|
on Game Theory |
| By: | Francesc Dilmé (University of Bonn) |
| Abstract: | We study finite-player normal-form games with compact metric action spaces and bounded measurable payoffs. Our main theorem shows that every Nash equilibrium of such a game can be recovered as the limit, in the product weak topology, of Nash equilibria of finite games obtained by discretizing the action spaces and perturbing payoffs by a uniformly vanishing amount. The proof samples from the target equilibrium, uses concentration inequalities to control weak convergence and incentive constraints on a growing finite set, and then applies a payoff perturbation to convert the resulting approximate equilibrium into an exact one. We also provide an example of a continuous game with a Nash equilibrium that cannot be approximated through Nash equilibria of finite games without perturbing payoffs. |
| Keywords: | Nash equilibria, infinite games, finite approximations |
| JEL: | C72 C62 |
| Date: | 2026–04 |
| URL: | https://d.repec.org/n?u=RePEc:ajk:ajkdps:400 |
| By: | Sampat, Khushi (University of Warwick) |
| Abstract: | This paper studies how decentralised neural agents trained by regret minimisation learn equilibrium behaviour in static games and whether such learning can be extended beyond Nash equilibria. The analysis proceeds in two parts. The first chapter examines equilibrium selection in coordination games with multiple Nash equilibria. Building on recent evidence that neural agents trained across large distributions of games systematically favour risk-dominant equilibria, the chapter introduces a structured pre-training curriculum designed to instil a bias toward payoffdominant outcomes in Stag Hunt environments. While pre-training successfully induces effcient coordination in these games, the results show that this bias is rapidly eroded under subsequent adversarial training on heterogeneous games, where play reverts to mixed or risk-sensitive equilibria. The second chapter investigates whether decentralised learners can acquire correlated equilibrium behaviour when coordination requires conditioning on private signals. Initial experiments demonstrate that standard personal regret objectives lead agents to ignore mediator signals and converge to unconditional Nash strategies. This limitation is overcome by replacing personal regret with a squared obedience (swap) regret objective. Under this modified objective, neural agents successfully learn signal-contingent behaviour and generalise correlated equilibrium strategies to unseen coordination games. Together, the findings clarify the capabilities and limitations of regret-based learning as a mechanism for equilibrium formation in strategic environments. |
| Keywords: | Correlated Equilibrium ; Regret Minimisation ; Deep Reinforcement ; Learning ; Neural Networks ; Game Theory JEL classifications: C72 ; C63 ; C73 ; D83 ; C61 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:wrk:wrkesp:99 |
| By: | Amil Camilo Moore; Fabrizio Germano; Rosemarie Nagel |
| Abstract: | We study similarity in the complete set of one-shot 2×2 games with payoffs from {1, 2, 3, 4} without replacement. Similarity is defined geometrically via a neighborhood structure on games and continuity of behavior, and is applied to both theoretical rules (e.g., Nash equilibrium, level-k reasoning) and experimental data. This produces a partition of the games into (theoretical or empirical) similarity classes. We run a large-scale experiment in which each subject plays all 78 games within our class without feedback. We find that empirically inferred similarity classes diverge sharply from those predicted by Nash equilibrium and dominance reasoning. Instead, the empirical similarity classes align closely with the theoretical classes of a level-k variant, with deviations reflecting fairness and efficiency concerns. At the individual level, subjects’ play can be classified according to primary and secondary rules, conforming with either level-k variant (0 ≤ k ≤ 5) or a fairness and efficiency-based heuristic. The main insights extend to strategic settings beyond our 2 × 2 games. |
| Keywords: | Similarity of games; level-k reasoning; equity and efficiency; experiments; topology of games |
| JEL: | C52 C70 C72 C81 C90 C93 D91 |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:upf:upfgen:1942 |
| By: | Sam Ganzfried |
| Abstract: | We present a new solution concept called evolutionarily stable Stackelberg equilibrium (SESS). We study the Stackelberg evolutionary game setting in which there is a single leading player and a symmetric population of followers. The leader selects an optimal mixed strategy, anticipating that the follower population plays an evolutionarily stable strategy (ESS) in the induced subgame and may satisfy additional ecological conditions. We consider both leader-optimal and follower-optimal selection among ESSs, which arise as special cases of our framework. Prior approaches to Stackelberg evolutionary games either define the follower response via evolutionary dynamics or assume rational best-response behavior, without explicitly enforcing stability against invasion by mutations. We present algorithms for computing SESS in discrete and continuous games, and validate the latter empirically. Our model applies naturally to biological settings; for example, in cancer treatment the leader represents the physician and the followers correspond to competing cancer cell phenotypes. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.18385 |
| By: | Eric Yan |
| Abstract: | High-stakes auctions are often preceded by nonbinding communication between bidders and the seller. Motivated by these practices, this paper examines a two-period model in which two bidders send private cheap talk messages to the seller about their valuations, and the seller decides in the second period whether to run a mechanism or take an outside option that disappears if she chooses to run the auction. The seller has commitment within any mechanism she chooses to run, but no commitment over how she uses any information communicated. Despite having potentially asymmetric posteriors after the communication stage, the seller cannot run discriminatory auctions in equilibrium. Under some natural restrictions, any bidder-symmetric perfect Bayesian equilibrium of this model is a threshold equilibrium where the seller runs a second-price auction with a single reserve if and only if both bidders are above the threshold. The seller is better off being able to commit to the restricted class of mechanisms where she must choose a single reserve price. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.17733 |
| By: | Paramahansa Pramanik |
| Abstract: | We develop a Euclidean path-integral control to characterize optimal firm behavior in an economy governed by Walrasian equilibrium, Pareto efficiency, and non-cooperative Markovian feedback Nash equilibrium. The approach recasts the problem as a Lagrangian stochastic control system with forward-looking dynamics, thereby avoiding the explicit construction of a value function. Instead, optimal policies are obtained from a continuously differentiable Ito process generated through integrating factors, which yields a tractable alternative to conventional solution methods for complex market environments. This construction is useful in settings with nonlinear stochastic differential equations where standard Hamilton-Jacobi-Bellman (HJB) formulations are difficult to implement. Consistent with Feynman-Kac-type representations, the resulting solutions need not be unique. In economies with a large number of firms, the analysis admits a natural comparison with mean-field game formulations. Our main contribution is to derive a noncooperative feedback Nash equilibrium within this path-integral setting and to contrast it with outcomes implied by mean-field interactions. Several examples illustrate the method's applicability and highlight differences relative to solutions based on the Pontryagin maximum principle generated by HJB. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.25086 |
| By: | Thomas Pitz; Vinicius Ferraz |
| Abstract: | In many applications of cooperative game theory -- from corporate governance and cartel formation to parliamentary voting -- not all winning coalitions are feasible. Ideological distances, institutional constraints, or pre-electoral agreements may render certain coalitions implausible. Classical power indices ignore this and weight all winning coalitions equally. We introduce cohesion structures to quantify coalition feasibility and axiomatically characterize two families of cohesion-sensitive power indices, represented as expected marginal contributions under Luce-type distributions. In the Banzhaf branch, coalition weights are a power transformation of cohesion; in the Shapley branch, additional axioms separate size from cohesion, recovering the classical size weights with cohesion acting within each size class. All results have been mechanically verified in Lean 4 with Mathlib. We illustrate the framework on the German Bundestag and the French Assembl\'ee Nationale, where cordon sanitaire and double cordon scenarios produce sharp, interpretable power shifts. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.27220 |
| By: | Deepal Basak; Joyee Deb; Aditya Kuvalekar |
| Abstract: | Algorithmic content targeting homogenizes information, with implications for strategic interactions. For example, this increased homogenization was arguably responsible for the run on the Silicon Valley Bank. We argue that existing measures of similarity are inappropriate for studying games -- especially coordination games -- because they do not discipline agents' conditional beliefs. We propose a class of stochastic orders, Concentration Along the Diagonal (CAD), built on agents' conditional beliefs. In canonical binary-action coordination games, greater CAD-similarity is both necessary and sufficient for strategic similarity -- agents adopt the same strategy. We further demonstrate CAD's applicability in congestion games, collective action, and second-price auctions. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.28190 |
| By: | Nora Paulus (DF, Université du Luxembourg); Weihua Ruan (Purdue University Northwest, USA); Benteng Zou (DEM, Université du Luxembourg) |
| Abstract: | "Corporate tax competition has driven statutory rates downward for decades, eroding fiscal capacity and raising concerns about global equity. The OECD/G20 Global Minimum Tax (GMT) seeks to halt this “race to the bottom, ” yet its dynamic implications remain unclear. We study the GMT with a differential game of international tax competition with mobile capital. Governments set corporate tax rates while multinational firms reallocate capital in response to effective taxwedges created by the minimum tax and the substance-based income exclusion. We distinguish between Markovian behavior, in which governments adjust tax rates in response to current capital allocations, and open-loop behavior, in which they commit to tax paths in advance. We also compare enforcement through Qualified Domestic Minimum Top-up Taxes (QDMTT) and the Income Inclusion Rule (IIR). In the Markovian game, the GMT does not pin down a unique long-run outcome: a continuum of steady states arises under both enforcement regimes, including low-tax configurations. By contrast, under open-loop commitment the dynamic system is saddle-point stable, implying convergence to a unique transition path for given initial conditions. Commitment therefore acts as a dynamic selection device. Whether the economy converges to high- or low-tax configurations depends on enforcement: under QDMTT, a race to the bottom may emerge when public revenue is used inefficiently and the minimum tax is sufficiently high, whereas under IIR such dynamics are ruled out. Overall, the GMT can stabilize tax competition under commitment but does not, in general, eliminate downward pressure on statutory rates." |
| Keywords: | "Dynamic Tax Competition; Qualified Domestic Minimum Top-up Taxes; Income InclusionRule; Global Minimum Taxation." |
| JEL: | C73 F21 H21 H87 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:luc:wpaper:26-07 |
| By: | Ginger Zhe Jin; D. Daniel Sokol; Liad Wagman |
| Abstract: | We study a novel dynamic inspection game in which a regulator commits to a detection technology, and a regulated firm chooses whether to engage in harmful conduct and, if detected and sanctioned, whether to incur a redesign cost that generates a modified violation requiring renewed detection. In this adaptive environment, bounded sanctions give rise to three Markov perfect equilibria: full compliance, harm until recognition, and persistent redesign. We show that AI-augmented monitoring can shift the equilibrium from persistent redesign to harm until recognition, but only when regulatory investment crosses a regime-shifting threshold. Below this threshold, greater monitoring intensity may increase enforcement workload without reducing aggregate harm, as the regulator repeatedly detects adaptive violations while the firm continues to redesign. Thus, partial investments in AI monitoring can generate congestion rather than deterrence. When the firm can also adopt AI to reduce its redesign cost, the regulator’s deterrence threshold rises, reinforcing the strategic interaction between enforcement and evasion technologies. Moreover, congestion becomes particularly salient when AI-flagged violations require human review and regulatory review capacity is binding. In this case, the precision of AI triage—especially its false positive rate—matters as much as detection intensity. Enforcement effectiveness therefore depends not only on expanding detection, but also on allocating scarce human review resources efficiently. |
| JEL: | D82 K21 K42 L40 |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35010 |
| By: | Nguyen, Duc Manh (Monash University) |
| Abstract: | This study investigates how the experience of conflict and the framing of post-conflict identity affect trust. In a pre-registered laboratory experiment in Vietnam, implemented shortly after the celebration of the 50th anniversary of the end of the war, 534 partici- pants were randomly assigned to either a treatment group that engaged in a multi-round competitive game intended to simulate conflict (called the “Attacker/Defender†game) (Gross et al., 2022) before playing the Trust game under four identity framings: paired with someone from the opposing group of the conflict, the same group, with no information about partner’s prior group, or with a new, neutral group identity designed to symbolically represent an absence of relation with conflict, or a control group which only take part in the Trust Game. We find that playing the Attacker/Defender game (i.e., being exposed to conflict in the lab) lowers trust by 13–21%, regardless of which side participants were in the conflict. |
| Keywords: | Trust ; Intergroup Conflict ; Identity Framing ; Laboratory Experiment JEL classifications: D91 ; Z13 ; D83 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:wrk:wrkesp:94 |
| By: | Noriaki Matsushima (Osaka School of International Public Policy, the University of Osaka); Kazuki Nishikawa (Graduate School of Economics, the University of Osaka); Jiaying Qiu (Graduate School of Economics, the University of Osaka) |
| Abstract: | A binding minimum wage can raise the regulated firm's profits when labor-market power interacts with product-market competition. We develop a duopoly model in which firms compete in the same product market but hire workers from distinct, geographically segmented labor markets. Because the minimum wage applies only to one firm's labor market, it does not directly raise its rival's costs. With monopsony power, the minimum wage reduces the regulated firm's marginal cost and induces it to expand output, forcing its rival to contract through strategic interaction. Under Cournot competition, this mechanism also increases total employment and consumer surplus. |
| Keywords: | Minimum wage, Monopsony power, Segmented labor markets, Product-market competition |
| JEL: | J38 J42 L13 J23 C72 |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:osp:wpaper:26e003 |
| By: | Clayton, Christopher; Maggiori, Matteo (Stanford University); Schreger, Jesse |
| Abstract: | We build a model of two hegemons that have valuable trading relationships with each other and at the same time compete in exerting geoeconomic power over countries in the rest of the world. When the hegemons trade with each other their optimal policy is shaped both by classic economic considerations - the profitability of that specific trade - and by geoeconomic competition - how the trade affects the power of each hegemon vis-a-vis the rest of the world. We show that containment, a policy mix in which an hegemon attempts to limits sales of its inputs to the rival hegemon and uses its power to demand that the rest of the world shifts away from sourcing from the rival hegemon, arises when the two hegemons offer relatively substitutable exports since a stronger rival would offer a better outside option to the targeted countries. Accommodation between the hegemons, instead, occurs when power motives are small and the two hegemons focus on purely economic profit motives. We characterize how the rest of the countries welfare depends on the contain/accommodate regime of the hegemons. |
| Date: | 2026–03–24 |
| URL: | https://d.repec.org/n?u=RePEc:osf:socarx:4sy2k_v1 |
| By: | Khaw, Rachel (Monash University) |
| Abstract: | This thesis develops a theoretical model of digital platform competition in which moderation choices endogenously generate ideological differentiation. Competing platforms decide which content providers to host, trading off advertising revenues against moderation costs, while consumers sort by ideological proximity and content variety. In equilibrium, breadth competition cancels out, leaving ideological tilt as the key dimension of differentiation. Polarisation emerges as the most robust equilibrium, maximising platform profits but welfare-reducing for moderates, while generalism is socially optimal but privately fragile. By modelling ideology as the outcome of moderation intensity rather than an exogenous stance, the paper clarifies how moderation incentives shape polarisation, welfare, and regulatory trade-offs. |
| Keywords: | Digital platforms ; content moderation ; ideological differentiation ; polarisation ; welfare ; industrial organisation. JEL classifications: L13 ; L82 ; D43 ; D72 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:wrk:wrkesp:98 |
| By: | Hattori, Keisuke |
| Abstract: | This paper shows that entry can raise each individual firm's profit---not merely industry profits---when Cournot oligopolists finance working capital through a contestable banking sector. Under average-cost pricing of loans, entry dilutes fixed banking costs across greater lending volume, lowering loan rates. Entry raises per-firm profits if and only if equilibrium output lies in an intermediate range where financing relief dominates intensified competition. Bank-side and firm-side frictions play opposing roles: firm-side frictions facilitate profit-increasing entry by amplifying cost relief as firms shrink, while bank-side frictions suppress it by raising funding costs as aggregate lending expands. |
| Keywords: | Profit-increasing entry, Cournot oligopoly, Contestable banking, Financial frictions |
| JEL: | D43 G21 G32 L13 |
| Date: | 2026–01–31 |
| URL: | https://d.repec.org/n?u=RePEc:pra:mprapa:127926 |
| By: | Murat Erkurt; Emre Ozdenoren |
| Abstract: | We study contests in which players sequentially search for a high score at a cost per draw, with unlimited opportunities, no recall, and the best score wins a prize. In the unique symmetric equilibrium, the acceptance probability depends only on the number of players, the cost, and the prize, not on the distribution, and total expenditure equals the prize. These properties extend to multiple prizes and hierarchical team competition. Efficiency relative to a planner is determined by the hazard rate of the distribution. With a finite horizon, a selectivity effect can dominate the discouragement effect when search costs are low. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.20683 |
| By: | Nigar Hashimzade; Haoran Sun |
| Abstract: | Industrial policy has returned to the centre of economic governance, particularly in the high-tech sectors where positive network externalities in demand make market dominance self-reinforcing. This paper studies the welfare effects of an industrial policy targeting a sector with network externalities in a two-country model with strategic trade and R&D investment. We show how the welfare consequences of this policy are determined by the interaction between the strength of the externality, the type of R&D, and the degree of product differentiation between the home and the imported goods. When externalities are weak or the goods are close substitutes, the business-stealing effect produces a race to the bottom that dissipates more surplus than it creates. Under sufficiently strong externalities and weak substitutability or complementarity of the goods, industrial policy competition can make both countries simultaneously better off compared to the laissez-faire outcome because of the mutual business-enhancement effect. The case is stronger for the product innovation than for the process innovation, as the former directly affects the demand and triggers a stronger network effects than the latter which operates indirectly through the supply. Thus, the network externalities create an opportunity for a win-win industrial policies, but its realisation depends on the market structure and the nature of innovation. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.29542 |
| By: | Zhonghong Kuang; Sanxi Li; Yi Liu; Yang Yu |
| Abstract: | In light of prevailing data regulations, consumer mobility across diverse markets inherently endogenizes market segmentation. Considering such strategic interactions, we define a market segmentation as strategy-proof when no consumer (with positive measure) has an incentive to deviate to another market. We show that in every strategy-proof market segmentation, the producer surplus remains at the uniform monopoly level, and the consumer surplus is bounded between the buyer-optimal level and the uniform monopoly level. Remarkably, no consumer is worse off than in the case of a uniform monopoly. We also construct a family of strategy-proof segmentations to realize every possible welfare outcome. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.20609 |
| By: | Alma Cohen; Alon Klement; Zvika Neeman; Eilon Solan |
| Abstract: | In many institutional settings, $k$ items are selected with the goal of representing the underlying distribution of claims, opinions, or characteristics in a large population. We study environments with two adversarial parties whose preferences over the selected items are commonly known and opposed. We propose the Quantile Mechanism: one party partitions the population into $k$ disjoint subsets, and the other selects one item from each subset. We show that this procedure is optimally representative among all feasible mechanisms, and illustrate its use in jury selection, multi-district litigation, and committee formation. |
| Date: | 2026–03 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2603.24727 |