Games and Economic Behavior最新文献

We introduce behavioral diversity to an otherwise standard signaling model, in which a fraction of agents choose their signaling actions according to an exogenous distribution. These behavioral agents provide opportunities for strategic low-type agents to successfully emulate higher types in equilibrium, which in turn reduces the cost for strategic high-type agents to separate from lower types. Behavioral diversity thus improves the equilibrium payoffs to all types of strategic agents. The model also exhibits a convergence property, which is intuitively more appealing than the least-cost separating equilibrium of the standard setting.

We consider a set of agents who have claims on an endowment that is not large enough to cover all claims. Agents can form coalitions but a minimal coalition size θ is required to have positive coalitional funding that is proportional to the sum of the claims of its members. We analyze the structure of stable partitions when coalition members use well-behaved rules to allocate coalitional endowments, e.g., the well-known constrained equal awards rule (CEA) or the constrained equal losses rule (CEL). For continuous, (strictly) resource monotonic, and consistent rules, stable partitions with (mostly) θ-size coalitions emerge. For CEA and CEL we provide algorithms to construct such a stable partition formed by (mostly) θ-size coalitions.

The stable marriage and stable roommates problems have been extensively studied due to their applicability in various real-world scenarios. However, it might happen that no stable solution exists, or stable solutions do not meet certain requirements. In such cases, one might be interested in modifying the instance so that the existence of a stable outcome with the desired properties is ensured. We focus on three different modifications.

1. In the stable roommates problem, we show that finding a smallest subset of agents whose removal results in an instance with a stable matching is NP-complete if the capacities are greater than one, or the deleted agents must belong to a fixed subset of vertices. We further show that analogous results hold for the stable marriage problem when one would like to achieve the existence of a stable and perfect matching through the deletion of vertices.

2. We investigate how to modify the preferences of the agents as little as possible so that a given matching becomes stable. The deviation of the new preferences from the original ones can be measured in various ways; here, we concentrate on the ${\ell }_1$-norm. We show that, assuming the Unique Games Conjecture, the problem cannot be approximated within a factor smaller than 2. By relying on bipartite-submodular functions, we give a polynomial-time algorithm for the bipartite case. We also show that a similar approach leads to a 2-approximation for general graphs.

3. Last, we consider problems where the preferences of agents are not fully prescribed, and the goal is to decide whether the preference lists can be completed so that a stable matching exists. We settle the complexity of several variants, including cases when some of the edges are required to be included or excluded from the solution.

This paper investigates the dynamics of product adoption under incomplete information regarding the product quality. A new agent observes a small sample of product choices within the population and receives a noisy private signal regarding the realized state of the world. Using simple heuristics, the agent estimates both the product quality and the distribution of product choices in the population. The agent then chooses a product that is a best response to this estimate. We show that fast adoption of the optimal product in the population occurs if and only if the strength of the private signal exceeds a certain threshold.

The literature on punishment and prosocial behavior has presented conflicting findings. In some settings, punishment crowds out prosocial behavior and backfires; in others, however, it promotes prosociality. We examine whether the punisher's motives can help reconcile these results through a novel experiment in which the agent's outcomes are identical in two environments, but in one the pre-emptive punishment scheme is self-serving (i.e., potentially benefits the punisher), while in the other it is other-regarding (i.e., potentially benefits a third party). We find that self-serving punishment reduces the social stigma of selfish behavior, while other-regarding punishment does not. Self-serving punishment is thus less effective at encouraging compliance and is more likely to backfire. We further show that the normative message is somewhat weaker when punishment is less costly for the punisher. Our findings have implications for the design of punishment mechanisms and highlight the importance of the punisher's motives in expressing norms.

This paper studies the implications of agents signaling their moral type in a lying game. In the theoretical analysis, a signaling motive emerges where agents dislike being suspected of lying and where some lies are more stigmatized than others. The equilibrium prediction of the model can explain experimental data from previous studies, particularly on partial lying, where individuals lie to gain a non-payoff maximizing amount. I discuss the relationship with theoretical models of lying that conceptualize the image concern as an aversion to being suspected of lying and provide applications to narratives, learning, the disclosure of lies, and the selection into lying opportunities.

An election with full turnout is supposed to achieve an outcome that perfectly reflects the majority's preference. This result requires voters to be perfectly informed about their preferences and to vote accordingly. I show that incentivizing participation with an abstention fine does not necessarily incentivize information acquisition. While a small abstention fine always increases information acquisition compared to Voluntary Voting, a high abstention fine that achieves full turnout increases information acquisition only if voting costs are high. If voting costs are low, the opposite is true: Less individuals acquire information under Compulsory Voting with full turnout than under Voluntary Voting.

Five decades after Harsanyi and Selten's seminal work on equilibrium selection, we remain unable to predict the outcomes of real-life coordination even in simple cases. One reason is that experiments have struggled to quantify the effects of payoff- and risk-dominance and to separate them from context factors like feedback, repetition, and complexity. This experiment is the first to demonstrate that both payoff- and risk-dominance significantly and independently impact coordination decision-making. Three innovations characterize the design: First, payoff- and risk-dominance are disentangled using orthogonal measures of strategic incentives and welfare externalities. Second, a no-feedback, choice-list task format minimizes deviations from one-shot incentives. Third, beliefs about others' behavior are elicited. Strikingly, heterogeneous beliefs across the population rationalize not only reactions to risk dominance but also most reactions to payoff dominance. In addition, deviations from expected-value maximization in specific games suggest a minor role for social projection or other-regarding preferences.

We study a fair division setting in which participants are to be fairly distributed among teams, where not only do the teams have preferences over the participants as in the canonical fair division setting, but the participants also have preferences over the teams. We focus on guaranteeing envy-freeness up to one participant (EF1) for the teams together with a stability condition for both sides. We show that an allocation satisfying EF1, swap stability, and individual stability always exists and can be computed in polynomial time, even when teams may have positive or negative values for participants. When teams have nonnegative values for participants, we prove that an EF1 and Pareto optimal allocation exists and, if the valuations are binary, can be found in polynomial time. We also show that an EF1 and justified envy-free allocation does not necessarily exist, and deciding whether such an allocation exists is computationally difficult.