Remarks on Utility in Repeated Bets

Nimrod Megiddo

2023-06-07

Abstract:The use of von Neumann -- Morgenstern utility is examined in the context of multiple choices between lotteries. Different conclusions are reached if the choices are simultaneous or sequential. It is demonstrated that utility cannot be additive.

Artificial Intelligence,Probability

What problem does this paper attempt to address?

Expected multi-utility representations of preferences over lotteries

Paolo Leonetti

2024-01-13

Abstract:Let $\succsim$ be a binary relation on the set of simple lotteries over a countable outcome set $Z$. We provide necessary and sufficient conditions on $\succsim$ to guarantee the existence of a set $U$ of von Neumann--Morgenstern utility functions $u: Z\to \mathbf{R}$ such that $$ p\succsim q \,\,\,\Longleftrightarrow\,\,\, \mathbf{E}_p[u] \ge \mathbf{E}_q[u] \,\text{ for all }u \in U $$ for all simple lotteries $p,q$. In such case, the set $U$ is essentially unique. Then, we show that the analogue characterization does not hold if $Z$ is uncountable. This provides an answer to an open question posed by Dubra, Maccheroni, and Ok in [J. Econom. Theory~\textbf{115} (2004), no.~1, 118--133]. Lastly, we show that different continuity requirements on $\succsim$ allow for certain restrictions on the possible choices of the set $U$ of utility functions (e.g., all utility functions are bounded), providing a wide family of expected multi-utility representations.

Classical Analysis and ODEs,Theoretical Economics,Functional Analysis
Expected Utility from a Constructive Viewpoint

Kislaya Prasad

DOI: https://doi.org/10.48550/arXiv.2303.08633

2024-02-27

Abstract:This paper introduces a space of variable lotteries and proves a constructive version of the expected utility theorem. The word ``constructive'' is used here in two senses. First, as in constructive mathematics, the logic underlying proofs is intuitionistic. In a second sense of the word, ``constructive'' is taken to mean ``built up from smaller components.'' Lotteries as well as preferences vary continuously over some topological space. The topology encodes observability or verifiability restrictions -- the open sets of the topology serve as the possible truth values of assertions about preference and reflect constraints on the ability to measure, deduce, or observe. Replacing an open set by a covering of smaller open sets serves as a notion of refinement of information. Within this framework, inability to compare arises as a phenomenon distinct from indifference and this gives rise to the constructive failure of the classical expected utility theorem. A constructive version of the theorem is then proved, and accomplishes several things. First, the representation theorem uses continuous real-valued functions as indicators of preference for variable lotteries and these functions reflect the inability to compare phenomenon. Second, conditions are provided whereby local representations of preference over open sets can be collated to provide a global representation. Third, the proofs are constructive and do not use the law of the excluded middle, which may not hold for variable lotteries. Fourth, a version of the classical theorem is obtained by imposing a condition on the collection of open sets of the topology which has the effect of making the logic classical.

Theoretical Economics,Logic
Some Reflections on Utility

Oskar Morgenstern

DOI: https://doi.org/10.1007/978-94-015-7629-1_6

1979-01-01

Abstract:The von Neumann-Morgenstern ‘expected utility hypothesis’ as presented in Chapter I of The Theory of Games and Economic Behavior (1944) was the first attempt at a rigorous axiomatization of a fundamental concept of economics, namely, a general theory of utility. There is no a priori notion of ‘rationality’ in the theory; rather, that is a concept to be derived and given meaning to from the theory. The theory is ‘absolutely convincing’ in the sense that if behavior deviates from that predicated by the theory, after explanation of the theory, behavior would be adjusted. The von Neumann-Morgenstern ‘expected utility theory’ was the first approximation to an undoubtedly much richer and far more complicated reality. Critics of the theory, such as Allais and others, have centered their criticism of the theory not on the internal consistency of the axioms of which the theory is composed but rather on the validity of the axioms themselves. In that sense, the attacks on the theory are misplaced. It is common knowledge throughout scientific circles that axiomatic systems are often modified as axioms are more precisely specified, qualified, or, at times, replaced, as our understanding of physical laws is expanded. Not unexpectedly, the evolution of the von Neumann-Morgenstern ‘expected utility theory’ over the past thirty years is in complete concordance with scientific progress.
The same but different: an empirical investigation of the reducibility principle

David V. Budescu,Ilan Fischer

DOI: https://doi.org/10.1002/bdm.372

2001-01-01

Journal of Behavioral Decision Making

Abstract:According to classical utility theory the valuation of any lottery should depend only on its outcomes and their respective probabilities, and should not be affected by the nature, complexity and structure of the chance mechanism. Previous research has documented systematic violations of this principle. For example, most subjects prefer lotteries in which the payoffs are contingent on the joint occurrence of multiple (high probability) events to simple lotteries, and lotteries in which the earlier stages offer higher probabilities than the later stages. We review the various violations of this principle and suggest a classification into two major types associated with misunderstanding of chance mechanisms and attitude towards the chance mechanism and process. In the present study 40 subjects were presented with 30 pairs of binary gambles. In any given pair the lotteries had identical outcomes and equal ‘reduced’ probabilities (and thus equal expected values). However, the chance mechanisms varied along a variety of factors such as the size of the sample space, the number of stages, temporal ordering, order of probabilities, their transparency and time constraints. Half the subjects saw lotteries involving gains and the other half considered only losses. After choosing one lottery in each pair, the subjects were asked to explain and justify their choices. The findings revealed systematic violations of the reducibility principle: subjects displayed a preference for lotteries with larger sample spaces, and for lotteries that allow quicker resolutions in the earlier stages. A clear distinction between some patterns of preferences in the gains and loss domains was revealed. In gambles involving gains subjects preferred to have the highest probability on the first stage (and the lowest probability on the last stage), but displayed the opposite preferences for losses. A content analysis of the subjects' stated reasons for their choices identified eight major categories. The most frequently invoked were hope, fun, simplicity, stress and time. Copyright © 2001 John Wiley & Sons, Ltd.

psychology, applied
Large compound lotteries

Zvi Safra,Uzi Segal

DOI: https://doi.org/10.1016/j.jmateco.2024.102994

IF: 0.747

2024-05-24

Journal of Mathematical Economics

Abstract:Extending preferences over simple lotteries to compound (two-stage) lotteries can be done using two different methods: (1) using the Reduction of compound lotteries axiom, under which probabilities of the two stages are multiplied; (2) using the compound independence axiom, under which each second stage lottery is replaced by its certainty equivalent. Except for expected utility preferences, the rankings induced by the two methods are always in disagreement and deciding on which method to use is not straightforward. Moreover, sometimes each of the two methods may seem to violate some kind of monotonicity. In this paper we demonstrate that, under some conditions, the disagreement disappears in the limit and that for (almost) any pair of compound lotteries, the two methods agree if the second stage lotteries are replicated sufficiently many times.

economics,social sciences, mathematical methods,mathematics, interdisciplinary applications
Combinatorial auctions with decreasing marginal utilities

Benny Lehmann,Daniel Lehmann,Noam Nisan

DOI: https://doi.org/10.1016/j.geb.2005.02.006

IF: 1.265

2006-05-01

Games and Economic Behavior

Abstract:In most of microeconomic theory, consumers are assumed to exhibit decreasing marginal utilities. This paper considers combinatorial auctions among such submodular buyers. The valuations of such buyers are placed within a hierarchy of valuations that exhibit no complementarities, a hierarchy that includes also OR and XOR combinations of singleton valuations, and valuations satisfying the gross substitutes property. Those last valuations are shown to form a zero-measure subset of the submodular valuations that have positive measure. While we show that the allocation problem among submodular valuations is NP-hard, we present an efficient greedy 2-approximation algorithm for this case and generalize it to the case of limited complementarities. No such approximation algorithm exists in a setting allowing for arbitrary complementarities. Some results about strategic aspects of combinatorial auctions among players with decreasing marginal utilities are also presented.

economics
Risk Preferences in Time Lotteries

Yonatan Berman,Mark Kirstein

DOI: https://doi.org/10.48550/arXiv.2108.08366

2021-08-18

Theoretical Economics

Abstract:An important but understudied question in economics is how people choose when facing uncertainty in the timing of events. Here we study preferences over time lotteries, in which the payment amount is certain but the payment time is uncertain. Expected discounted utility theory (EDUT) predicts decision makers to be risk-seeking over time lotteries. We explore a normative model of growth-optimality, in which decision makers maximise the long-term growth rate of their wealth. Revisiting experimental evidence on time lotteries, we find that growth-optimality accords better with the evidence than EDUT. We outline future experiments to scrutinise further the plausibility of growth-optimality.
Expected utility theory with probability grids and preference formation

Mamoru Kaneko

DOI: https://doi.org/10.1007/s00199-019-01225-4

2019-08-28

Economic Theory

Abstract:Abstract We reformulate expected utility theory, from the viewpoint of bounded rationality, by introducing probability grids and a cognitive bound; we restrict permissible probabilities only to decimal ( $$\ell $$ l -ary in general) fractions of finite depths up to a given cognitive bound. We distinguish between measurements of utilities from pure alternatives and their extensions to lotteries involving more risks. Our theory is constructive from the viewpoint of the decision maker. When a cognitive bound is small, the preference relation involves many incomparabilities, but these diminish as the cognitive bound is relaxed. Similarly, the EU hypothesis would hold more for a larger bound. The main part of the paper is a study of preferences including incomparabilities in cases with finite cognitive bounds; we give representation theorems in terms of a 2-dimensional vector-valued utility functions. We also exemplify the theory with one experimental result reported by Kahneman and Tversky.

economics
Utility Theory for Sequential Decision Making

Mehran Shakerinava,Siamak Ravanbakhsh

DOI: https://doi.org/10.48550/arXiv.2206.13637

2022-06-28

Abstract:The von Neumann-Morgenstern (VNM) utility theorem shows that under certain axioms of rationality, decision-making is reduced to maximizing the expectation of some utility function. We extend these axioms to increasingly structured sequential decision making settings and identify the structure of the corresponding utility functions. In particular, we show that memoryless preferences lead to a utility in the form of a per transition reward and multiplicative factor on the future return. This result motivates a generalization of Markov Decision Processes (MDPs) with this structure on the agent's returns, which we call Affine-Reward MDPs. A stronger constraint on preferences is needed to recover the commonly used cumulative sum of scalar rewards in MDPs. A yet stronger constraint simplifies the utility function for goal-seeking agents in the form of a difference in some function of states that we call potential functions. Our necessary and sufficient conditions demystify the reward hypothesis that underlies the design of rational agents in reinforcement learning by adding an axiom to the VNM rationality axioms and motivates new directions for AI research involving sequential decision making.

Artificial Intelligence,Computer Science and Game Theory,Machine Learning
Mixture independence foundations for expected utility

Jingyi Meng,Craig S. Webb,Horst Zank

DOI: https://doi.org/10.1016/j.jmateco.2023.102938

IF: 0.747

2024-01-12

Journal of Mathematical Economics

Abstract:An alternative preference foundation for expected utility is provided. Our segregated approach considers four logically independent implications of the classic von Neumann-Morgenstern independence axiom. The monotonicity principle is, for a transitive relation, equivalent to monotonicity with respect to first-order stochastic dominance. Rank-dependent separability is similar to the comonotonic sure-thing principle used in the ambiguity literature. The remaining two properties are weak formulations of the independence principle which invoke the latter only for probability mixtures with the extreme, that is the best, respectively, the worst outcome. These four implications of independence, together with completeness, transitivity and continuity of a preference relation, characterize expected utility. Furthermore, if rank-dependent separability is dropped, expected utility still holds on each subset of three-outcomes lotteries that give positive probability to both the best and the worst outcomes.

economics,social sciences, mathematical methods,mathematics, interdisciplinary applications
Utility maximization in models with conditionally independent increments

Jan Kallsen,Johannes Muhle-Karbe

DOI: https://doi.org/10.48550/arXiv.0911.3608

2009-11-19

Abstract:We consider the problem of maximizing expected utility from terminal wealth in models with stochastic factors. Using martingale methods and a conditioning argument, we determine the optimal strategy for power utility under the assumption that the increments of the asset price are independent conditionally on the factor process.

Portfolio Management,Optimization and Control,Probability,Computational Finance
Probabilistic Outcomes Are Valued Less in Expectation, Even Conditional on Their Realization

Gabriele Paolacci,Quentin André

DOI: https://doi.org/10.1287/mnsc.2021.02284

IF: 5.4

2023-12-31

Management Science

Abstract:Most theories of decision making under risk assume that payoffs and probabilities are separable. In the context of a lottery, the subjective value of a prospective outcome (the payoff) is assumed to be independent of the likelihood that the outcome will occur (the probability). In violation of this assumption, we present eight experiments showing that people anticipate less utility from uncertain outcomes than from certain outcomes, even conditional on their realization. The devaluation of uncertain outcomes is observed across different measures of utility (willingness to spend money or time; choice between different options), different populations (student and online samples), and different manipulations of uncertainty. We show that this result does not simply reflect a misunderstanding of the instructions or people’s aversion toward a “weird” transaction with unexplained features. We highlight the implications of this phenomenon for empirical investigations of risk preferences and conclude with a discussion of the psychological mechanisms that might drive the devaluation of probabilistic outcomes. This paper was accepted by Yuval Rottenstreich, behavioral economics and decision analysis. Supplemental Material: The data files are available at https://doi.org/10.1287/mnsc.2021.02284 .

management,operations research & management science
The Utility Concept Applied to the Theory of Insurance

Karl Borch

DOI: https://doi.org/10.1017/s0515036100009685

1961-07-01

Astin Bulletin

Abstract:In some recent papers ((1), (2) and (3)) about reinsurance problems I have made extensive use of utility concepts. It has been shown that if a company follows well defined objectives in its reinsurance policy, these objectives can be represented by a utility function which the company seeks to maximise. This formulation of the problem will in general make it possible to determine a unique reinsurance arrangement which is optimal when the company's objectives and external situation are given. More than 50 years ago Guldberg (4) wrote (about the probability of ruin): “Wie hoch diese Wahrscheinlichkeit gegriffen werden soil, muss dent subjektiven Ermessen oder von Aussen kommenden Bedingungen überlassen bleiben”. This is the traditional approach to reinsurance problems. It does obviously not lead to a determinate solution. Most authors taking this approach conclude their studies by giving a mathematical relation between some measure of “stability”, such as the probability of ruin, and some parameter, for instance maximum retention, to which the company can give any value within a certain range. Such studies do usually not state which particular value the company should select for this parameter, i.e. what degree of stability it should settle for. This question is apparently considered as being outside the field of actuarial mathematics.

economics,social sciences, mathematical methods,mathematics, interdisciplinary applications,statistics & probability
A connection between von Neumann-Morgenstern expected utility and symmetric potential games

Mehmet S. Ismail,Ronald Peeters

DOI: https://doi.org/10.1007/s11238-024-09991-x

2024-05-22

Theory and Decision

Abstract:This paper establishes a previously unexplored connection between expected utility theory and potential games. Starting with a decision problem with a complete preference relation over lotteries on a finite set of alternatives, we construct a two-person symmetric game using a payoff function that represents the preference relation, and show that if the preference relation satisfies the von Neumann-Morgenstern expected utility axioms then the constructed game is a potential game. Conversely, starting with a two-player symmetric game, we (uniquely) construct a (complete) preference relation over the lotteries using the first player's payoffs in the game, and show that if the game is a potential game then the resulting preference relation satisfies the expected utility axioms.

economics,social sciences, mathematical methods
Convergence of utility indifference prices to the superreplication price in a multiple-priors framework

Romain Blanchard,Laurence Carassus

DOI: https://doi.org/10.48550/arXiv.1709.09465

2020-10-02

Abstract:This paper formulates an utility indifference pricing model for investors trading in a discrete time financial market under non-dominated model uncertainty. The investors preferences are described by strictly increasing concave random functions defined on the positive axis. We prove that under suitable conditions the multiple-priors utility indifference prices of a contingent claim converge to its multiple-priors superreplication price. We also revisit the notion of certainty equivalent for random utility functions and establish its relation with the absolute risk aversion.

Mathematical Finance
A remark on utility streams

Karl Schlechta

DOI: https://doi.org/10.48550/arXiv.0709.0113

2007-09-02

Logic

Abstract:We consider arbitrarily long, but finite utility streams, and some appropriate axioms.
Subjective Expected Utility and Psychological Gambles

Gianluca Cassese

2023-10-27

Abstract:We obtain an elementary characterization of expected utility based on a representation of choice in terms of psychological gambles, which requires no assumption other than coherence between ex-ante and ex-post preferences. Weaker version of coherence are associated with various attitudes towards complexity and lead to a characterization of minimax or Choquet expected utility.

General Economics
Risk and Intertemporal Preferences over Time Lotteries

Minghao Pan

DOI: https://doi.org/10.48550/arXiv.2209.01790

2022-09-05

Abstract:This paper studies relations among axioms on individuals' intertemporal choices under risk. The focus is on Risk Averse over Time Lotteries (RATL), meaning that a fixed prize is preferred to a lottery with the same monetary prize but a random delivery time. Though surveys and lab experiments documented RATL choices, Expected Discounted Utility cannot accommodate any RATL. This paper's contribution is two-fold. First, under a very weak form of Independence, we generalize the incompatibility of RATL with two axioms about intertemporal choices: Stochastic Impatience (SI) and No Future Bias. Next, we prove a representation theorem that gives a class of models satisfying RATL and SI everywhere. This illustrates that there is no fundamental conflict between RATL and SI, and leaves open possibility that RATL behavior is caused by Future Bias.

Theoretical Economics
Stochastic Monotonicity and Random Utility Models: The Good and The Ugly

Henk Keffert,Nikolaus Schweizer

2024-09-01

Abstract:When it comes to structural estimation of risk preferences from data on choices, random utility models have long been one of the standard research tools in economics. A recent literature has challenged these models, pointing out some concerning monotonicity and, thus, identification problems. In this paper, we take a second look and point out that some of the criticism - while extremely valid - may have gone too far, demanding monotonicity of choice probabilities in decisions where it is not so clear whether it should be imposed. We introduce a new class of random utility models based on carefully constructed generalized risk premia which always satisfy our relaxed monotonicity criteria. Moreover, we show that some of the models used in applied research like the certainty-equivalent-based random utility model for CARA utility actually lie in this class of monotonic stochastic choice models. We conclude that not all random utility models are bad.

General Economics,Methodology
On the dual problem of utility maximization in incomplete markets

Lingqi Gu,Yiqing Lin,Junjian Yang

DOI: https://doi.org/10.1016/j.spa.2015.10.009

2015-11-26

Abstract:In this paper, we study the dual problem of the expected utility maximization in incomplete markets with bounded random endowment. We start with the problem formulated in the paper of Cvitanić-Schachermayer-Wang (2001) and prove the following statement: in the Brownian framework, the countably additive part $Q^r$ of the dual optimizer $Q\in (L^\infty)^*$ obtained in that paper can be represented by the terminal value of a supermartingale deflator $Y$ defined in the paper of Kramkov-Schachermayer (1999), which is a local martingale.

Probability,Optimization and Control

Remarks on Utility in Repeated Bets

Expected multi-utility representations of preferences over lotteries

Expected Utility from a Constructive Viewpoint

Some Reflections on Utility

The same but different: an empirical investigation of the reducibility principle

Large compound lotteries

Combinatorial auctions with decreasing marginal utilities

Risk Preferences in Time Lotteries

Expected utility theory with probability grids and preference formation

Utility Theory for Sequential Decision Making

Mixture independence foundations for expected utility

Utility maximization in models with conditionally independent increments

Probabilistic Outcomes Are Valued Less in Expectation, Even Conditional on Their Realization

The Utility Concept Applied to the Theory of Insurance

A connection between von Neumann-Morgenstern expected utility and symmetric potential games

Convergence of utility indifference prices to the superreplication price in a multiple-priors framework

A remark on utility streams

Subjective Expected Utility and Psychological Gambles

Risk and Intertemporal Preferences over Time Lotteries

Stochastic Monotonicity and Random Utility Models: The Good and The Ugly

On the dual problem of utility maximization in incomplete markets