Bibliography
risky utility u = transform of strength of preference v
Download 7.23 Mb.
|
risky utility u = transform of strength of preference v: people think that after some days of headache, an additional day of headache brings more extra suffering than the first day, so the suffering escalates and the utility function seems to be convex. Still, in risky choices, they are risk averse suggesting that the utility function is concave. Some might interpret this finding as a difference between risky and riskless utility. I would ascribe the risk aversion to taking numbers numerically rather than as values. %} Kahneman, Daniel & Jackie S. Snell (1990) “Predicting Utility.” In Robin M. Hogarth (ed.) A Tribute to Hillel J. Einhorn, 295–310, University of Chicago Press, Chicago. {% %} Kahneman, Daniel & Jackie S. Snell (1992) “Predicting a Changing Taste: Do People know What They Will Like?,” Journal of Behavioral Decision Making 5, 187–200. {% Authors distinguish between experienced and decision utility. Consider ways to optimize the perceived joy of receipt of income, suggesting it is maximized with gradually increasing income and now and then a bonus which does not change the perception of status quo. %} Kahneman, Daniel & Richard H. Thaler (1991) “Economic Analysis and the Psychology of Utility: Applications to Compensation Policy,” American Economic Review, Papers and Proceedings 81, 341–346. {% Cite evidence that people don’t predict their future tastes properly sometimes. %} Kahneman, Daniel & Richard H. Thaler (2006) “Utility Maximization and Experienced Utility,” Journal of Economic Perspectives 20, 221–234. {% %} Kahneman, Daniel, & Anne Treisman (1984) “Changing Views of Attention and Automaticity.” In Raja Parasuraman, D. Roy Davies, & Jean Beatty (eds.) Variants of Attention, 29–61, New York: Academic Press. {% %} Kahneman, Daniel, Anne Treisman, & Jacquelyn Burkell (1983) “The Cost of Visual Filtering,” Journal of Experimental Psychology: Human Perception and Performance 9, 510–522. {% %} Kahneman, Daniel, Anne Treisman, & Brian J. Gibbs (1992) “The Reviewing of Object Files: Object-Specific Integration of Information,” Cognitive Psychology 24, 175–219. {% %} Kahneman, Daniel & Amos Tversky (1972) “Subjective Probability: A Judgment of Representativeness,” Cognitive Psychology 3, 430–454. Abbreviated version as Ch. 3 in Daniel Kahneman, Paul Slovic, & Amos Tversky (1982, eds.) Judgment under Uncertainty: Heuristics and Biases, Cambridge University Press, Cambridge. {% %} Kahneman, Daniel & Amos Tversky (1973) “On the Psychology of Prediction,” Psychological Review 80, 237–251. Reprinted as Ch. 4 in Daniel Kahneman, Paul Slovic, & Amos Tversky (1982, eds.) Judgment under Uncertainty: Heuristics and Biases, Cambridge University Press, Cambridge. {% An early version of their 1979 Econometrica prospect theory paper. Use term uncertainty weight iso decision weight. P. 9 ff. lets isolation refer only to outcomes being changes w.r.t. reference point. 1979 paper will take isolation more general. P. 12: “Hence, it appears that over a reasonably wide range of assets the value function is approximately the same.” Remarkably, for pure prospects (x,p,y) with x > y > 0, they take CE(x,p,y) = y + CE(xy,p,0), which deviates from their 1979 version and is less satisfactory in the sense that it cannot be defined for nonquantitative outcomes. P. 14 is explicit on the “two-stage” model (term not used there) where first probability judgments are made and then these are transformed as objective probabilities would. This is not explicit in the 1979 version, only some text on p. 281 paragraph -2. uncertainty amplifies risk: p. 15 2nd para repeats the two-stage model, and writes: “In these cases, the regressiveness of uncertainty weights with respect to objective probabilities will be further enhanced by the tendency to overestimate low probabilities and underestimate high ones.” This is exactly the condition in my (oh well) 2004 psychological review paper. Wow! Pp. 19 ff is very remarkable on dynamic consistency, already containing the idea of Hammond (1988) and Burks (1977 Ch. 5), to derive independence from dynamic principles, and preceding both of these. Well, at least, they show it for the Allais paradox, but it is typical of any violation of independence. They first have choices between A1 and A2, and then their scaled-down versions B1 and B2. Then come the sequential C1 and C2. They first explain that from the prior perspective, due to RCLA, C1 and C2 are identical to B1 and B2. This is analyzing using dynamic consistency + RCLA, optimizing from the prior perspective. Then they explain that subjects, in a figure that presents the decision node in the 2nd stage, ignore the lower branch and take the 2nd stage choice in isolation. This is what Machina (1989) called consequentialism. Thus they show that to do the Allais paradox one either has to violate DC + RCLA or consequentialism. They claim it more generally, for the certainty effect (which in this paper they formally define as what they will later call subadditivity). They do not claim it in full generality for independence, but they are very close and deserve crediting. I realized this for their 1975 paper (preceding Ch. 5 in Burks 1977!) only 9 Dec. 2012, whereas for their 1979 published version I realized it around 2008. P. 23: “In this theory insurance and gambling occur in spite of the value function, not because of it.” Nice, very explicit, statement. P. 24: “Utility theory can be viewed as an attempt to eliminate the concept of attitude to risk or uncertainty and to explain risky choices solely in terms of attitudes to money or wealth.” P. 25: “Value theory does not purport to account for all forms of risk-seeking and risk-aversion. Many factors not included in this theory (e.g., regret., social pressure, superstition, magical thinking) probably play an important role in risky choices. Value theory is an attempt to modify those assumptions of utility theory that are most severely violated, so as to achieve a more realistic account of choice behavior.” second-order probabilities to model ambiguity: there is a whole section on Ellsberg (pp. 30-33) suggesting that Ellsberg is second-order probability (without RCLA), and then the somewhat far-fetched idea that people then treat 1st order probabilities of winning as outcomes and process them concavely, suggesting a kind of second-order-probability risk aversion. Note that this is a special version of the smooth ambiguity model of KMM (2005): it is already the smooth model when there are only two outcomes! (event/utility driven ambiguity model: utility-driven) Whereas the 1979 paper is explicit about expected utility being normative, this paper has a nice discussion on normative implications without ever committing to a normative status of expected utility. paternalism/Humean-view-of-preference; P. 35: “The observation that people’s preference vary with the formulation of problems underscores the need for decision aids to help people make more consistent and rational choices.” P. 36 has this argument that, for example, regret must be accepted if it cannot be avoided: “If man is constructed in such a way that he is much more sensitive to gains and losses than to absolute wealth, then any attempt to maximize human welfare must recognize this fact. More generally, a normative approach to decision must take into account the nature of man as a pleasure machine.” They call certainty effect what in their 1979 version they will call subadditivity. %} Kahneman, Daniel & Amos Tversky (1975) “Value Theory: An Analysis of Choices under Risk,” paper presented at the ISRACON conference on Public Economics, Jerusalem, 1975. Link to paper {% In 2006, this was the second-most cited of all economics papers published between 1970 and 2005 (only White, Econometrica 1980, had more). See Kim, Morse, & Zingales (2006, Table 2). PT: data on probability weighting; Risk averse for gains, risk seeking for losses; inverse-S; real incentives/hypothetical choice, p. 265; P. 276 Eq. 2: contrary to what many think, for prospects with two outcomes, both nonzero, and either both gains or both losses, the value of a prospect xpy is NOT w(p)v(x) + w(p)v(y). P. 276 l. 15: “The evaluation of strictly positive and strictly negative prospects follows a different rule.” What happens is that for such prospects, PT is RDU w.r.t. w for gains, and RDU w.r.t. the dual of w for losses. That is, for x > y > 0 it is w(p)v(x) + (1 w(p))v(y). For losses with x < y < 0 it is also w(p)v(x) + (1 w(p))v(y), meaning it is RDU with reflected w there. paternalism/Humean-view-of-preference: p. 277: The equations of prospect theory retain the general bilinear form that underlies expected utility theory. However, in order to accom[m]odate the effects described in the first part of the paper, we are compelled to assume that values are attached to changes rather than to final states, and that decision weights do not coincide with stated probabilities. These departures from expected utility must lead to normatively unacceptable consequences, such as inconsistencies, intransitivities, and violations of dominance. Such anomalies of preference are normally corrected by the decision maker when he becomes aware that his preferences are inconsistent, intransitive, or inadmissible. In many situations, however, the decision maker lacks the opportunity to discover that his preferences could violate decision rules that he wishes to obey. In these circumstances the anomalies implied by prospect theory are expected to occur.” Here they state that expected utility is normative. Kahneman (2003, American Economic Review, Papers and Proceedings, p. 163) will state the opposite. P. 277 explains reference dependence by saying that the utility function is a book, each page describing it for a difference reference point, which metaphor was also used by Edwards (1962, p. 116) be it not for utility. P. 277: many people erroneously think that, according to prospect theory, preference depends only on the differences of outcomes with the reference point, and not on the reference point otherwise. This is not so. For different reference points the value function and probability weighting function (and loss aversion) can be different. Here is what the authors write: “The emphasis on changes as the carriers of value should not be taken to imply that the value of a particular change is independent of initial position.” But they then point out that the dependence is weak: “However, the preference order of prospects is not greatly altered by small or even moderate variations in asset position. The certainty equivalent of the prospect (1,000, .50), for example, lies between 300 and 400 for most people, in a wide range of asset positions. Consequently, the representation of value as a function in one argument generally provides a satisfactory approximation.” (The last sentence finished on p. 278.) P. 277: decreasing ARA/increasing RRA: it suggests decreasing absolute risk aversion. Pp. 278-279: that utilities are locally nonsmooth. At wealth level where you can buy a house, you have a high marginal utility of money. P. 279 1st para: that concavity of utility for losses is more common than convexity for gains. P. 279: risk seeking for symmetric fifty-fifty gambles. The authors do not think this and speculate that people are highly averse to such risks. P. 281 top: what they call subadditivity in fact is subproportionality. uncertainty amplifies risk (for inverse-S probability weighting): p. 281, lines -6/-5: inverse-S: “In many real-life situations, overestimation and overweighting may both operate to increase the impact of rare events.” This relates to the preference condition in my 2004-Psych. Rev. paper! Similarly, p. 289 l. 5-6: “Consequently, subcertainty should be more pronounced for vague than for clear probabilities.” Pp. 282-283: that small probabilities can be overweighted or ignored. P. 283/284 point out that their theory may violate dominance and say that editing can prevent that, but then indirectly (through transitivity) it can still happen. utility concave near ruin: p. 279 says that utility for losses may have concave regions for large losses, that necessitate changes of life style. Do not explicitly relate it to ruin. P. 288 4th para claims that the extension of prospect theory to many-valued prospects is straightforward, but does not give the formulas and it might not be clear what they had in mind. P. 18 of their 1975 version does give the formula, and it says that for mixed prospect the separate-probability weighting formula of Edwards and others is to be used. This agrees with what most people have ascribed to prospect theory. biseparable utility: unlike what many think, biseparable utility is satisfied by the original prospect theory of this paper when restricted to gains or when restricted to losses. P. 289 l. 1: the text says that DECISION WEIGHTS primarily determine decision weights. This does not say that most of the nonadditivity is generated by cognitive factors, but goes a little bit in that direction. Here is the para: “The decision weight associated with an event will depend primarily on the perceived likelihood of that event, which could be subject to major biases [45] [Their Science 74 paper on heuristics and biases] In addition, decision weights may be affected by other considerations, such as ambiguity or vagueness. The work of Ellsberg [10] and Fellner [12] indeed implies that vagueness reduces decision weights. Consequently, subcertainty should be more pronounced for vague than for clear probabilities.” %} Kahneman, Daniel & Amos Tversky (1979) “Prospect Theory: An Analysis of Decision under Risk,” Econometrica 47, 263–291. {% %} Kahneman, Daniel & Amos Tversky (1979) “Intuitive Prediction: Biases and Corrective Procedures,” TIMS Studies in Management Science 12, 313–327. Reprinted as Ch. 30 in Daniel Kahneman, Paul Slovic, & Amos Tversky (1982, eds.) Judgment under Uncertainty: Heuristics and Biases, Cambridge University Press, Cambridge. {% %} Kahneman, Daniel & Amos Tversky (1982) “The Psychology of Preferences,” Scientific American 246 (1, Jan.), 160–173. {% paternalism/Humean-view-of-preference: p. 124 seems to write: “although errors of judgment are but a method by which some cognitive processes are studied, the method has become a significant part of the message” %} Kahneman, Daniel & Amos Tversky (1982) “On the Study of Statistical Intuitions,” Cognition 11, 123–141. Reprinted as Ch. 34 in Daniel Kahneman, Paul Slovic, & Amos Tversky (1982, eds.) Judgment under Uncertainty: Heuristics and Biases, Cambridge University Press, Cambridge. {% %} Kahneman, Daniel & Amos Tversky (1982) “Variants of Uncertainty,” Cognition 11, 143–157. Reprinted as Ch. 35 in Daniel Kahneman, Paul Slovic, & Amos Tversky (1982, eds.) Judgment under Uncertainty: Heuristics and Biases, Cambridge University Press, Cambridge. {% %} Kahneman, Daniel, & Amos Tversky (1982) “Judgment of and by Representativeness.” In Daniel Kahneman, Paul Slovic, & Amos Tversky (eds.) Judgment under Uncertainty: Heuristics and Biases. Cambridge University Press, Cambridge. {% %} Kahneman, Daniel & Amos Tversky (1982) “The Simulation Heuristic.” In Daniel Kahneman, Paul Slovic, & Amos Tversky (eds.) Judgment under Uncertainty: Heuristics and Biases, 201–208, Cambridge University Press, Cambridge. {% Seem to describe probability weighting function as “psychophysics of chance” on p. 344 %} Kahneman, Daniel & Amos Tversky (1984) “Choices, Values, and Frames,” American Psychologist 39, 341–350. {% %} Kahneman, Daniel, & Amos Tversky (1995) “Conflict Resolution: A Cognitive Perspective.” In Kenneth J. Arrow et al. (eds.) Barriers to Conflict Resolution, Ch. 3. Norton, New York. {% A long list of points on which the authors disagree with Gigerenzer’s critiques. Many are misunderstandings or different wordings. For example, if Gigerenzer criticizes the Linda example for ignoring context and content, he means that the question how likely it is that Linda is a feminist bank teller can be taken by subjects as referring to conditional probability rather than unconditional as it is meant. This is different than K&T use the term. K&T reply here that they tested for this confound, but then, this is less clear, and, … In short, hard to judge for outsiders. P. 582: “Similarly, the role of availability in frequency judgments can be demonstrated by comparing two classes that are equal in objective frequency but differ in the memorability of their instances.” P. 582, about their biases and heuristics: “However, it soon became apparent that “although errors of judgment are but a method by which some cognitive processes are studied, the method has become a significant part of the message” (Kahneman & Tversky, 1982a, p. 124).” P. 589, last sentence of paper, on Gigerenzer’s emphasizing of relative frequencies (reminds me also of the experienced-uncertainty approach of Erev at al.): The view that “both single-case and frequency judgments are explained by learned frequencies (probability cues), albeit by frequencies that relate to different reference classes” (Gigerenzer, 1991, p. 106) appears far too restrictive for a general treatment of judgment under uncertainty. First, this treatment does not apply to events that are unique for the individual and therefore excludes some of the most important evidential and decision problems in people’s lives. Second, it ignores the role of similarity, analogy, association, and causality. There is far more to inductive reasoning and judgment under uncertainty than the retrieval of learned frequencies. %} Kahneman, Daniel, & Amos Tversky (1996) “On the Reality of Cognitive Illusions: A Reply to Gigerenzer’s Critique,” Psychological Review 103, 582–591. {% Seem to write: “As with the fruit fly, we study gambles in the hope that the principles that govern the simple case will extend in recognizable form to complex situations” (p. xi). Lopes (1983) also used the metaphor of what she spelled in one word as a fruitfly. %} Kahneman, Daniel & Amos Tversky (2000, eds.) “Choices, Values, and Frames.” Cambridge University Press, New York. {% %} Kahneman, Daniel, Bernard Tursky, David Shapiro, & Andrew Crider (1969) “Pupillary Heart Rate and Skin Resistance Changes During a Mental Task,” Journal of Experimental Psychology 79, 164–167. {% %} Kahneman, Daniel & Carol A. Varey (1990) “Propensities and Counterfactuals: The Loser that Almost Won,” Journal of Personality and Social Psychology 59, 1101–1110. {% %} Kahneman, Daniel & Carol A. Varey, (1991) “Notes on the Psychology of Utility.” In John Elster & John E. Roemer (eds.) Interpersonal Comparisons of Well-Being. Studies in Rationality and Social Change, 127–163, Cambridge University Press, New York. {% utility = representational?; discounting normative; paternalism/Humean-view-of-preference A good dish is enjoyed three times: when happily anticipating, during the eating itself, and when remembering in complete satisfaction. %} Kahneman, Daniel, Peter P. Wakker, & Rakesh K. Sarin (1997) “Back to Bentham? Explorations of Experienced Utility,” Quarterly Journal of Economics 112, 375–405. Link to paper
Kahneman, Daniel, & Ruth E. Wolman (1970) “Stroboscopic Motion: Effects of Duration and Interval,” Perception and Psychophysics 8, 161–164. {% %} Kahneman, Daniel, & Patricia Wright (1971) “Changes in Pupil Size and Rehearsal Strategies in a Short-Term Memory Task,” Quarterly Journal of Experimental Psychology 23, 187–196. {% Mentions many applications of CEU (Choquet expected utility). %} Kaivanto, Kim (2000) “Endogenously Non-Additive Aggregate Probabilities: Syndicate Surrogate Functions and Composite Market Beliefs,” Warwick. {% Nicely describes neo-additive as linear-with-boundary-discontinuity. %} Kaivanto, Kim (2011) “Optimal Cutoff Threshold Placement in Signal Detection Theory under Cumulative Prospect Theory,” Warwick. {% Signal detection theory (“is this email genuine or malignent”) is reanalyzed using PT. Decentralized behavioral decisionmakers are biased toward underdetection, and system-level risk is consequently greater than in analyses predicated upon normative rationality. %} Kaivanto, Kim (2014) “The Effect of Decentralized Behavioral Decision Making on System-Level Risk,” Risk Analysis 34, 2121–2142. {% Nicely points out that St. Petersburg paradox very crucially depends on RCL, and on gamblers fallacy of people, after some tails, wrongly thinking that now heads must become more likely. %} Kaivanto, Kim & Eike B. Kroll (2012) “Alternative Bias and Reduction in St. Petersburg Gambles: An Experimental Investigation,” Lancaster University, Lancaster, UK. {% %} Kajii, Atsushi (1997) “On the Role of Options in Sunspot Equilibria,” Econometrica 65, 977–986. {% Consider forms of additivity between full-force and comonotonic additivity, and characterize various special cases of the Choquet integral. %} Kajii, Atsushi, Hiroyuki Kojima, & Takashi Uic (2007) “Cominimum Additive Operators,” Journal of Mathematical Economics 43, 218–230. {% %} Kajii, Atsushi & Stephen Morris (1997) “Common p-Belief: The General Case,”Games and Economic Behavior 18, 73–82. {% %} Kalai, Ehud & Meir Smorodinsky (1975) “Other Solutions to Nash’s Bargaining Problem,” Econometrica 43, 513–518. {% %} Kalai, Ehud & David Schmeidler (1977) “Aggregation Procedure for Cardinal Preferences: A Formulation and Proof of Samuelson’s Impossibility Conjecture,” Econometrica 45, 431–438. {% revealed preference; They consider choice functions that cannot be represented by one preference relation, but by a number r of preference relations. Present some numerical results, such as limiting and maxmin, on r. %} Kalai, Gil, Ariel Rubinstein, & Ran Spiegler (2002) “Rationalizing Choice Functions by Multiple Rationales,” Econometrica 70, 2481–2488. {% ranking economists %} Kalaitzidakis, Pantelis, Theofanis P. Mamuneas, & Thanasis Stengos (2003) “Rankings of Academic Journals and Institutions in Economics,” Journal of the European Economic Association 1, 1346–1366. {% HYE %} Kalant, Norman. (1991) “Ionic versus Nonionic Contrast Media: A Burden or a Bargain?,” Can. Med. Assoc. J. 144, 123–124. {% information aversion; of people with possibly Huntington’s disease, only 5% take the test! %} Kalb, Claudia (2006) “Healt for Life; Peering into the Future,” Newsweek December 11, 2006, 46–52. {% Experiment plus desire to link individual and group behavior. PT falsified: risk seeking for symmetric fifty-fifty gambles: they seem to find it. %} Kameda, Tatsuya & James H. Davis (1990) “The Function of the Reference Point in Individual and Group Risk Decision Making,” Organizational Behavior and Human Decision Processes 46, 55–76. {% Imagine Bayesian B1 can choose which signal to be revealed to another Bayesian B2, wanting to manipulate the latter. If this desire is common knowledge, can B1 still manipulate? The paper answers affirmatively. The signal can make B2’s preferred action, disfavorable to B1, more favorable in situations where it will be chosen anyhow, but make it more unfavorable in situations where this does change the choice. Concavity/convexity of utility also plays a role. I did not read the paper enough to see if meta-info considerations can play a role, with B2 guessing there may be signals making him go the other way but not revealed to him. %} Kamenica, Emir & Matthew Gentzkow (2011) “Bayesian Persuasion,” American Economic Review 101, 2590–2615. {% information aversion?? Games with incompete information, value of information %} Kamien, Morton I., Yair Tauman, & Shmuel Zamir (1979) “On the Value of Information in a Strategic Conflict.” {% foundations of probability %} Kamlah, Andreas (1983) “Probability as a Quasi-Theoretical Concept—J.V. Kries’ Sophisticated Account after a Century,” Erkenntnis 19, 239–251. {% Reviewed in JMP 34, 336-363, by Harold P. Lehmann, extensively and nicely %} Kanal, Laveen N. & John F. Lemmer (1986) “Uncertainty in Artificial Intelligence; Machine Intelligence and Pattern Recognition, Vol.4.” North-Holland, Amsterdam. {% %} Kanal, Laveen N., Todd S. Levitt, & John F. Lemmer (1989) “Uncertainty in Artificial Intelligence 3; Machine Intelligence and Pattern Recognition, Vol.5.” North-Holland, Amsterdam. {% Differences in optimal income taxation if analyzed using prospect theory iso EU. %} Kanbur, Ravi, Jukka Pirttilä, & Matti Tuomala (2008) “Moral Hazard, Income Taxation and Prospect Theory,” Scandinavian Journal of Economics 110, 321–337. {% decreasing ARA/increasing RRA: seem to give thought experiment criticizing constant RRA %} Kandel, Shmuel & Robert F. Stambaugh (1991) “Asset Returns and Intertemporal Preferences,” Journal of Monetary Economics 27, 39–71. {% %} Kaneko, Mamoru (1980) “An Extension of the Nash Bargaining Problem and the Nash Social Welfare Function,” Theory and Decision 12, 135–148. {% %} Kaneko, Mamoru (1994) “Axiomatic Considerations of Nash Equilibrium.” {% %} Kaneko, Mamoru & Takashi Nagashima (1988) “Players’ Deductions and Deductive Knowledge on Theorems,” Hitotsubashi University, Kunitachi, Tokyo 186(?? of in Blacksburg?), E88-02-01. {% completeness-criticisms: seems to give that %} Kannai, Yakkar (1963) “Existence of a Utility in Infinite Dimensional Partially Ordered Spaces,” Israel Journal of Mathematics 1, 229–234. {% %} Kannai, Yakkar (1977) “Concavifiability and Constructions of Concave Utility Functions,” Journal of Mathematical Economics 4, 1–56. {% %} Kannai, Yakkar (1981) “Concave Utility Functions, Existence, Constructions and Cardinality.” In Siegfried Schaible & William T. Ziemba (eds.) Generalized Concavity in Optimization and Economics, 543–611, Academic Press, New York. {% Seems to be a well known paper on total absence of information. ordering of subsets; show that a betweenness axiom for average-utility representation and the additivity axiom (called monotonicity) for qualitative probability are incompatible on sets of 5 or more elements. %} Kannai, Yakkar & Bezalel Peleg (1984) “A Note on the Extension of an Order on a Set to the Power Set,” Journal of Economic Theory 32, 172–175. {% conservation of influence: Seems to open with: “All of nature, as far as it is within the reach of his power, is subjected to the will of man, with the exception of other men and reasonable beings. From the point of view of reason, the things in nature can only be regarded as means to ends, but man alone can himself be regarded as an end. … Animals, as well [as unreasonable things], have no value in themselves, since they have no consciousness of their existence – man is the purpose of creation; nevertheless, he can also be used as a means by other reasonable beings. However, man is never merely a means; rather he is at the same time an end. For example: If a mason serves me as a means to building a house, I serve him, in turn, as a means to acquire money. … The world, as a system of ends, finally has to contain a purpose, and this is the reasonable being. If there existed no end, the means would serve no purpose and would have no value. — Man is an end. It is therefore contradictory that he should be a mere means. — If I am making a contract with a servant, he has to be an end as well, just as I am, and not merely a means.” %} Kant, Immanuel (1785/ 2002). “Groundwork for the Metaphysics of Morals.” Translated into English by Allen Wood. Yale University Press, New Haven, CT. {% free-will/determinism: seems to have written that you have to act under the presupposition, even if illusion, of free will. Seems to have written on free will being only our imagination: “Daher ist Freiheit nur eine Idee der Vernunft, deren objekive Realität in sich zweifelhaft ist, Natur aber ein Verstandesbegriff, der seine Realität an Beispielen der Erfahrung beweiset und notwendig beweisen muss.” Translation: [Therefore freedom is only an idea of “Vernunft,” whose intrinsic objective reality is questionable, nature however is a concept of “Verstand,” which proves, and necessarily has to prove, its reality by examples of experience.] Here Vernunft and Verstand are two different terms for rationality with subtle differences, Verstand being more practically oriented. %} Kant, Immanuel (1961) “Grundlegung zur Metaphysik der Sitten;” edn. of 1961. Reclam, Sittingen, Germany. {% foundations of probability. %} Kaplan, Mark (2010) “In Defense of Modest Probabilism,” Synthese 176, 41–55. {% %} Kaplan, Robert M. (1993) “Quality of Life Assessment for Cost/Utility Studies in Cancer,” Cancer Treat. Rev. 19 suppl A, 85–93. {% foundations of probability; %} Kaplan, Stan (1988) “Will the Real Probability Please Stand Up?,” Reliability Engineering and System Safety 23, 285–292. {% By measuring how much people are willing to pay for reducing mortality risk, the income elasticity of the value of a statistical life can be measured. Note here how utility is measured through probability of survival = 1 mortality risk, very similar through the modeling of utility through the probability of gaining a prize (Roth & Malouf 1979). The income elasticity of statistical life must then also be 1 power of utility of income; i.e., the RRA index of the utility function of income. .Income elasticities of statistical lives typically found in the literature ranges around 0.5. The author now only refers to RRA indexes found in finance and macroeconomics, which are around 2, and considers the discrepancy a paradox. However, in individual choice experiments in laboratories, RRA indexes of 0.5 are typically found, and the paradox is resolved! %} Kaplow, Louis (2005) “The Value of a Statistical Life and the Coefficient of Relative Risk Aversion,” Journal of Risk and Uncertainty 31, 23–34. {% equity-versus-efficiency: if criteria other than individual utility, such as equity, are considered, then sometimes some of individual utility must be sacrificed to equity. By reshifting and continuity this can lead to a situation where, for equity considerations, all individuals sacrifice some utility, which violates the Pareto principle defined in a narrow sense. %} Kaplow, Louis & Steven Shavell (2001) “Any Non-Welfarist Method of Policy Assessment Violates the Pareto Principle,” Journal of Political Economy 109, 281–286. {% equity-versus-efficiency %} Kaplow, Lowis & Steven Shavell (2002) “Fairness versus Welfare.” Harvard University Press, Cambridge. {% risky utility u = transform of strength of preference v: in §1 the authors adopt the assumption that intertemporal utility Z(.) is a composition W(U(.)), with U a risky vNM utility and W something like a welfare function. It is remiscent of the Dyer-Sarin risky-riskless utility difference, although the authors do not cite this strand of literature but work from scratch. The authors blame other authors who use different models, such as the cynical “in excellent company” on p. 126 middle. Then there follow many discussions of the chosen composition, again criticizing everyone who did it differently. %} Kaplow, Louis & David Weisbach (2011) “Discount Rates, Social Judgments, Individuals’ Risk Preferences, and Uncertainty,” Journal of Risk and Uncertainty 42, 125–143. {% %} Kapteyn, Arie (1985) “Utility and Economics,” De Economist 133, 1–20. {% questionnaire versus choice utility Abstract. Since the work of Pollak and Wales (1979), it is well known that demand data are insufficient to identify a household cost function. Hence, additional information is required. For that purpose I propose to employ direct measurement of feelings of well-being, elicited in surveys. In the paper I formally establish the connection between subjective measures and the cost function underlying the AID system. The subjective measures fully identify cost functions and the expenditure data do this partly. This makes it possible to test the null hypothesis that both types of data are consistent with one another; i.e., that they measure the same thing. I use two separate data sets to set up a test of this equivalence. The outcomes are somewhat mixed and indicate the need for further specification search. Finally, I discuss some implications of the outcomes. %} Kapteyn, Arie (1994) “The Measurement of Household Cost Functions: Revealed Preference versus Subjective Measures,” Journal of Population Economics 7, 333–350. {% dominance violation by pref. for increasing income; Use panel data, so no real incentives and hypothetical choice, to do an alternative to Barsky et al. (1997). Model with habit formation suggests more utility curvature than without (so additive separability over time). P. C147: under assumption of intertemporal separability, they find power (= 1 relative-risk-aversion index) of about 0.94, and if they allow for violation of intertemporal separability then they get 3.8 (p. C150 Tables 3 and 4, where = 1 power and they give ln()) intertemporal separability criticized: p. C151: “The main finding of our empirical analysis may be the rejection of intertemporal additivity.” %} Kapteyn, Arie & Federica Teppa (2003) “Hypothetical Intertemporal Consumption Choices,” Economic Journal 113, C140–C152. {% %} Kapteyn, Arie & Tom J. Wansbeek (1982) “Empirical Evidence on Preference Formation,” Journal of Economic Psychology 3, 137–154. {% %} Kapteyn, Arie & Tom J. Wansbeek (1985) “The Individual Welfare Function,” Journal of Economic Psychology 6, 333–363. {% risky utility u = strength of preference v (or other riskless cardinal utility, often called value) %} Kapteyn, Arie & Tom J. Wansbeek (1985) “The Individual Welfare Function, A Rejoinder,” Journal of Economic Psychology 6, 375–381. {% %} Kapteyn, Arie, Tom J. Wansbeek, & Jeannine Buyze (1980) “The Dynamics of Preference Formation,” Journal of Economic Behavior and Organization 1, 123–157. {% %} Kareev, Yaakov (1992) “Not That Bad after All: Generation of Random Sequences,” Journal of Experimental Psychology: Human Perception and Performance 18, 1189–1194. {% %} Karlsson, Goran & Magnus Johannesson (1996) “The Decision Rules of Cost-Effectiveness Analysis,” PharmacoEconomics 9, 113–120. {% inverse-S %} Karmarkar, Uday S. (1974) “The Effect of Probabilities on the Subjective Evaluation of Lotteries,” Working paper No. 698–74, MIT. {% utility elicitation: p. 65 points out that utility curve, elicited under EU calculation, depends on probability used. (This was posed as a research question by Swalm 1966, p. 134 last para.) Karmarkar (1974) describes the experiment. inverse-S: underprocessing of information versus overprocessing of information (latter if it would not be inverse-S but regular S) %} Karmarkar, Uday S. (1978) “Subjectively Weighted Utility: A Descriptive Extension of the Expected Utility Model,” Organizational Behavior and Human Performance 21, 61–72. {% %} Karmarkar, Uday S. (1979) “Subjectively Weighted Utility and the Allais Paradox,” Organizational Behavior and Human Performance 24, 67–72. {% state-dependent utility %} Karni, Edi (1983) “Risk Aversion for State-Dependent Utility Functions: Measurement and Applications,” International Economic Review 24, 637–647. {% state-dependent utility %} Karni, Edi (1985) “Decision-Making under Uncertainty: The Case of State-Dependent Preferences.” Harvard University Press, Cambridge, MA. {% state-dependent utility %} Karni, Edi (1987) “Generalized Expected Utility Analysis of Risk Aversion with State-Dependent Preferences,” International Economic Review 28, 229–240. {% %} Karni, Edi (1989) “Generalized Expected Utility Analysis of Multivariate Risk Aversion,” International Economic Review 30, 297–305. {% state-dependent utility %} Karni, Edi (1992) “Subjective Probabilities and Utility with State-Dependent Preferences,” Journal of Risk and Uncertainty 5, 107–125. {% state-dependent utility; utility depends on probability %} Karni, Edi (1992) “Utility Theory with Probability Dependent Outcome Valuation,” Journal of Economic Theory 57, 111–124. {% Does it for AA; state-dependent utility %} Karni, Edi (1993) “A Definition of Subjective Probabilities with State-Dependent Preferences,” Econometrica 61, 187–198. {% state-dependent utility; does it for Savage %} Karni, Edi (1993) “Subjective Expected Utility with State-Dependent Preferences,” Journal of Economic Theory 60, 428–438. {% state-dependent utility %} Karni, Edi (1996) “Probabilities and Beliefs,” Journal of Risk and Uncertainty 13, 249–262. {% state-dependent utility: assumes in Harsanyi-style model that best and worst state of each agent have the same utility, and, thus, can compare utility units. The importance weights that can now be derived, should all be the same under impartiality. The probability, under the veil of ignorance, of being some future individual is not objectively given, but is to be inferred as subjective from the social planner’s preferences. %} Karni, Edi (1998) “Impartiality: Definition and Representation,” Econometrica 66, 1405–1415. {% Assumes bounded state-dependent utility. Utility is then normalized, it is assumed that the range of utility is the same across different states of nature. That is, extreme outcomes have state-independent utility. They can then be used to elicit probability. P. 482: “This definition of subjective probability involves a convention, namely, the normalization of the event-dependent utility functions … so that their least upper bounds and the largest lower bounds coincide.” %} Karni, Edi (1999) “Elicitation of Subjective Probabilities when Preferences Are State-Dependent,” International Economic Review 40, 479–486. {% Tradeoff method is used for theoretical purposes, in variation of Karni, Schmeidler, & Vind. %} Karni, Edi (2003) “On the Representation of Beliefs by Probabilities,” Journal of Risk and Uncertainty 26, 17–38. {% %} Karni, Edi (2003) “Impartiality and Interpersonal Comparisons of Variations in Well-Being,” Social Choice and Welfare 21, 95–111. {% criticisms of Savage’s basic model: people usually follow Savage routinely in taking states-consequences-acts as he does, and don’t seem to be aware that there is quite some arbitrariness in it, first, in how we define what as function of what mathematically, but second, to what extent things are independent from each other causally. I like Luce’s work in the sense that he models these things in a provocatively different way. Karni also challenges these foundational aspects. The present paper makes things tangible because it does not just say things, but it formalizes and axiomatizes. The primary point of the paper is, therefore, for me that it brings new and different insights into the primitives of decision under uncertainty. Given each action, there is a traditional model with effects playing much the role of states of nature, not influenced by what the decision maker does (given the action chosen!). At the same time, there is place for influence of the decision maker on resolutions of uncertainty, and this is through the influence of actions on the effects. Accordingly, effects can also carry value, and not only be sources of uncertainty. This is clear by the general model plus a specification where they “happen” not to carry value.) %} Karni, Edi (2006) “Subjective Expected Utility Theory without States of the World,” Journal of Mathematical Economics 42, 325–342. {% %} Karni, Edi (2007) “Foundations of Bayesian theory,” Journal of Economic Theory 132, 167–188. {% %} Karni, Edi (2007) “Archimedean and Continuity,” Mathematical Social Sciences 53, 332–334. {% Action-dependence and effect-dependence are used to avoid the use of states of nature. %} Karni, Edi (2007) “A New Approach to Modeling Decision-Making under Uncertainty, Economic Theory 33, 225–242. {% Karni’s action-dependent theory is used to analyze the principal-agent problem and the common prior assumption. %} Karni, Edi (2008) “Agency Theory: Choice-Based Foundations of the Parametrized Distribution Formulation,” Economic Theory 36, 337–351. {% For an event E, the well-known matching probability p is defined through 100E0 ~ 100p0. This paper discusses the well-known Becker-DeGroot-Marschak method for eliciting this p. Karni & Safra (1987) discussed the general BDM (Becker-DeGroot-Marschak) mechanism too from a theoretical perspective. §30.5 of Holt (2007) used BDM to elicit matching probabilities as recommended by this paper, and did experiments with it. %} Karni, Edi (2009) “A Mechanism Design for Probability Elicitation,” Econometrica 77, 603–606. {% Maxmin expected utility is applied to a principal-agent situation. %} Karni, Edi (2009) “A Reformulation of the Maxmin Expected Utility Model with Application to Agency Theory,” Journal of Mathematical Economics 45, 97–112. {% The action-dependent model of the author is applied with medical interpretations. Interestingly, the model could be taken as axiomatization of willingness to pay for health. %} Karni, Edi (2009) “A Theory of Medical Decision Making under Uncertainty,” Journal of Risk and Uncertainty 39, 1–16. {% His model has bets which are a sort of side payments. This makes it possible to measure and axiomatize all kinds of dependencies that cannot be so in classical models, such as act-dependent probabilities and dependence of decisions on information set. Tradeoff method: used theoretically. %} Karni, Edi (2011) “A Theory of Bayesian Decision Making with Action-Dependent Subjective Probabilities,” Economic Theory 48, 125–146. {% Assume transitivity and nontriviality throughout. Schmeidler (1971) showed, for connected topological spaces, that continuity (both for open and closed sets) implies completeness. Dubra (2011) & Galaabaatar (2010) showed similar results in the vNM EU context. This paper does so too, combining all the above, and showing that it matters much if and how one take weak or strict preference as primitive. It also gives new results on indifference versus incomparability. %} Karni, Edi (2011) “Continuity, Completeness and the Definition of Weak Preferences,” Mathematical Social Sciences 62, 123–125. {% %} Karni, Edi (2011) “Subjective Probabilities on a State Space,” American Economic Journal: Microeconomics 3, 172–185. {% Generalizes his 2011 ET paper by incorporating effect-dependent risk attitudes that can also depend on their actions. Tradeoff method: used theoretically. %} Karni, Edi (2013) “Bayesian Decision Making with Action-Dependent Probabilities and Risk Attitudes,” Economic Theory 53, 335–356. {% Uses the Anscombe-Aumann model, studying conditional incompletenesses, where familiar events have conditional completeness. Also considers sources of events, citing Chew & Sagi (2008). %} Karni, Edi (2014) “Familiarity Breeds Completeness,” Economic Theory 56, 109–124. {% survey on nonEU %} Karni, Edi, Fabio Maccheroni, & Massimo Marinacci (2014) “Ambiguity and Nonexpected Utility.” In Peyton H. Young & Shmuel Zamir (Eds.) Handbook of Game Theory, Vol. 4, North-Holland, Amsterdam. {% %} Karni, Edi & Mark J. Machina (1987) “Multivariate Risk Aversion for Nonexpected Utility Preferences,” Working paper no. 185, The Johns Hopkins University, Department of Political Economy. {% %} Karni, Edi & Philippe Mongin (1997) “On the Determination of Subjective Probability by Choices,” {% End shows that for BDM (Becker-DeGroot-Marschak), for every nonEU there exists a lottery where BDM does not give right certainty equivalent if subject does RCLA. %} Karni, Edi & Zvi Safra (1987) “Preference Reversal and the Observability of Preferences by Experimental Methods,” Econometrica 55, 675–685. {% dynamic consistency; nicely described in Epstein (1992, p. 51); according to Karni & Schmeidler (1991, p. 407), they assume RCLA and forgone-branch independence (often called consequentialism) implicitly. %} Karni, Edi & Zvi Safra (1989) “Dynamic Consistency, Revelations in Auctions and the Structure of Preferences,” Review of Economic Studies 56, 421–434. {% dynamic consistency %} Karni, Edi & Zvi Safra (1989) “Ascending Bid Auctions with Behaviorally Consistent Bidders,” Annals of Operations Research 19, 435–446. {% dynamic consistency: favors abandoning time consistency, so, favors sophisticated choice (what they call behavioral consistency); Best ref. for defense sophisticated choice. %} Karni, Edi & Zvi Safra (1990) “Behaviorally Consistent Optimal Stopping Rules,” Journal of Economic Theory 51, 391–402. {% inverse-S %} Karni, Edi & Zvi Safra (1990) “Rank-Dependent Probabilities,” Economic Journal 100, 487–495. {% dynamic consistency; introduction strongly suggests that they consider “behavioral consistency” (which is sophisticated behavior) to satisfy dynamic consistency. They use DC (dynamic consistency) in a weak sense. Behavioral consistency entails forgone-branch independence, time neutrality, weak DC, RCLA, and violates strong DC; i.e., DC à la Machina. %} Karni, Edi & Zvi Safra (1994) “Unbounded Behaviorally Consistent Stopping Rules,” Journal of Risk and Uncertainty 9, 231–238. {% %} Karni, Edi & Zvi Safra (1995) “The Impossibility of Experimental Elicitation of Subjective Probabilities,” Theory and Decision 38, 313–320. {% %} Karni, Edi & Zvi Safra (1998) “The Hexagon Condition and Additive Representation for Two Dimensions: An Algebraic Approach,” Journal of Mathematical Psychology 42, 393–399. {% A theorem reminiscent of Karni, Schmeidler, & Vind (1983), state-dependent expected utility, with conceivable every probability distribution over the state space. %} Karni, Edi & Zvi Safra (2000) “An Extension of a Theorem of von Neumann and Morgenstern with an Application to Social Choice Theory,” Journal of Mathematical Economics 34, 315–327. {% %} Karni, Edi & Zvi Safra (2002) “Individual Sense of Justice: A Utility Representation,” Econometrica 70, 263–284. {% %} Karni, Edi & Zvi Safra (2008) “Moral Sentiments and Social Choice,” Social Choice and Welfare 30, 427–446. {% is a set of states of mnd , and for every , is a preference relation. Acts are in the Anscombe-Aumann model. Preferences are over menus; i.e., subsets of acts. An act induced by a menu assigns to each the best act from the menu accordint to . Acts induced by menus are evaluated by having a subjective probability on and then take the probability-weighted average EU given each , where the EU is dependent. Preferences over hypothetical acts are involved where acts conditioned on different moods are compared, where the authors take them as hypothetical and not revealed-preference based. The model is related to many random-choice models and menu-models in the literature. The paper extends many results of Karni & Schmeidler (1980, working paper) and Karni (1985), linking those to modern models. %} Karni, Edi & Zvi Safra (2016) “A Theory of Stochastic Choice under Uncertainty,” Journal of Mathematical Economics 63, 164–173. {% Test preference for fairness if it concerns probabilistic fairness. %} Karni, Edi, Tim Salmon, Barry Sopher (2008) “Individual Sense of Fairness: An Experimental Study,” Experimental Economics 11, 174–189. {% utility depends on probability %} Karni, Edi & Edward E. Schlee (1995) “Utility Theory with Probability-Dependent Outcome Valuations: Extensions and Applications,” Journal of Risk and Uncertainty 10, 127–142. {% SIIA/IIIA; revealed preference %} Karni, Edi & David Schmeidler (1976) “Independence of Nonfeasible Alternatives, and Independence of Nonoptimal Alternatives,” Journal of Economic Theory 12, 488–493. {% %} Karni, Edi & David Schmeidler (1990) “Fixed Preferences and Changing Tastes,” American Economic Review, Papers and Proceedings 80, 262–267. {% survey on nonEU %} Karni, Edi & David Schmeidler (1991) “Utility Theory with Uncertainty.” In Werner Hildenbrand & Hugo F. Sonnenschein (eds.) Handbook of Mathematical Economics 4, Ch. 33, 1763–1831, North-Holland, Amsterdam. {% dynamic consistency; see Alias-literature %} Karni, Edi & David Schmeidler (1991) “Atemporal Dynamic Consistency and Expected Utility Theory,” Journal of Economic Theory 54, 401–408. {% Savage model, only there is a finite partition of S, and P4 holds only within each element of the partition. %} Karni, Edi & David Schmeidler (1993) “On the Uniqueness of Subjective Probabilities,” Economic Theory 3, 267–277. {% DOI 10.1007/s11238-016-9545-0 state-dependent utility In AA framework, preferences over acts and state-prize lotteries, both maximizing vNM EU, and monotonicity are assumed. This is necessary and sufficient for state-dependent EU, with P unique up to states with trivial state-prize preferences. This is similar to Arrow (1951 pp. 431-432). %} Karni, Edi & David Schmeidler (2016) “An Expected Utility Theory for State-Dependent Preferences,” Theory and Decision 81, 467–478. {% state-dependent utility %} Karni, Edi, David Schmeidler & Karl Vind (1983) “On State Dependent Preferences and Subjective Probabilities,” Econometrica 51, 1021–1031. {% criticisms of Savage’s basic model; Acording to the traditional Bayesian model, every new observation is a subset of the universal state space, which shrinks and shrinks. In this paper, new observations enlarge the state space and open new possibilities not thought of before. Hence the nice title. They give an axiomatization. They do not use the usual Savage model where states and consequences are given as primitives, but take acts and consequences as primitives, and then all states are all maps from acts to consequences (à la Schmeidler & Wakker 1987). Thus, discovering new outcomes or new acts enlarges the state space. It can be taken to model unforeseen events or unawareness. They use the Anscombe-Aumann model. An invariance axiom (awareness consistency) ensures that expanding the model does not affect the preferences already there. %} Karni, Edi & Marie-Louise Vierø (2013) “Reverse Bayesianism: A Choice-Based Theory of Growing Awareness,” American Economic Review 103, 2790–2810. {% Generalize their 2013 AER paper from EU to probabilistic sophistication, while, in particular, maintaining the updating results. %} Karni, Edi & Marie-Louise Vierø (2015) “Probabilistic Sophistication and Reverse Bayesianism,” Journal of Risk and Uncertainty 50, 189–208. {% Use the reverse Bayesianism approach and get preference-based utility of, for instance, unimaginable or even nonexisting outcomes. %} Karni, Edi & Marie-Louise Vierø (2017) “Awareness of Unawareness: A Theory of Decision Making in the Face of Ignorance,” Journal of Economic Theory 168, 301–328. {% Harsanyi’s aggregation %} Karni, Edi & John A. Weymark (1996) “An Informationally Parsimonious Impartial Observer Theorem.” {% information aversion: higher anxiety seems to give lower compliance with self-examination guidelines in woman with a family history of breast cancer. (decision under stress) %} Kash, Kathryn M., Jimmy C. Holland, Marilyn S. Halper, & Daniel G. Miller (1992) “Psychological Distress and Surveillance Behaviors of Women with a Family History of Breast Cancer,” Journal of the National Cancer Institute 84, 24–30. {% %} Kass, Robert E. & Adrian E. Rafferty (1995) “Bayesian Factors,” Journal of the American Statistical Association 90, 773–795. {% For loss aversion, Peeters & Czapinski (1990) and others discussed whether people really suffer more under losses than they are happy under gains, or whether this is not so but people still overweight losses, and tested it using introspective measurements. This paper does the same for discounting, whether people (think they) feel less in the future (“anhedonia”), or feel the same but still weigh the future less. The novelty is not in putting up this question, unlike the suggestion in the abstract, because the authors give many references, but it is in testing it. So the authors conjecture that people underestimate future feelings. In other studies they have investigated the “impact bias,” claiming that people overestimate future effects. Footnote 1 on p. 1534 explains that these are “fully consistent” because we may be overestimating future effects but, simply, be overestimating all present effects even more. Experiment 1b tries to demonstrate anhedonia by seeing if WTP in the fuure will be smaller than now. I wonder if WTP in the future is not also subject to anhedonia. In experiment 2a the authors show that not all subjects are completely driven by one bias, which however does not show that the bias would be completely absent. %} Kassam, Karim S., Daniel T. Gilbert, Andrew Boston, & Timothy D. Wilson (2008) “Future Anhedonia and Time Discounting,” Journal of Experimental Social Psychology 44, 1533–1537. {% Dutch book; Consider a version of book making between regular book making and comonotonic book making, where comonotonicity is imposed on the acts of one side of the book but not the other. The condition is necessary and sufficient for Choquet expected utility with linear utility and a convex capacity. It is the linear-in-payment analogue of the linear-in-probabilistic-mixing results of Wakker (1990, Journal of Economic Theory). %} Kast, Robert & André Lapied (2003) “Comonotonic Book Making and Attitudes to Uncertainty,” Mathematical Social Sciences 46, 1–7. {% dynamic consistency; Do what title says, for uncertainty (not risk). Do CEU (Choquet expected utility) with linear utility, DC (dynamic consistency) with violation of weak consequentialism (forgone-event independence), has updating of weighting functions. P. 32 bottom: one can consider discounted expectation or expected discounting. %} Kast, Robert & André Lapied (2010) “Valuing Future Cash Flows with Non-Separable Discount Factors and Non-Additive Subjective Measures: Conditional Choquet Capacities on Time and on Uncertainty,” Theory and Decision 69, 27–53. {% Reviews and compares the performance of several optimization theories and several heuristics in several contexts, depending on information available and so on. Pleas for a mixed use of both approaches. %} Katsikopoulos, Konstantinos V. (2011) “Psychological Heuristics for Making Inferences: Definition, Performance, and the Emerging Theory and Practice,” Decision Analysis 8, 10–29. {% %} Katz, Leonard (1964) “Effects of Differential Monetary Gain and Loss on Sequential Two-Choice Behavior,” Journal of Experimental Psychology 68, 245–249. {% %} Katzenstein, Herbert & William S. Sachs (1992) “Direct Marketing;” 2nd edn. New York: MacMillan. {% %} Katzner, Donald W. (1970) “Static Demand Theory.” MacMillan, London. {% %} Kauder, Emil (1965) “A History of Marginal Utility Theory.” Princeton University Press, Princeton, NJ. {% Necessary and sufficient condition for stochastic maximization of utility, being SARSP, strong axioms of revealed stochastic preference. %} Kawaguchi, Kohei (2017) “Testing Rationality without Restricting Heterogeneity,” Journal of Econometrics 197, 153–171. {% N = 25,000 subjects aged 18 to 79. Onhline survey; hypothetical. They measured what they call loss aversion through the following Samuelson-colleague-type question: “Suppose that, if you invested 100,000 yen, you would either get a capital gain of 20,000 yen or a capital loss of 10,000 yen at a 50% probability. What would you do?” Here 100 Yen is about €1. 78.6% replied that they would not invest and 21.4% that they would. %} Kawamura, Noriaki for Central Council for Finanical Services Information (2016) “Financial Lieracy Survey,” Public Relations Department, Bank of Japan; working paper. {% This paper, and many others in this issue of this journal, devoted to use of probabilistic evidence in jurisdiction %} Kaye, David H. & Jonathan J. Koehler (1991) “Can Jurors Understand Probabilistic Evidence?,” Journal of the Royal Statistical Society (Series A) 154, 75–81. {% free-will/determinism: beginning nicely summarized different views. The author argues for being agnostic on it. %} Kearns, Stephen (2015) “Free Will Agnosticism,” Nous 49, 235–252. {% information aversion %} Keasy, Kevin (1984) “Regret Theory and Information: A Note,” Economic Journal 94, 645–648. {% N = 240 subjects. Did individual decisions under ambiguity, decisions after discussions, and group decisions. The interactions with others generated moves in the direction of ambiguity neutrality, which can be interpreted as moves towards rationality. Certainty equivalents were obtained for binary gambles, with degrees of ambiguity manipulated by providing probability intervals. The actual compositions were determined by randomly and uniformly drawing the probabilities from the intervals, which is the same as having the midpoint of the interval as objective probability. But subjects were not told this, and were only told that the true composition was “determined by chance” (p. 63). They used random incentive system for real payment. P. 63 explains that they did not really control for suspicion other than tell subjects that the compositions of the ambiguous urns had really been determined by chance (which had not been specified further), and citing two references that it should be no problem. P. 64 Table 3 gives the data with average CEs for all the Bayesian-probability (interval-midpoints) levels used: p = 0.20, 0.50, 0.80, with also some risky choices at p = 0.35 and p = 0.65. As the seize of the interval increases, so does ambiguity. Decreasing CEs as ambiguity increases (so ambiguity aversion) happens mostly at p = 0.5, but maybe rather than looking at those absolutely we should look at them relatively to risk premium. For p = 0.20 it is close to ambiguity neutrality, more than for others, but things are not very clear or pronounced (ambiguity seeking for unlikely). Table 5 gives similar things with percentages of subjects/groups being ambiguity averse/seeking. %} Keck, Steffen, Enrico Diecidue, & David V. Budescu (2014) “Group Decisions under Ambiguity: Convergence to Neutrality,” Journal of Economic Behavior and Organization 103, 60–71. {% Z&Z, time preference; classical reference to argue that discounting for costs should be the same as for benefits, the “Keeler-Cretin paradox” %} Keeler, Emmett B. & Shan Cretin (1983) “Discounting of Life-Saving and Other Nonmonetary Effects,” Management Science 29, 300–306. {% Z&Z %} Keeler, Emmett B., Daniel T. Morrow, & Joseph P. Newhouse (1977) “The Demand for Supplementary Health Insurance, or Do Deductibles Matter?,” Journal of Political Economy 85, 789–801. {% Z&Z %} Keeler, Emmett B., Joseph P. Newhouse, & Charles E. Phelps (1977) “Deductibles and Demand for Medical Care Services: The Theory of a Consumer Facing a Variable Price Schedule under Uncertainty,” Econometrica 45, 641–655. {% Kimball showed that v is more prudent than u if the derivative v´ is a transform of u´ with positive second derivative (so convex). This paper shows that v is more downside risk averse than u iff v itself is a transform of u itself that has positive third derivative. %} Keenan, Donald C. & Arthur Snow (2009) “Greater Downside Risk Aversion in the Large,” Journal of Economic Theory 144, 1092–1101. {% U´´´/U´ (3/2)(U´´/U´)2, previously shown to be a good index of aversion to downside risk, has been known before in the maths literature as the Schwarzian derivative. It is discussed in this paper. %} Keenan, Donald C. & Arthur Snow (2012) “The Schwarzian Derivative as a Ranking of Downside Risk Aversion,” Journal of Risk and Uncertainty 44, 149–160. {% Seems to have argued that failures of independence indicate poor structuring of the attributes. Parnell et al. (2013) review papers resulting from Keeney’s book. %} Keeney, Ralph L. (1992) “Value-Focused Thinking.” Harvard University Press, Cambridge, MA. {% Argues that structuring is more important than the quantitative analysis (abstract). P. 195 argues that of 10,000 decisions, 10 can benefit from quantitative decision analysis as things are today. P. 196 writes that it should become 1000 out of 10,000. The paper presents an enthusiastic plea for decision analysis. Keeney is most known for his 1976 textbook with Raiffa, explaining expected utility, utility independence axioms for multiattribute utility, and applied utility measurements. Expected utility is for decision under risk/uncertainty, a small part of our decisions and life. The quantitative techniques provided by it, and the multiattribute utility measurements, using simple choices to derive more complex ones, and they provide particular quantitative tradeoff techniques that are only of some use in very particular situations. Many researchers too much think, and suggest, that their particular work is relevant to too much in life. This paper went too far that way too (own little expertise = meaning of life). Although the author nicely clarifies that of 10,000 decisions in our life, most don’t need decision analysis, he still too much puts the EU techniques forward as important. Again and again he overly quickly goes for his EU-multiattribute techniques (with probabilities to be assessed, for instance) as the one and only thing to do. To illustrate my criticism, I give three citations: (1) ”To analyze alternatives, one typically requires a list of key uncertainties, assessments of probabilities for these uncertainties, a decision tree, value tradeoffs, and a quantified attitude toward risk [risk tolerance]. Subjective judgment is necessary to specify each of these.” (p. 198 2nd column 3rd para ) (2) “Decision analysis should guide all of our thinking about decisions.” P. 200 3rd para (3) “Decision analysis is useful for resolving all decisions worth thinking about.” P. 201 2nd para There are many other texts like the above ones. Had the author not worked on uncertainty all his life, but on intertemporal choice, then he would have written, instead of the above citation (1): ”To analyze alternatives, one typically requires a list of future gains and losses, assessments of approximate times points of receipts of those gains and losses, a decision tree, value tradeoffs, and a quantified attitude toward discounting. Subjective judgment is necessary to specify each of these.” Had the author worked in game theory, it would have been: ”To analyze alternatives, one typically requires a list of key opponents, assessments of their strategies and interests, a game tree, noncredible threats, and a quantified utlity function for each opponent. Subjective judgment is necessary to specify each of these.” As the saying goes, if all you have is a hammer, then everything looks like a nail.” The broadenings in §7 help but stay too close to the techniques. %} Keeney, Ralph L. (2004) “Making Better Decision Makers,” Decision Analysis 1, 193–204. {% %} Keeney, Ralph L. & Timothy L. McDaniels (1999) “Identifying and Structuring Values to Guide Integrated Resource Planning at BC Gas,” Operations Research 47, 651–662. {% Apply some multiattribute utility techniques from Keeney & Raiffa (1976) to the case where attributes are different persons, to get a weighted average of individual utilities. %} Keeney, Ralph L. & Robert F. Nau (2011) “A Theorem for Bayesian Group Decisions,” Journal of Risk and Uncertainty 43, 1–17. {% decreasing ARA/increasing RRA??? check out real incentives/hypothetical choice: §1.4.3, p. 18, discusses the necessity for decision analysis to use hypothetical choice, so as to clarify real choice. substitution-derivation of EU: very concisely, on pp. 133-134, §4.1.1. utility families parametric: Table 4.5, p. 173 risky utility u = transform of strength of preference v, latter doesn’t exist, because they let value function be ordinal; Digression in §4.4.1, p. 150, makes it very clear that they think so. They say very explicitly that vNM utility and economists’ utils are very different, adding on utils: “which are never explicitly defined.” real incentives/hypothetical choice: §1.4.3 explains that hypothetical choice is crucial in decision analysis. Download 7.23 Mb. Share with your friends:
|