George Box’s Realization, That All Models, Especially Statistical Models, Are Wrong Means That It Is Impossible for There to Be Any True Probabilities, True Models, True Theories, True Expectations or True (Accepted) Hypotheses: The Claims, Made By Rational Expectations Proponents About True Objective Probabilities, True Models, True Expectations, True Expected Values, etc., Have No Scientific Foundation

Abstract

R. Muth’s 1961 paper in Econometrica is an example of an academic economist suffering from extreme and extraordinary ignorance about very basic scientific, statistical, methodological, epistemological, philosophical, and logical tenets. These tenets were all discussed rigorously by Haavelmo in his 1944 paper in Econometrica on methodological and epistemological issues related to the question about exactly what conclusions a statistical model using economic data could support as far as sound(never valid, right ,correct, true). Haavelmo’s entire 115 page article can be summarized briefly by George Box’s brief statement noted above .Muth’s belief ,that economists had derived true theories and true models is simple nonsense. What Muth should have argued in his 1961 article, but did not, is that rational expectations, when compared to its rivals,was a better model compared to the rival models in terms of minimizing its forecasting error.Of course, Muth would have had to have supported this claim of forecast superiority with explicit proper scoring rule studies showing that the rational expectations statistical model did in fact yield superior forecast results. No such study exists at the present time because no proponent of the rational expectations hypothesis has ever demonstrated such forecasting superiority. Two additional, very severe problems result from Muth’s confused understanding of what scientific modeling is capable of regarding the specification of objective and subjective probabilities or probability distributions. First,It is simply impossible to assume the truth of the objective theory of a limiting frequency interpretation of probability and then claim that decision makers can learn from experience what the convergence properties of such a distribution are in the short run when such properties will only manifest themselves in the far ,distant(infinite)long run. It is impossible for any human ,unless rational expectations proponents are claiming that there are actual decision makers who live to infinity,to learn of such properties .Kolmogorov made it quite clear that the strengthened limiting frequency interpretation he had provided for R.Von Mises law of large numbers,which resulted from the substitution of the assumption of countable additivity for finite additivity “ …does not contribute anything to substantiate the application of the results of probability theory to real practical problems where we always have to deal with a finite number of trials”(Kolmogorov). Second, given L. J. Savage’s clearly expressed and emphasized restrictions concerning the applicability of the theory of subjective probability and expected subjective utility to the short run at the micro level ,any attempt to apply Savage’s theory at the macro level or to the intermediate to long run or intertemporally over infinite time is “utterly preposterous” and complete”nonsense according to Savage.This means that Muth’s claim about subjective probability transforming itself into objective probability might only be considered a subject for possible study and scrutiny in the short run at the microeconomic level. Economists and econometricians are good number crunchers only. Their subjective claims and beliefs about infinitely lived decision makers, who are able to learn about so called true ,objective probability distributions ,are “utterly preposterous “ and anti scientific. William McChesney Martin ,FRS chair of the BOD in Washington,D.C.from 1950 and 1970 ,correctly kept the econometricians ,economists, and Paul Volcker in the basement of the NY Federal Reserve Bank, where they had no input or say in actual or proposed monetary policy decisions. Martin realized that the extremely narrow understanding of what was required in the real world of monetary policy on the part of economists necessitated this setup. Alan Greenspan, in his 2004 (May) paper in the AER ,also rejected the applicability of the theory of rational expectations in monetary policy from 1987 to 2005,given his emphasis on the importance of taking into consideration all models, including those dealing with uncertainty where the knowledge base ranged from complete ignorance to complete knowledge. On the other hand ,A. Burns, P.Volcker,B. Bernanke, and T .Geithner did believe that their mathematical and econometric models were true. One can compare the great success of Martin and Greenspan with the failures of Volcker and Bernanke(Geithner) and reach the conclusion that it is time to return the economists and econometricians back to the basement of the FRBNY where they belong as effective and useful number crunchers only.