Neoclassical Economics and the Psychology of Risk and Uncertainty
- 92 Downloads
The prospect of writing a chapter on the psychology of uncertainty and its potential implications for the economics of uncertainty left me at once excited and worried. The excitement was prompted by the fact that the psychology of economics has a direct bearing on the hard core of the neoclassical research program. The apprehension was related to effectively the same thing. The dilemma that I was faced with is the following: can one do psychology of economics (or more specifically, uncertainty) without at the same time giving up the neoclassical research program? Or is it necessarily true that de gustibus non est disputandum in neoclassical economics?
KeywordsRational Expectation Hard Core Prospect Theory Expect Utility Theory Neoclassical Economic
Unable to display preview. Download preview PDF.
- 2.Sahlin (1983) refers to literature that extends the Bayesian decision theory to take account of higher order probabilities. He makes the astute observation that “. . . in the realm of second order estimates we are dealing with ‘unverifiable’ events . . . , i.e. we don”t know, and will never know, which first order estimate is the ‘true’ one” (1983, p. 97). Even this Bayesian theory is ad hoc in its treatment of the infinite regress problem caused by third, fourth and higher order probabilities (1983, p. 96).Google Scholar
- 3.There are some econometric queries relating to their empirical work. First, in fitting their model to individual subjects they find that 5 out of the 32 subjects behave in a manner inconsistent with the theory. In the final preferred results these 5 “recalcitrant” subjects have been omitted which would seem to suggest that the estimation is somewhat “cooked” in favor of acceptance of the model (see Einhorn and Hogarth, 1985a, p. 444, and their table 3, p. 446). Second, although they explicitly acknowledge the dependence of the parameters b and t on individual and situational factors, they fail to model this dependence for example by using a varying parameter model (see Maddala, 1977, chapter 17). Therefore it is impossible to judge how well the model fits the data overall. Inspection of their table 3 (p. 446) furthermore reveals that the parameters b and t do vary enormously across individuals, so that the results reported in their table 1 (p. 444) that ignore these differences are not very revealing.Google Scholar
- 4.This section relies heavily upon Heijdra and Lowenberg (1986a). The MSRP has been applied to psychology and its relation to economics by Coats (1976) and Simon (1976), among others.Google Scholar
- 5.See, for example, Green (1977) and Grossman and Stiglitz (1980) for a discussion of existence problems in models with perfect competition and costly information. The reason for the nonexistence of a competitive equilibrium is quite simple. By definition, in a competitive equilibrium, arbitrage opportunities are exhausted. This implies that arbitrageurs buy costly information without getting a return; clearly not an equilibrium situation. A major defect of the rational expectations concept is the multiplicity problem. This and other problems are discussed in Azariadis (1981) and Frydman and Phelps (1983). A survey and further discussion of these matters can be found in Lowenberg and Heijdra (1985).Google Scholar