing of a bird and he says this looks like a bird, but no bird looks like this. Russ: So, let's talk about the models in finance, particularly ones that you worked with for that long period of time on Wall Street. I recently interviewed Eugene Fama, who is associated with the efficient markets hypothesis. Then you also talk about Black-Scholes, and you talk about the Capital Asset Pricing Model [CAPM]. Let's talk about each of those in turn, if you would--what you like about them and what you are not so crazy about. The efficient market model--it seems to me everyone has a slightly different definition of what they mean. I know Burton Malkiel reviewed my book in the Wall Street Journal, and he liked it generally, he was complimentary, but he claims in one paragraph that I was putting too much weight--he sort of claimed that all the efficient market model says is, I don't know what's going to happen next, as opposed to saying that current prices are right. Russ: The first claim is a modest claim. Guest: Yes, it is a modest claim. And I think maybe he is technically correct, but I think in everyday parlance people strive for a stronger version of it. Russ: Yeah; I think the way you wrote it in the book is that prices reflect all publically available information. Guest: Yes. Russ: It's an attractive idea. And there's something to it. Guest: Yes. And I'm getting off topic a little, but I'm always a little disappointed by what's happened in Behavioral Finance and economics because whatever people do, it's sort of interesting in a small way, but it doesn't seem to have the overall breadth that the whole Efficient Market aspiration of the model had. So, I like the model; but when people put it into practice--and here I'm thinking of CAPM--they assume geometric Brownian motion for stock moves. And so I think what's wrong with it--I think it's probably right in some sense, or at least defensible, that current stock prices reflect at least all available information. But I think the bad, the faulty part of the model is saying that stock prices satisfy, geometric stock returns satisfy geometric Brownian motion, and that risk is this well-defined thing you can tap into by just giving a standard deviation or a volatility. It's got the qualities of something risky, which is nice, but it doesn't have all the attributes of real-world risk. Russ: Going back to Hayek, it's using, say, Brownian motion or other high-level mathematical concepts to fancy-up the model, is scientism. Not science. Guest: Yes. Russ: It gives the patina of science. I like the Gilbert and Sullivan quote: It gives verisimilitude to what is otherwise a bald and unconvincing narrative. It gussies it up and makes it look fancy; but maybe it isn't true. Guest: Maybe it's kind of a valiant effort. I don't agree with Nassim Taleb; I think everybody who invented these things deserves whatever Nobel Prizes they got. But at the same time, the real world is much more complicated. The real world of human beings and stock returns and markets is much more complicated than can be captured in saying stocks' risk can be captured by their standard deviation. Russ: But I think the plus side of it--and again I think there's a big difference between how I use the efficient market hypothesis and how Wall Street people use it--but for me, when somebody says: I have a great idea for a stock, because I saw that this company just invented blah-blah-blah, I always say: It's too late. Now, it's possible it's not too late. Maybe the price hasn't already risen. But as an inducer of caution, it's a very effective idea. And I think it's a very powerful idea. Where it seems to go wrong, of course, is that things don't always act as we expect. Arbitrage conditions don't always hold in every instant. And there is often psychology that intervenes in ways that make markets work a little differently than you expect. Guest: Yeah, exactly. And stocks jump. People get panicky, in this contagion. None of that violates the strict mathematical assumptions behind the efficient market model. Russ: And that may mean there is money to be made that wouldn't otherwise be available. But the odds that you are going to make it can maybe still be very small. Which is why I think it's a good, sobering thing to keep in mind.
28:08Russ: But the other two models are more complicated. And you spend a lot of time on Black-Scholes. Talk about what Black-Scholes tries to do. Guest: I was sort of amazed by Black-Scholes when I first saw it. And then for a while, 10 years ago or so, I thought it wasn't so good. But I've sort of come around to thinking that it really is admirable. And the way Fischer Black and Myron Scholes originally derived it, the derivation you read now is always the stochastic calculus one, but the way they derived it was by saying qualitatively that the [William] Sharpe ratio of a stock, and the Sharpe ratio of an option, and the Sharpe ratio of the underlying stock that the option is written on has to be the same; that the excess return per unit of risk for both of them should be the same when the market is in equilibrium. And I think that's a completely reasonable statement. You are saying the option and the stock have the same underlying source of riskiness. And if you use one or the other, then if everybody is sensible, then you'll get the same excess return per unit of risk, whichever one you decide to employ. Russ: So, back up a little bit. First, explain what an option is. And then you can try to get to the Sharpe ratio without a blackboard. But you might not be able to do that. But the intuition. Guest: I can try. Everybody knows what a stock is. And a stock price can go up or down at any instant, and nobody generally knows which way it is going to go. So, an option is a derivative or a contingent contract on a stock. And for example, a call option says: you only pay for a contract that gives you an upside, but not the downside. So, if a stock goes up, you make a dollar that a stock goes up eventually; but if the stock goes down, you don't lose anything. And so it's a-- Russ: And you pay for that privilege. Guest: You pay for that privilege. And the big puzzle of Black-Scholes was how much do you pay for the upside risk but no downside risk; in this case. And what Black and Scholes came up with, from my point of view of looking at it, was saying--it's tricky. But the source of risk for both the option and the stock is the same. But the mathematical characteristics are different because the one is on the upside and the other has symmetric risk. And you can find the risk of the option in terms of the stock, mathematically. Russ: So, to think about it again, without the math: if the stock price goes up, I make money if I hold the stock. If the stock price goes up, I make money with the option. But if the stock price goes down, I lose money. But not with the option. So, there's got to be a value to that. They both have risk. Because in both cases, the stock can go down. In which case you lose--in the case of the option, you lose what you paid for it. You forfeit that. But you don't lose more than that. So, obviously the value of those two things has to be interconnected. And that's what this tries to calculate, right? Guest: Right. And trying not to be mathematical--I'll have to think more about how to say it--but Black and Scholes show that if you borrow money and buy the stock, you can buy something that's very much like the option for a moment. Russ: That's the clever part. Guest: That's the clever part. I like to separate the science and engineering. The science is--the ostensible science is defining what the risk of a stock is like, and the geometric Brownian motion, and that's scientific, but it's not accurate. The theory is wrong, or the model is wrong; it's not a real description of the world. But then the second idea is their engineering idea, which is that you can create the risk of an option by borrowing money and buying the stock. That's a very great idea, and they are sort of building a recipe for creating an option, for creating the risk of an option out of borrowing money and buying a stock; and that's the sort of engineering/modeling idea. And I think that's completely right and everything they derive after that is right if you assume geometric Brownian motion; but it's the geometric Brownian motion, the risk of the stock, that's badly modeled in the efficient market model. Russ: I have to say, the part I find strange about this--not being a finance person--is the focus of Wall Street on getting the right price. We don't do that anywhere else in the economy. If you hadn't, if it was a long time ago and we hadn't had this scientific revolution in mathematical finance and financial engineering, if you had said to me: What's the right price of the option? I'd have said: Well, sell it. Put it on the market. See what people pay for it. That's the right price. That's how we price everything else in the world. We don't say--you give the examples, it's a nice example, the example of apartment buildings. We don't say: Well, this apartment has so many square feet, it's so many miles from this subway station and it's got this school district, so therefore the right price is $2300/month. And if somebody says: Well, I'll pay $2400, we don't say we hear a fool. That's what people value it at. Because we don't pretend to have scientific weights on the attributes of the apartment. We understand that it has something to do with square footage and something to do with location, but we don't pretend to try to weight those, because everyone weights them differently. And similarly, with financial products, people weight risk differently. Of course they can diversify some of it away. And I think maybe this is your point: that you can't diversify all of it away and pretend that you can, pretend you can create a portfolio of something that's "the same" when it's not quite the same. Maybe it's an illusion. Guest: Yeah, I agree with you. I wish I knew more economics. Maybe one day off the interview, you can tell me what's a good way to get educated in economics. Russ: You just have to listen every Monday morning, Emanuel. EconTalk comes out every Monday. An hour a week! I don't know. In two or three years, you'll be a genius! Like my other listeners. Guest: Well, I still don't know much about economics, but I sort of learnt everything the wrong way around. I learned finance first and economics later; and I'd really like to learn more economics now. It seems to have more importance. Russ: Well, yeah, more and more people seem to feel the same way. Maybe we ought to pay more attention to it. Of course, a little learning is a dangerous thing, as folks have said. Guest: Alexander Pope? Russ: Yes. He said:
A little learning is a dangerous thing;
drink deep, or taste not the Pierian spring:
there shallow draughts intoxicate the brain,
and drinking largely sobers us again.
So, either drink a lot or don't drink. Don't pretend you are an expert when you have only had a little bit of drinking, a little bit of knowledge.
35:01Russ: So, let's talk briefly about CAPM. CAPM is an attempt to extend some of the insights of pricing to a wider range of stuff. Give it a shot. Guest: So, I start with the efficient markets coming first, meaning that you don't know if stocks are going to go up or down, and therefore the simplest model is this model of Brownian motion, where stock prices can diffuse somewhat like smoke diffusing from a cigarette. And then Sharpe and Lintner and what's the name, Fischer as collaborator--Treynor--come up in various version with CAPM. Which, the way I like to look at it is, they are saying that: You should only get paid for taking risk that you have to take. And any risk you can diversify away or hedge away, you shouldn't get paid for because you didn't have to take it. It's only [?], obligatory risk that you should get paid for. And then they analyze how you can diversify, how you can diversify away risk or hedge away risk and end up with a price for the--the only risk you can't get rid of is the whole market's risk; because all the risk of individual stocks you can get rid of by putting together a big portfolio where the risks mostly cancel out. And they come up with a formula that relates the expected return of the stock to the expected return of the market. Russ: And what turns out to be important is some stocks are very correlated with the market--they move up when the market goes up in lock step--and others don't move that way. Guest: Yes, and so it depends on their correlation, their beta to the market, as finance professionals call it. And Black-Scholes, or actually Fischer-Black, was very in love with the whole CAPM market, and he just applied CAPM to options and derived the Black-Scholes model, which was what I was explaining imperfectly before. But I think Black-Scholes is much better than CAPM. Although it is based on the same idea. Because different stocks really have such different risk characteristics that the assumptions of geometric Brownian motion and the assumptions in CAPM don't hold very well for individual stocks. Whereas--and I think the model is much more flawed. Whereas when you apply it to options, you are saying the stock and the option on top of it have to have the same bang per unit of buck or the same bang per unit of bang per unit of risk. And you are talking about a much more confined world when you are talking about a stock and an option than when you are talking about CAPM, which refers to all the stocks in the universe. Russ: Black-Scholes isn't as ambitious. And that's no fun. Guest: So it's much more--in fact, yeah--that's a very good way you put it. It isn't as ambitious and it's actually able to go a lot farther. And the same is sort of true of I think fixed income financial engineering or quantitative finance compared to equities, because fixed income instruments have regular coupons and have a much more mathematical behavior, you can write down much more truer things about them than you can for stocks, where anything goes. Russ: And it's hard not to, as Nassim Taleb says: It's hard to not extend your scientific knowledge into areas where it may not be as applicable. It's hard to fight that urge.
38:37Russ: At the end of the book you talk about the crisis that we are still in the middle of. And you say that you are unsurprised by the meltdown, in some dimension. Talk about that. You say you are not surprised that economics didn't do so well, and that finance didn't do so well. For example--you don't mention it specifically in this book, but that the value-at-risk models that people are using that attempted to look the riskiness of an overall portfolio--they turned out to be ghastly and incorrect. You say you are not surprised by that. Why not? Guest: I think there was a time that I would have been surprised. When I first came to finance in 1985 from physics, I only saw the similarities between finance or mathematical finance and physics. And I really imagined, as I said in my first book, I thought you could write down some grand, unified theory of finance. As it did in physics. But over the years I got to realize that financial models--and they are mostly models, not theories--are really, are really glorified ways of interpolation. And I don't mean that in a bad sense. But what they do is: They take some analogy between financial markets and smoke or hydraulics or something that you understand better, and then apply that. And make predictions on the basis of it. And it works in a very narrow range when the world doesn't change too much. But when the world undergoes a crisis, your model is fundamentally wrong and inapplicable. And all bets are off. All the assumptions you make about individual prices being independent of what everybody else in the whole wide world does or having the reservoir of a market, all of that just drops away when people get panicky. And your whole interpolation mechanism based on some analogy breaks down when you have big moves happening. I think all of these models only work as long as you stay close to the regime you started out in. Russ: And of course, an even simpler model: That model is going to be like today; and today is like yesterday. You don't need all the firepower that these models typically bring. And then, what you want the firepower for is when you have a big change. Guest: And then they don't work. Russ: Yeah. I think about that a lot when I think about the Keynesian multiplier. We have got all these time series models in economics, we estimate these relationships; they are very stable. And then all of a sudden they don't hold and the model isn't useful any more. Guest: Yeah, I'm always amazed when I read on the Internet the battles between the different camps of economists; the effect of the multiplier and how big it will be. And nobody seems to agree. Russ: Well, it's difficult when you can't verify it. The biggest thing I've learned about macroeconomics since the crisis started, in a scientific framework or trying to; and the reason I don't think economics is a science, is that it's one thing to say your predictions aren't very accurate. It's another thing to think that you can't verify them. So, when people say that the stimulus created 3.3 million jobs--how would you know? It's an interesting idea. But there is no independent way to verify that the prediction is accurate. Other than the same model you used. So, it's like saying: We're sending a rocket to Mars. When it should be halfway there, you are asked: Is it going to end up there? You sure: My model predicts that it will. And then when it's supposed to land there, you say: Where is it? You say: Well, it's there! Because my model said that's where it's going. But if you can't see Mars and you can't get a broadcast back from the ship, you are fooling yourself. Guest: That's very well put. I always in retrospect admire Fischer Black. I was trying to find this quote, he was saying that reminded me: he sort of understood that models were not reality. I'm trying to find this quote. I think I've got it here. He says he is going to build a model and he is going to rely on stylized facts and introspection. Because he sort of understands that you are trying to come up with something plausible. Russ: That's what we do in economics. Guest: Here he says; sorry, this is actually Newton, actually Keynes quoting Newton. Newton had this great intuition, more than anything else. Russ: It reminds me--who was it? Maybe it was Alfred Marshall; maybe it was Keynes quoting Marshall. You had these great results, and somebody said: He hid the tools that did the work. Guest: Oh, oh! Russ: And I think--maybe it's not Marshall. I'll see if I can find it. Guest: Similar to what I was saying Maxwell was saying about Ampere. Russ: Right, sure. Guest: Murray Gilman [?] has some kind of French cooking in which you bake something between slabs of pork and afterward remove the pork, but not the taste.
44:08Guest: What sort of shocked me more was the breakdown of what I thought of as capitalist morals, somehow. Russ: Talk about that. Guest: Somehow I always, I don't want to say justified, but I always thought that capitalism was kind of a little brutal, but it's symmetric. You know, you make money if you take risk, and it takes human ingenuity and human endeavor to take risk, and that's what the market runs on; but if you want to get the upside, you have to get the downside. And what was horrifying for me during the last four years was seeing how some people had gotten the upside and said, praised risk, big corporations; and then they ran for government help and got government help when they were bailed out by taxpayers. Russ: And other people weren't bailed out. It was a very asymmetric system. People who borrowed a lot of money to buy a house they couldn't afford, they are struggling. But the people in large institutions that borrowed a lot of money to buy assets that weren't worth very much--the lenders got all their money back. It's a total mystery to me. And I used to argue with friends about it, and they would all say: Well, you want to just punish them and destroy the whole world in order to seek revenge? I don't know how close we came to destroying the whole world. I mean, the economic world does run on credit; I sort of more and more understand. But nevertheless I think it's a terrible example to see people or corporations literally saved by taxpayers and go back to record profits the next year. Russ: It's not a matter of vengeance. It's a matter of the costs of those incentives. I don't want to punish them. There's no evil when you make bad decisions. But if you reward bad decisions, you get more of them. That's the part I find troubling. Guest: You mean if you allow people to get away with them. Russ: Yeah. It's not that I want to see General Motors or AIG punished. But I don't want to see them rewarded. Guest: Yeah, I'm exactly the same. I can't quite believe it. Russ: Did you have any conversations with former colleagues about that? Guest: At Goldman, you mean? Russ: Yes. Guest: No; I've really been gone for 10 years, so not really. Most of the people I know are gone. But I ran into people around the time when Goldman and Morgan got access to the Fed window, commercial banks. I think people were really scared at the time. Russ: They were. Guest: I mean, it was a race to the bottom to see whose stock was going to be gotten rid of next, and they were heading for the bottom when they were saved. It's sort of like everybody thinks there's only price risk; but all of these places have terrible funding, overnight funding risk. I think Goldman didn't own trash the way Merrill or Lehman did; but nobody is going to lend them money any more. Russ: But they created that world. They chose to live in that world. Overnight funding risk is a crazy way to run a business. It's the same thing we were talking about before. If everything is going fine, it makes total sense to lend enormous sums of money overnight on assets that seem to be worth something. If they turn out not to be worth something and all of a sudden people start to be uneasy, I think people are aware that that could happen. But they didn't bother planning for an alternative strategy or an insurance plan, because they figured if the whole thing melted down, they'd all be taken care of. And they were. Guest: I agree with you. This is where I differ with my friends, who say, for example, Goldman and Merrill were smarter and they were only hurt because of overnight funding. But that's one of the risks you take. I agree with you. And it's a bad example to see people saved from that by-- Russ: That's why they were able to make so much money. That's the real problem. If you are going to say: This world is a good world because you take risk and you earn profit, but there's a chance you'll get losses. If you don't have the losses--boy! Where are we going next? Guest: I agree with you. I'm getting out of my depth when I'm talking about the Fed, but I don't really understand what they are doing, either. Russ: You are not alone. Don't worry about that. Let's talk a little bit about that ethical issue. The way I think about it, which is really not the way economists think about these things at all, is there is a certain groupthink that pardons ethical lapses in certain settings. Nobody was murdering people on Wall Street to make money. What they were doing is selling something that they maybe weren't so confident that was what they said it was. You said something in the book about the importance of transparency. It's hard to be introspective about morality when it's the norm of your culture and people around you do what you are doing and accept it as it is. It's a shame. But I think that's the way we are as human beings. Guest: Yeah. There are a lot of things that aren't acceptable now. Politically correct point of view that we were completely okay 20-30 years ago. Russ: Do you think there is any evidence that things are changing on Wall Street? Guest: I'm not close to--I sort of deal with people on the buy-side but I'm not close to people in investment banks any more. But I would say not. They got away with sort of metaphorical murder, and now it's back to business. I don't see any real change. Russ: Obviously if you don't let them suffer, they are not going to learn that lesson. Guest: I think you are right. It's not so much punishment as just setting an example for the future. You can have all the regulations in the world, but I think seeing people who did something that deserves to fail actually fail would have a much bigger effect. Russ: There are people that argue--I'm not one of them--that Wall Street did pay a price. It's smaller; a lot of people lost their jobs. But it looks to me at the higher level where decisions were made, there was pain; but it was pain that was compensated for by other gains. So, people point to Richard Fuld of Lehman or Jimmy Cayne of Bear Stearns losing about a billion dollars on stock. Guest: But they were still left with $60 or $80 million. Russ: $500 million. Guest: It's hard to feel sorry for them. People do point them out as examples, but I don't think it's a very good example. Russ: Yeah, I don't get that.
51:12Russ: So, what do you teach? Guest: I teach two courses. One is called Introduction to the Volatility [?], which is about models that go beyond Black-Scholes, that try to explain the nature of option pricing in equity derivatives. And Black-Scholes doesn't work quite right; it doesn't describe the way volatility behaves. So, I teach one course on that. And another course, more like a discussion course, where each week--it's called Discussion Papers in Quantitative Financial Engineering--we pick one paper or one or two papers and the first couple of weeks I make a presentation to let the students get up to speed, and then they start making presentations in small groups for the rest of the class, and we discuss it. Actually, this year we did a lot of financial stuff, but this year, since the crisis, I'm trying to incorporate some economics in there, to educate myself. So, we did one on the financial crisis last week and read, used as a basis a paper by Metrick--two guys from Yale, Metrick and somebody, weekend's worth of reading on what happened in the financial crisis and various people's theories about it. And we did convertible bonds, when we could have used some more traditional stuff. We are going to do something on the theory of money, so I'm trying to get some economics in there for my sake. Russ: So, at the end of your book, you talk about--the models have usefulness. They are not use-less. But they have to be used with caution. How do you integrate that into the classroom? Or, more generally, what do you think young people going into finance should have in mind as they go forward? Given that when we are young, just like physics was very seductive for you, I think we love when we are 19 and 23, we love certainty and we love equations and we love results and beauty; and if they are not quite right we tend to say: Well, they are close enough and we keep going. Guest: In the course on teaching options, which was the thing I did on Wall Street; in the papers I wrote, I try to bring a lot of intuition into it. I use the mathematics, but I try hard to explain that all of these are just ways of trying to get a handle on how to price something. In fact, what I like to say, is the only way to price something in finance, the only law of finance, is if you want to know what something is worth, you have to find a way to make a recipe to create it out of other things using values you already know. Sort of if you want to know what fruit salad is worth, you tell me what fruit is worth and how to make fruit salad, and I'll tell you what fruit salad is worth. I try to use that principle to price options and exotic options and CAPM and all kinds of stuff, all based on this common-sense principle that if you want to know what something complicated is worth, figure out how to make it. And then I try to point out all the flaws in that, as well. I'm careful to try to stress--even before the crisis--that all of these are approximations, and when the world goes crazy, and equality disappears, none of this is going to be very accurate. Russ: Do you think they listen? Guest: Yeah, up to a point. The general tendency--and I think I'm anomalous in a sense--in all finance is to get more mathematical and less and less intuitive or real world. I think that does seem to overwhelm them, although I try to combat it. They are all in jobs, which is reasonable. Russ: It reminds me--I used to teach in a business school, and I taught economics; and I taught it as a very intuitive, non-mathematical class. And one of my students complained; she said: I wish it were harder. And I said: I think it's incredibly hard. And she had struggled in the class. She meant: more equations. More calculus. Calculus is harder. Intuition, by definition, is easier. But to me it's the other way around. Calculus is easy. You get answers and you know what the right answer is. But intuition is harder; artful and harder. Guest: Yeah, it is. I never want to come across as saying--I mean, I think finance is harder than physics, not easier than physics. I got an email today. One student is going to work for the Securities and Exchange Commission [SEC] in their risk division, just got a job there. Russ: Just like Wall Street has its own incentives, so does the SEC. Hard to stay pure in both places, I suspect. A question; I'll pretend it's a question of personal advice. My oldest son is 17; he's very good at math and he's very good at hypothesis testing and using data to think about analytical questions. Five years ago, if you'd said, 6 years ago, where should I encourage him to go study and what should he study, one of the things I would have thought about is finance. It's a natural thing for him. It's financially rewarding; it's got an intellectual aspect to it. I would have said: That's a good life. But I'm not so sure any more. It seems to me there is an aspect to it that is less noble, less pure, and a little bit tawdry, I'm afraid to say. Dirty, corrupt. That the links between Wall Street and Washington are not so healthy. Maybe the practitioners themselves are not so aware of it, but as an outsider it gives me the creeps. I'm not sure I want my son in that world. He may go there anyway. I can't control him. But in the past I might have encouraged him to think about it. Now, I'm not so sure. Do you think that's an unfair-- Guest: No, I think it's fair. I went into it when I was 40 and I wanted to do something interesting; and it was. But I kind of agree with you a little. I think it's changed. I've kind of become much more aware of all the sort of political ramifications, or the sociological ramifications. Which go beyond just the math. I think it's fair. I also think it's a bad idea for students to do this stuff as undergraduates, although we have that at Columbia. I think it's better to spend your undergraduate career doing things that are not evanescent. And I think financial models are mostly evanescent. Not very real. I think it's better to do arts or science or physics or music or something as an undergraduate and only do the stuff in graduate school if you are going to do it, and then maybe first get a job for a couple of years in the industry to see how you like to get a sense of what's important, rather than imagining that what they teach in graduate school is what people do. Russ: Well, that's good advice. I'll pass that along.

More EconTalk Episodes

Search Econlib