A Non-Bayesian Theory of State-Dependent Utility

This study analyzes TV program content for viewers with respect to the following three properties: primary utility (intrinsic value by genre, cast, and so on), macro externality, such as the effect of others' past behavior (audience ratings), and state dependence, or the viewer's past actions that determine whether to watch the program. Focusing on the effects of macro externality and state-dependence on viewer behavior, the authors construct a hierarchical Bayesian model and derive an individual utility function. Applying their method to the choice of TV programs in Japan, they estimate the effects of macro externality and state dependence on an individual's viewing behavior, and illustrate the distribution of these effects. The authors find that macro externality has a strong overall effect on the total number of viewers and the state-dependence effect has a slightly positive effect. Moreover, comparing the attributes revealed that male viewers tend to show a lower effect of ratings and higher state dependence than females. They find that the older the viewers are, the higher the effect of state dependence. Further, using utility function data obtained by our estimation, they conduct a simulation to measure the dynamic effects of macro externality and state dependence.

Time Aggregation and State Dependence in Welfare Receipt

Price Adjustments in a General Model of State-Dependent Pricing

We formalize a consumption–investment–insurance problem with the distinction of a state-dependent relative risk aversion. The state dependence refers to the state of the finite state Markov chain that also formalizes insurable risks such as health and lifetime uncertainty. We derive and analyze the implicit solution to the problem, compare it with special cases in the literature, and illustrate the range of results in a disability model where the relative risk aversion is preserved, decreases, or increases upon disability.

The proliferation of novel preference theories in financial economics is hampered by a lack of non-experimental evidence and by the theories’ additional complexity which has not been shown to be critical in applications. In this article I present arguments in support of preferences with rank dependency. Using the Survey of Consumer Finances data, I document two widespread patterns inconsistent with expected utility: (i) many households simultaneously invest in well-deversified funds and in poorly-diversified portfolios of stocks; and (ii) some households with substantial savings do not invest anything in equities. I show that portfolio choice models with rank-dependent preferences, plausibly parameterized and under fully rational assumptions, are quantitatively consistent with the observed diversification. These results call for further efforts to integrate the models of rank-dependent preferences in portfolio theory and asset pricing. The decision theoretic literature has long acknowledged the shortcomings of the expected utility theory of Von Neumann and Morgenstern (1944). Several generalizations of the theory are inspired by the experimental evidence and aimed at improving its descriptive accuracy. However, there are few applications of new models of preferences in finance, and expected utility continues to dominate the field. 1 Two principal hurdles slow the proliferation of novel preferences in financial economics: first, the lack of empirical (non-experimental or ‘‘field’’) data confirming that the behavior motivating non-expected utilities is widespread outside the laboratory settings; second, the additional and often significant complexity of the alternative preferences which has not been shown to be critical in applications where the models based on expected utility fall short of reasonable predictions. This article presents empirical and theoretical This paper was previously circulated under the title ‘‘Household Portfolio Diversification.’’ The paper has benefited from the comments of Shlomo Benartzi, Luca Benzoni, Markus Brunnermeier, John Dickhaut, Maria Perozek, Raj Singh, Hersh Shefrin, and workshop participants at Minnesota Finance bag lunch, 2003 Wharton conference ‘‘Household Financial Decision Making and Portfolio Choice’’ and 2003 summer meetings of Econometric Society. I would like to thank Cam Harvey (the editor) and an anonymous referee for many helpful comments and suggestions. A special thanks to Jack Kareken for comments and editorial help on an earlier draft. I am grateful to Donna Zerwitz for editorial assistance. The views expressed herein are those of the author and not necessarily of the Federal Reserve Bank of Minneapolis or the Federal Reserve System. All errors and omissions are the sole responsibility of the author. Address correspondence to Valery Polkovnichenko, Finance 3-122 CSOM, 319 19 th

We study information aggregation in a dynamic trading model. We show theoretically that separable securities, introduced by Ostrovsky in the context of Expected Utility, no longer aggregate information if some traders have imprecise beliefs and are ambiguity averse. Moreover, these securities are prone to manipulation as the degree of information aggregation can be influenced by the initial price set by the uninformed market maker. These observations are also confirmed in our laboratory experiment using prediction markets. We define a new class of strongly separable securities, which are robust to the above considerations and show that they characterize information aggregation in both strategic and non-strategic environments. We derive several testable predictions, which we are able to confirm in the laboratory. Finally, we show theoretically that strongly separable securities are both sufficient and necessary for information aggregation but, strikingly, there does not exist a security that is strongly separable for all information structures.

Quantum electrodynamics predicts that heavy atoms () will undergo the process of atomic collapse where electrons sink into the positron continuum and a new family of so-called collapsing states emerges. The relativistic electrons in graphene exhibit the same physics but at a much lower critical charge () which has made it possible to confirm this phenomenon experimentally. However, there exist conflicting predictions on the effect of a magnetic field on atomic collapse. These theoretical predictions are based on the continuum Dirac–Weyl equation, which does not have an exact analytical solution for the interplay of a supercritical Coulomb potential and the magnetic field. Approximative solutions have been proposed, but because the two effects compete on similar energy scales, the theoretical treatment varies depending on the regime which is being considered. These limitations are overcome here by starting from a tight-binding approach and computing exact numerical results. By avoiding special limit cases, we found a smooth evolution between the different regimes. We predict that the atomic collapse effect persists even after the magnetic field is activated and that the critical charge remains unchanged. We show that the atomic collapse regime is characterized: (1) by a series of Landau level anticrossings and (2) by the absence of scaling of the Landau levels with regard to magnetic field strength.

Deception is a fundamental issue across a diverse array of settings, from cybersecurity, where decoys (e.g., honeypots) are an important tool, to politics that can feature politically motivated “leaks” and fake news about candidates. Typical considerations of deception view it as providing false information. However, just as important but less frequently studied is a more tacit form where information is strategically hidden or leaked. We consider the problem of how much an adversary can affect a principal's decision by “half-truths”, that is, by masking or hiding bits of information, when the principal is oblivious to the presence of the adversary. The principal's problem can be modeled as one of predicting future states of variables in a dynamic Bayes network, and we show that, while theoretically the principal's decisions can be made arbitrarily bad, the optimal attack is NP-hard to approximate, even under strong assumptions favoring the attacker. However, we also describe an important special case where the dependency of future states on past states is additive, in which we can efficiently compute an approximately optimal attack. Moreover, in networks with a linear transition function we can solve the problem optimally in polynomial time.

Existing preference models are generally based on utility functions and study rational behaviors of humans. However human preferences are actually accompanied with irrational factors. Psychologists Tversky and Kahneman have given some examples that indicate humans are simply irrational by the standards of the utility axioms. How to depict some non-rational decision of human — that is, by what function this is done — is a crucial component of this subject of human (ir)rationality. In this paper, we apply filter and ultrafilter tools in nonstandard analysis to represent human preferences on two levels, intuitively and subconsciously. We indicate how this preference modeling using ultrafilter notion can reflect the combination of human rationality and irrationality. In particular, we show in what cases human preferences can be called non-rationality.

Chapter 10 - Preference functions and risk-adjusted performance measures

Two models of ambiguity preferences that permit comparative statics analysis of greater ambiguity aversion yield definite predictions concerning propensities for self-insurance and self-protection: The levels of both activities that are optimal for an ambiguity-averse decision maker are higher in the presence of ambiguity than in its absence, and demands for both activities increase with greater ambiguity aversion. The reason is that, at levels optimal for one decision maker, an increase in either activity results in a mean-preserving contraction in the distribution of expected utility in the presence of ambiguity, which is valuable to anyone with the same risk preferences who is more ambiguity averse.

In the spring of 2009, a novel strain of the H1N1 influenza virus, containing a never before witnessed combination of gene segments from human influenza, two forms of swine influenza, and avian influenza, 1 was declared a global pandemic. The UK Government had to decide whether to undertake, at a cost of £1.2 billion (USD 1.9 billion at the time, equivalent to 1 percent of that year’s health budget), an extensive set of preparatory measures, including the purchase of both antiviral medication and a novel vaccine in quantities sufficient to cover the entire UK population, or whether instead to take substantially less costly measures, which would involve having only a limited supply of these medicines and vaccines at hand.2 The possible.

For decisions whose consequences accrue over time, there are several possible techniques to compute total utility. One is to discount utilities of future consequences at some appropriate rate. The second is to discount per-period certainty equivalents. And the third is to compute net present values (NPVs) of various possible streams and to then apply the utility function to these net present values. We find that the best approach is to first compute NPVs of various possible income streams and then take the utility of such NPVs. We show the drawbacks of other alternative models of evaluating income streams. The article discusses the advantages of the power and logarithmic forms in the modeling of time preference. These are the only forms for which utility of income and utility of consumption are strategically equivalent. Further, these forms permit the flexibility in the choice of a time period (e.g., monthly or quarterly) without modifying the utility function, thus simplifying analysis.

Many recorded roll calls in Congress each year are votes on bills that have no chance of becoming law, or are purely symbolic, or are procedural without policy content. Yet models of voting and measurement models of member preferences make assumptions that vote choices are largely about utility derived from policies. We consider the possibility that votes plausibly connected to policy and votes not plausibly connected to policy may have different data-generating processes and rely on different utility functions. Substantively, similarity across different contexts for policy change implies an importance of messaging over policy. Methodologically, similarity across these contexts is necessary to avoid biasing estimates of member preferences. We find that members’ voting patterns are highly stable across contexts in which policy change is credible and not credible. This indicates that existing measures of ideal points are likely not dramatically biased by the inclusion of policy-irrelevant votes.

A previously reported positive influence of a distractor alternative on decision accuracy was driven by an unintended covariation between key variables, beyond which the distractor has a modest negative influence on accuracy, possibly mirroring asymmetric multiattribute context effects.