John Hull wrote: >"Example 3: Subjective Utility
>"Most of the utility 'functions' occurring in >neoclassical microeconomics...are not well defined--as >Henri Poincare pointed out to Leon Walras. In fact, >the only conditions required of them is that they be >twice differentiable, the first derivative being >positive and the second negative. Obviously, >infinitely many functions satisfy these mild >requirements. THIS OFTEN SUFFICES IN SOME BRANCHES OF >PURE MATHEMATICS.... BUT THE FACTUAL (OR EMPIRICAL) >SCIENCES ARE MORE DEMANDING: HERE ONE USES ONLY >FUNCTIONS THAT ARE DEFINED EXPLICITLY...OR IMPLICITLY. > Finally, experimental studies have shown that >preferences and subjective estimates of utility and >risk do not satisfy the assumptions of expected >utility theory. >"In short, THE USE OF UTILITY FUNCTIONS IS OFTEN >MATHEMATICALLY SLOPPY AND EMPIRICALLY UNWARRANTED. >Now, rational choice models make heavy use of both >subjective utilities and subjective probabilities, as >well as of the simplistic hypothesis that selfishness >is the only motivation of human behavior. Not >suprisingly, NONE OF THESE MODELS FITS THE FACT. >Hence, although at first sight they look scientific, >as a matter of fact they are pseudoscientific." My only comment on this is: it is a silly, uninformed criticism of economics. Economics may or may not be "pseudoscientific", but not for these reasons. First, it has been shown long ago, by Gerard Debreu and others, that utility functions are not needed in order arrive at many of the important results in microeconomic theory. Second, the claim that economists use the hypothesis that selfishness is the only motivation of human behaviour is simply wrong, wrong, wrong. Finally, yes, economists often use models that are false and don't fit some data. So do physicists when they assume away the existence of any frictions. Is physics also a pseudoscience for this reason? Alex
