Customer Reviews

The Economic World View: Studies in the Ontology of Economics

5 star:		(0)
4 star:		(0)
3 star:		(0)
2 star:		(1)
1 star:		(0)

 

 

The editor of this volume of collected essays,U Maki,made the mistake of republishing an error filled paper of J.Runde that originally appeared as an invited paper in The Monist(1995).Runde corrected several of his errors regarding Frank Knight,but none of his errors regarding John Maynard Keynes.Runde's essay is then used ,implicitly and/or explicitly,by the other essay authors whenever issues about Keynes and probability,decision making,rationality,expectations,etc.,arises as an issue in their essays.Runde's major error is that he bases his entire discussion of Keynes's approach to the quantification and measurement of probabilities on chapters 3 and 4 of Keynes's A Treatise on Probability(1921;TP).This is precisely the same error made by Frank P. Ramsey, in 1922 and 1926, in two error filled reviews of Keynes's book that Bertrand Russell correctly described as having the least value of any paper written by Ramsey during his brief lifetime(1903-1930).In chapter 3 of the TP,p.37,Keynes warns the reader that his approach to the measurement and quantification of estimating probabilities would not be completed until after Part II of the TP had been finished.In chapter 15(pp.160-163) and chapter 17(p.186-194)of his TP,Keynes makes it crystal clear that by the terms "nonnumerical" and "nonmeasurable" he means not by a single numeral(number)but by TWO numerals.Keynes is the founder of the interval approach to probability.This is directly contrary to Runde's claim that Keynes's theory is strictly a qualitative (comparative)approach involving the use of ordinal rankings,combined with some very special cases of numerical measurement involving frequencies and applications of Keynes's principle of indifference.This would make Keynes's theory inoperative in the real world of actual decision making.Fortunately,Ramsey and Runde are wrong.For Keynes,it is the intervals themselves,if they overlap in any way, which are incomparable,nonrankable,and incommensurable.A simple example is to take the two probability interval estimates [.3,.5] and [.4,.6].These two probabilities can't be ranked or compared.They are "nonnumerical" probabilities.Runde wastes the reader's time by presenting diagrams purporting to represent intervals without the use of any numbers. This conception has absolutely nothing to do with the approach to probability estimation presented in 1921 by Keynes in his TP.We can now summarise Keynes's approach. All probabilities are either point estimates ,involving single numbers, or they are interval estimates, involving two numbers.These latter proabilities can be called nonadditive,imprecise or indeterminate probabilities,as long as the Bayesian interpretation of imprecise is not imputed to Keynes.In chapter 26 of the TP,Keynes incorporated his interval estimate approach into a method for technically dealing with expectations and decision theory.Keynes presented his conventional coefficient of risk and weight,c.The goal of the decision maker is to maximize cA,as opposed to pA,expected value or pU(A),expected utility.c=p(1/(1+q)[2w/(1+w)].Both c and w,which measures the completeness of the potential and actual amount of relevant evidence upon which the decision maker will base his calculation of an estimated probability,are normalized on the unit interval [0,1].Naturally,one will obtain a lower bound and an upper bound when using the c coefficient formula.This would then lead to a "nonnumerical" weighted value and/or "nonnumerical" mathematical expectation or expected value.All of this ,however,is omitted in both Ramsey and Runde.The lesson to be learned is that papers should be carefully examined for errors by editors before they are accepted for publication.The only possible reason for purchasing this book would be if a potential reader wanted to compare the original 1995 Runde article with Runde's 2001 contribution in this volume and see if any learning from experience has occurred.My own opinion is that the reader would be better off just reading chapters 15 and 17 of the TP.