Optimal Risk Sharing with Optimistic and Pessimistic Decision Makers

Similar documents
Could Nash equilibria exist if the payoff functions are not quasi-concave?

The Existence of Equilibrium in. Infinite-Dimensional Spaces: Some Examples. John H. Boyd III

Documents de Travail du Centre d Economie de la Sorbonne

Multidimensional inequalities and generalized quantile functions

Choice under Uncertainty

Notes on Recursive Utility. Consider the setting of consumption in infinite time under uncertainty as in

Regularity of competitive equilibria in a production economy with externalities

About partial probabilistic information

Competitive Equilibria in a Comonotone Market

Existence of Equilibria with a Tight Marginal Pricing Rule

Chapter 1 - Preference and choice

Random sets. Distributions, capacities and their applications. Ilya Molchanov. University of Bern, Switzerland

Risk Measures in non-dominated Models

Dual utilities under dependence uncertainty

Zero sum games Proving the vn theorem. Zero sum games. Roberto Lucchetti. Politecnico di Milano

Choquet Integral and Its Applications: A Survey

Patryk Pagacz. Equilibrium theory in infinite dimensional Arrow-Debreu s model. Uniwersytet Jagielloński

Journal of Mathematical Economics. Coherent risk measures in general economic models and price bubbles

Lectures for the Course on Foundations of Mathematical Finance

Financial asset bubble with heterogeneous agents and endogenous borrowing constraints

g 2 (x) (1/3)M 1 = (1/3)(2/3)M.

Robust Comparative Statics of Non-monotone Shocks in Large Aggregative Games

Intertemporal Risk Aversion, Stationarity, and Discounting

Multivariate comonotonicity, stochastic orders and risk measures

Noncooperative Games, Couplings Constraints, and Partial Effi ciency

A SHORT INTRODUCTION TO BANACH LATTICES AND

Inverse Stochastic Dominance Constraints Duality and Methods

Anti-comonotone random variables and Anti-monotone risk aversion

On Kusuoka Representation of Law Invariant Risk Measures

Pareto Optimal Allocations for Law Invariant Robust Utilities

Elimination of Arbitrage States in Asymmetric Information Models

Ambiguity Aversion and Trade

Documents de Travail du Centre d Economie de la Sorbonne

When does aggregation reduce risk aversion?

Measurable Ambiguity. with Wolfgang Pesendorfer. August 2009

(c) For each α R \ {0}, the mapping x αx is a homeomorphism of X.

Near-Potential Games: Geometry and Dynamics

Aliprantis, Border: Infinite-dimensional Analysis A Hitchhiker s Guide

Asymptotics of minimax stochastic programs

On a Class of Multidimensional Optimal Transportation Problems

Reward-Risk Portfolio Selection and Stochastic Dominance

Conditional and Dynamic Preferences

A Note on Robust Representations of Law-Invariant Quasiconvex Functions

Convex Analysis and Economic Theory AY Elementary properties of convex functions

Lecture 1. Stochastic Optimization: Introduction. January 8, 2018

Near-Potential Games: Geometry and Dynamics

Functional Analysis. Martin Brokate. 1 Normed Spaces 2. 2 Hilbert Spaces The Principle of Uniform Boundedness 32

Integral Jensen inequality

of the set A. Note that the cross-section A ω :={t R + : (t,ω) A} is empty when ω/ π A. It would be impossible to have (ψ(ω), ω) A for such an ω.

Decision-making with belief functions

Valid Inequalities and Restrictions for Stochastic Programming Problems with First Order Stochastic Dominance Constraints

Exhaustible Resources and Economic Growth

Equilibrium Theory under Ambiguity

SOME REMARKS ON THE SPACE OF DIFFERENCES OF SUBLINEAR FUNCTIONS

Decomposability and time consistency of risk averse multistage programs

Economics 201B Second Half. Lecture 12-4/22/10. Core is the most commonly used. The core is the set of all allocations such that no coalition (set of

Portfolio optimization with stochastic dominance constraints

Mathematics II, course

PROBLEMS. (b) (Polarization Identity) Show that in any inner product space

Aggregate Risk. MFM Practitioner Module: Quantitative Risk Management. John Dodson. February 6, Aggregate Risk. John Dodson.

WORKING PAPER NO. 443

Generalized quantiles as risk measures

Market Equilibrium and the Core

Decision principles derived from risk measures

Introduction to Functional Analysis

Allais, Ellsberg, and Preferences for Hedging

Columbia University. Department of Economics Discussion Paper Series. The Knob of the Discord. Massimiliano Amarante Fabio Maccheroni

Lecture Notes October 18, Reading assignment for this lecture: Syllabus, section I.

Lecture 2. We now introduce some fundamental tools in martingale theory, which are useful in controlling the fluctuation of martingales.

(1) Consider the space S consisting of all continuous real-valued functions on the closed interval [0, 1]. For f, g S, define

In the Name of God. Sharif University of Technology. Microeconomics 1. Graduate School of Management and Economics. Dr. S.

Strategic Uncertainty and Equilibrium Selection in Discontinuous Games.

Social Welfare Functions for Sustainable Development

Bayesian Persuasion Online Appendix

MULTIPLE-BELIEF RATIONAL-EXPECTATIONS EQUILIBRIA IN OLG MODELS WITH AMBIGUITY

Theoretical Foundation of Uncertain Dominance

Robust Comparative Statics for Non-monotone Shocks in Large Aggregative Games

Representation theorem for AVaR under a submodular capacity

RISK AND RELIABILITY IN OPTIMIZATION UNDER UNCERTAINTY

MULTI-BELIEF RATIONAL-EXPECTATIONS EQUILIBRIA: INDETERMINACY, COMPLEXITY AND SUSTAINED DEFLATION

Core equivalence and welfare properties without divisible goods

Lower semicontinuous and Convex Functions

Certainty Equivalent Representation of Binary Gambles That Are Decomposed into Risky and Sure Parts

Finite-dimensional spaces. C n is the space of n-tuples x = (x 1,..., x n ) of complex numbers. It is a Hilbert space with the inner product

The Ohio State University Department of Economics. Homework Set Questions and Answers

Differentiable Welfare Theorems Existence of a Competitive Equilibrium: Preliminaries

B. Appendix B. Topological vector spaces

MATH 426, TOPOLOGY. p 1.

Persuading a Pessimist

Strictly monotonic preferences on continuum of goods commodity spaces

TWO MAPPINGS RELATED TO SEMI-INNER PRODUCTS AND THEIR APPLICATIONS IN GEOMETRY OF NORMED LINEAR SPACES. S.S. Dragomir and J.J.

The newsvendor problem with convex risk

A NICE PROOF OF FARKAS LEMMA

Contents: 1. Minimization. 2. The theorem of Lions-Stampacchia for variational inequalities. 3. Γ -Convergence. 4. Duality mapping.

Random Process Lecture 1. Fundamentals of Probability

Notes on Distributions

January 29, Introduction to optimization and complexity. Outline. Introduction. Problem formulation. Convexity reminder. Optimality Conditions

USING FUNCTIONAL ANALYSIS AND SOBOLEV SPACES TO SOLVE POISSON S EQUATION

Second Welfare Theorem

Choice under uncertainty with the best and worst in mind: Neo-additive capacities

Transcription:

Optimal Risk Sharing with Optimistic and Pessimistic Decision Makers A. Araujo 1,2, J.M. Bonnisseau 3, A. Chateauneuf 3 and R. Novinski 1 1 Instituto Nacional de Matematica Pura e Aplicada 2 Fundacao Getulio Vargas 3 Paris School of Economics, Université Paris 1 Panthéon-Sorbonne Centre d Economie de la Sorbonne, UMR CNRS 8174 Economic Theory Conference University of Kansas December 9-10, 2011 Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 1

Outline 1 Introduction 2 3 4 Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 2

Outline 1 Introduction 2 3 4 Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 3

Introduction Main motivation: we aim at proving that under mild conditions individually Pareto rational optima will exist even in presence of non-convex preferences. We essentially consider preferences on l + hence accommodating the behavior of DM in front of a countable flow x of payoffs, or else choosing among financial assets x whose outcomes depend on the realization of a countable set of states of the world. Therefore, our conditions for the existence of P.O. can be interpreted as myopia in the first context and as pessimism or reasonable optimism in the second context. Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 4

Introduction (continued) Replacing the uncertain context by a risky one, we then derive some necessary properties of risk sharing. For strong risk averters these conditions are exactly the usual ones of pairwise comonotonicity, for strong risk lovers exclusively corner conditions or pairwise comonotonicity can appear. A crucial result for our work is the proof that the closed unit ball B(0, 1) for the normed dual l of the normed space l 1 is compact in the Mackey topology τ(l, l 1 ). Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 5

Outline 1 Introduction 2 3 4 Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 6

Theorem Let B(S) be the normed space of bounded real mappings on S endowed with the sup-norm. Then: Corollary 1) B(S) is the norm dual of the space rca(s) of all regular and bounded Borel measures for S endowed with the discrete topology. 2) The closed unit ball of B(S) is compact for the Mackey topology τ(b(s), rca(s)). The closed unit ball B(0, 1) for the normed dual l of the normed space l 1 is compact in the Mackey topology τ(l, l 1 ). Proof. Taking S = N, clearly l = B(S). Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 7

Outline 1 Introduction 2 3 4 Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 8

Existence of individually rational Pareto efficient allocations under Mackey upper-semicontinuous preferences Assumption (1) For every i = 1,..., m, i is a transitive reflexive preorder on l +. (ii) For every i and every y i l +, {x i l + x i i y i } is τ-closed where τ is the Mackey topology τ(l, l 1 ). Theorem Under Assumptions (1) and (2), individually rational Pareto efficient allocations exist. Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 9

Examples of Mackey upper-semicontinuous preferences Definition Let (S, A) be a measurable space and B (A) be the space of bounded A-measurable mappings from S to R. v is a capacity on A if v( ) = 0, v(s) = 1, A, B A, A B v(a) v(b). x B (A), the Choquet integral of x with respect to v, denoted xdv is defined by 0 (v(x t) 1)dt + + 0 v(x t)dt. Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 10

>From Araujo et al. General Equilibrium, Wariness and efficient bubbles", JET forthcoming, we know that preferences on l represented by a Choquet integral is Mackey usc if and only if is strongly myopic where: Definition A preference relation on l, is strongly myopic if (x, y) l l x y, z l +, implies that there exists n large enough such that x y + z En where z En (m) = 0 if m < n and z(m) if m n. Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 11

Pessimistic Mackey upper-semicontinuous preferences Definition A Choquet DM will be said pessimistic if v is convex, i.e., A, B A, v(a B) + v(a B) v(a) + v(b). Recall that if v is convex, then xdv = min{e P (x) P additive probability measure v} Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 12

Proposition For a pessimistic decision maker, the following assertions are equivalent: (1) is Mackey upper-semicontinuous; (2) v is outer continuous, i.e., A n, A A = P(N), A n A v(a n ) v(a). Example I(x) = inf N {x(n)} or equally I(x) = xdv where v(a) = 0 A N, v(n) = 1 Remark that I is not Mackey lower semicontinuous. Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 13

Optimistic Mackey upper-semicontinuous preferences Definition A Choquet DM will be said optimistic if v is concave, i.e., A, B A, v(a B) + v(a B) v(a) + v(b). Recall that if v is concave, xdv = max{e P (x) P additive probability measure v} or else x, y B (A), x y, α ]0, 1[ αx + (1 α)y y. Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 14

Proposition For an optimistic decision maker the following assertions are equivalent: (1) is Mackey upper-semicontinuous; (2) is Mackey lower-semicontinuous; (3) v is outer continuous at, i.e., A n P(N), A n v(a n ) 0. Remark. I(x) = sup N {x(n)}, i.e., I(x) = xdv with v(a) = 1 A, v( ) = 0 is not Mackey upper-semicontinuous. Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 15

Example where no existence of nontrivial individually rational Pareto optima if such a too optimistic" DM is present Consumption space X = l +, p s = 1 2. s DM1 U 1 (x) = s N p sx s, initial endowments ω 1 = (2 1 s+1 ) s N DM2 U 2 (x) = sup s N x s, ω 2 = ω 1. The unique Pareto optimal allocation is ( x 1 = ω 1 + ω 2, x 2 = 0). No individually rational Pareto optimal allocation. Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 16

Outline 1 Introduction 2 3 4 Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 17

for risk lovers and risk averse decision makers Interpret l as the set of all bounded A measurable mappings x from N to R where A = P(N) and assume that a σ-additive probability P is given on (N, P(N)). So any x can now be interpreted as a random variable. Definition x is less risky than y for the second order stochastic dominance denoted x SSD y if t P(x u)du t P(y u)du for all t R. x is strictly less risky than y denoted x SSD y if one of the previous inequality is strict. Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 18

Definition Let i be the preference relation of a DM. i is a strict (strong) risk averter if x SSD y x i y and x SSD y x i y. i is a strict (strong) risk lover if x SSD y x i y and x SSD y x i y. Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 19

Examples of Mackey upper-semicontinuous strict risk lover of strict risk averter Consider a Choquet decision maker where v = f P, i.e., v is a distortion of a probability P, i.e., f : [0, 1] [0, 1], f (0) = 0, f (1) = 1, f strictly increasing. If f is strictly concave and continuous, then x l I(x) = xdv f with v f = f P defines a Mackey upper-semicontinuous strict risk lover. If f is strictly convex and continuous, then x l I(x) = xdv f with v f = f P defines a Mackey upper-semicontinuous strict risk averter. Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 20

The main result Theorem Consider m strict risk averters i = 1,..., m with Mackey upper-semicontinuous preferences i and n strict risk lovers j = 1,..., n with Mackey upper-semicontinuous preferences j with initial endowments ω i l ++, ω j l ++. Then individual rational Pareto efficient allocations (in (l + ) m+n ) exist. Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 21

Theorem continued Theorem For such PO (x i, i = 1,..., m; y j, j = 1,..., n) we have: 1) The allocations of risk averters are pairwise comonotonic, i.e., (x i1 (s) x i1 (t))(x i2 (s) x i2 (t)) 0 (s, t) N 2, (i 1, i 2 ) {1,..., m} 2. 2) For two risk lovers j 1, j 2 : (y j1 (s) y j1 (t))(y j2 (s) y j2 (t)) < 0 (y j1 (s) y j1 (t))(y j2 (s) y j2 (t)) = 0 (i.e. in at least one state, one of these two risk lovers gets 0). 3) If all allocations of risk lovers are strictly positive, then these allocations are pairwise comonotonic. Araujo, Bonnisseau, Chateauneuf and Novinski Risk Sharing with Diverse Decision Makers 22