Book contents
- Frontmatter
- Contents
- Preface
- 1 Introduction
- 2 Models of convergence
- 3 Behavior in the short run
- 4 Uncertainty
- 5 An application to state lottery games
- 6 An application to the problem of search behavior
- 7 Inflationary disequilibrium
- 8 Advertising and imitation
- 9 An application to migration
- 10 Conclusions
- References
- Index
- Frontmatter
- Contents
- Preface
- 1 Introduction
- 2 Models of convergence
- 3 Behavior in the short run
- 4 Uncertainty
- 5 An application to state lottery games
- 6 An application to the problem of search behavior
- 7 Inflationary disequilibrium
- 8 Advertising and imitation
- 9 An application to migration
- 10 Conclusions
- References
- Index
Summary
In 1921, Frank H. Knight was careful to draw a distinction between the concepts of “risk” and “uncertainty.” He maintained that “risk” describes circumstances under which one can associate a known probability with each of a series of alternative states of the world. When this is true, expected-return calculations are possible, and maximization of expected payoff provides an appropriate guide for decision making. “Uncertainty,” on the other hand, was a term he reserved for circumstances under which outcome probabilities are either unknown (such as the likelihood of a disaster at a nuclear power plant) or unknowable (such as personal well-being in the afterlife). In such cases, expected-return computation is impossible, and one must seek for some alternative decision rule.
The problem of decision making under uncertainty generated a considerable amount of speculation during the 1940s and 1950s, and a number of proposals evolved for the specification of behavior that might be appropriate for dealing with it. These ranged from a pessimistic “maximin” criterion (which proposes that “nature” is bent on one's destruction, so that one's best choice is that which least exposes one to potential damage) to the naive “principle of insufficient reason” (which places equal likelihoods on all distinct outcomes). A good survey has been provided by Luce and Raiffa (1957). None of these has achieved general approval, however, and in recent years Knight's distinction itself has lost currency as economists have turned to the development of Bayesian models of decision making and to an acceptance of Savage's view (1954) that one can formulate subjective probability distributions that will successfully explain behavior using the expected-utility procedure no matter what the objective circumstances may be.
- Type
- Chapter
- Information
- A Theory of Adaptive Economic Behavior , pp. 59 - 83Publisher: Cambridge University PressPrint publication year: 1983