Methodology from the Viewpoint of an Economic Theorist: Fifty Years On
(1)
Department of Economics, Saarland University, Saarbrücken, Germany
Revised version of a paper presented at the Research Seminar of Dinko Dimitrov, University of Saarland, Saarbrücken July 22nd, 2014. I wish to thank Max Albert (Gießen), Ulrich Schlieper (Mannheim), Egon Richter (Braunschweig) and Dieter Schmidtchen (Saarbrücken) for their advice, criticism, and suggested corrections.
1 Introductory Remarks
1.
As part of social sciences, modern political economics has been started by Scottish moral philosophers, among them David Hume with his Treatise of Human Nature (1739/1740). He treats, what is called today, the problem of social control under the assumption of egoistic behaviour of individuals.1 What got lost in classical economics (as defined by Keynes 1936) was that Hume started essentially from the problem of future uncertainties of individual human beings. How are human individuals able to deal with future uncertainties despite their physical weaknesses?2 Comparing humans in the state of nature with animals, Hume (1739/1740, 1984, 536 ff.) writes:
Of all animals, with which the globe if peopled, there is none towards whom nature seems, at first sight, to have exercis’d more cruelty than towards man, in the numberless wants and necessities, with which she has loaded him, and the slender means, which she affords tot he relieving these necessities. …
′Tis by society alone he is able to supply his defects, and raise himself up to an equality with his fellow-creatures, and even acquire a superiority above them. By society all his infirmities are compensated; …
…
This can be done by no other manner, than by a convention enter’d into by all the members of the society to bestow stability on the possession of those external goods and leave everyone in the peaceable enjoyment of what he may acquire by his fortune and industry.
[The rule for individual property is:] I observe, that it will be for my interest to leave another in the possession of his goods, provided he will act in the same manner with regard to me.
2.
By contrast, Léon Walras (1874) was supposedly guided by Isaac Newton’s (1687/1999) celestial mechanics in developing general equilibrium theory,3 assuming (among others) no uncertainty, i.e., all risks can be shifted through the market. It became an essential part of Neoclassical Microeconomics whose mathematics is based on the differential calculus of Newton’s analytic mechanics.4 Before long, other mathematical theories developed for physicists originally were also used by micro economists—among them also the calculus of variations (applied for dynamic optimization in the theory of optimal growth5) or stochastic differential equations (of statistical physics) as used in the “New Keynesian Framework” of macroeconomics. In the latter case, coefficients are no more estimated statistically but “calibrated” by researchers who take parameters that have been estimated for similar models into their models, solve them numerically and use them for simulation purposes. Sims (1996, p. 113) argues that
…what dynamic, stochastic, general equilibrium modellers in economics are doing not only resembles Kuhn’s normal science, it is normal science. Macroeconomists are said to have available a “well-tested”, or “standard” theory. They do (computational) “experiments.” These experiments usually result in “established theory becoming stronger,” but occasionally discover an extension of the existing theory that is useful, and thereby “established theory” is “improved.”
Sims continues, “these analogies with established physical sciences are strained,” and what Kydland and Prescott
…call computational experiments are computations, not experiments. In economics, unlike experimental sciences, we cannot create observations designed to resolve our uncertainties about theories; no amount of computation can change that (Sims ibid.).
Nevertheless, economists love examples from natural sciences—not only from physics but also from biology. Thus, e.g., the new developments in cognitive neurosciences become increasingly popular—both among marketing economists like Kroeber-Riel (1992) and scholars of financial markets like Camerer et al. (1994).6 However, what remains alien to the natural sciences is the economic problem of uncertainty of the future in the sense, as Keynes put it, “… there is no scientific basis on which to form any calculable probability [of future events] whatever. We simply do not know.” (Keynes 1937, p. 114).7
2 Dominant Methodology of Economic Theory 50 Years Ago: An Outline
Methodology of economics was a fashionable topic in Germany 50 years ago.8 The German Economic Association (Verein für Socialpolitik) had organized a workshop on methodological problems of economics (Giersch and Borchardt 1962), on which the Popper Criterion of physical science (Popper 1959) that just Hans Albert had thrown into the debate caused some unrest. Despite all its problems, German economic theorists (including myself) were marked by a desire to apply the analytic style of theoretical physics to their subject area. As a result, we observe a strong move towards the methodology of physics, a fact illustrated by the main contents of my inaugural lecture at the University of Saarland in December 1964.9 I presented the methodology of economic theory in short as follows:
Economics is part of Real Sciences, not of Humanities. To explain real phenomena means to derive them from other already known phenomena. For that purpose, a logical structure is required of whatever type—verbally or mathematically—into which we map that section of the real world, which we wish to explain—a process called “measuring.” We imagine, in this context, the real world to consist of objects, properties of objects, and relationships between objects. To measure then means to map those objects (together with their properties), and the relationships between them into some logical structure, e.g. a mathematical theory. Of particular significance are the relationships between objects, for it is they, which permit us to make inferences as to the phenomena we wish to explain. Relationships between objects that appear continuously are called general hypotheses or “laws.” Only the presumption that particular relations between objects that have been previously found to exist will also exist in the future, enables us to derive one phenomenon from others, i.e., to explain or predict it. Empirical tests of theories take place by, so to speak, some reverse mapping from model to reality. It is in this context in which the Popper Criterion applies (Popper 1959). It declares that an empirical theory only says something about reality when it is possible to refute it by empirical test. Popper’s criterion demands to reject a falsified theory. The problem is that economic theories fail to fulfil this criterion.
So much, in a nutshell, on the methodology of economic theory of the years following World War II. What to say in retrospect? Deidre McCloskey (1983, p. 484) answers that above methodological position is a poor, arrogant and pretentious way of thinking that no economist would apply in his research work.10 Rather, economists would employ the rhetorical method, which we’ll illustrate shortly by an example. However, if this is true, which method do economists use to assess the quality of their research? None replies Ms McCloskey. Economists rather follow the rules of rhetoric—the classical method to convince other people of the correctness of one’s opinion. They do this by use of such common rhetoric tricks as analogies, appeal to authority, simplifications for the sake of argument and so forth. In fact, the language used by representatives of the many quarrelling schools of economics illustrate the rhetoric character of our science: Keynesianism, Monetarism, German Historic School, Austrian School, German Ordnungspolitik, New Institutional Economics and others—Though the rhetoric element plays an important role in our field, we need some standard of assessment for our professional claims or predictions. According to my experience we do this on the basis of the majority opinion among economic experts or “peers”. But this means that an economic expert’s opinion is of quite a different quality compared with that of, say, an experimental physicist. Whereas theories of natural sciences can be quite precisely confirmed or rejected, economic theories are at most very general or not at all certifiable. As a result, decisions based on economic theories are much more “speculative” than those based on theories of the natural sciences.
I’ll now return again to the economic problem of uncertainty of the future as emphasized at the beginning of my paper. With all the enthusiasm for the usefulness and elegance of analytic economic mechanics à la Isaac Newton, we must not ignore that this approach—simply from it’s posing the question—is unable to answer our problem of future uncertainties. Things become complicate by the fact that economics is a special field of social science—the science of a society of intelligent individuals that is. Robbins (1932, p. 16), as mentioned above in this chapter (footnote 8), describes economics as the analysis of “human behaviour as a relationship between ends and scarce means which have alternative uses”—to which must be added “in the face of uncertainty.” Against uncertainty does only help above indicated “socialisation” of individuals, based on principles of individual behaviour such as David Hume’s principles of Natural Law.11 However, already from the way of putting the question, that problem is incomparable with those of Newton’s mechanics.
Regardless of this, it is worth the trouble to apply Newton’s style of research to problems of economics as done in neoclassical microeconomics. It helps us to structure the problem of the allocation of scarce resources among a great number of people with different needs and wealth levels without interposition of a central planning unit. However, it provides no more than the model of an ideal world in the state of equilibrium without explaining the dynamic process of searching for that equilibrium. It is not made for applying as forecasting model—even though Arrow’s time-state preference theory seems to have played some role among modern financial market theorists.12 In any case, we must not forget always to mention the ideal assumptions of microeconomic models—as is regularly done by serious economic theorists.13
3 Economics as Rhetoric Science: Irving Fisher’s Debt-Deflation Theory and Its Present Defence
One example must suffice: The Debt-Deflation Theory of Irving Fisher as explanation of the Great Depression and its present application by Ben Bernanke in support of monetary policy of the FED (and later of the ECB under Mario Draghi).
Note: Monetary policy is foreign to neoclassical microeconomics in which money is neutral and for which uncertain expectations (as defined by Keynes) are unknown. With the assumption of non-neutral money and uncertain expectations, we enter the field of what became Keynesian macroeconomics, though Irving Fisher’s Debt-Deflation Theory (1933) appeared already some years before Keynes’s General Theory (1936).
Fisher justifies his theory partly by business style arguments, partly on the basis of historic experiences. He describes his theory by a sequence of nine states of the economy—methodologically by use of the concept of “understanding” (Dilthey) of the humanities (Geisteswissenschaft):