Julio Huato wrote: > I don't know how else to put this but frankly. You mix up so many > things it's hard to even disentangle them.
Perhaps it's because it's so difficult to get beyond such dichotomies as "reject math" vs. "embrace math" to understand such notions as "math should be tamed" or that "math should be used as a means to an end rather than as an end in itself." It's important to remember that math is a metaphor or simile for real-world phenomena and does not correspond to those phenomena themselves. Down with reification! > Let me take up a sliver of your post: > You say rational expectations [RATEX] is "nothing but" an equilibrium > condition. Next you say that it "has a very very weak story for why > and how equilibrium is achieved." Then you say it's "not like supply > and demand." So here it is your supply-demand equilibrium condition: > Q_S = Q_D. Now, tell me: Where's *the story* about how equilibrium is > achieved? The mechanism leading to or away from equilibrium is > separate from the equilibrium condition. You can perfectly have the > form of the functions Q_S=S(.) and Q_D=D(.) be such that there's no > tendency to equilibrium at all. How can your equilibrium condition > Q_S = Q_D change that? You define "supply and demand" in totally mathematical terms, and extremely narrowly at that. Let's stick with the math and ignore the tendency to let math restrict thinking. In grad school, I learned two stories about equilibration. (1) Walras: the change in price = f(excess demand or x) = f(Q_D – Q_S), where df/dx > 0 and f(0) = 0. This is the standard textbook story: a shortage causes prices to rise, while a glut causes prices to fall. Market clearing causes no (endogenous) change in prices. (2) Marshall: the change in the quantity in the market = g(excess price or y) = g(P_D – P_S), where dg/dy > 0 and g(0) = 0. This says that if the demand price (the highest price that purchasers will pay for the currently available quantity) exceeds the supply price (the lowest price that sellers will accept), then the suppliers will bring more to the market during a specific market day. And vice-versa. One or the other of these is just as much part of the supply & demand story as the equilibrium conditions. If people forget the equilibration processes, they're falling for the economics profession's obsession with equilibrium (which spawned such crap as "New Classical macroeconomics"). In reality, we could see both equilibration processes happening simultaneously. But I understand that this makes the math intractable. So those who let math limit their thinking drop this possibility. > You say that the advocates of the EMH "take this purely mental test > (is the model internally consistent?)." Just so that it's clear: EMH > is not the only possible use of rational expectations, but anyway, > Don't logical or mathematical models need to be internally consistent? _Of course_ mathematical models have to be internally consistent; that's what makes them mathematical or logical. It's also true that EMH is not the only use of RATEX. (I have never said anything different on either point.) The key is that neither mathematical nor logical models -- including the EMH -- can ever correspond directly to empirical reality. They are metaphors or similes. Worse, EMH like most (if not all) RATEX models is profoundly dependent on its assumption of perfect markets (lacking non-convexities, etc.) It's extremely easy to show that in non-perfect markets, there can be more than one equilibrium in a model. (One version of the Keynesian model has an infinite number of equilibria.) With multiple equilibria, RATEX leads to nonsensical results. How can anyone expect specific results (plus random error, ignoring true uncertainty) that correspond to the model's equilibrium if there's more than one equilibrium? Further, RATEX produces a kind of internal consistency that usually does not make sense. Under the continually-changing and irreversible process that characterizes real life in historical time, it should be the exception (not the rule) that individuals expect what actually happens (even on average). Life is a series of surprises. By the time we've adjusted to the surprise, the economic conditions that economists try to understand using formal models have changed. This suggests that adaptive expectations or Bayesian learning are more likely to apply than RATEX. The only way that people could have little economic models in their heads that roughly correspond to the phenomenon being modeled is via learning. That is, the attainment of RATEX equilibrium requires adaptive expectations (ADEX). RATEX may apply in the setting that Muth described in his original presentation, i.e., a stable and unchanging perfectly competitive market. But it doesn't apply anywhere else I can think of, except mixed with ADEX. Simple analysis shows that mixing RATEX with ADEX leads to the latter dominating. (Here and above, I rely on my friend George Evans' Ph.D. dissertation at UC-Berkeley, circa 1978.) > *Anything* can be deduced from a self-contradictory reasoning. How > does that help you? It seems that almost anything can also be derived from from internally-consistent reasoning. In response to a lay criticism of the idea of the Oedipus Complex, a psychoanalyst allegedly once said that "obviously your rejection is based on the fact that you've got that Complex and cannot handle it emotionally." This presents a nice internally-consistent argument that cannot be falsified in any way. That does not say that internally-consistent reasoning is wrong (which would be silly). Rather, it may be necessary but it's not sufficient. > Then you you shift the argument to empirical validity. The issue of > the empirical validity of a hypothesis like rational expectations has > to be predicated on a particular empirical application of the > hypothesis. I wouldn't say that RATEX is a hypothesis as much as an assumption. Then, the empirical validity of any simplifying assumption is incomplete, because the point of assumptions is to simplify an extremely complex empirical reality in order to gain some understanding of it. By simplifying the world, we not only gain understanding but we lose something. Contrary to Milton Friedman's "Positive Economics," there are _costs_ to simplification. We have to judge each simplifying assumption using cost/benefit analysis rather than jumping feet-first into making assumptions and constructing models. The key issue is whether RATEX represents the main empirical tendency about how individual expectations of the future are determined. In that case, the simplification's benefits may actually exceed its costs. But do people really have little economic models in their heads (which by coincidence are the same as those of the economic theorist) which allow them to figure out what's going to happen in the future (on average)? As mentioned above, the use of the RATEX assumption is highly dependent on assumptions of market perfection. Small deviations from perfection can make the assumption nonsensical. > Particular empirical applications typically refer to > specific sets of stylized facts or data. What particular application > are you talking about? What particular facts or data sets? See, for example, Mishkin's discussion of the EMH in his Money & Banking textbook. BTW, perhaps I made a typo, but the empirical validity bit applied to EMH, not RATEX. RATEX should be seen as a simplifying assumption (one small piece of a larger model) not a model as a whole. > So, if rational expectations is only "an equilibrium condition," What > kind of empirical test would be required to reject equilibrium > condition Q_S = Q_D? An equilibrium condition is, like RATEX, an assumption. The key issue is the usefulness of that assumption in understanding the world. Does it contribute to larger stories that help us understand what's going on? Most economists would see market equilibrium, unlike RATEX, as a useful assumption. As mentioned in a previous missive, even Marx used concepts of market equilibrium even though he didn't think that actual (real-world) attainment of equilibrium would ever occur. (For Marx, when "supply equals demand," the price equals the price of production in a competitive market. I read the phrase in quotation makres as saying "when the quantity supplied equals the quantity demanded at the price of production" in modern lingo.) > If I showed you that at my local apple market, > at the given price and point in time, Q_S is not equal to Q_D, would > that suffice to you? I don't think that you can test a simplifying assumption in that way. By definition, a simplifying assumption is simplifying, i.e., strictly-speaking empirically untrue. The key point is whether or not there is a center of gravity for prices or a tendency of quantities supplied and demanded to get nearer to each other over time. Most conceptions of market equilibrium help us understand this center of gravity. > And then, would you reject this or that > proposition in Capital because to derive it Marx *explicitly* assumed > Q_S = Q_D? No. It would be silly. BTW, as far as I can tell he only made that assumption in order to understand the conditions necessary to continued reproduction of the capitalist system in historical time. He never made that assumption as a description of how the world works at any single point in time. > You say that empirically, rational expectations has "been soundly > rejected." Then you say that it's "empirically irrelevant." If you > care to refute it, then you're admitting that it's relevant. No? If > it were irrelevant, there'd be no need to refute it. Some other people think RATEX is relevant. Since those other people have power and influence in the world, it seems worthwhile to refute their nonsense. (Lucas has been "adopted" as a leader of the profession by the SRPE committee.) > It's "the real world that matters" -- you say -- as if postulating a > hypothesis to "close" a model implied that it doesn't. Models are not the same as the real world. They are tools, not true representations of the world. > How do people > transform the real world anyway? Can we transform the world > intentionally without transforming ourselves and our conceptions of > the world? How do we appropriate the real world conceptually? How do > we grapple with it in our minds? Directly in its concretion? Have we > ever been able to do anything human without abstractions, logic, > concepts? I was NOT arguing against abstractions, logic, or concepts. Please do not assume that I am a fool! It's a serious mistake to jump from someone's criticism of ONE abstraction to the assertion that he opposes ALL abstraction. (I don't like this Granny Smith apple I am eating. That does not mean that I don't like all Granny Smiths.) Rather, I was saying that (1) not all abstractions, logical arguments, or concepts are created equal. Some abstractions are better (more useful, more revealing) than others. and (2) empirically and logically speaking, RATEX is a poor abstraction. > The trees are not the forest, but without trees there's no forest. > According to Deutscher, Trotsky used to say that there's nothing more > practical than a good theory. Hey wasn't Lucas once a leftist of some sort? do you think he followed Trotsky? I don't see how anyone could defend RATEX using Trotsky. The former involves a totally idealist and static vision of the world, while Trotsky was an historical materialist. > ... I know where this argument leads. Nowhere. You can type faster than > I can think. I just wanted possible young readers to know that > there's a different way to look at these things. That's all. I want "young" readers to know that you don't have to think the way the hegemonic schools of economics think (idealism, reification, etc.) It may be necessary to succeeding in the econ. biz, but if we want clear thinking, we should avoid that way of understanding the world. Maybe it's because I have tenure, but I think clear thinking is more important than professional success. -- Jim Devine / "Segui il tuo corso, e lascia dir le genti." (Go your own way and let people talk.) -- Karl, paraphrasing Dante. _______________________________________________ pen-l mailing list [email protected] https://lists.csuchico.edu/mailman/listinfo/pen-l
