Centre d'Economie de la Sorbonne, Salle S/17
Franz Dietrich (Université Paris 1 - PSE- CNRS)
Title : Fully Bayesian Aggregation
Abstract : I study the aggregation of preferences under uncertainty by requiring social preferences to be Bayesian in the full, i.e., static and dynamic, sense: society should hold expected-utility preferences and revise them in the rational Bayesian way. I characterize the class of preference aggregation rules which generate fully Bayesian social preferences and satisfy two regularity conditions (continuity and a mild Pareto axiom). This class consists of rules which aggregate utilities linearly but aggregate beliefs geometrically.
Centre d'Economie de la Sorbonne, Salle S/17
Guillaume Vigeral (Université Paris-Dauphine)
Title : Sets of equilibrium payoffs of finite games
Abstract : It is well known that the set of equilibrium payoff of any finite $N$-player game is a nonempty semialgebraic and compact subset of $R^N$. We show that for $N\geq 3$ the converse is true : any nonempty semialgebraic and compact subset of $R^N$ is the set of equilibrium payoffs of some finite $N$-player game. The proof is constructive and the size of the constructed game is polynomial in the size of the semi-algebraic set. We give implications of this result on the complexity and computability of some problems involving Nash equilibria.
Centre d'Economie de la Sorbonne, Salle S/17
Sebastiano Della Lena (Universite of Venice)
Title : Non-Bayesian Social Learning and the Spread of Misinformation in Networks
Abstract :
People are exposed to constant flows of information; nevertheless, misinformation is ubiquitous in society. This paper studies the spread of misinformation when agents receive new information each period and update their opinions taking into account both their experience and neighborhood’s ones. I consider two types of misinformation: permanent and temporary. Permanent misinformation is modeled through stubborn agents and produces long-run effects. The distortion induced by stubborn agents in social learning depends on a novel centrality measure that identifies the key agents of a social learning process, the “updating centrality”. Conversely, temporary misinformation, represented by shocks of rumors or fake news, has only short-run effects on the opinion dynamics. Using spectral graph theoretical techniques, I show that the consensus time is increasing with respect to the “bottleneckedness” of the underlying network, while the learning time is decreasing with respect to agents’ reliance on their private signals.
Centre d'Economie de la Sorbonne, Salle S/17
Marc Barthelemy (CNRS, IPHT)
Title : Modeling car traffic in cities: effect of density and other critical factors (joint with Vincent Verbavatz)
Abstract : Car traffic in urban systems has been studied intensely in past decades but its analysis is often limited to empirical observations and agent-based modelling, and despite the importance and urgency of the problem we have a poor theoretical understanding of the parameters controlling urban car use and congestion. Here, we combine economical and transport ingredients into a statistical physics approach and propose a parsimonious model that predicts the share of car drivers, the CO2 emitted by cars and the average commuting time. We confirm these analytical predictions on 25 major urban areas in the world, and our results suggest that urban density is not the relevant variable controlling car-related quantities but rather are the city's area size and the density of public transport. Mitigating the traffic (and its effect such as CO2 emissions) can then be obtained by reducing the urbanized area size or, more realistically, by improving either the public transport density or its access. In general, increasing the population at fixed area would increase the emission of CO2 in sharp contrast with the commonly accepted paradigm that increasing the density leads to a reduction of car traffic.
Pas de Séminaire (CTN Workshop)
Centre d'Economie de la Sorbonne, Salle S/17
Markus Pivato (Université de Cergy-Pontoise)
Title : Does deliberation improve the reliability of epistemic democracy? (joint with Huihui Ding)
Abstract : We study the effects of deliberation on epistemic social choice, in two settings. In the first setting, the group faces a binary epistemic decision analogous to the Condorcet Jury Theorem. In the second setting, group members have probabilistic beliefs arising from their private information, and the group wants to aggregate these beliefs in a way that makes optimal use of this information. During deliberation, each agent discloses private information to persuade the other agents of her current views. But her views may also evolve over time, as she learns from other agents. This process will improve the performance of the group, but only under certain conditions; these involve the nature of the social decision rule, the group size, and also the presence of “neutral agents” whom the other agents try to persuade.
Vacances
Centre d'Economie de la Sorbonne, Salle S/17
Olivier Bos (Université Paris 2)
Title: Auctions with signaling concerns (joined with Tom Truyts)
Abstract:
We study a symmetric private value model with signaling, in which the auction outcome is used by an outside observer to infer the bidders’ types. We elicit conditions under which an essentially unique D1 equilibrium bidding function exists in four auction formats. We establish the revenue equivalence does not hold between these auction designs. This is because signaling affects differently the bidders’ incentives to overbid their types given the payment rule and in the presence – or not – of an increasing price clock. We show this boost more revenues in the second-price auction than in the English auction, and more in the English auction than the first-price and all-pay auctions. Applications are art auction houses and charity auctions.
Centre d'Economie de la Sorbonne, Salle S/17
Peter Sudhölter (University of Southern Denmark)
Title: Welfare egalitarianism in surplus-sharing problems and convex games (joint with Pedro Calleja and Francesc Llerena)
Abstract: We show that the constrained egalitarian surplus-sharing rule, which divides the surplus so that the poorer players' resulting payoffs become equal but not larger than any remaining player's status quo payoff, is characterized by Pareto optimality, path independence, both well-known, and less first (LF), requiring that a player does not gain if her status quo payoff exceeds that of another player by the surplus. This result is used to show that, on the domain of convex games, Dutta-Ray's egalitarian solution is characterized by aggregate monotonicity (AM), bounded pairwise fairness, resembling LF, and the bilateral reduced game property (2-RGP) à la Davis and Maschler. We show that 2-RGP can be replaced by individual rationality and bilateral consistency à la Hart and Mas-Colell. We prove that the egalitarian solution is the unique core selection that satisfies AM and bounded richness, requiring that the poorest players cannot be made richer within the core. Replacing “poorest” by “poorer” allows to eliminate AM.
Centre d'Economie de la Sorbonne, Salle S/17
Olivier Beaude (EDF R&D)
Title: Applications of game-theory in a changing electricity system
Abstract: Historically, optimization tools have been largely applied to design and operate electricity systems. A typical problem is the so-called "Unit-Commitment". It consists in scheduling at a minimal cost electricity production units to satisfy the equality constraint between production and consumption at all time. Now, the context in electricity systems is rapidly changing, leading to more competition between – newly created – stakeholders, a "more local" management of the production-consumption constraint, and stochastic aspects with the integration of – local – renewable energy sources with very uncertain generation. In this context, many new problems can be tackled using the tools of game theory. This talk will propose a few of such illustrative examples, linking practical crucial issues of electricity systems operators to well-known classes of games (routing, potential, matching, auctions). This will highlight the vitality of the field of "game-theory applied to Smart Grids", and potential avenues to improve this link.
Centre d'Economie de la Sorbonne, Salle S/17
René Van den Brink (School of Business and Economics , Tibergen Institute)
Title: The Degree Ratio Ranking Method for Directed Networks (joint work with Agnieszka Rusinowska)
Abstract: One of the most famous ranking methods for digraphs is the ranking by Copeland score. The Copeland score of a node in a digraph is the difference between its outdegree (i.e. its number of outgoing arcs) and its indegree (i.e. its number of ingoing arcs). In the ranking by Copeland score a node is ranked higher, the higher is its Copeland score. In this paper, we deal with an alternative to rank nodes according to their out- and indegree, namely ranking the nodes according to their degree ratio, i.e. the outdegree divided by the indegree. In order to avoid dividing by a zero indegree, we implicitly take the out- and indegree of the reflexive digraph. We provide an axiomatization of the ranking by degree ratio using a clone neutrality which says that cloning a node (in a specific way) does not change the ranking among the original nodes. We also provide a new axiomatization of the ranking by Copeland score using the same axioms except that this ranking method satisfies a different cloning neutrality. Finally, we modify the ranking by degree ratio by not considering the reflexive digraph, and by definition assume nodes with indegree zero to be ranked higher than nodes with a positive indegree. We provide an axiomatization of this ranking by modified degree ratio using yet another clone neutrality and a maximal property. In this way, we can compare the three ranking methods by their respective clone neutrality.
Centre d'Economie de la Sorbonne, Salle S/17
Annick Laruelle (University of the Basque Country and Ikerbasque research professor)
Title: “Not this one” Experimental use of the approval and disapproval ballot
Abstract: In the last few decades there has been a decrease in electoral turnout in most democratic countries. It is crucial to understand why citizens do not vote. One of the many explanations for abstaining may be that voters are not offered the possibility to express dissatisfaction explicitly. The right to say no to candidates in elections is vindicated and tested. A field experiment with a ballot that allows voters to vote against candidates was conducted during the 2017 French presidential elections. The experiment demonstrates that participants appreciate the additional expressive possibilities offered by the experimental ballot, in particular the possibility of casting a vote against candidates. Moreover a richer picture of the political landscape is obtained. If this ballot were used, clearer signals could be sent to political actors. The utilization of this ballot in two popular consultations (also reported in the paper) proves that it could be easily implemented in practice. The experiment and the popular consultations offer empirical evidence for debate on social choice. In theory it is easy to show differences in outcomes due to different voting rules, but it is not clear that these differences occur in real voting situations.
JEL classification: D72, C93
Keywords: Two-round electoral systems; In situ voting experiment; Expressions of dissatisfaction; Approval voting.
Vacances
Vacances
Centre d'Economie de la Sorbonne, Salle S/17
Dino Borie (THEMA, Université Cergy-pontoise )
Title: Description-dependent Choices (joint with Dorian Jullien)
Abstract: We propose a theory of choice that accounts for the class of framing effects where a preference reversal is due to different descriptions of the same alternatives. These framing effects violate the principle of “description invariance” which is implicit in the standard theory of choice. We make this principle
explicit by taking descriptions of alternatives (instead of alternatives per se) as the primary objects of choice. We then explore several ways to relax description invariance to see the different preferences that are generated and we pay specific attention to the role played by the Weak Axiom of Revealed Preference in description-invariant and description-dependent choices. Finally, we extend our theory under risk to show how it can generate risk attitudes that are dependent on descriptions.
Centre d'Economie de la Sorbonne, Salle S/17
Maël Letreust (Université Cergy-Pontoise, ENSEA, CNRS )
Title: Persuasion with limited communication capacity (joint work with Tristan Tomala, HEC Paris)
Abstract: We consider a Bayesian persuasion problem where the persuader and the decision maker communicate through an imperfect channel which has a fixed and limited number of messages and is subject to exogenous noise. Imperfect communication entails a loss of payoff for the persuader. We establish an upper bound on the payoffs the persuader can secure by communicating through the channel. We also show that the bound is tight: if the persuasion problem consists of a large number of independent copies of the same base problem, then the persuader can achieve this bound arbitrarily closely by using strategies which tie all the problems together. We characterize this optimal payoff as a function of the information-theoretic capacity of the communication channel.
Centre d'Economie de la Sorbonne, Salle S/17
No Seminar
Centre d'Economie de la Sorbonne, Salle S/17
Clara Macedo (Université Paris 1)
Title: Multiplier Stabilization Applied to Two-Stage Stochastic Programs
Abstract:
In many mathematical optimization applications dual variables are an important output of the solving process, due to their role as price signals. When dual solutions are not unique, different solvers or different computers, even different runs in the same computer if the problem is stochastic, often end up with different optimal multipliers. From the perspective of a decision maker, this variability makes the price signals less reliable and, hence, less useful. We address this issue for a particular family of linear and quadratic programs by proposing a solution procedure that, among all possible optimal multipliers, systematically yields the one with the smallest norm. The approach, based on penalization techniques of nonlinear programming, amounts to a regularization in the dual of the original problem. As the penalty parameter tends to zero, convergence of the primal sequence and, more critically, of the dual is shown under natural assumptions. The methodology is illustrated on a battery of two-stage stochastic linear programs and the example of the hydro-energy in North Europe.
Centre d'Economie de la Sorbonne, Salle S/17
Ron Peretz (Bar Ilan University)
Title: The Speed of Innovation Diffusion in Social Networks
Abstract :
New technologies typically gain a foothold through the actions of a few innovators, and then diffuse more rapidly as more and more people come into contact with prior adopters. We model this process as a network game and derive topology-free bounds on the expected waiting time until a given fraction of the population has adopted the innovation. The bounds depend on the payoff gain from using the innovation instead of the status quo, and on the noisiness of the players' response functions, but they do not depend on the network structure per se. In particular, the bounds hold for directed and undirected networks of arbitrary size whose structure may be evolving over time.
Centre d'Economie de la Sorbonne, Salle S/17
Anna Bogomolnaia (University of Glasgow- National Research University Higher School of Economics)
Title: Fair division of a random object (joint with Herve Moulin and Fedor Sanodmirskiy )
Abstract:
Objects arrive randomly, and must be allocated on the spot between a fixed set of beneficiaries. We explore the tradeoff between the concerns for Fairness and the utilitarian measure of Efficiency:
Even a Bayesian manager (who knows in each period the full probability distribution of utility profiles) faces a steep (and well known) Price of Fairness: in the worst case implementing Fair Share allows her to capture only a O((1/(√n))) fraction of the efficient surplus.
A Prior-free manager who only knows the expectation of individual utilities in each period (or just the ratios of these expectations), but neither the actual distribution in any period, nor the number of periods, ensures Fair Share by the simple Proportional rule: agents get the object with probabilities proportional to their (normalized) utilities (or disutilities).
We define the equally simple one-dimensional family of Top Heavy rules and show that they capture the optimal tradeoffs between fairness and efficiency: any other prior-free rule meeting Fair Share is less efficient ex post (for every realization of utilities) than one of our rules. In particular the Proportional rule is substantially less efficient than one of our rules. Moreover the Top Heavy rules pay the same Price of Fairness as the best Bayesian rules.
Centre d'Economie de la Sorbonne, Salle S/17
Mariya Teteryatnikova (Higher School of Economics - University of Moscow)
Title: Cautious Farsighted Stability in Network Formation Games with Streams of Payoffs
Abstract:
We propose a new notion of farsighted pairwise stability for dynamic network formation which includes two notable features: consideration of intermediate payoffs and cautiousness. This differs from existi ng concepts which typically consider either only immediate or nalpayoffs, and which often require that players are optimistic in any environment without full communication and commitment. For arbitrary (and possibly heterogeneous) preferences over the process of network formation, a non-empty cautious path stable (CPS) set of networks always exists. Strongly efficient networks do not always belong to a CPS set but a Pareto dominant network, if it exists, is the unique prediction of the CPS. Furthermore, some general relationships exist between CPS and other farsighted concepts.
Centre d'Economie de la Sorbonne, Salle 114
Cancelled
Centre d'Economie de la Sorbonne, Salle 114
Rim Lahmandi-Ayed (ESSAI and UR MASE-ESSAI, Université de Carthage. )
Title: When Do Imperfectly Competitive Firms Maximize Profits ? The Lessons from A Simple General Equilibrium Model with Shareholders' Voting. (joint work with Didier Laussel)
Abstract:
We consider a general equilibrium model with vertical preferences, where workers and consumers are differentiated respectively by their sensitivity to effort and their intensity of preference for quality. We consider a monopoly the shares of which are owned by a fraction of the general population. The price is determined through a vote among all the shareholders. We identify necessary and sufficient conditions for (i) an absolute (relative) majority to vote for the profit maximizing price; (ii) an absolute (relative) majority to vote for a different price. We argue that the more concentrated the ownership the more likely it is that the firm charges the profit-maximizing price.
Centre d'Economie de la Sorbonne, Salle 114
Ugo Bolletta (Université Aix-Marseille)
Title: A model of peer effects in school
Abstract: When designing interventions aiming to foster peer effects in schools, knowledge on the endogenous sorting of individuals into groups is key. We propose a theoretical model where agents form groups endogenously, and their outcomes are affected accordingly. Using a popular payoff structure, we show that equilibrium outcomes have a direct correspondence with the linear-in-means model, used to study empirically peer effects and we characterize the set of stable networks. The model can be fitted with real data, providing the tools to i) infer the network in case we do not have this information, ii) match theoretical outcomes with real ones, to infer the motives behind network formation and iii) simulate effects of interventions that manipulate the composition of classrooms. Simulating data we show that the model can consistently reproduce the results of Carrell et al. (2013)
Centre d'Economie de la Sorbonne, Salle 114
Ana Carolina Tereza (Insper Institute of Research and Teaching )
Title: Updating Variational (Bewley) Preferences
Abstract: In this paper first we study the problem of characterizing an update rule for the class of variational Bewley preferences. We show that under the standard dynamic consistency the updated preference still in the same class of preferences and the corresponding ambiguity index follows a generalization of the full Bayesian update. Then we study the problem of updating preferences when the decision maker is characterized by two binary relations. Given the empirical evidence for incomplete preference in the sense of variational Bewley preferences, the first unconditional relation satisfies a set of behavioral conditions consis- tent with this class of preferences. Motivated by the fact that on many occasions, the decision maker must be able to compare any pair of acts (forced choices), we investigate the problem of finding an ex post complete, transitive, continuous and monotone prefer- ence that preserves the ambiguity attitude of the original unconditional one. We find that the ex post relation must be given by a variational preference represented by the same affine utility function and the ambiguity index follows the same updating rule of variational Bewley preference. This extension is unique and characterizes a weak completion of the unconditional ex ante preference. Also, our result can be viewed as a novel foundation for the full Bayesian updating of variational preferences.
No seminar
Centre d'Economie de la Sorbonne, Salle 115
Raimunda Saona (Universidad Chile)
Title: When to stop searching in the future something we hope to be better?
Abstract:
How to apply price discrimination to a stream of customers? When should one stop searching for job offers and take the one which is now available? In this talk we will present a setting where we can accurately answer these questions, a performance metric for every strategy and deduce the optimal one. How good does this strategy performs? We still do not know, upper and lower bounds are presented, analizing simpler strategies to solve this problem.
To be more precise we revise the "prophet secretary problem", where a principal must asign an item to only one of n possible agents, who has different valorization for it. The principal knows the distribution of each valorization, but not the value. Moreover, it's sequentially revealed the actual valorization of each agent to the principal in uniform random order, but he must decide the owner of the item online, ie: he can give the item only to the last agent who showed his valorization. The problem consists in maximizing the expected value of the social welfare, comparing it with the expectation of the maximum of all valorization.
Centre d'Economie de la Sorbonne, Salle 115
Laura Kasper (Maastricht University and Saarland University)
Title: Farsighted Rationality (joint work with Dominik Karos)
Abstract: An abstract game consists of a set of states and specifies what coalitions are allowed to replace one state by another one. Agents are called farsighted if they compare the status quo to the long term outcome following their deviation rather than to the status they actually deviate to. What the literature has ignored so far is that if a coalition does not move out of the status quo, they might still expect another coalition to do so. Specifically, above definition of farsightedness implies that agents ignore this possibility in their reasoning. So, in fact, agents are not fully farsighted. Expectation functions assign to each state a (potentially) different state and a coalition that moves from the former to the latter, thereby creating paths between states. This endows agents with an expectation about what any potential deviation entails---namely the path of the prescribed further moves. We extend these functions by capturing coalitions' expectations about the consequences of not moving out of a state. We impose three stability and optimality axioms on extended expectation functions that reflect full farsightedness and rationality. We then show that an expectation function satisfies our axioms if and only if it can be associated with a non-cooperative equilibrium of the abstract game. % that is stable with respect to coalitional deviations. We finally apply our solution to games in characteristic function form and matching problems.
Centre d'Economie de la Sorbonne, Salle 115
Gaëtan Fournier (Université Aix-Marseille)
Title: Spatial competition and dispersion of sellers
Abstract: A location game is a competition between players who select a location in order to attract as many consumers as possible. Buyers obviously prefer when sellers are dispersed, but it is not always the case: during the talk I describe different models and show what are the important ingredients to observe differentiation between players.
Centre d'Economie de la Sorbonne, Salle 114
Title: Political violence and social instability
Abstract: We provide a model of protests and revolutions within a social network. The equilibrium concept is closely related to Peer Confirming Equilibrium meaning that each agent best reacts to the strategies of her neighbours and uses the latter to gather information about other parts of the network. When we allow protesters to randomly coordinate, the probability that an anti-government protest turns into a successful revolution is higher under more repressive regimes. This is true for arbitrary social networks with heterogeneous agents. The implications of the provided model are illustrated using data on protests, revolutions, and political terror worldwide between 1976 and 2014.
Centre d'Economie de la Sorbonne, Salle 114
Title: Some strategic foundations of preferential attachment (with Xavier Venel)
Abstract: There exists a wide gap between game-theoretic models and empirical observations of the dynamics of network formation. Game theoretic models offer a detailed representation of individuals’incentives and behaviors but they predict the emergence of much simpler structures than these observed empirically. Random network formation processes, such as preferential attachment, provide a much better fit to empirical observations but generally lack micro-foundations. In order to bridge this gap, we focus on extensive games as models of network formation and characterize which strategic settings give rise to equilibrium dynamics consistent with random network formation processes. In general, if there is imperfect information and some form competition between players, these can have incentives to introduce a stochastic dimension in their link formation strategies. In this perspective we analyze a class of games of network formation through sequential competition in which a player competes with his predecessor and his successor for the benefits induced by socio-economic relationships. Such games can represent a number of strategic economic interactions such as oligopolistic competition in supply networks or diffusion of influence in opinion networks. We show that the focal equilibrium that emerge in this setting is one where players use probability distributions with full support and target the whole network with probabilities inversely proportionally to the utility of each node. Notably, when the utility of a node is inversely proportional to its degree, equilibrium play induces a preferential attachment process.
Centre d'Economie de la Sorbonne, Salle 115
Arnaud Dragicevic (IRSTEA)
Title: Exchange Networks with Stochastic Matching
Abstract: We consider an exchange network with stochastic matching between the pairs of players and analyze the dynamics of bargaining in such a market. The cases of homogeneous expectations, heterogeneous expectations and of social preferences are studied. The results show that, in all three cases, the dynamics converge to the solution concept of balanced outcome or Nash bargaining solution, which is an equilibrium that combines notions of stability and fairness. In the first two scenarios, the numerical simulations reveal that the convergence toward a fixed point is achieved at the value of the outside option. In the third scenario, the fixed point converges to the value of the outside option supplemented by the surplus share.
Centre d'Economie de la Sorbonne, Salle 115
Alexandre Voisin et Pierre Vallois (Université de Lorraine)
Title: Intégrale de Choquet avec des entrées aléatoires
Abstract: On s'intéresse à l'intégrale de Choquet par rapport à une mesure floue donnée et avec $n$ entrées aléatoires. On suppose que ces $n$ composantes sont des variables aléatoires indépendantes et admettant la même densité de probabilité. Dans ces conditions, on montre que l'intégrale de Choquet (qui est une variable aléatoire) a également une densité que l'on calcule explicitement. On a aussi déterminé ses moments d'ordre un et deux. De plus des simulations ont été réalisées avec un exemple de capacité donnée et des entrées de lois gaussiennes, uniforme et exponentielle afin de comparer les différents résultats.