Rational constrained Choice là gì

 1. Preferences are continuous if they do not present sudden reversals. More formally, an individual has continuous preferences when her preferring each element in a sequence of options {x n } to the corresponding element in the sequence {y n } implies that she also prefers the limiting option of the first sequence to the limiting option of the second sequence. On continuity and the other properties of preferences, see Mas-Colell, Whinston and Green (1995 Mas-Colell, A., Whinston, M.D. and Green, J.R. 1995. Microeconomic Theory, New York: Oxford University Press.  [Google Scholar], chaps 1 and 3).

 2. Stochastic models were introduced by Georgescu-Roegen (1950 Georgescu-Roegen, N. 1950. The Theory of Choice and the Constancy of Economic Laws. Quarterly Journal of Economics, 64: 125–138. [Crossref], [Web of Science ®] , [Google Scholar]), Quandt (1956 Quandt, R.E. 1956. A Probabilistic Theory of Consumer Behavior. Quarterly Journal of Economics, 70: 507–536. [Crossref], [Web of Science ®] , [Google Scholar]), Luce (1958 Luce, R.D. 1958. A Probabilistic Theory of Utility. Econometrica, 26: 193–224.  [Google Scholar], 1959 Luce, R.D. 1959. Individual Choice Behavior, New York: Wiley.  [Google Scholar]) and Marschak (1959 Marschak, J. 1959. “Binary Choice Constraints and Random Utility Indicators”. In Mathematical Methods in the Social Sciences, Edited by: Arrow, K.J., Karlin, S. and Suppes, P. 312–329. Stanford, CA: Stanford University Press.  [Google Scholar]). For a discussion of more recent contributions to the research program on stochastic models, see Loomes (2005 Loomes, G. 2005. Modelling the Stochastic Component of Behaviour in Experiments: Some Issues for the Interpretation of Data. Experimental Economics, 8: 301–323.  [Google Scholar]) and Wilcox (2008 Wilcox, N.T. 2008. Stochastic Models for Binary Discrete Choice under Risk: A Critical Primer and Econometric Comparison. Research in Experimental Economics, 12: 197–292. [Crossref] , [Google Scholar]). The classification of stochastic models used here follows Loomes and Sugden (1995 Loomes, G. and Sugden, R. 1995. Incorporating a Stochastic Element into Decision Theories. European Economic Review, 39: 641–648.  [Google Scholar], 1998 Loomes, G. and Sugden, R. 1998. Testing Different Stochastic Specifications of Risky Choice. Economica, 65: 581–598. [Crossref], [Web of Science ®] , [Google Scholar]).

 3. In applied demand analysis, McFadden (1974 McFadden, D. 1974. “Conditional Logit Analysis of Qualitative Choice Behavior”. In Frontiers of Econometrics, Edited by: Zarembka, P. 105–142. New York: Academic Press.  [Google Scholar]) and others have introduced stochastic models of the Fechner type, usually labeled Random Utility Models. In these models, the random component ϵ is also meant to accommodate measurement errors on the part of the observer.

 4. For a presentation of RCT in contexts involving uncertainty, strategy, and time, see Mas-Colell et al. (1995 Mas-Colell, A., Whinston, M.D. and Green, J.R. 1995. Microeconomic Theory, New York: Oxford University Press.  [Google Scholar], chaps 6, 8, 9 and 20). An example of extended RCT that incorporates experiences and social forces into the theory is provided by Becker (1996 Becker, G.S. 1996. Accounting for Tastes, Cambridge, MA: Harvard University Press. [Crossref] , [Google Scholar]).

 5. A consumer has locally non-satiated preferences when for any consumption bundle x, there exists another bundle y arbitrarily close to x which is strictly preferred to x by the consumer.

 6. This kind of compensation is called Slutsky's compensation. With Hicksian compensation, by contrast, the consumer is compensated so that her utility level is kept constant when prices change. Both kinds of compensations induce negative substitution effects, but Slutsky's are those used in empirical studies because they can be determined even without knowing the consumer's utility function.

 7. This is the so-called ‘law of market demand.’ Exceptions to the law are represented by Giffen goods, which are, however, rare for individual demand and extremely implausible for market demand. For an analysis of Giffen goods in individual and market demand, see Battalio, Kagel and Kogut (1991 Battalio, R.C., Kagel, J.H. and Kogut, C.A. 1991. Experimental Confirmation of the Existence of a Giffen Good. American Economic Review, 81: 961–970. [Web of Science ®] , [Google Scholar]).

 8. For a history of the early experiments on demand behavior, see Moscati (2007 Moscati, I. 2007. Early Experiments in Consumer Demand Theory: 1930–1970. History of Political Economy, 39: 359–401. [Crossref], [Web of Science ®] , [Google Scholar]).

 9. An earlier strand of experimental research employed pigeons and rats as experimental subjects, and checked whether their demand displayed a negative substitution effect. This research is summarized in Kagel, Battalio and Green (1995 Kagel, J.H., Battalio, R.C. and Green, L. 1995. Economic Choice Theory: An Experimental Analysis of Animal Behavior, New York: Cambridge University Press. [Crossref] , [Google Scholar]).

10. GARP was introduced by Varian (1982 Varian, H.R. 1982. The Nonparametric Approach to Demand Analysis. Econometrica, 50: 945–973. [Crossref], [Web of Science ®] , [Google Scholar]) and is a modification of the Weak Axiom of Revealed Preference (WARP) proposed by Samuelson (1938 Samuelson, P. 1938. A Note on the Pure Theory of Consumer's Behaviour. Economica [NS], 5: 61–71. [Crossref] , [Google Scholar]) and the Strong Axiom of Revealed Preference (SARP) proposed by Houthakker (1950 Houthakker, H.S. 1950. Revealed Preference and Utility Function. Economica [NS], 17: 159–174. [Crossref], [Web of Science ®] , [Google Scholar]). WARP allows for cyclical choices, which are excluded by RCT, and excludes indifference curves with straight segments, which are compatible with RCT. SARP rules out cyclical choices but still excludes straight indifference curves. GARP rules out cycles and allows for straight indifference curves, thereby providing a complete behavioral characterization of RCT.

11. The intuition is rough since cyclical choices, which are ruled out by GARP, may materialize only when at least three commodities and three budget/price situations are involved.

12. On the separability assumption, see Varian (1983 Varian, H.R. 1983. Non-Parametric Tests of Consumer Behaviour. Review of Economic Studies, 50: 99–110. [Crossref], [Web of Science ®] , [Google Scholar], 1988 Varian, H.R. 1988. Revealed Preference with a Subset of Goods. Journal of Economic Theory, 46: 179–185.  [Google Scholar]).

13. For a detailed discussion of the power of the GARP test, see Bronars (1987 Bronars, S.G. 1987. The Power of Nonparametric Tests of Preference Maximization. Econometrica, 55: 693–698. [Crossref], [Web of Science ®] , [Google Scholar]) and Andreoni and Harbaugh (2008 Andreoni, J. and Harbaugh, W. 2008. Power Indices for Revealed Preference Test, Mimeo San Diego: University of California.  [Google Scholar]).

14. In addition to the percentage of random agents that violate GARP, one could also employ the percentage of GARP violations as a measure of GARP's power; indeed such a measure is computed by some experimenters. The main problem with it is that there are different ways to count GARP violations. For instance, choices like those in Figure 2(d) count as one violation in some experiments, and are regarded as two violations in others. We focus on the percentage of random agents that violate GARP because this measure is univocally determined, because it was calculated in all experiments reviewed in Section 3, and because no significant new insight is gained by combining it with the percentage of GARP violations.

15. For a critical discussion of the implications of this restriction, see Tubaro (2009 Tubaro, P. 2009. Is Individual Rationality Essential to Market Price Formation? The Contribution of Zero-Intelligence Agent Trading Models. Journal of Economic Methodology, 16: 1–19. [Taylor & Francis Online] , [Google Scholar]).

16. The goods were: cigarettes, coffee, two types of candy, cookies, soda, milk, meal deal with a cigarette (category one); private dormitory room, private locker, grounds pass to leave the ward for a fixed period of time (category two); repeated use of the ground pass, clothes, weekly dance, breakfast, different rights such as right to use cash for packages from home (category three).

17. The goods were: Coca-Cola, orange juice, coffee, licorice, snacks, music video clips, computer games, and magazines.

18. In experiment 1, the goods were: milk chocolate, salted peanuts, biscuits, text markers, ballpoint pens, plastic folders, writing pads, and post-it notes. In experiment 2, milk chocolate, biscuits, orange juice, iced tea, writing pads, plastic folders, diskettes, and post-it notes. In experiment 3, milk chocolate, biscuits, orange juice, iced tea, post-it notes, audio cassettes, ballpoint pens, and batteries.

19. Scientific realism is the view that scientific theories describe the world as it really is. Conventionalism contends that scientific theories rationalize experience in a simple, systematic and possibly conventional way, rather than describing the world as it truly is. Instrumentalism views scientific theories as mere instruments for prediction. For a discussion of realism, conventionalism, and instrumentalism, see Bird (1998 Bird, A. 1998. Philosophy of Science, Abingdon: Routledge. [Crossref] , [Google Scholar], chap. 4), Chalmers (1999 Chalmers, A.F. 1999. What is This Thing Called Science?, Maidenhead and New York: Open University Press.  [Google Scholar], chap. 15) and Psillos (1999 Psillos, S. 1999. Scientific Realism: How Science Tracks Truth, London: Routledge.  [Google Scholar]). For a realist critique of Friedman's as-if argument, see Blaug (1980 Blaug, M. 1980. The Methodology of Economics, Cambridge: Cambridge University Press.  [Google Scholar]), Caldwell (1980 Caldwell, B.J. 1980. A Critique of Friedman's Methodological Instrumentalism. Southern Economic Journal, 47: 366–374. [Crossref], [Web of Science ®] , [Google Scholar]), Musgrave (1981 Musgrave, A. 1981. “Unreal Assumptions” in Economic Theory: The F-Twist Untwisted. Kyklos, 34: 377–387. [Crossref], [Web of Science ®] , [Google Scholar]) and Hausman (1992 Hausman, D. 1992. The Inexact and Separate Science of Economics, Cambridge: Cambridge University Press. [Crossref] , [Google Scholar], chap. 9).

20. See, for example, Tversky and Thaler (1990 Tversky, A. and Thaler, R.H. 1990. Anomalies: Preference Reversals. Journal of Economic Perspectives, 4: 201–211. [Crossref], [Web of Science ®] , [Google Scholar]), Kahneman, Knetsch and Thaler (1991 Kahneman, D., Knetsch, J.L. and Thaler, R.H. 1991. Anomalies: The Endowment Effect, Loss Aversion, and Status Quo Bias. Journal of Economic Perspectives, 5: 193–206. [Crossref], [Web of Science ®] , [Google Scholar], 2008 Kahneman, D., Knetsch, J.L. and Thaler, R.H. 2008. “The Endowment Effect: Evidence of Losses Valued more than Gains”. In Handbook of Experimental Economics Results, Edited by: Plott, C.R. and Smith, V.L. Vol. 1, 939–948. North-Holland: Amsterdam.  [Google Scholar]) and Seidl (2002 Seidl, C. 2002. Preference Reversal. Journal of Economic Surveys, 16: 621–655.  [Google Scholar]).

21. For a discussion, see among others Thagard (1978 Thagard, P.R. 1978. The Best Explanation: Criteria for Theory Choice. Journal of Philosophy, 75: 76–92. [Crossref], [Web of Science ®] , [Google Scholar]) and Lipton (2004 Lipton, P. 2004. Inference to the Best Explanation, London: Routledge.  [Google Scholar]).

22. In game theory, the psychological realism of random behavior has been discussed in relation to mixed strategies. Under the interpretation proposed by Harsanyi (1973 Harsanyi, J.C. 1973. Games with Randomly Disturbed Payoffs: A New Rationale for Mixed Strategy Equilibrium Points. International Journal of Game Theory, 2: 1–23. [Crossref] , [Google Scholar]), players do not randomly choose between different pure strategies but choose rather a single pure strategy in a game of incomplete information associated with the original game of complete information. For a discussion, see Osborne and Rubinstein (2001 Osborne, M.J. and Rubinstein, A. 2001. A Course in Game Theory, Cambridge, MA: The MIT Press.  [Google Scholar], chap. 3).

23. On the Duhem–Quine problem, see Hands (2001 Hands, D.W. 2001. Reflection without Rules, Cambridge: Cambridge University Press. [Crossref] , [Google Scholar], chap. 3).

24. Using the terminology introduced by Cubitt (2005 Cubitt, R. 2005. Experiments and the Domain of Economic Theory. Journal of Economic Methodology, 12: 197–210. [Taylor & Francis Online] , [Google Scholar]), we are assuming that the choices recorded in the experiments belong to the ‘Testing-domain’ of RCT. For a discussion of the Testing-domain of individual choice theory under conditions of uncertainty, see Bardsley et al. (2010 Bardsley, N., Cubitt, R., Loomes, G., Moffatt, P., Starmer, C. and Sugden, R. 2010. Experimental Economics. Rethinking the Rules, Princeton, NJ and Oxford: Princeton University Press.  [Google Scholar], chap. 2).

25. Satz and Ferejohn (1994 Satz, D. and Ferejohn, J. 1994. Rational Choice and Social Theory. Journal of Philosophy, 91: 71–87. [Crossref], [Web of Science ®] , [Google Scholar]) discuss the variance of RCT's explanatory power and its dependency on the particular environment that the theory is applied to. In particular, they suggest that rational-choice explanations are most plausible ‘under conditions of scarcity, where human choice is severely constrained’, while ‘in environments without strong constraints, agents will not generally behave as the theory predicts’ (p. 81). Their conclusions, however, do little to shed light upon RCT's violations and the variance of RCT's explanatory power across the six experiments, in which subjects faced almost the same choice situation and the Satz–Ferejohn scarcity condition for rational choice was satisfied because the subjects' choices were always budget-constrained.

26. Actually, Cox (1997 Cox, J.C. 1997. On Testing the Utility Hypothesis. Economic Journal, 107: 1054–1078.  [Google Scholar], p. 1076) suggested a rough test to compare RCT and the random-choice model, and argued that the individual's choices in his experiments were more consistent with RCT than with random choice. However, not even Cox has examined the issue in detail.

27. In ‘Categorize Then Choose’ (CTC), the agent decides in two stages. First, she categorizes the alternatives in broad classes and focuses on one class; then she chooses an alternative from that class. For example, a CTC agent categorizes restaurants by type of cuisine and focuses on, say, Mexican restaurants; then she chooses the preferred Mexican restaurant. CTC agents may violate RCT.