"The social object of skilled investment should be to defeat the dark forces of time and ignorance which envelop our future. The actual, private object of the most skilled investment today is ‘to beat the gun’, as the Americans so well express it, to outwit the crowd, and to pass the bad, or depreciating, half-crown to the other fellow."
– John Maynard Keynes, 1936
I’m about to share with you a little excerpt (only some 1,750 words) from one of the many magnificent papers by David Felix, an Emeritus Professor of Economics at Washington University in St. Louis, USA.
Philosophically speaking, anything that is not harebrained (e.g., the "invisible hands," not of Adam Smith, but of the occultists and the action-at-a-distancers and the panpsychists), and that with some degree of plausibility demotes the scope of human knowledge, is attractive to me. My purposes in reproducing this excerpt therefore are (a) to show you how David Felix demotes the scope of human knowledge among the financial, insurance, and real estate collective ("The coup de grace," he writes, "came when it was shown that the…predictions of all these models were worse than those produced by a naïve ‘random walk’ model"), while also (b) promoting the work of this remarkable but under-appreciated economist.
Endnotes 2 and 4 are taken directly from Felix’s original, which is something more than 20,000 words long, with nine tables and an extensive bibliography and footnotes. Endnote 1 derives from the same. I myself have added endnote 3 by way of using Keynes’s imagery to illustrate Felix’s point about the "innately unstable dynamics" of market relations, and the limited scope of human knowledge when it comes to understanding these dynamics — beyond the fact that they are real, are bad for everybody’s health, and cannot be controlled and counted on by their participants to "defeat the dark forces of time and ignorance which envelop our future."
(Excerpted from David Felix, "Why International Capital Mobility Should Be Curbed, and How it Could Be Done," February, 2001, pp. 27-32. Analysis prepared on behalf of the International Confederation of Free Trade Unions (
"The Empirical Fallacy of Milton Friedman’s ‘Rational Expectations’ Model"
Contrary to Milton Friedman’s prediction [that floating foreign-exchange rates plus free capital mobility offer multiple advantages to national economies as well as to the international system,] rapidly rising short-term capital movements accompanied the rising exchange rate volatility [that followed the dismantling in 1971 of Bretton Woods' fixed foreign exchange system that had pegged member exchange rates to the gold-convertible U.S. dollar].
By the early 1980s, it was evident that short-term capital movements were far outstripping the flows needed to finance international trade in commodities and non-financial services, such as shipping, international travel and tourism. The
The explanation sparked a succession of econometric "new models" of exchange rate determination to test its empirical validity. That is, they tested whether the exchange rate volatility reflected in fact the response of financial markets to new relevant information, with relevancy determined by the "true model" that was supposed to guide the reactions. To pass the test, the "news" model would have to show that as new information kept being added to the information set, the model traced out exchange rate movements whose volatile path closely mimicked that of the actual exchange rate, both within the period covered by the sample data and outside of it. When the initial "true model" — Friedman’s quantity theory extended to monetary interactions between countries — produced poor predictions of the "out-of-sample" exchange rates, Ratex-embedded "true models" with different structural features were devised and tested econometrically. The coup de grace to this general line of empirical research came when it was shown that the out-of-sample predictions of all these models were worse than those produced by a naïve "random walk" model. The latter merely predicts that today’s exchange rate is always equal to yesterday’s rate, plus or minus random deviations that have a constant variance and cancel each other out. The fact that such a minimal information model predicted better than models purporting to provide more structural information on exchange rate dynamics was devastating. The shock was compounded by subsequent econometric investigation, which showed the random walk model to be dominated by the "martingale" model. In that model, today’s rate is predicted to equal yesterday’s plus or minus deviations of varying size, whose variance is neither stable nor cancelled out. Such a model provides exchange market traders with even less useful information than the random walk model.
"The Demolition of the Logical Validity of the ‘Rational Expectations’ Postulate"
The demolition involved two mathematical criticisms. The first showed that the solutions to the linear equations used in the Ratex-embedded econometric models of asset prices, including exchange rates, necessarily contain a constant term that can move the predicted asset prices to anywhere between plus and minus infinity. This explosive constant, dubbed the "rational speculative bubble," is arbitrarily set at zero in the Ratex models on the premise that rational speculators, aware that in reality such bubbles must sooner or later burst, would time the burst correctly and make huge profits from cashing in before the explosion. But since all speculators were by assumption equally rational and knowledgeable, they would do likewise. Collectively they would move the burst closer and closer to the present, leading to the conclusion that a rational expectations bubble could never get started. But this implies that the speculators draw on information that is not contained in the Ratex models, which predict everlasting price bubbles.
This problem is a very general one and appears in all rational expectations models. In all these models there is an infinity of solutions, most of which are unstable. The need then arises to select one particular solution. This selection will necessarily be based on information not contained in the model. Thus even in rational expectations models, ad hoc assumptions are necessary. Fully consistent expectations appear to be impossible.
The second logical criticism comes from "chaos mathematics," which deals with the strange properties of non-linear systems of equations. Chaos mathematics shows that when the equations are iterative — that is, when the time paths of the variable(s) being measured are determined by the preceding value(s) of the variables(s) — the paths can become extremely sensitive to very small differences in the constant terms that define "initial conditions" of the system. And for some initial values no stable path or solution exists.
[Passage omitted. Felix writes, for example: "[W]hen k equals or exceeds 3.57, called the Feigenbaum bifurcation, after Mitchell Feigenbaum the physicist who discovered the patterns through computer simulations, the trajectories begin bounding around with no discernible pattern, other than that minute differences in the value of k produce major divergences between the bounding trajectories."]
The implication? In markets for very frequently traded assets, such as for shares of stock or foreign exchange, where asset prices change minute to minute, the reliability of price forecasts of up to only about a week is relatively insensitive to small informational or computational errors. Beyond that time span, sensitivity to minute errors of either type shoots up explosively. The Ratex assumption that rational participants in financial markets can and do operate with accurate long-term forecasts is a pipe dream.
"Back to Reality"
The logical demolition of Ratex allows more reality to enter the modeling of financial markets. For example, chartists — short-term market traders who time their buying and selling by searching past price and volume data for hidden behavioral patterns that indicate to them whether the ongoing direction of asset prices and volume of transactions will persist or change — are a permanent and sizeable component of financial markets. Yet trading losses should have driven them from the market according to Friedman type models, and they could never appear according to Ratex models. Freed of these blinkered spectacles, many current modeling efforts are now trying to explain financial market dynamics as generated by interaction between "momentum" and "fundamentals" traders, with the relative importance of the two groups shifting in response to recent results.
The demolition of Ratex also requires abandoning its "representative agent" simplification. Analysts cannot avoid dealing with how asset prices get formed in markets composed of traders who disagree on the "true model," and thus on what information is relevant, as well as on market timing strategies. In such markets, the neoclassical premise that rationality means maximization goes out the window, at least for rational market participants. This is because the processing of "news" about fundamentals by rational traders has to include assessing how the rational traders with other strategies will react to the "news," along with awareness that they are also assessing reaction to the "news." Since trying to form accurate a priori deductions of each others’ reactions involves an infinite regress into subjectivity, rational traders have no choice but to resort to inductive reasoning. They validate their investment positions by evaluating the payoffs, but they lack the essential information for determining a priori which positions produce the maximum payoffs. Heterogeneous traders arriving at a common set of expectations would be a special case, based on reaching consensus not on the "true model," but on likely reactions to "news." That’s most likely to occur during bubbles, when the extended upsurge makes momentum investing for a time "the only game in town," and during crashes, when all try to beat it to the exit at once. All this is a general validation of Keynes’s perspective on investor behavior under uncertainty, as put forth in Chapter 12 of his magnum opus. The policy implication is that free capital mobility, which unleashes the innately unstable dynamics of financial markets on the global economy, lacks normative support from economic theory, whether neoclassical or Keynesian. To claim otherwise is false.
—- Endnotes —-
—- Endnotes —-
 Paul De Grauwe et al., Exchange Rate Theory: Chaotic Models of Foreign Exchange Markets (Oxford: Blackwell Publishing, 1993), p. 69.
 The major foreign exchange markets are dominated by momentum traders. Thus a recent survey of the trading strategies of dealers in the London foreign exchange market, the world’s largest, indicates that 90% used charting for short period trading because, they said, it provided more reliable short-period forecasts than econometric models that emphasize fundamentals. For longer period trading, information about fundamentals help shape their position [Macdonald and Taylor, "Exchange Rate Economics: A Survey," IMF Staff Papers, Vol. 39, March, 1992]. The latter, however, is small beer, since over 80% of foreign exchange trades now involve round-trips of a week or less.
 See John Maynard Keynes, The General Theory of Employment, Interest, and Money (of 1936) (Harcourt Brace Jovanovich, Inc., 1964), Ch 12, "The State of Long-Term Expectations," pp. 147-164. — In Keynes’s famous illustration of market psychology and the "extreme precariousness of the basis of knowledge" by which speculators attempt to navigate its high-seas, "It is as though a farmer, having tapped his barometer after breakfast, could decide to remove his capital from the farming business between 10 and 11 in the morning and reconsider whether he should return to it later in the week" (p. 151). "Or," Keynes adds later, "to change the metaphor slightly, professional investment [i.e., Wall Street and hedge-fund-caliber speculation] may be likened to those newspaper competitions in which the competitors have to pick out the six prettiest faces from a hundred photographs, the prize being awarded to the competitor whose choice most nearly corresponds to the average preferences of the competitors as a whole; so that each competitor has to pick, not those faces which he himself finds prettiest, but those which he thinks likeliest to catch the fancy of the other competitors, all of whom are looking at the problem from the same point of view. It is not a case of choosing those which, to the best of one’s judgment, are really the prettiest, nor even those which average opinion genuinely thinks the prettiest. We have reached the third degree where we devote our intelligences to anticipating what average opinion expects the average opinion to be. And there are some, I believe, who practice the fourth, fifth and higher degrees" (p. 156). — Catastrophically, and to paraphrase Richard Nixon, we are all speculators now.
 A promising way of modeling stock market behavior along these lines is reported in [Brian] Arthur et al. in 1997. The method is to vary the parameters of a non-linear model of heterogeneous investors that allows them to react to disappointing payoffs by modifying their trading strategies. The model, which has an expectational equilibrium built into it, was tested on a computerized, simulated stock market. Whether the traders converged on the equilibrium turned out to depend on the nature and speed of their reactions to the changing prices, volumes and yields generated by the computer runs. The general finding was that the more quickly the traders adjusted their strategies to the market outcomes, the more the market self-organizes into a complex regime. "A rich market psychology — a rich set of expectations — becomes observable. Technical [i.e., chartist] trading emerges as a profitable activity, and temporary bubbles and crashes occur from time to time. Trading volume is high, with times of quiescence alternating with intense market activity. The price time series shows persistence in volatility…and in trading volume…individual behavior evolves continuously and does not settle down." [Brian Arthur et al., "Asset Pricing Under Endogenous Expectations in an Artificial Stock Market," Economic Notes, Vol. 26, No. 2, 1997, p. 203.]