DON'T FOLLOW ORDERS
Sunday, October 4, 2009
Every week CUIP's president Jacqueline Salit and strategist/philosopher Fred Newman watch the political talk shows and discuss them. Here are excerpts from their dialogues compiled on Sunday, October 4, 2009 after watching selections from "Hardball with Chris Matthews," “This Week with George Stephanopoulos,” and “The Charlie Rose Show.”
Salit: We watched a Charlie Rose interview with Paul Volcker, chairman of Obama’s Economic Recovery Advisory Board, about the state of the economy. Volcker says we need new revenue streams, but it’s too early to tackle the tax thing. Volcker says “we” should have been more alert to various weaknesses and vulnerabilities: credit default swaps, sub-prime mortgages, and so forth. Charlie Rose says ‘Well, people didn’t see these things.’ Volcker says ‘Well, actually, some people did see them, but no one wanted to act.’ Then, we read a piece in the New York Review of Books about economic development by William Easterly, professor of economics at NYU. The author’s angle was that when you get down to it, it’s not known how to promote growth. Growth is real. It occurs in different places and in different ways and economists try endlessly to analyze it and identify the factors that contribute to growth, but they don’t seem to be able to do that in any reliable or replicable way. So, Easterly suggests a number of methodological problems that account for not being able to pinpoint how growth occurs. I wanted to talk to you about some of the methodological problems that he identifies.
First, he explains that a survey of top economists showed they identified a total of 145 separate factors found to be associated with economic growth. But, when they were “field-tested,” these “patterns” – so-called – failed to hold up. So he raises the question, “how can there still be so many writers, [meaning economists] who claim to know how to promote growth?” He answers this question by paraphrasing a theme from the book “The Drunkard’s Walk,” claiming “Humans are suckers for finding patterns where none really exist, like seeing the shapes of lions and giraffes in the clouds.” So, tell me your thoughts about the quest for seeing patterns and identifying patterns in human life.
Newman: I don’t know if I’d say that people are suckers. People make the mistake of thinking that their capacity to, or their interest in, ordering things has anything to do with the things that they order, and so – even beyond the logical errors that are pointed to in these articles – is that logical error. The assumption is that because you can order things, it follows that there is an order that you’ve discovered in the ordering of things. But there isn’t. Ken Gergen and I wrote a paper for the American Psychological Association convention in 1995 about this very topic called “Diagnosis: The Human Cost of the Rage to Order.”*
Salit: Yes, of course.
Newman: And so things proceed on their merry way and when you come back next time, they have gone about doing or being what they are, which may or may not match the pattern you came up with. So, there is a more fundamental, not so much logical, though it’s related to logic, but ontological problem that human beings simply don’t want to deal with.
Salit: And that is?
Newman: Well, that the world isn’t the way in which they’ve ordered it.
Salit: OK, so to what do you attribute what you and Gergen called the Rage to Order, what Easterly and Leonard Mlodinow, the author of “The Drunkard’s Walk,” call finding patterns where none really exist? In the case of developmental economics, the economists are presumably looking for patterns because they want to be able to replicate the circumstances where growth occurs.
Salit: That’s what they’re trying to discover.
Newman: And there’s only one consistently valid rule or if you will, law, to guide them on that matter.
Salit: And that is?
Newman: That growth occurs where growth occurs.
Newman: That doesn’t have very much predictive value.
Salit: No, it doesn’t.
Newman: But, that’s life.
Salit: OK, growth occurs where growth occurs.
Salit: And that has little or no predictive value.
Newman: As far as I can see.
Salit: Maybe I’m making a huge leap here, but if you’re a developmental economist, or if you’re a government, or if you’re a world citizen and you’re concerned about the issue of poverty, for example, and you want to discover how to promote growth in the so-called Third World, how do you go about doing that? I guess there are two different issues here and maybe this goes to what you’re pointing to. There’s the question of what you do about that. And then there’s the question of whether you can create models to do something about that, and are those the same thing?
Newman: Well, we don’t know. The final section of the article you read to me by Easterly is about how an obscure Korean car shop owner did more to impact on the Korean economy than all the studies of Asian miracles put together.
Salit: Yes. He’s the guy who built Hyundai after many of his businesses failed.
Newman: So the message is, if you want to grow an economy, you’re better off being the guy who controls Hyundai, than you are being the academic who’s identifying the 75 factors that produce growth. We all know that Ju-Yung Chung is richer than all those academics put together.
Salit: No doubt.
Newman: So, let’s not make another logical error here. The point is not that there aren’t factors which produce growth – at the individual level, the emotional level, the international level, whatever. The point is that the models that we develop are no guarantee or even close to a guarantee. Or may not even be relevant to the growth. That’s the point.
Salit: So, why don’t we stop fooling ourselves?
Newman: Who knows? Because we have this rage to order and this rage to, therefore, glorify science and to want to think that things that aren’t science, like economics, really are. And so we persist in producing endless papers and institutes and studies – which may be useful – but they also may not be useful. But that’s a hard fact to accept, because human beings want to know that something is really useful. They want a sure bet at Las Vegas.
Newman: But, there are no sure bets at Las Vegas. The house doesn’t have to cheat to make money. They make money because of the human need to suppose that we know something that we can’t possibly know. Red Seven has got to win. How do you know? I called up 30 people and they all said play Red Seven. But it doesn’t come up. Baseball fans sit at home rooting madly for someone who’s supposed to get a hit, the assumption being that their rooting has something to do with whether he will or won’t get a hit.
Salit: So, you’re saying something more than, or other than, that models can be helpful but not determinative or predictive.
Newman: Well, with due respect, I’m not saying anything about determinative. You’re using traditional language. I’m trying to speak a simpler language. But, I guess you’re saying I’m not succeeding so you have to translate my language into a fancier language, which is halfway down the slippery slope into thinking you can make economics into a science, even if it isn’t. But you can’t.
Salit: I was using “fancy language” because I wanted to ask you what’s the activity that happens other than the model?
Newman: Well, it’s the activity that happens. That’s the activity that happens other than the model.
Salit: In Easterly’s article, he talks about growth being unstable. Growth is unstable, he says, and there’s no way to predict its patterns.
Newman: In whose eye is growth unstable? Does “unstable” mean the models, roughly speaking, that we put together are not coherent with what happens?
Salit: Well, it means that things happen that you have no way of predicting.
Newman: Well, how do you know that? I can tell you a way to predict them.
Newman: Predict them.
Salit: But, if there’s no way of knowing that, then how can you predict them?
Newman: There are lots of things that people predict where they don’t have any way of knowing them. It’s what keeps people playing the horses at the track day after day after day.
Newman: And they spend hours going over the racing sheets and so forth. And then they look frustrated when the two-year-old kid took a crayon and picked nine winners. They’re like, Who is this kid?
Salit: He’s a prodigy.
Newman: And they insist, in an amusing way, that there’s an “order” that can be discerned. People assume that they can create something orderly and coherent based upon appropriate information and that means they will have a higher percentage of right answers than other people, and that will serve them better than people who have wrong answers.
Newman: And certainly, if you get that result, it does serve you better. But the fallacy that we make up as human beings, and this is the Rage part of the Rage to Order, is that this very process, even though it can work on some occasions, is sufficient to justify that the model is going to work all the time. And there’s no reason why it should. And it doesn’t. And interestingly, science, by which, roughly speaking, I mean physics, understands that. In some respects, it’s the basis, methodologically speaking, of much of contemporary physics.
Salit: The understanding of the unreliability of models.
Newman: The understanding of the undecidability of this whole process, including physics itself. A lot of pseudo-sciences, for all of their fancy language and mathematical formulas, are really what I would call Baconian in their essence, and I’m talking here about Francis Bacon, in that they are concocted more by human subjectivity than by anything which really turns out to be useful to anybody. That might be an unfair claim to make against Bacon, but nonetheless.
Salit: Unfortunately, I’m not in a position to defend him. But I do want to ask you about Easterly’s claim that “when you study the effect of a particular policy on growth, you also have to control for some of the other factors affecting it, for which there are practically endless possibilities.” I take it you would agree with that.
Newman: Well, I agree and I don’t agree, since you can’t. It’s well and good to say You also have to control for lots of other factors. But you can’t.
Salit: You can’t?
Newman: It’s too complex. You can fool yourself into thinking that you’ve taken into account the more important ones. And that might get you a grant from the National Science Foundation.
Newman: But you might have left out either one very important one or 75,463 unimportant ones that, taken together, make a difference.
Salit: So you can’t control for it.
Newman: That’s what I’m saying. He’s saying that you have to. I’m saying far from you have to, you can’t.
Salit: One of the reasons that he’s saying you have to, as I understand it, is that he’s trying to go up against what he calls the Confirmation Bias, which he sees as another methodological problem in economics. Confirmation Bias, roughly, is when you’re looking to prove something and you select your evidence, your models, accordingly. And, by the way, this may be the case in his article too, as he seems to be trying to prove that when push comes to shove, free market capitalism is a better formula for growth than other alternatives.
Newman: Might be.
Salit: Okay. Confirmation Bias. It sounds hifalutin. Fancy language, which you accused me of before. But basically it means you select the data that proves your point. I learned to do this in fifth grade.
Salit: It’s how you write an essay.
Salit: But, it’s one thing if you do that in fifth grade when you submit your paper to Mrs. Levy on “Causes of the War of 1812.” But it’s another thing if you’re running the world’s largest economy that affects the health and well-being of billions of people.
Newman: There’s no difference at all.
Salit: Well, that’s a little bit frightening, if the methodological primitiveness is the same. When you say there’s no difference between the essay you write in fifth grade on the causes of the War of 1812 and the fundamental methodological approach that economists use in coming up with the solution to the economic crisis, that’s not exactly good news.
Newman: Right, there’s no difference relative to the validity of the methodology. There is a difference in terms of the power the heads of government have to effect and transform policy, including making it seem as if what they’re saying is sensible or valid or so on. The fifth grader in Mrs. Levy’s class presumably doesn’t have that much control of things and so she or he is more subject to the authoritarian classroom rule of Mrs. Levy.
Salit: Actually, I liked Mrs. Levy.
Newman: Fine. But chances are your essay on the War of 1812 wasn’t going to serve as a template for U.S. policy in the War of 2012.
Salit: Certainly not! Unless we’re going to war with Britain.
Salit: Next up on Easterly’s checklist is The Law of Large Numbers and the Law of Small Numbers. Easterly says “This is a sarcastic reference to the Law of Large Numbers, in which you can have a high degree of confidence in the average value of a sample if the sample includes a very large number of observations. The Law of Small Numbers is when you stop way short of having ‘enough’ observations and show high confidence anyway.” So he gives an example from economic history. If you looked at the 1930s, he explains, the U.S. economy crumbled, while the Soviet economy grew. That led a whole school of economists to conclude Well, socialism or state-planned centralized economies grow at a faster rate than capitalist economies, which are vulnerable to these kinds of collapses, etc. So, that’s an example of a Law of Small Numbers, because it later turned out that the Soviet Union unraveled, in no small part because of the stagnation of its economic model, while capitalism rebounded. So, I guess the Law of Small Numbers is another way of saying there wasn’t enough data to make a judgment call.
Newman: It wasn’t a question of “enough.” I don’t think it’s even worthy of being considered a law, even a sarcastic one. These were just bad observations. Ordinary human beings recognize that you don’t serve the roast until it’s done. Otherwise, people won’t like it.
Newman: It wasn’t a question of numbers. If you want to get relative certainty on predicting the time of someone’s death…
Newman: Make the prediction roughly ten seconds before they die. And you’ll get it right every time. But of what value is it?
Salit: Virtually none.
Newman: There you go. But, you’ll be right. So, I don’t think it’s ultimately a question of numbers. I think it’s a question of which numbers. Underlying that kind of thinking is a qualitative analysis of the importance or relevancy of these numbers to the outcome you’re looking to predict. So, in some cases, one kind of thing can happen and it’s enough to make extraordinary predictions and maybe to make it with extraordinary success. In many cases, you can have a million numbers and they’re not particularly relevant. So, they make no difference. It’s not a quantitative issue so much as it is an issue of selection. And that’s a more qualitative indicator. There are all kinds of examples of this in the centuries of responses to David Hume’s “Theory of Constant Conjunction,” namely, that it’s an analysis of causality which shows that there’s a profound fallacy in that certain things can be constantly conjoined, virtually forever, and that doesn’t show that there’s a causal connection.
Newman: But whether they’re of value, whether it’s a million, ten or even one, it’s not about the numbers. That’s what the Bible tells us. After all, in western culture the Bible remains the most accepted book of understanding of how things got going. But there’s only one thing cited there.
Salit: One thing that got the Universe going?
Newman: It was God. There’s only one. There’s not even any effort to show that this happened 10 of 18 times. It happened exactly one time.
Salit: I see your point.
Newman: And, it’s still the most accepted notion in western culture.
Salit: That certainly is a Law of Small Numbers.
Newman: Right, you can’t get much smaller than that.
Salit: You certainly can’t.
Newman: No. And if you put forth alternatives to God having run the whole show, the response is That’s blasphemy. God is not only a small number. He’s the only allowable number.
Newman: Beyond that, there are also all kinds of known problems with probability, rationality, and utility. In fairness, I find myself having to be much more of a pragmatist than some of these people. Utility is filled with a similar set of problematic or contradictory features. So, does that have to lead you to existential despair?
Salit: It doesn’t have to lead you anywhere.
Newman: It could lead you to a recognition of the limited capacity for human beings to understand. Does that lead you to saying that we shouldn’t keep trying to understand? No. It doesn’t lead me there. Actually, I think there are some good arguments to show that it makes you more strongly committed to trying to understand. Since, in the understanding of this, we become clearer on the depth and degree of our ignorance, which depending on certain attitudinal issues makes you more or less eager to pursue knowing something. But I don’t think it necessarily means you give up in hopeless despair. In some ways, it doesn’t lead to anything that fits into a “rational picture” of this whole thing called living. But, I don’t think that it leads you into despair.
Salit: So in the Rage to Order, and this is going to be a broad oversimplification…
Newman: Ahhhh. The Rage to Simplify.
Salit: The Rage to Simplify, there you go. You and Gergen are talking about the drive to systematize an “understanding” of emotional problems.
Newman: Emotional disorders. So-called.
Salit: And, there is a connection between that, between the Rage to Order, and the shortcomings, or inability of psychology, as a field, to effectively treat mental illness and emotional pain.
Salit: Would you say that the Rage to Order in economics is connected to the inability to end poverty?
Newman: Yeah, I’d be glad to say it, or affirm your saying it, whichever you like. I’m saying there’s an epistemological issue, a methodological and epistemological issue at the root of all this and it’s important to comprehend it or them, however many there are. Your question is a simple question with a complex answer and I don’t know if we have time to get into the details of that. But, yes I would.
Salit: Thanks, Fred.
# # #