Bibliography
§§5 ff. become more informal
Download 7.23 Mb.
|
| §§5 ff. become more informal.
§7 refers to studies where probability judgments were used to predict economic actions. %} Manski, Charles F. (2004) “Measuring Expectations,” Econometrica 72, 1329–1376. {% Derives equilibrium result that price reflects a particular quantile of the beliefs of the agents. %} Manski, Charles F. (2006) “Interpreting the Predictions of Prediction Markets,” Economics Letters 91, 425–429. {% A follow-up paper on his minimax-type ambiguity decision model, without references to the decision-theory literature such as Gilboa & Schmeidler. Seems to recommend diversified treatment of identical persons (…) so as to turn unknown ambiguous probabilities into known probabilities, reminiscent of Raiffa (1961; not cited). %} Manski, Charles F. (2009) “The 2009 Lawrence R. Klein Lecture: Diversified Treatment under Ambiguity,” International Economic Review 50, 1013–1041. {% Argues, right so, that in prescriptive decision analysis only one choice situation is for real, and the rest is hypothetical to improve the real decision. If choice axioms are imposed only on the real situation then not much more than dominance can be thought of. %} Manski, Charles F. (2011) “Actualist Rationality,” Theory and Decision 71, 297–324. {% DOI: http://dx.doi.org/10.1146/annurev-economics-061109-080359 The abstract writes that the paper reviews recent work on ambiguity, but the intro adds that it is recent work only by the author himself. %} Manski, Charles F. (2011) “Choosing Treatment Policies under Ambiguity,” Annual Review of Economics 3, 25–49. {% probability communication: argues that probability estimates should also report error. %} Manski, Charles F. (2015) “Communicating Uncertainty in Official Economic Statistics: An Appraisal Fifty Years after Morgenstern,” Journal of Economic Literature 53, 631–653. {% %} Mansour, Selima Ben, Elyès Jouini, & Clotilde Napp (2006) “Is there a “Pessimistic” Bias in Individual Beliefs? Evidence from a Simple Survey,” Theory and Decision 61, 363–371. {% A single-valued choice function is derived from two binary relations, where maximization is lexicographically. If the first relation is incomplete then new things can occur. They relate it to Tversky’s elimination by aspects and characterize it. %} Manzini, Paola & Marco Mariotti (2007) “Sequentially Rationalizable Choice,” American Economic Review 97, 1824–1839. {% %} Manzini, Paola & Marco Mariotti (2008) “On the Representation of Incomplete Preferences over Risky Alternatives,” Theory and Decision 65, 303–323. {% Consider revealed preference, but choices can depend on the psychological state of mind of the agent. %} Manzini, Paola & Marco Mariotti (2015) “State Dependent Choice,” Social Choice and Welfare 45, 239–268. {% real incentives/hypothetical choice: for time preferences: use real incentives, students themselves have to pick up money; longest period is 9 months. %} Manzini, Paola, Marco Mariotti, & Luigi Mittone (2010) “Choosing Monetary Sequences: Theory and Experimental Evidence,” Theory and Decision 69, 327–354. {% Extend Ryan’s result to unbounded random variables. %} Mao, Tiantian & Taizhong Hu (2012) “Characterization of Left-Monotone Risk Aversion in the RDEU Model,” Insurance: Mathematics and Economics 50, 413–422. {% Gives many arguments for why inconsistency of preference may be rational. That it may be rational to maintain ambiguity about “true” preferences. There are many nice sentences. On advantage of representative agent for decision models: P. 588: “So long as we use individual choice models to predict the behavior of relatively large numbers of individuals or organizations, some potential problems are avoided by the familiar advantages of aggregation.” This paper uses term ambiguity in sense of uncertainty/variability of consumer’s preference relation. Pp. 591-592 discuss aspiration levels and step-function-tastes. P. 593: posterior rationality: intentions discovered as interpretation of action afterwards; evaluation after the fact. The central focus of the paper is paternalism/Humean-view-of-preference: p. 594: Simon showed that actual human choice behavior is more intelligent than it appeared, and that conforming it more to normative theory may be bad. The paper lists many arguments in favor of not-well-specified preferences. I interpret the author’s arguments in favor of ambiguity as criticisms of completeness more than of other consistency conditions. Such as p. 597: to avoid being manipulated by others in game situations (suspicion under ambiguity). And pp. 598-599 that list five reasons. And P. 603: “And precision in objectives does not allow creative interpretation of what the goal might mean (March, 1978). Thus, the introduction of precision into the evaluation of performance involves a tradeoff between the gains in outcome attributable to closer articulation between action and performance on an index of performance and the losses in outcomes attributable to misrepresentation of goals, reduced motivation to development of goals, and concentration of effort on irrelevant ways of beating the index.” (completeness-criticisms) P. 597: “We do not believe that what we do must necessarily result from a desire to achieve preferred outcomes.” P. 597 properly criticizes Stigler & Becker (1977): “to trivialize the issue into a ‘definitional problem.’ By suitably manipulating the concept of tastes, one can save classical theories of choice as ‘explanations’ of behavior in a formal sense, but probably only at the cost of stretching a good idea into a doubtful ideology (Stigler & Becker, 1977).” The text then immediately continues with a nice statement of the point that a normative theory can be useful only if it sometimes !deviates! from actual behavior, the point also stated nicely by Raiffa (1961): “More importantly from the present point of view, such a redefinition pays the cost of destroying the practical relevance of normative prescriptions for choice. For prescriptions are useful only if we see a difference between observed procedures and desirable procedures.” P. 602, on inconsistency: “the other problems probably require a deeper understanding of contradiction as it appears in philosophy and literature” %} March, James G. (1978) “Bounded Rationality, Ambiguity and the Engineering of Choice,” Bell Journal of Economics 9, 587–608. {% %} March, James G. & Zur Shapira (1992) “Variable Risk Preferences and the Focus of Attention,” Psychological Review 99, 172–183. {% Criticizes the often-used misleading interpretation of invariance w.r.t. scale as if this concerned only a rescaling of the modeling of outcomes without empirical meaning. %} Marchant, Thierry (2008) “Scale Invariance and Similar Invariance Conditions for Bankruptcy Problems,” Social Choice and Welfare 31, 693–707 (Erratum pp. 709–710.) {% Study the Balloon Analogue Risk Task (BART), surveying it, and relating it to sensation seeking and impulsivity for 2120 subjects in a meta-analysis. The authors are throughout very positive on the predictive power of risk-attitude measurements. P. 30: “Borrowing from Appelt et al. (2011), we strongly believe that only measures with a theoretical tie with risky decision making are likely to result in consistent findings both inside and outside the laboratory setting.” P. 27 2nd column: find positive relation between age and risk aversion in age range 11-23 years and no relation with gender (gender differences in risk attitudes). %} Marco Lauriola, Angelo Panno, Irwin P. Levin, & Carl W. Lejuez (2014) “Individual Differences in Risky Decision Making: A Meta-analysis of Sensation Seeking and Impulsivity with the Balloon Analogue Risk Task,” Journal of Behavioral Decision Making 27, 20–36. {% Whereas many axiomatizations of generalized means use associativity plus symmetry, this paper shows that the weaker strong decomposability suffices. %} Marichal, Jean-Luc (2000) “On an Axiomatization of the Quasi-Arithmetic Mean Values without the Symmetry Axiom,” Aequationes Mathematicae 59, 74–83. {% Characterizes the Choquet integral through linear-minimum conditions in terms of the Möbius transform. %} Marichal, Jean-Luc (2000) “An Axiomatic Approach of the Discrete Choquet Integral as a Tool to Aggregate Interaction Criteria,” IEEE Transactions on Fuzzy Systems 8, 800–807. {% qualitative probability. %} Marinacci, Massimo (1992) “A Note on Comparative Probability Structures.” Dept. of Economics, Northwestern University. {% %} Marinacci, Massimo (1996) “Decomposition and Representation of Coalitional Games,” Mathematics of Operations Research 21, 1000–1015. {% %} Marinacci, Massimo (1997) “Vitali’s Early Contribution to Non-Additive Integration,” Rivista di Matematica per le Scienze Economiche e Sociali 20, 153–158. {% Characterizes infinite sequences with zero discounting. %} Marinacci, Massimo (1998) “An Axiomatic Approach to Complete Patience and Time Invariance,” Journal of Economic Theory 83, 105–144. {% %} Marinacci, Massimo (1999) “Limit Laws for Non-Additive Probabilities and Their Frequentist Interpretation,” Journal of Economic Theory 84, 145–195. {% %} Marinacci, Massimo (1999) “Upper Probabilities and Additivity,” Sankhya: The Indian Journal of Statistics 61, 358–361. {% Assumes multiple priors ( min + (1) max) for 1/2. An event is defined to be unambiguous if the binary acts w.r.t. the event can be represented by SEU (i.e., all probability measures in the set of priors assign same probability to the event). Note that this definition is not in terms of prefs. Pref. defs can be given by using existing axiomatizations of SEU. The paper shows under regularity assumptions that, if multiple priors holds, probabilistic sophistication holds, and there exists one unambiguous event in the above sense, then SEU must hold throughout. It is like a continuous strictly increasing function w from [0,1] to [0,1] with w(0) = 0 and w(1) = 1, if it is convex and if there is a p with w(p) + w(1p) = 1 (implying that not both w(p) and w(1p) can be below the diagonal), then w must be linear. Without the assumption of an unambiguous event the implication need not hold. Any RDU with convex probabibility weighting is maxmin ( = 1) multiple priors, not SEU, and there exists no nontrivial unambiguous event (there is such an example at the end of §3). Although this model is multiple priors in a formal sense, it is not “in spirit.” My interpretation is not so much that we may study ambiguity attitudes in the multiple priors model while assuming SEU as ambiguity-neutrality benchmark, so not so much that we may assume probabilistic risk attitude away without loss of generality (see p. 756 3rd para). Instead, my interpretation is that multiple priors (in the classical sense with EU for each probability measure) may only be appropriate if we have extraneous prior reasons to believe that probabilistic risk attitude plays no role; i.e., that people do EU for given probabilities. So it is not a consequence but a prior requirement. %} Marinacci, Massimo (2002) “Probabilistic Sophistication and Multiple Priors,” Econometrica 70, 755–764. {% Seems to provide conditions under which ambiguity fades away in sampling with replacement from the same ambiguous urn. %} Marinacci, Massimo (2002) “Learning from Ambiguous Urns,” Statistical Papers 43, 145–151. {% survey on nonEU: Many examples of decision under uncertainty/ambiguity, with calculations in all kinds of models added, on p. 1030 ff., 1037 ff., 1042 ff., 1057 ff., 1070 ff. This paper can take the space to well and didactically explain the author’s general views on ambiguity comprehensively. It has many valuable historical references. The author throughout assumes that there is a true objective (physical) but unknown probability model. He uses here, as in other papers, the broad term model (uncertainty) for (uncertainty only about) that true probability measure. The author mostly assumes a two-stage setup, where in the lowest, first, stage ucertainty is objective, captured by the objective probability measure. The author uses the term generative (or data generating) mechanism (p. 1024 3rd para) and calls.this uncertainty physical. (So they cannot be purely and only subjective as in Savage, 1954, as I understood them to possibly be in the smooth model.) However, the end of the para confuses me because it writes that this uncertainty is epistemic, apparently because the decision maker knows/has info about this uncertainty (reiterated on p. 1040). But if I understand this right then every physical object can be called epistemic. As often done today, a two-stage model is introduced by first specifying a set of possible 1st stage probabilized uncertainties on Savage (outcome-relevant) states, and only then imposing second-order uncertainty over them, thus appealing to the popular concept of sets of priors. The set of possible priors is then the support of the second-order uncertainty (expressed through a probability measure ). P. 1024 last line (also p. 1037 2nd para) takes the set of possible probabilities as datum, so, exogenous. I assume that is not datum, but subjective. The author refers here, as in other places (p. 2037 footnote 32), to statistics. Classical statistics has a same informational structure, with probabilities over observations objective, but a second stage (what the true hypothesis/statistical parameter) with the uncertainty unprobabilized, and this is the only analogy that the author is after. Outcome relevance is different. In statistics the second-stage uncertainty, what is the true hypothesis, is outcome-relevant, and the first stage (statistical observation) is only informational. In this paper it is opposite, and the first-stage uncertainty (about the Savage-state) with objective probabilities specified is outcome relevant, with the second stage only informational. But there is a big difference with statistics, discussed next. [DIFFERENCE STATISTICS AND AUTHOR’S MODEL] In the author’s model, the 1st stage generative mechanism with objective probabilities (“physical”) is about (Savage) states and they are outcome relevant. The 2nd stage epistemic (“model”) uncertainty is only instrumental, to give info about the 1st stage uncertainty. Once you know which Savage state is true, you know which outcome you get and you don’t care anymore about the 2nd stage epistemic uncertainty. In statistics these things are the other way around. The 1st stage generative mechanism with objective probabilities (“physical”) is about observations of statistics and they are NOT outcome relevant. They are only instrumental to give info about the epistemic 2nd stage uncertainty. The 2nd stage epistemic (“model”) uncertainty is about the statistical hypotheses (using the author’s term; or statistical parameters in estimations) and these are outcome relevant. Once you know which hypothesis is true, you know which outcome you get and you don’t care anymore about the 1st stage objective uncertainty. P. 1025: “The often-made modeling assumption that a true generative mechanism exists is unverifiable in general and so of a metaphysical nature” P. 1045 bottom repeates this. This interpretation gets closer to the smooth model as I understand it. If no explicit physical mechanism (such as unknown composition of an urn) can be specified that underlies the physical probabilities, then I think that this physical interpretation is not very useful and I would just leave it at subjective (so epistemic). P. 1025: In the sentence “In any event, the assumption underlies a fruitful causal approach that facilitates the integration of empirical and theoretical methods—required for a genuine scientific understanding.” the intimidating “required” can only refer to the latter integration and not to the former modeling assumption. P. 1025: “We assume that the DMs’ ex-ante information also enables them to address model uncertainty through a subjective prior probability over models” restricts the author’s approach to cases where reduction of compound lotteries (probability multiplication) is abandoned. Although later he will consider alternatives such as alpha maxmin in which no such 2nd order distribution is assumed, as then explained. P. 1025: “The second layer is ignored by classical statistics.” is negative about approaches that do not assume a 2nd order probability distribution. P. 1026: “The two layers of analysis motivated by such a distinction naturally lead to two-stage decision criteria: actions are first evaluated with respect to each possible probability model, and then such evaluations are combined by means of the prior distribution.” This is in fact assuming backward induction, giving up reduction of compound lotteries, which in nonEU is controversial. Machina (1989) argued against it for decision under risk. P. 1027 discusses physical probability as propensity, citing Popper. I find Ramsey (1931) the best text on this point. criticisms of Savage’s basic model: well, a discussion. P. 1029, §2.2, presents the Wald decision model, with act set A, state space S, and a consequence function mapping each (a,s) to a consequence c from a consequence space C. P. 1035: the term consequentialism used here is not to be confused with the dynamic decision principle of Machina (1989). P. 1036 then presents Savage’s model as a special case, suggesting and citing people arguing that Wald’s model is more natural. I feel that of acts, states, and consequences, acts are most basic (first thing is that a decision is faced), and states and consequences are next equally (non)basic, but this gives me no preference between the two models. P. 1038 health insurance example, and other examples similarly: if four experts each give one guess of the physical probability measure, then the set of priors is taken as the set of those four. But I think that in such situations many other probabilies are possible too, and it is not the case that one of the four estimates of the experts must be THE exact true physical probability measure. Experts are not generative mechanisms. The author writes in the overview that model misspecification will be ignored, and thus can justify treating expert opinions this way. Probably model misspecification would magnify issues without making them fundamentally different. P. 1039 footnote 37 cites the evidentialist view of probability, with several references. I know this under the name logical view of probability, with Carnap the main advocate. Carnap is not cited here, but in footnote 14 (p. 1027), and I am not sure if for the author evidentialist and logical view are the same, as they are for me. P. 1042: crisp acts are not affected by ambiguity. P. 1045 reiterates that the generative mechanism, taken as physical, may be unobservable and then the second-order probability measure can only be observed from hypothetical betting behavior. Section 4 is useful in describing the smooth model including the assumptions underlying it. Footnote 52, p. 1050, cites works related to the smooth model. I would also cite Kahneman & Tversky (1975, p. 30 ff.), who had the smooth model for ambiguity for two outcomes, Dobbs (1991) who had a different but similar model, and recursive EU by Neilson (cited in Footnote 56), and Kreps & Porteus (1978) who used the same functional form. P. 1051 in §4.1 is especially useful in discussing portability. In the smooth model, , the utility transformer in the second stage, is assumed to depend only on the subject and to be invariant across different decision situations. So, it is assumed that once has been elicited using, say, Ellsberg urns, then it applies to all situations of uncertainty. This is a strong assumption, giving the richness of uncertain events, but it is then very good in being tractable. SEU = risk: it is discussed on p. 1051 bottom ff. P. 1061 will cite someone arguing that it is not intuitive to treat objective and subjective (in the sense of comprising ambiguity) probabilities the same way. P. 1052 ff. discusses source dependence, but takes source differently than I do. P. 1052 2nd para 2nd sentence: “We distinguished two sources of uncertainty, physical and epistemic.” So this is how the author takes it: one source is the epistemic 2nd stage uncertainty, the attitude to which is captured by the utility-transformer . The other source(s) are the generative mechanisms the attitude to which is captured by the risk-attitude function u. So it is categorical and dichotomous, with only two (kinds of) sources, epistemic or physical, and there are two kinds of attitudes, and u respectively. The only thing that can bring changes, and a gradual path from one kind of uncertainty to the other, is the 2nd order distribution . My work on the source idea is different. There can be many kinds of sources of uncertainty, inducing many kinds of attitudes with higher and lower degrees of pessimism (and insensitivity). In this respect my use of sources is more general and accordingly less tractable. P. 1052 footnote 54: Smith (1969) did not have the idea of source. He did discuss the competence effect, one of the several factors that impact ambiguity attitude, several of which were studied by Yates and co-authors in several papers. But mentioning a factor that impacts ambiguity attitude is too far a cry from the source idea, being way more general. Tversky has the priority of the source concept, in Heath & Tversky (1991), mentioning it briefly in Tversky & Kahneman (1992), where I usually take Tversky & Fox (1995) as the main reference for introducing it in a mature form. P. 1053 2nd para refers to the “negative attitude” of the DM towards ambiguity, focusing on ambiguity aversion. P. 1055, §4.2, middle, discusses CARA and CRRA . In the former case, it can be argued that ambiguity attitude is constant and in this way independent of outcomes. But it is constant only in an absolute sense then. In the second case, it is constant and independent of outcomes only in relative sense. Footnote 60 there discusses constant absolute ambiguity aversion w.r.t. utility units, by Grant & Polak (2013) and others.| §4.1, p. 1057 ff. discusses the Ellsberg two-urn paradox, assuming all priors. P. 1072 ff., §4.6, discusses the maxmin EU model of Gilboa & Schmeidler (1989). P. 1063 points out that here the set of priors can be taken subjectively, to induce ambiguity aversion. P. 1063 middle points out that priors are only in or out, and the 2nd stage plays no role. P. 1082 2nd para ff. repeats the point. P. 1066 gives examples and calculations, interpreting some quantities as ambiguity premiums. The analysis of ambiguity attitudes on p. 1070 focuses on ambiguity aversion. %} Marinacci, Massimo (2015) “Model Uncertainty,” Journal of the European Economic Association 13, 1022–1100. {% Presents and derives many mathematical properties of nonadditive set functions. %} Marinacci, Massimo & Luigi Montrucchio (2004) “Introduction to the Mathematics of Ambiguity.” In Itzhak Gilboa (ed.) Uncertainty in Economic Theory: Essays in Honor of David Schmeidler’s 65th Birthday, 46–107, Routledge, London. {% %} Marinacci, Massimo & Luigi Montrucchio (2004) “A Characterization of the Core of Convex Games through Gateaux Derivatives,” Journal of Economic Theory 116, 229–248. {% %} Marinacci, Massimo & Luigi Montrucchio (2005) “Ultramodular Functions,” Matematics of Operations Research 30, 311–332. {% three-prisoners problem %} Marinoff, Louis (1996) “A Reply to Rapoport,” Theory and Decision 41, 157–164. {% %} Mariotti, Marco (1995) “The Subjective Probabilities and Non-Expected Utilities of Cautious von Neumann-Morgenstern Expected Utility Maximizers,” {% %} Mariotti, Marco (1995) “Is Bayesian Rationality Compatible with Strategic Rationality?,” Economic Journal 105, 1099–1109. {% Nash bargaining solution %} Mariotti, Marco (1998) “Nash Bargaining Theory when the Number of Alternatives Can Be Finite,” Social Choice and Welfare 15, 413–421. {% revealed preference %} Mariotti, Marco (2008) “What Kind of Preference Maximization Does the Weak Axiom of Revealed Preference Characterize?,” Economic Theory 35, 403–406. {% real incentives/hypothetical choice: stated preference is what mainstream economists call hypothetical choice. Revealed preference is then called real choice (market data and so on). The “data fusion literature” investigates how to combine them, and use one to predict the other. The paper gives references. %} Mark, Tami L. & Joffre D. Swait (2004) “Using Stated Preference and Revealed Preference Modeling to Evaluate Prescribing Decisions,” Health Economics 13, 563–573. {% %} Markle, Alex, George Wu, Rebecca White, & Aaron Sackett (2014) “Goals as Reference Points in Marathon Running: A Novel Test of Reference-Dependence,” working paper. {% Proposed reference point; risky utility u = strength of preference v (or other riskless cardinal utility, often called value); Note that he uses the terms convex and concave conversely than is done nowadays. I will use the terms in their current sense. Predicts concave utility (risk aversion) for small losses (roughly, the threshold is somewhere between $100 and $10,000) and convex utility (risk seeking) for small gains (threshold somewhere between $100 and $10,000), exactly opposite to the predictions of prospect theory. P. 154, I assume that the first inequality sign below Fig. 5 is a typo and should be reversed. P. 154 following Fig. 5 he seems to suggest loss aversion, P. 157 top of 2nd column, however, suggests that there is an inflection point with almost linear utility around 0, strangely enough. P. 155, top of first column, explicitly discusses variation in reference point through prior endowment subtracted from the gamble outcomes. P. 156 mentions the tendency to take more risk after prior wins (now called the house money effect). P. 157 makes explicit that it is a weak point that there is no theory about the location of the reference point. %} Markowitz, Harry M. (1952) “The Utility of Wealth,” Journal of Political Economy 60, 151–158. {% %} Markowitz, Harry M. (1959) “Portfolio Selection: the Efficient Diversification of Investments,” Yale UP, New Haven. {% A book by Jason Zweig (“Your money or your brain”) seems to give the following citation of Markowitz: “I should have computed the historical co-variances of the asset classes and drawn an efficient frontier. I visualized my grief if the stock market went way up and I wasn’t in it — or if it went way down and I was completely in it. So I split my contributions 50/50 between stocks and bonds” Markowitz, Harry M. {% Reviews literature on relating mean-variance to EU. %} Markowitz, Harry (2014) “Mean–Variance Approximations to Expected Utility,” European Journal of Operational Research 234, 346–355. {% %} Marley, Anthony A.J. & R. Duncan Luce (2001) “Ranked-Weighted Utilities and Qualitative Convolution,” Journal of Risk and Uncertainty 23, 135–163. {% %} Marley, Anthony A.J. & R. Duncan Luce (2005) “Independence Properties vis-à-vis Several Utility Representations,” Theory and Decision 58, 77–143. {% Uses event commutativity and some other natural and structural axioms to axiomatize biseparable utility. Assume solvability both for events and for outcomes. P. 44 points out that Axiom A.10, requiring existence of a particular mapping on events, is not behavioral in the usual sense. P. 43, following Eq. 10, points out that this approach needs independent repetitions of events. Axiom A.4 seems to imply that there are no nonempty null events. The event space must be infinite though because of denseness. P. 45, Theorem 4.1: I think that axiom A.10 is not necessary for general RDU because the set of events need not be sufficiently rich. %} Marley, Anthony A.J. & R. Duncan Luce (2002) “A Simple Axiomatization of Binary Rank-Dependent Utility of Gains (Losses),” Journal of Mathematical Psychology 46, 40–55. {% %} Marley, Anthony A.J., R. Duncan Luce, & Imre Kocsis (2008) “A Solution to a Problem Raised in Luce & Marley (2005)” Journal of Mathematical Psychology 52, 64–68. {% Christiane, Veronika & I %} Marques, J. Frederico (1999) “Changing ‘Europe’—The Euro as a New Subject for Psychological Research in Numerical Cognition,” European Psychologist 4, 152–156. {% Z&Z; Re-analyze hypothetical choices in the famous RAND data using prospect theory. Find support for loss aversion and risk seeking for losses (Risk averse for gains, risk seeking for losses). Unfortunately, there are so many unclear points in their modeling of prospect theory that the results are not clear to me. They do not consider probability weighting (which in itself can be an OK working hypothesis for pragmatic reasons, made by many) but do consider the certainty effect. The latter is, however, typically modeled through probability weighting. Apparently they have some utility-of-gambling model in mind such as for instance Diecidue, Schmidt, & Wakker (2004), but this is not clear. They do what they call seggregation, where they do not integrate the riskless and risky payments but evaluate them separately and additively, as a kind of additive version of Luce’s joint receipt. Kahneman & Tversky (1979) considered something in this spirit but, for a positive prospect that yields $x >0 as minimal outcome and with probability p y > x, took as evaluation, where I write U for utility=value function, U(x) + w(p)(U(y) U(x)) which is just the regular rank-dependent evaluation and y is the INTEGRATED payment to be added to the reference point r so that final wealth is r+y. These authors further seggregate payments added to the reference point as a kind of mental accounting, which is a fundamental deviation from PT. Pp. 422-423 has nice distinction between initial wealth and reference point. utility concave near ruin: p. 423 takes utility for losses first convex but for large losses concave. P. 425: their parameter estimates find CONSTANT utility on [200,0] which is as much against loss aversion as one can think of and is clearly absurd. They do not consider loss aversion. P. 423 footnote 7 points out that they tested their model not only with status quo as reference point but also with complete insurance, but this fitted the data worse. %} Marquis, M. Susan & Martin R. Holmer (1996) “Alternative Models of Choice under Uncertainty and Demand for Health Insurance,” Review of Economics and Statistics 78, 421–427. {% Contrary to a claim of Machina, Marschak does not object, on p. 320, to EU; he doesn’t even mention it. He objects to EV only, and says that other moments will be relevant as well. %} Marschak, Jacob (1938) “Money and the Theory of Assets,” Econometrica 6, 311–325. {% %} Marschak, Jacob (1948) (Title unknown), Cowles Commission Discussion Paper, Economics No. 226 (hectographed), July, 1948. {% %} Marschak, Jacob (1949) “Measurable Utility and the Theory of Assets” (abstract), Econometrica 17, 63–64. {% P. 193 mentions maxmin over expected values, but prefers Savage’s minimax regret. %} Marschak, Jacob (1949) “Role of Liquidity under Complete and Incomplete Information,” American Economic Review, Papers and Proceedings 39, 182–195. {% dynamic consistency: favors abandoning RCLA when time is physical, because of Utility of gambling. Mentions two 1948 working papers. %} Marschak, Jacob (1950) “Rational Behavior, Uncertain Prospects, and Measurable Utility,” Econometrica 18, 111–141. {%:Luce says: p. 176 gives the independence axiom; this lecture was given on Dec.6, 1950. %} Marschak, Jacob (1951) “Why “Should” Statisticians and Businessmen Maximize “Moral Expectation”?” In Jerzy Neyman (1951, ed.) Proceedings of the Second Berkeley Symposium on Mathematical Statistics and Probability, University of California Press, Berkeley. {% %} Marschak, Jacob (1954) “Probability in the Social Sciences.” In Paul F. Lazersfeld (ed.) Mathematical Theory in the Social Sciences, 166–215, The Free Press, New York. {% second-order probabilities %} Marschak, Jacob (1975) “Personal Probabilities of Probabilities,” Theory and Decision 6, 121–153. {% risky utility u = transform of strength of preference v First sentence of §II.9, p. 168: “If one could assume that, by good luck, the functions s and B to coincide” (here s is the psychological utility function, B the vNM %} Marschak, Jacob (1979) “Utilities, Psychological Values, and the Training of Decision Makers.” In Maurice Allais & Ole Hagen (eds.) Expected Utility Hypotheses and the Allais Paradox, 163–174, Reidel, Dordrecht. {% %} Marschak, Jacob & Roy Radner (1972) “Economic Theory of Teams.” Yale University Press, New Haven. {% Independence as monotonicity w.r.t. a variabele partition. %} Marschak, Thomas (1987) “Independence versus Dominance in Personal Probability Axioms.” In Walter P. Heller, Ross M. Starr, & David A. Starrett (eds.) Uncertainty, Information and Communication, Essays in Honor of Kenneth J. Arrow, Vol. III, 129–171, Cambridge University Press, Cambridge. {% Stigler Footnote 250 refers to p. 94 of 8th edn. for example of carpet to cover floor where last yard has more utility than yards before. To preserve diminishing marginal utility, Marshall says that whole carpet should be taken as one object. marginal utility is diminishing: pp. 398-400: risk aversion is ascribed to diminishing marginal utility; Footnote IX in Mathematical Appendix proves that risk aversion iff u concave, well he derived it only for two-outcome gambles. Marshall seems to have been the first to demonstrate this point. Bernoulli 1738 §13 also suggests it and §14 first sentence claims it in general, but does not really prove it. Citation: “The argument that fair gambling is an economic blunder is generally based on Bernoulli’s or some other definite hypothesis. But it requires no further assumption than that, firstly the pleasure of gambling may be neglected; and, secondly, '' is negative for all values of x, where (x) is the pleasure derived from wealth equal to x” (Then Marshall gives a proof, only for two-outcome gamble. He continues: “It is true that this loss of probable happiness need not be greater than the pleasure derived from the excitement of gambling, and we are thrown back upon the induction that pleasures of gambling are in Bentham’s phrase “impure;” since experience shows that they are likely to engender a restless, feverish character, unsuited for steady work as well as for the higher and more solid pleasures of life.” (Marshall, 1920: 843). linear utility for small stakes: this is crucial for Marshall to obtain cardinal utility. Seems to be in Book III. %} Marshall, Alfred (1890) “Principles of Economics.” 8th edn. 1920 (9th edn. 1961), MacMillan, New York. {% decreasing/increasing impatience: seem to find that utility of life duration has increasing risk aversion, which indirectly implies increasing impatience. %} Martin, Andrew J., Paul Glasziou, R. John Simes, & Thomas Lumley (2000) “A Comparison of Standard Gamble, Time Trade-off, and Adjusted Time Trade-Off Scores,” International Journal of Technology Assessment in Health Care 6, 137–147. {% Bayes’ formula intuitively %} Martin, Scott L. & William Terris (1991) “Predicting Infrequent Behavior: Clarifying the Impact on False-Positive Rates,” Journal of Applied Psychology 76, 484–487. {% Compare matching and choice, where opaque means that choice questions leading to an indifference are interspersed with other questions so that subjects do not know. Then preference reversals can be avoided. It adds to Bostic, Herrnstein, & Luce (1990). %} Martinez, Fernando I. Sanchez, José Luis Pinto, Jose María Abellán Perpiñan, & Murcia Jorge Martínez Pérez (2014) “Avoiding Preference Reversals with Opaque Methods,” in preparation. {% conservation of influence: discusses and cites several books and works by Ian Hacking on the differences between natural and social sciences. Seems to be mainly that the construction of social sciences and of our image of man is interactive with our construction work. %} Martínez, María Laura (2009) “Hacking's Proposal for the Distinction between Natural and Social Sciences,” Philosophy of the Social Sciences 39, 212–234. {% Relate concavity of utility of income to concavity properties of utility of commodity bundles to be bought for the income. %} Martínez-Legaz, Juan E. & John K. -H. Quah (2007) “A Contribution to Duality Theory, Applied to the Measurement of Risk Aversion,” Economic Theory 30, 337–362. {% conservation of influence: seems to write, according to Georgescu-Roegen (1954, QJE, p. 511), in “Equivalent form of value (pp. 64ff) that all commodities must have a common facto (pp. 43-45)r: %} Marx, Karl (1932) “Capital. Vol. I.” Kerr & Co, Chicago. {% %} Mas-Colell, Andreu (1974) “An Equilibrium Existence Theorem without Complete or Transitive Preferences,” Journal of Mathematical Economics 1, 237–246. {% Shows that every continuous consumer is limit of differentiable consumers. No experiment can prove nondifferentiability. %} Mas-Colell, Andreu (1974) “Continuous and Smooth Consumers: Approximation Theorem,” Journal of Economic Theory 8, 305–336. {% revealed preference %} Mas-Colell, Andreu (1978) “On Revealed Preference Analysis,” Review of Economic Studies 45, 121–131. {% Has been most popular textbook for teaching micro for many years. I find that amazing because whatever I read in it was dry, no ideas at all, just the formalities, and those in inefficient manners, needlessly complex. I taught game theory from it for one year and was unsatisfied so I switched to Peters’ (2008) textbook. All students I spoke expressed negative judgments about the Mas-Colell et al. book. P. 185 presents the St. Petersburg paradox in utility units. That one would be willing to give up all of one’s wealth for it is called “patently absurd.” inverse-S: the classical economists’ view of p. 185: “The concept of risk aversion provides one of the central analytical techniques of economic analysis, and it is assumed in this book whenever we handle uncertain situations.” %} Mas-Colell, Andreu, Michael D. Whinston, & Jerry R. Green (1995) “Microeconomic Theory.” Oxford University Press, New York. {% %} Masatlioglu, Yusufcan & Efe A. Ok (2005) “Rational Choice with Status Quo Bias,” Journal of Economic Theory 121, 1–29. {% They show that for Köszegi-Rabin the choice-acclimating personal equilibrium (CPE), when taken in its most popular form with gain-loss utility that has a kink at 0 but is linear otherwise, is exactly the intersection of quadratic and rank-dependent utility. Proposition 3: then loss aversion 1 iff mixture averse, so, under RDU, iff w convex (this uses my 1994 theorem), and 1 iff mixture loving, so w concave. In the proof, p. 2780, the authors indicate a generalization of my 1994 result: it also holds if w is not increasing, with no change in the proof required. Proposition 7 shows that now loss aversion iff first-order risk aversion under RDU, consistent with a claim by Köbberling & Wakker (2005) that most of first-order risk aversion is due to loss aversion. %} Masatlioglu, Yusufcan & Collin Raymond (2016) “A Behavioral Analysis of Stochastic Reference Dependence,” American Economic Review 106, 2760–2782. {% Experimentally examine reference dependence in multiattribute choice. They compare the well-known model of Tversky & Kahneman (1991) with a model by Masatlioglua & Ok developed in some papers. In the latter model, the decision maker has two selves, and an alternative is preferred to the status quo only if both selves agree. The two models correctly predict choices if one alternative dominates the status quo but the other does not. They do not in other cases, and there the model of Masatlioglua & Ok, which predicts no reference effect there, is confirmed. %} Masatlioglua, Yusufcan & Neslihan Uler (2013) “Understanding the Reference Effect,” Games and Economic Behavior 82, 403–423. {% Seems to describe wishful thinking: assigning higher likelihood to preferred outcome %} Mascaro, Guillermo F. (1969) “ ‘Wishful Thinking’ on the Presidential Election,” Psychological Reports 25, 357–358. {% %} Maschler, Michael, Eilon Solan, & Shmuel Zamir (2013) “Game Theory.” Cambridge University Press, Cambridge. {% Use choice list to measure risk aversion. Groups are more risk averse than individuals. %} Masclet, David, Nathalie Colombier, Laurent Denant-Boemont, & Youenn Lohéac (2009) “Group and Individual Risk Preferences: A Lottery-Choice Experiment with Self-Employed and Salaried Workers,” Journal of Economic Behavior and Organization 70, 470–484. {% Characterizes maximization of sum on n. Every xj in (x1,…,xn) is interpreted as utility level of individual i, is taken as empirical primitive, and the sum is interpreted as utilitarianism. Elimination of indifferent individuals is Debreu’s (1960) separability. Full comparability amounts to both constant relative and constant absolute risk aversion and, jointly with separability, generates the linear representation. %} Maskin, Eric (1978) “A Theorem on Utilitarianism,” Review of Economic Studies 45, 93–96. {% %} Maskin, Eric (1979) “Decision Making under Ignorance with Implications for Social Choice,” Theory and Decision 11, 319–337. {% losses from prior endowment mechanism: discuss it in footnote 4, p. 189. Consider choices between loss-prospects, and find some deviations from expected utility when there are small-probability losses. Argue that, in view of such deviations, policy decisions based on expected utility can be wrong. Do not use prospect theory to analyze it. %} Mason, Charles F., Jason F. Shogren, Chad Settle, & John A. List (2005) “Investigating Risky Choices over Losses Using Experimental Data,” Journal of Risk and Uncertainty 31, 187–215. {% Gives common psychophysical measurement methods. Noted that upward matching gives different results than downward. %} Massaro, Dominic W. (1975) “Experimental Psychology and Information Processing.” Rand McNally, Chicago. {% Survey of neural networks, suited for mathematicians. %} Masson, Egill & Yih-Jeou Wang (1990) “Introduction to Computation and Learning in Artificial Neural Networks,” European Journal of Operational Research 47, 1–28. {% %} Matheson, James E. & Robert L. Winkler (1976) “Scoring Rules for Continuous Probability Distributions,” Management Science 22, 1087–1096. {% CBDT; %} Matsui, Akihiko (2000) “Expected Utility and Case-Based Reasoning,” Mathematical Social Sciences 39, 1–12. {% This paper considers additive conjoint measurement for a preference relation on a product set X1 x … x Xn. It assumes that every Xj is endowed with an operation oj. It imposes the usual Hölder-type axioms to get an additive representation uj for every oj. Then additive representation u1(x1) + … + un(xn) can be obtained the same way as p1x1 + … + pnxn is axiomatized by the de Finetti additivity type axiom (a ~ b ==> a + c ~ b + c) where now addition is in terms of oj; i.e., each aj of de Finetti is replaced by uj(xj) and so on. This is Definition 5. %} Matsushita, Yutaka (2010) “An Additive Representation on the Product of Complete, Continuous Extensive Structures,” Theory and Decision 69, 1–16. {% Pp. 57-58 on Sébastien le Prestre de Vauban (1633-1707, French military engineer, politically influential and writer on many topics including forestry: “This vision notwithstanding, Vauban recognized that few prorietors could afford to wait decades — lifetimes, even, depending on the tree’s type and intended purpose — before realizing a return on their investment. Fewer still would embark on what might only amount to ancestral largesse compared with the annual returns from grain or even coppices. He resigned himself to hoping that landowners would “do their best,” while conceding that plantations were really “an activity of the King,” for only the crown had the authority and incentive to cultivate timber of the long term.” %} Matteson, Kieko (2015) “Forests in Revolutionary France: Conservation, Community, and Conflict, 1669-1848.” Cambridge University Press, Cambridge UK. {% revealed preference; related to paper Hans Peters and me. %} Matzkin, Rosa L. (1991) “Axioms of Revealed Preference for Nonlinear Choice Sets,” Econometrica 59, 1779–1786. {% %} Matzkin, Rosa L. & Marcel K. Richter (1991) “Testing Strictly Concave Rationality,” Journal of Economic Theory 53, 287–303. {% %} Maule, A. John, G. Robert J. Hockey, & Larissa Bdzola (2000) “Effects of Time-Pressure on Decision Making under Uncertainty: Changes in Affective State and Information Processing Strategy,” Acta Psychologica 104, 283–301. {% This paper shows one thing: [rewriting lotteries by collapsing outcomes should not affect evaluation] implies EU-maximization. %} Maxwell Christopher (1990) “Decision Weights and the Normal Form Axiom,” Economics Letters 32, 211–215. {% free-will/determinism: free will seems to rule out determinism but also does not sit well with chance. %} May, Joshua (2014) “On the very Concept of Free Will,” Synthese 191, 2849–2866. {% Argues for Intransitive %} May, Kenneth O. (1954) “Intransitivity, Utility, and the Aggregation of Preference Patterns,” Econometrica 22, 1–13. {% Argues for Intransitive %} May, Regine M. (1987) “Realismus von Subjektiven Warscheinlichkeiten: Eine Kognition-Psychologische Analyse Inferentieller Prozesse beim Overconfidence.” Peter Lang, Frankfurt am Main. {% Formulate it in context of multi-criteria decision making. P. 298 1/3: that capacity is exponentially complex. Considers a form of ordinal information, with only finitely many preferences expressed, and then characterizes 2-additive capacities. Surprisingly, belief functions can be captured by a 2-additive capacity. %} Mayag, Brice, Michel Grabisch & Christophe Labreuche (2011) “A Representation of Preferences by the Choquet Integral with Respect to a 2-Additive Capacity,” Theory and Decision 71, 297–324. {% foundations of statistics; has ensuing discussion; DOI: 10.1214/13-STS457 %} Mayo, Deborah G. (2014) “On the Birnbaum Argument for the Strong Likelihood Principle,” Statistical Science 29, 227–239. {% foundations of statistics; %} Mayo, Deborah G. & Aris Spanos (2006) “Severe Testing as a Basic Concept in a Neyman–Pearson Philosophy of Induction,” Philosophy of Science 57, 323–357. {% foundations of statistics: c ollection of discussions of Bayesian versus classical statistics. %} Mayo, Deborah G. & Aris Spanos (2012) “Error and the Growth of Experimental Knowledge.” University of Chicago Press, Chicago. {% proper scoring rules: they do not only use the properness condition of de Finetti in terms of preferences (they call this pragmatic) but also an epistemic criterion, referring to distance from true measure in some sense; may be distance from true state of nature. They get impossibility results for sets of priors extending preceding results in the literature. %} Mayo-Wilson, Conor & Gregory Wheeler (2016) “Scoring Imprecise Credences: A Mildly Immodest Proposal,” Philosophy and Phenomenological Research 93, 55–78. {% Seems to have proposed hyperbolic discounting over the interval [t,t+d] (time t and duration d) as (1 + kt)/(1 + k(t+d)). %} Mazur, James E. (1987) “An Adjusting Procedure for Studying Delayed Reinforcement.” In Michael L. Commons, James E. Mazur, John A. Nevin, & Howard Rachlin (eds.) Quantitative Analyses of Behavior 5, 55–73, Lawrence Erlbaum, Hillsdale NJ. {% Analyzes saving behavior of family, by relating its risk aversion and prudence to that of its members. Paper shows that, paradoxically, insurance component of risk sharing can raise saving, and that increased prudence of one individual can lower family prudence and, hence, household saving. Hara utility plays an important role, with paradoxes avoided iff all members have same HARA. %} Mazzocco, Maurizzo (2004) “Saving, Risk Sharing, and Preferences for Risk,” American Economic Review 94, 1169–1182. {% Assume expected utility with HARA utility, and also intertemporal separability and separability between consumption and leisure. Show that assumption of homogenous risk preferences can lead astray. Do empirical testing in rural India. Eficient risk sharing is rejected in villages, but accepted in castes. Mazzocco, Maurizio & Shiv Saini (2012) “Testing Efficient Risk Sharing with Heterogeneous Risk Preferences,” American Economic Review 102, 428–468. {% %} McCabe, Kevin, Daniel Houser, Lee Ryan, Vernon Smith, & Theodore Trouard (2001) “A Functional Imagining Study of Cooperation in Two-Person Reciprocal Exchange,” Proceedings of the National Academy of Sciences 98, 11832–11835. {% probability communication: graphical ways to communicate small probabilities. %} McCaffery, Kirsten J., Ann Dixon, Andrew Hayen, Jesse Jansen, Sian Smith, & Judy M. Simpson (2012) “The Influence of Graphic Display Format on the Interpretations of Quantitative Risk Information among Adults with Lower Education and Literacy: A Randomized Experimental Study,” Medical Decision Making 32, 532–544. {% %} McCaffery Edward J., Daniel Kahneman, & Matthew L. Spitzer (1995) “Framing the Jury: Cognitive Perspectives on Pain and Suffering Awards,” Virginia Law Review 81, 1341–1420. {% %} MacCallum, Robert C., Shaobo Zhang, Kristopher J. Preacher, & Derek D. Rucker (2002) “On the Practice of Dichotomization of Quantitative Variables,” Psychological Methods 7, 19–40. {% proper scoring rules: Theorem 1: imagine a forecaster reports subjective probabilities q = (q1, …, qn) of events E1,…,En, and gets paid fi(q), where forecaster wants to maximize subjective expected value w.r.t. subjective probabilities p1,…,pn. Then f is a proper scoring rule, giving qj = pj in the optimum, if and only if fj(q) is the partial derivative w.r.t. qi of a convex function f(q) that is homogeneous of the first degree. %} McCarthy, John (1956) “Measures of the Value of Information,” Proceedings of the National Academy of Sciences 42, 654–655. {% %} McCarthy, John & Patrick J. Hayes (1969) “Some Philosophical Problems from the Standpoint of Artificial Intelligence.” In Bernard Meltzer & Donald Michie (eds.) Machine Intelligence Vol. 4, 463–502, Edinburgh University Press, Edinburgh, UK. {% %} McCauley, Clark, Nathan Kogan, & Allan I. Teger (1971) “Order Effects in Answering Risk Dilemmas for Self and Others,” Journal of Personality and Social Psychology 20, 423–424. {% survey on belief measurement: survey of calibration; follow-up of Lichtenstein, Fischhoff, & Phillips (1982). %} McClelland, Alastair & Fergus Bolger (1994) “The Calibration of Subjective Probabilities: Theories and Models 1980–1994.” In George Wright & Peter Ayton (eds.) Subjective Probability, 453–481, Wiley, New York. {% Risk averse for gains, risk seeking for losses: for small losses in insurance framework, people are risk neutral for moderate probabilities, for small probabilities some (25% for $4, 15% for $40) ignore the risk but most become risk averse. small probabilities: seem to show that there are two types of persons, one type fully ignoring small probabilities and the other overweighting them. Nice reference for this point. %} McClelland, Gary H., William D. Schulze, & Don L. Coursey (1993) “Insurance for Low-Probability Hazards: A Bimodal Response to Unlikely Events,” Journal of Risk and Uncertainty 7, 35–51. {% dynamic consistency %} McClennen, Edward F. (1983) “Sure-Thing Doubts.” In Bernt P. Stigum & Fred Wendstøp (eds.) Foundations of Utility and Risk Theory with Applications, 117–136, Reidel, Dordrecht. {% dynamic consistency; discusses resolute choice. P. 100/101 describes sophisticated choice. This paper is, to the best of my knowledge, the first to introduce resolute choice. He says that, if prior agent did planning, then posterior agent prefers following that because of the very fact of prior planning. P. 103: “For such agents, the ex post situation is different from what it would have been if there had been no ex ante resolve.” Proposes that because of that the prior agent can get it his way in the dynamic Allais paradox, that Ulysses can sail past the Syrens without extraneous things such as being tied up by his men. Final paragraph suggests that not only prefs but also consequences themselves, can have been changed as a result of the very fact of prior planning; i.e., that prior planning can be an attribute of a consequence. Also discusses prisoner’s dilemma but I will not discuss that here. %} McClennen, Edward F. (1985) “Prisoner’s Dilemma and Resolute Choice.” In Richmond Campbell & Lanning Sowden (eds.) Paradoxes of Rationality and Cooperation, 94–104, University of British Columbia Press, Vancouver. {% dynamic consistency: favors abandoning forgone-event independence, so, favors resolute choice, mostly in context of prisoners dilemma where it is part of the defended cooperative solution. It is argued that by cooperating the opponent is also made to cooperate so that it is really for higher monetary benefits that one is resolute and cooperative. The term context-sensitive preferences (e.g. §6) and the text show that McClennen thinks, à la Machina, that preferences at some moment depend on counterfactual forgone events. Argues on p. 110/11 that resoluteness can do the same, endogenously, as precommitment, but cheaper. §11 discusses forgone-branch independence (often called consequentialism) and deliberately wants to deviate from it. %} McClennen, Edward F. (1988) “Constrained Maximization and Resolute Choice,” Social Philosophy and Policy 5, 95–118. {% dynamic consistency %} McClennen, Edward F. (1988) “Dynamic Choice and Rationality.” In Bertrand R. Munier (ed.) “Risk, Decision and Rationality,” 517–536, Reidel, Dordrecht. {% dynamic consistency: favors abandoning forgone-event independence, so, favors resolute choice Describe a.o. history of ?independence? in Chs 3 and ??, Par.3.5 and Chrs. 7,8 tell about role of forgone-branch independence with descriptions of contributions by Ramsey and others (Chernoff?) de novo tree (cut off prehistory); normal form tree (prior choice, choose from strategies) Separability of McClennen = consequentialism of Machina = what I like to call forgone-branch independence dynamic consistency + consequentialism of McClennen = dynamic consistency of Machina Myopic: SEP + CON of McClennen, not dynamic consistency Sophisticated (Strotz schijnt ’t): SEP + DC of McClennen, not CON Resolute: DC + CON of McClennen, not SEP Cubitt (1996) mentions “NEC” (normal-extensive coincidence), suggesting it is vague because normal and extensive have not been defined, but suggesting it comprises reduction + Machina-DC (minus Cubitt-DC?) %} McClennen, Edward F. (1990) “Rationality and Dynamic Choice: Foundational Explorations.” Cambridge University Press, Cambridge. {% foundations of statistics; important criticism; That people look too much at statistical significance and ignore substantive significance. That, for large samples, one can detect with high significance a minor and fully unimportant difference, gives nice historical examples, e.g., Meehl (1970) with 55,000 high-school students where about everything correlated with everything significantly. Closing sentence of §III: “The siren song of “significance” is a hazard to navigation.” %} McCloskey, Donald N. (1985) “The Loss function has Been Mislaid: The Rhetoric of Significance Tests,” American Economic Review, Papers and Proceedings 75, 201–205. {% Discusses (claimed) misunderstandings of Coase’s intentions with his theorem. %} McCloskey, Deirdere (1998) “Other Things Equal: The So-Called Coase Theorem,” Eastern Economic Journal 24, 367–371. {% real incentives/hypothetical choice: for time preferences; Seem to use dated checks/vouchers; use random incentive system with one choice per person played for real. Choices with only future rewards involve only cortex, the analytic part of our brains. Choices with one present and one future reward involve both cortex and limbic system; latter is emotional part of brains that we share with virtually all animals. For - (quasi-hyperbolic) model, it is argued that concerns lymbic system and the cortex. N = ? If they do more difficult choices then visual and motoric parts of brains do not become more active than for simple choices, but analytic parts do. DC = stationarity in very explicit and annoying manner. P. 504 2nd para: “It is well accepted that rationality entails treating each moment of delay equally, thereby discounting according to an exponential function” %} McClure, Samuel M., David I. Laibson, George F. Loewenstein, & Jonathan D. Cohen (2004) “Separate Neural Systems Value Immediate and Delayed Monetary Rewars,” Science 306, 503–507. {% %} McCord, Mark R. & Richard de Neufville (1983) “Empirical Demonstration that Expected Utility Analysis is Not Operational.” In Bernt P. Stigum & Fred Wendstøp (eds.) Foundations of Utility and Risk with Applications, Reidel, Dordrecht. {% utility elicitation; p. 281 states Raiffa’s 1961 argument that a normative theory can be useful only if it sometimes !deviates! from actual behavior, but in a way expressing that the authors don’t like the argument. risky utility u = strength of preference v (or other riskless cardinal utility, often called value): p. 295 observes that differences between utility and value are of same magnitude as various utility functions assessed in different ways . %} McCord, Mark R. & Richard de Neufville (1983) “Fundamental Deficiency of Expected Utility Decision Analysis.” In Simon French, Roger Hartley, Lyn C. Thomas, & Douglas J. White (eds.) Multi-Objective Decision Making, 279–305, Academic Press, New York. {% utility elicitation; Use 10 specialist subjects. Inductively defining xj+1 ~ (p,xj; 1p, 0) they calculate vNM utilities under SEU. Utilities depend on p, rejecting SEU. The higher p, the higher the utility. %} McCord, Mark R. & Richard de Neufville (1984) “Utility Dependence on Probability: An Empirical Demonstration,” Large Scale Systems 6, 91–103. {% utility elicitation; discrepancies between utility elicitations are greatly reduced if certain outcomes and, therefore, the certainty effect are avoided. %} McCord, Mark R. & Richard de Neufville (1985) “Assessment Response Surface: Investigating Utility Dependence on Probability,” Theory and Decision 18, 263–285. {% utility elicitation; recommend not using ~ (p:, 1p:) but (, 1:c) ~ (p:, (1p): , 1:c) but for utility elicitation, to avoid the certainty effect. Find that otherwise utilities depend on probability used in elicitation. Officer & Halter (1968) argued before (e.g. bottom of p. 259) that a method that does not invoke riskless gambles (called “Ramsey method” in their paper) is better. Davidson, Suppes, & Siegel (1957) did the same, to improve on Mosteller & Nogee (1951) who had used sure outcomes. A nice theoretical follow-up is Cerreia-Vioglio, Dillenberger, & Ortoleva (2015). They show that their cautious utility model holds iff M&d always give more risk aversion %} McCord, Mark R. & Richard de Neufville (1986) “ “Lottery Equivalents”: Reduction of the Certainty Effect Problem in Utility Assessment,” Management Science 32, 56–60. {% Participants have inconsistencies between choosing and ranking. When confronted with it, all participants wanted to correct. %} MacCrimmon, Kenneth R. (1968) “Descriptive and Normative Implications of the Decision-Theory Postulates.” In Karl H. Borch & Jan Mossin (eds.) Risk and Uncertainty, 3–23, St. Martin’s Press, New York. {% As there existed almost no experimental papers in those days, the authors set their own standards for what an experimental paper is supposed to do. They set their standards high, leading to an impressive comprehensive test of virtually all relevant preference conditions related to EU. P. 370 Rule 19: be ambiguity averse for large stakes, but ambiguity seeking for small. Here ambiguity attitude is outcome dependent. second-order probabilities to model ambiguity: p. 379 P. 380: do Ellsberg with slightly higher outcomes for ambiguous events, to rule out indifference natural sources of ambiguity: p. 382: “Our general interest, though, is how people treat real situations of uncertainty. … To obtain some information about this, we included the two stock price bets corresponding to the earlier MacCrimmon study, i.e., X´: the price of Pierce Industries goes down (x´) or does not go down (x;). P. 390: Newcomb’s problem; find that Ellsberg paradox induces more violations of EU than Allais paradox. %} MacCrimmon, Kenneth R. & Stig Larsson (1979) “Utility Theory: Axioms versus “Paradoxes” .” In Maurice Allais & Ole Hagen (eds.) Expected Utility Hypotheses and the Allais Paradox, 333–409, Reidel, Dordrecht. {% %} MacCrimmon, Kenneth R. & David M. Messick (1976) “A Framework for Social Motives,” Behavioral Science 2l, 86–l00. {% Have theory of random preference; extensively discussed by Butler & Loomes (2007, AER). %} MacCrimmon, Kenneth R. & Maxwell Smith (1986) “Imprecise Equivalences: Preference Reversals in Money and Probability.” University of Columbia Working Paper 1211. {% Nice brief didactical paper on which statistical tests to use. %} McCrum-Gardner, Evie (2008) “Which is the Correct Statistical Test to Use?,” British Journal of Oral and Maxillofacial Surgery 46, 38–41. {% A Bonetti paper has argued against the systematic prohibition of deception. These authors argue in favor of such a prohibition. %} McDaniel, Tanga & Chris Starmer (1998) “Experimental Economics and Deception: A Comment,” Journal of Economic Psychology 19, 403–409. {% %} McDaniels, Timothy L. (1995) “Using Judgment in Resource Management: A Multiple Objective Analysis of a Fisheries Management Decision,” Operations Research 43, 415–426. {% natural-language-ambiguity: seems to argue that tolerance of ambiguity (in general natural-language sense) is truly related to individual personality traits rather than a situation-dependent/content-specific expression of psychological stress. %} MacDonald, A.P. Jr. (1970) “Revised Scale for Ambiguity Tolerance: Reliability and Validity,” Psychological Reports 26, 791–798. {% %} MacDonald, Don H, John H. Kagel, & Raymond C. Battalio (1991) “Animals’ Choices over Uncertain Outcomes: Further Experimental Results,” Economic Journal 101, 1065–1084. {% error theory for risky choice %} McFadden, Daniel L. (1974) “Conditional Logit Analysis of Qualitative Choice Behavior.” In Paul Zarembka (ed.) Frontiers of Econometrics, 105–142, Academic Press, New York. {% error theory for risky choice %} McFadden, Daniel L. (1976) “Quantal Choice Analysis: A Survey,” Annals of Economic and Social Measurement 5, 363–390. {% error theory for risky choice; good reference on representative agent model %} McFadden, Daniel L. (1981) “Econometric Models of Probabilistic Choice.” In Charles F. Manski & Daniel L. McFadden (eds.) Structural Analysis of Discrete Data and Econometric Applications, 198–272, MIT Press, Cambridge, MA. {% P. 97: a large reference list on WTP (“value non-use public goods”) and its discrepancies P. 98: “…arbitragers are pervasive only in a limited number of highly organized markets, such as financial markets.” P. 98: second-most expensive wine is the one mostly sold. P. 99: “Economics needs to catch up to marketing to understand the extent to which the mix and presentation of products reflects anomalies in consumer behavior.” P. 110, concluding sentence, on constructive preference: “Then, careful attention to the processes that consumers use to define tasks … and construct preferences …may allow one to look behind the superficial errors to uncover stable principles, attitudes, and preferences upon which a new economic analysis might be built.” (See also p. 97.) %} McFadden, Daniel L. (1999) “Rationality for Economists?,” Journal of Risk and Uncertainty 19, 73–105. {% %} McFadden, Daniel L. (2001) “Economic Choices,” American Economic Review 91, 351–378. {% Z&Z & paternalism/Humean-view-of-preference: end of paper, §VI, will discuss the privatization of Medicare in the US starting Jan 01 2006, and an empirical investigation into consumer choices. The first five sections discuss that people often don’t take optimal decisions because of the many biases, and to what extent they need assistance, referring to libertarian paternalism of Thaler & Sunstein (2003). P. 12 has nice citation of owner of restaurant who, when told to reposition his wine list so as to increase profits based on biases, replied: “tell me something I didn’t learn in hotel school.” Download 7.23 Mb. Share with your friends:
|