Feed on

In a working paper two students and I write:

Stevans argues that, “most academic economists are concerned with studying such obscure topics as backward induction among chess players and the existence of monotone pure-strategy equilibrium in Bayesian games.” Does “most” mean more than half? Eighty-percent? Is there a right amount? He is onto something, of course. As Paul Krugman has written in many places, economic confusion comes about because modelers have little conceptual notion of what it is they are trying to model. But this does not mean modeling qua modeling is wrong, or even mindless drudgery. There is modeling for modeling’s sake – the kind that perhaps should be scrutinized. And then: there is good modeling. Models are not just neat ways of illustrating sometimes complex ideas. They are essential.

Applying formal theoretical and mathematical techniques to economic history (referred to as “cliometrics”) has revolutionized the way we understand the history of slavery, transportation, agriculture and other areas without making those topics inaccessible. In defending the formalization of economic history, Nobel Laureate Robert Fogel famously stated that, “The belief that the older economic history is solidly grounded in fact is an illusion … it is permeated with untested covert models and subliminal mathematical assumptions.” Graduate education focuses on model building because models prevent us from saying whatever it is we want to say about the world. Writing down a model of human behavior and interaction is the only way we can understand whether or not what we are saying is logical – with propositions within the models and outside of them. Perhaps the crisis in modern economics is not that we build too many models, or that the models of academic economists are unnecessarily complex, but perhaps that we have yet to discover a model to help non-economists model the world themselves.

A student came to me the other day and mentioned, sort of out of the blue, that private agencies should not deliver water to cities, and that rural and suburban water ought not be priced because it would cause the poor to suffer. I did not push the issue. But this person, if asked what model of the world he was operating under, would very likely not have been able to offer one. And that is a problem, as we’ll elaborate on in the future. But he certainly was operating under a model – and I wonder how he’d defend it. The simple model? That the elasticity of demand for water is zero. If that were indeed true, then when private agencies charged market prices for water that reflected its true social cost, we’d only see transfers from customers to firms – and no economizing on water use at all. But we know that perfectly inelastic demand curves cannot exist, and we can also ask our student what conditions would have to prevail in order to generate such a curve in the water market. And he’d have seen that those conditions are extraordinarily unlikely.

So, while I am not a huge fan of fancy, technocractic model building, that does not mean I am not a strong supporter of logical and systematic explanations of behavior. To say that you can dispassionately look at facts, examine history or understand the social, political and cultural influences on people and that means you do not need an economic model to describe behavior is implausible. The point is, when making such judgments you are still using an economic model, it’s just that you are not articulating what it is. I have taken this to heart lately. It doesn’t make sense trying to reason with people, so what I do is simply ask questions about how they came to the positions they are taking.

5 Responses to “There’s Always, Always, Always a Model Beneath”

  1. jb says:

    This is a great topic. You quoted Stevans “Perhaps the crisis in modern economics is not that we build too many models, or that the models of academic economists are unnecessarily complex, but perhaps that we have yet to discover a model to help non-economists model the world themselves.”

    I beg to differ, the Capital Asset Pricing Model (CAPM) has been enormously beneficial to both insitutional and individual investors. While there is no shortage of CAPM detractors, many of them fail to realize that it is the ongoing debate over CAPM that in effect establishes its usefulness.

    This elegant model essentially simply establishes the logic behind the notion that capital assets must be priced according to the risk they present. Without the model we would have nothing but a nebulous idea; the benefits of diverfication could not be quantified. Thanks to the model we have a vigorous and healthy debate about market efficiency and whether anyone can really add value with regard to picking stocks or market timing. As I see it, CAPM has put the “burden of proof” on those who seem to think that markets are somehow defective and fail to price risk accurately.

    Look around. There are now hundreds if not thousands of “passive” advisors and money managers who employ the model , or some variation of it, explicilty, by buying index funds for their clients. Our “value added” is essentially the disclipline to maintain target allocations that the client establishes. But even many “active” managers who claim to provide risk-adjusted returns in excess of the market admit that they have to show they have “added alpha”. They have adopted the vocabulary of the model.

    Even the behavioralists, who are willing to question the assumption of the ratioanl investor, must acknowledget the logic of the model in order to test this assertion. Without the model there would be nothing to frame the debate, and no way for financial theory to move forward. Perhaps one day CAPM will be replaced by something better. That’s the way (social) science progresses, as I understand it.

    But my opinion regarding CAPM or EMH in this matter is not what is important. Rather my point is that without this formal model, financial theory and investing would be in the dark ages.

    I think there are 2 types of CAPM critics; those who are highly skeptical but constantly test it, and those who lazily reject the idea, and have pronounced it dead, but without offering an altnerative. The former are social scientists. The latter are charlatans who are worse than useless, especially when they provide phony statistics touting returns without measuring the risk they are assuming, and charging exhorbitant fees for talent that very likely does not exist.

    Those who condemn modeling per se do not understand that a model is not reality. If it were, we would call it reality.

  2. Rod says:

    Back when I was an investment advisor, I spent a whole day studying an oil and gas limited partnership and came to the conclusion that the general partners got the money and the limited partners (of which my client was one) got the losses. At any rate, one of my colleagues in the company told me to pitch the papers on the limited partnership in the wastebasket, and he handed me a quarterly report for American Home Products, then a company with nil long term debt and ten to twenty percent increased earnings for years on end. It was a simple story: Occam’s Razor applied to investments.

    Said another way, if you have to apply arcane formulas and models to tell whether a stock is a worthwhile investment, it’s probably not, or at least it’s a waste of time burrowing through all that manure in search of a pony.

  3. Dan says:

    I’ve had too narrow of an understanding of the term ‘model’, with the impression that they imply a certain level of rigor with graphs and fancy math. Can you define ‘model,’ in your terms? And we’ll appreciate any other interesting models you have in mind, perhaps ones that connect more generally than pricing a good.

  4. Harry says:

    Great topic. You have to have a model.

    Many of us non-degreed commentators are tempted to criticize practitioners of the dismal science, often unfairly, because it lacks the same precision as, say, Newtonian physics. But medicine is also imprecise in a similar way, dealing with uncounted known and unknown variables, and we rely on trusted doctors to make a judgement about that thing they detect, and the remedy they propose.

    I think the difficult part is knowing the limitations of the models, and I am not going to launch a critique of anybody, even those who may in my opinion might stray from their areas of expertise, even if they are messing with my rice bowl, like Paul Krugman, who believes the multiplier is 1.86.

  5. Rod says:

    Sure, you have to have a model, especially in science: it’s called a hypothesis, something that gets either confirmed or rejected by empirical fact — as in the “hockey stick” model in climatology. The problem with the hockey stick model is that it’s both simpleminded and without any statistical basis. Al Gore reads up a little about the climate on Venus and the disappearance of the Snows of Kilimanjaro and suddenly he is able to predict the Destruction of Life as We Know it unless carbon gets taxed a whole bunch.

    Similarly,, some economists claim they can predict what interest rates will be in the next quarter and whether the stock market will break 14,000 before the 2012 presidential election. Like the climate, the direction of interest rates depends on a lot of things, including what’s inside Ben’s head, and it probably does not make a whole lot of sense to put a bet on it either way.

    Last week I got to watch Trading Places on the satellite, and there are several scenes in there I find funny.

    The first one is where Randolph Duke is explaining to Billy Ray Valentine about how futures work. “Mortimer and I deal in commodities. We have frozen orange juice, which you may have had for breakfast; wheat, which may be used to make a loaf of bread; bacon, as in a bacon, lettuce and tomato sandwich [Eddie Murphy looks into the camera, nonplussed] and gold. …. The good thing is that whether our clients make money or lose money, Duke and Duke always makes money. “Oh, I see,” says Valentine, “you guys are a couple of bookies.”

    The second scene is one in which Valentine tells Mortimer Duke not to go long on gold futures: he says the commodity brokers are all out of the office trying to find a G.I Joe with the Kung Fu Grip, and that it would be wise to wait until the price takes a dip. Eddie is right: the price does go lower.

    And the third scene is the one where the Dukes, Valentine and the Dukes’ friends are all having dinner at the Bellevue Stratford, with Louis Winthorpe standing outside looking through the window, and Valentine starts to say what he’d do in an investment, and everyone swings their heads around like they are in an EF Hutton commercial. In the same scene, this judge character, cracks a good joke about an “S” model Cadillac, with the punch line being, “How fast can that S car go?

    Hey! Don’t nobody put their Kools out on my floor!

Leave a Reply