|
on Game Theory |
| By: | Shaun Hargreaves Heap; Mehmet Mars Seven |
| Abstract: | We study whether zero-sum decision rules, maximin and minimax, harm agents' interests in positive-sum strategic environments relative to Nash equilibrium behavior or, more generally, than best response behaviour. Contrary to an influential evolutionary view, we give illustrations where maximin serves an agent's interests better than Nash equilibrium behaviour. We also show that these illustration are not atypical or idiosyncratic because, in our main result, the class of such games where a maximin profile strictly Pareto dominates all Nash equilibria has the same cardinality as the class of games in which a Nash equilibrium strictly Pareto dominates all maximin profiles. Thus, neither behavior is generally superior. We further identify additional mechanisms favoring maximin over Nash equilibrium, including coordination failures under multiple equilibria, where maximin can outperform Nash play in realised-pay-off terms. A systematic analysis of strictly ordinal symmetric 3x3 games shows that these effects arise with non-trivial frequency. Our findings, therefore, suggest that the observed rise in zero-sum thinking in many rich countries, when associated with a maximin decision rule, will not be readily displaced through its generation of inferior pay-offs. |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2604.19359 |
| By: | Cheikbossian, Guillaume; Peña, Jorge |
| Abstract: | We study the evolution of other-regarding preferences in a public goods game where the production function exhibits varying degrees of complementarity between individual efforts. Individuals are rational agents who play a Nash equilibrium, but differ in the weight they assign to others’ payoffs, capturing varying degrees of prosocial or anti-social preferences. This preference trait evolves through payoff-based biased social learning, modeled within an adaptive dynamics framework. Because material payoffs induced by the equilibrium contributions may be non-concave in the preference parameter, evolutionary branching can arise. We show that monomorphic populations are evolutionarily stable only when complementarity between individual efforts is sufficiently strong, in which case preferences converge toward either prosociality or anti-sociality depending on the nature of strategic interactions between players. By contrast, when contributions are highly substitutable, monomorphic populations can become unstable, giving rise to polymorphic populations in which multiple preference types coexist. These results highlight how the structure of the public goods environment shapes the evolution and diversity of other-regarding motivations in culturally evolving populations. |
| Keywords: | Adaptive dynamics; other-regarding preferences; public goods games |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:tse:wpaper:131681 |
| By: | Polachek, Solomon (Binghamton University, New York); Romano, Kenneth (State University of New York at Binghamton (Binghamton University)); Tonguc, Ozlem (State University of New York at Binghamton (Binghamton University)) |
| Abstract: | This study examines how large language models (LLMs) respond to varying stake sizes in the Dictator and Ultimatum games using the high-stakes design introduced by Andersen et al. (2011). We test ten leading LLMs chosen for their accessibility, prominence, and differences in reasoning capabilities. Results reveal substantial variation across models: Only 5 of 10 models exhibit strategic behavior by offering more in the Ultimatum Game (UG) than in the Dictator Game (DG). Relative to humans, 4 models are consistently more generous, 2 consistently less, and 4 vary with stake size. Only 1 model shows a monotonic decline in UG offers as stakes increase; the remaining 9 are non-monotonic or stable. Unlike humans, most models reduce UG offers when endowed with wealth. Prompting for "human-like†decisions generally increases generosity in the UG. These findings are important for evaluating whether LLMs can serve as realistic proxies for human subjects in behavioral experiments and highlight key limitations and future directions for model development. |
| Keywords: | ultimatum game, dictator game, fairness, payoff stakes, artificial intelligence |
| JEL: | D01 C72 C90 |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:iza:izadps:dp18545 |
| By: | Bardey, David |
| Abstract: | This paper studies competition between television channels in a two-sided market with asymmetric firms. Motivated by a competition case in Colombia, we consider an oligopoly with three channels—two large and one small—that compete for viewers and advertisers. Advertising affects viewers both directly and indirectly through content quality, which is endogenously determined by the share of revenues that channels reinvest rather than distribute to shareholders. We first characterise the equilibrium of the subgame between viewers and advertisers and derive comparative statics linking audience levels to prices and payout policies. We then analyse the equilibrium of the game between channels, which jointly choose advertising prices and payout rates. While equilibrium prices are characterised implicitly, the model delivers closed-form solutions for payout decisions. Our main result is that asymmetries in audience size translate into asymmetric competitive pressure on the advertising side, which weakens the smaller channel. This effect is amplified when advertisers are restricted to single-homing, as in the presence of exclusivity clauses. By concentrating advertising demand on dominant channels, exclusivity reduces the smaller channel’s revenues and its incentives to invest in content quality, thereby limiting its ability to compete. These findings provide a novel mechanism through which exclusivity can generate exclusionary effects in two-sided media markets by affecting both demand allocation and endogenous investment decisions. We find that exclusivity reduces social welfare, mainly due to a decline in advertisers’ surplus that is not offset by improvements on the viewers’ side. |
| Keywords: | Two-sided markets; free-TV; ad-financed business model; competitive bottleneck; exclusivity contracts |
| JEL: | D43 L11 L13 L82 L86 M37 |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:tse:wpaper:131679 |
| By: | Kevin Michael Frick |
| Abstract: | Artificial intelligence algorithms are increasingly used by firms to set prices. Previous research shows that they can exhibit collusive behaviour, but how quickly they can do so has so far remained an open question. I show that a modern deep reinforcement learning model deployed to price goods in a repeated oligopolistic competition game with continuous prices converges to a collusive outcome in an amount of time that matches empirical observations, under reasonable assumptions on the length of a time step. This model shows cooperative behaviour supported by reward-punishment schemes that discourage deviations. |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2604.15825 |
| By: | Anastasia Burkovskaya; Egor Starkov |
| Abstract: | We propose a model of causal persuasion, in which a sender selectively discloses a set of variables together with their true joint distribution and proposes a subjective causal model that binds them. A receiver is persuaded by this model only if the data conclusively identifies the causal link of interest. We characterize when such persuasion succeeds or fails, and how easily it can be achieved. We further show that if the receiver holds a pre-existing subjective model, debunking it is similar to persuading a receiver without one. To establish a true causal link, the sender often needs to disclose only one or two well-chosen variables. But to dispel a perceived link -- to persuade the receiver there is no causal relationship -- every common cause must be disclosed. Our results highlight a fundamental asymmetry in causal persuasion: Establishing causality is often much easier than ruling it out. |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2604.20664 |
| By: | Sebastian Galiani; Franco Mettola La Giglia |
| Abstract: | This paper models double agents—individuals coerced into simultaneously serving two rival intelligence agencies—as a finite-horizon trilateral game between Agency A, Organization B, and the Mole M. The double agent persists within a corridor of survival: a set of bilateral beliefs under which both principals continue the relationship. Under a linear-quadratic-Gaussian benchmark, we obtain a closed-form characterization of the corridor geometry, analytical upper bounds on expected duration, and three main results: (i) existence of a double-agent equilibrium, (ii) structural transitoriness— belief updating, terminal unraveling, and compounding survival risk ensure inevitable collapse, and (iii) comparative statics linking monitoring technology, punishment severity, and protection costs to expected duration. The key mechanism is self-destructive experimentation: B learns about the mole’s type through both the type-dependence of effort and the traceability channel γ, which amplifies this learning. Extensions establish existence under general specifications and show that duration decays exponentially in the number of rival agencies. Predictions are consistent with historical patterns from Kim Philby to the Cold War mole hunts. The setting—bilateral coercion, existential participation constraints, and Bayesian learning in a finite-horizon trilateral structure— defines what we term antagonistic common agency. |
| JEL: | C70 |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35085 |
| By: | Dirk Bergemann; Soheil Ghili; Xinyang Hu; Chuanhao Li; Zhuoran Yang |
| Abstract: | Bilateral bargaining under incomplete information provides a controlled testbed for evaluating large language model (LLM) agent capabilities. Bilateral trade demands individual rationality, strategic surplus maximization, and cooperation to realize gains from trade. We develop a structured bargaining environment where LLMs negotiate via tool calls within an event-driven simulator, separating binding offers from natural-language messages to enable automated evaluation. The environment serves two purposes: as a benchmark for frontier models and as a training environment for open-weight models via reinforcement learning. In benchmark experiments, a round-robin tournament among five frontier models (15, 000 negotiations) reveals that effective strategies implement price discrimination through sequential offers. Aggressive anchoring, calibrated concession, and temporal patience correlate with the highest surplus share and deal rate. Accommodating strategies that concede quickly disable price discrimination in the buyer role, yielding the lowest surplus capture and deal completion. Stronger models scale their behavior proportionally to item value, maintaining performance across price tiers; weaker models perform well only when wide zones of possible agreement offset suboptimal strategies. In training experiments, we fine-tune Qwen3 (8B, 14B) via supervised fine-tuning (SFT) followed by Group Relative Policy Optimization (GRPO) against a fixed frontier opponent. These stages optimize competing objectives: SFT approximately doubles surplus share but reduces deal rates, while RL recovers deal rates but erodes surplus gains, reflecting the reward structure. SFT also compresses surplus variation across price tiers, which generalizes to unseen opponents, suggesting that behavioral cloning instills proportional strategies rather than memorized price points. |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2604.16472 |
| By: | Dazhong Wang; Ruqu Wang; Xinyi Xu |
| Abstract: | This paper studies optimal auction design when valuations depend endogenously on post-auction collaboration between the seller and the winning bidder. Both parties exert non-contractible efforts after the auction, generating a double moral hazard problem alongside adverse selection. We analyze two role structures -- winner-pivotal and seller-pivotal collaboration -- and characterize optimal direct mechanisms using linear payment schemes that combine cash transfers with proportional value sharing. The optimal mechanism allocates the asset to the bidder with the highest virtual surplus, employs a deterministic value-sharing rule, and achieves full type revelation through the signal realization rule. Comparing the two scenarios yields three main findings. First, regarding value sharing, the seller secures a strictly higher share under seller-pivotal collaboration: for sufficiently low-type winners, the seller extracts the entire value, whereas under winner-pivotal collaboration every winner must retain a positive share to sustain his critical effort. Second, regarding effort exertion, the pivotal party always exerts higher post-auction effort than the supporting party, and each party exerts greater effort when pivotal than when providing support. Third, seller-pivotal collaboration yields strictly higher seller revenue than winner-pivotal collaboration for any type distribution. Finally, these optimal mechanisms can be implemented through ascending auctions with endogenously determined linear contracts. |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2604.17923 |
| By: | Adam Hallengreen Joergensen (Department of Economics, University of Copenhagen); Thomas H. Joergensen (Department of Economics, University of Copenhagen); Annasofie M. Olesen (Department of Economics, University of Copenhagen) |
| Abstract: | Dynamic household bargaining models are growing in popularity but are computationally demanding to solve and estimate. We propose a modification to the endogenous grid method (EGM) that allows this class of models to reap the benefits of the EGM. The method, which we refer to as interpolated EGM (iEGM), precomputes optimal intertemporal consumption as a function of the expected marginal value of wealth before solving the dynamic bargaining model. We illustrate the implementation of the iEGM in a simple example, where the iEGM is around 20 times faster than standard value function iterations. In a more complex quantitative model it is about 50 times faster, without compromising accuracy. We apply the iEGM to a rich household model to study how productivity shocks affect consumption inequality. Our results suggest that the degree of commitment of household members can affect the consequences of individual labor productivity shocks, such as e.g. skill-biased technological change. |
| Keywords: | Household Bargaining, limited commitment, life cycle, couples, numerical dynamic programming. |
| JEL: | D13 D15 C61 C63 C78 |
| Date: | 2026–04–17 |
| URL: | https://d.repec.org/n?u=RePEc:kud:kucebi:2605 |
| By: | Jullien, Bruno; Bedre Defolie, Özlem; Biglaiser, Gary |
| Abstract: | We study a startup’s choice of its “direction of innovation, ” how well the technology fits alternative acquirers, and the effects on acquisition outcomes and market dominance. Two horizontally differentiated firms bid to acquire the innovation and then compete in the product market. Firms differ in initial quality stock and in “absorption capabilities, ” how effectively the acquired innovation is integrated into their stock. The innovator designs the innovation to intensify bidding by putting firms on a more equal footing, thereby favoring the initially lower-quality firm. As a result, “increasing dominance” is less likely than under exogenous fit. The winner of the innovation is driven primarily by relative absorption capabilities rather than initial quality: the f irm with higher absorption capability is more likely to win. The equilibrium innovation direction minimizes industry profit and consumer surplus. In a two-period model, decreasing dominance becomes more likely when the low-quality firm has stronger absorption capabilities. |
| Date: | 2026–04–20 |
| URL: | https://d.repec.org/n?u=RePEc:tse:wpaper:131684 |