This paper studies a social-norm model in which agents choose effort in response to their own productivity and the efforts of their neighbours. We analyse how links between groups with different productivity levels affect equilibrium effort and derive the intergroup linking intensity that maximises aggregate effort. We work in a stylised two-group stochastic block setting that allows for a tractable characterisation of the effort-maximising intergroup linking probability. As is shown, adding links between the clusters of agents with different productivities can increase the aggregate effort but, at the same time, reduce social welfare. The analytical results are confirmed by simulation experiments demonstrating the influence of intergroup links on agents’ interaction.
We study the computation of pairwise Granger causality networks in vector autoregressive (VAR) systems. Exploiting a finite-sample identity between the Granger F-statistic and the partial correlation of residualized lagged regressors, we test all N candidate sources for a fixed target simultaneously, reducing the loop count of the textbook pairwise procedure from N(N-1) to N while preserving the F-statistic at machine precision. The construction extends from VAR(1) to an arbitrary lag order p≥ 1 by replacing the inner partial correlation with a multiple correlation, obtained by solving N batched p× p linear systems. Verified against a pairwise-OLS baseline and two statsmodels routines, the implementation reproduces their F-statistics to floating-point precision; against a like-for-like hand-coded NumPy baseline, it runs several times faster, with margin widening as the network grows. Monte Carlo experiments show that the empirical size remains close to the nominal level and that overspecification of the lag order, rather than under-specification, is the primary source of finite-sample power loss in our design. Applied to a panel of N=30 large US financial institutions from 2017 to 2024, a 250-day rolling window is updated in tens of milliseconds, making sub-second monitoring feasible. The false-discovery rate-controlled network density rises sharply around the COVID-19 sell-off of March 2020, and a stationary block bootstrap separates the stress window from the surrounding calm. The method places the linear pairwise test on the same operational footing as nonlinear or penalized alternatives, while retaining its closed-form null distribution and exact pairwise restriction.
Who gets to decide what AI systems optimize for? Current debates frame the risks of AI as a conflict between humans and machines. This brief argues instead that the central conflicts are between different groups of people, over the choice of the objectives that AI systems are built to maximize. Control over these objectives rests with those who control the inputs to AI, that is, the means of prediction: data, compute, expertise, and energy. To shed light on this control, I discuss the production function of AI, which maps data and compute into predictive performance, drawing on statistical learning theory and on the empirical scaling laws that have driven the industry’s costly scramble for scale and the resulting concentration of power. I then argue that market-based governance fails: individual property rights over data cannot address AI’s harms and benefits, because machine learning is fundamentally about data externalities, and because platform network effects are artificially maintained. I conclude with proposals for democratic control of the means of prediction, through institutions such as sortition and liquid democracy, to give those affected by algorithmic decisions a say over the objectives that AI pursues.
In politics your best move depends on what others do, and theirs on what you will do. This paper studies that interdependence in a model of distributive legislative voting: Each member of an N -member legislature decides, simultaneously and without knowing how the others will vote, whether to join a winning coalition for a spending bill. Because a legislator wants to join only if the coalition is not “too” large, her vote depends on her forecast of the Yes-total. I model that forecast explicitly: Each legislator has a small random set of predictors—algorithms mapping past vote totals to a forecast—monitors their performance and uses her most accurate one. In simulation, the aggregate settles into a stable band around the threshold ( n^*=85 of 101), while individual legislators churn constantly beneath it, switching votes and predictors. The focal point is robust to the choice of predictors, loss function, monitoring window, and learning rule. An extended specification, where legislators also dislike backing a failing bill, supports an all-No equilibrium, exact- n^* coalitions, and a fragile two-period cycle, all selected by the inherited voting history. Disagreement among rational agents who cannot take common knowledge for granted is not only sensible but a source of rich, computable structure.
This paper investigates the optimal public strategies to mitigate the spread of fake news by integrating education and fact-checking into a behavioral compartmental model. Building on an SVIR framework, we incorporate endogenous learning and reputational feedback, while explicitly modeling super-spreaders who disproportionately amplify misinformation. We formulate a dynamic optimal control problem in which the government chooses the optimal investment rate in education and fact-checking to minimize societal costs, which include the vulnerable population, reputational penalties, and the economic costs of interventions. Numerical simulations illustrate the interplay between learning, super-spreader dynamics, and public interventions, highlighting how targeted educational policies can effectively reduce misinformation diffusion and optimize resource allocation for long-term societal benefits. The results offer both theoretical and practical insights into designing policies to combat the viral spread of fake news in the digital age.
Recommendation letters are a ubiquitous yet puzzling institution in professional labor markets. While criticized as uniformly positive and uninformative signals, they remain central to hiring. This paper argues that their value lies not in their content alone but in their position within a relational network, offering a formal account of one mechanism through which relational information may mitigate information asymmetry. I adopt category theory as a modeling language to represent the job market as a category interconnecting applicants, referees, and firms through morphisms like applications and recommendations. My central claim, informed by the Yoneda Lemma, is that an applicant’s evaluative identity can be formally represented through the pattern of incoming relations from others, rather than through intrinsic attributes alone. This relational-first account explains how letters may acquire evaluative significance by composing into a structured web of endorsements that standard network analysis can miss. The framework generalises graph-based models, clarifies coherence conditions for evaluation, and formalises criteria shifts as natural transformations. I discuss limitations–including the static nature of the model–and sketch extensions to dynamic hiring processes and fairness-oriented institutional design. The aim is to provide economists with a rigorous lens for analyzing evaluation, signaling, and coordination in relational economic systems.
This paper proposes a new agent-based model grounded in the minority‑game framework to reveal the underlying mechanism of excess comovement. We model two key information diffusion behaviors of investors on social media: common attention to different stocks and information interaction about a single stock. The simulation results show that both behaviors significantly influence excess comovement, but their roles differ contextually. For stock pairs with historically positive return correlations, the impact of common attention dominates excess comovement when information interactions are infrequent, and a higher ratio of co-investors amplifies this effect. In contrast, for pairs with historically negative correlations, information interaction becomes the dominant driver of excess comovement when the ratio of co-investors is low, especially during periods of high market herding. Furthermore, the model provides accurate forecasts of excess comovement for both the next day and week.
The advancement of computer science has transformed the study of labor market dynamics, moving it away from traditional equation-based neoclassical theory toward more realistic, agent-based models. Agent-based modeling (ABM)—as a tool for describing complex systems—has become increasingly used in economics and business research, including the labor market. In this paper, we assess the applicability of agent-based modeling and simulation to the study of labor market dynamics, adopting a distinctive approach to the economics literature. We adopt a computer science perspective to evaluate how well existing models satisfy the software requirements of agent-based models. Accordingly, we identify the key components of an ABM and provide a recipe for developing agent-based models in labor market research. We highlight the advantages of ABMs over traditional econometric models and emphasize the practical implications of this bottom-up approach, including increased realism and an improved capacity to assess the effectiveness of economic policy tools.
This paper reviews the transmission channels of both conventional and unconventional monetary policy, including income composition, earnings heterogeneity, interest rate exposure, savings redistribution, inflation tax, portfolio composition, and household debt, with a focus on their impact on income and wealth inequality. The survey highlights the crucial role of household heterogeneity in shaping the transmission of monetary policy and its differential effects on income and wealth distribution. These effects operate through several factors, including the primary source of income, employment status, net debtor or net saver positions, differences in portfolio composition, and debt levels. The paper also discusses the shadow banking sector to illustrate the relationship between monetary policy and inequality in a complex financial system. Overall, the paper provides policymakers with a guide to the multifaceted nature of the monetary policy–inequality nexus, given the pervasiveness of household heterogeneity. It also suggests avenues for further research on the combined effects of different transmission channels using macroeconomic models suited to this purpose, such as agent-based models (ABM) with heterogeneous interacting agents.
Punishment plays a crucial role in driving norm change by incentivizing compliance with evolving shared normative expectations. Punitive measures typically encompass two components: a communicative effect, which delineates inappropriate conduct, and an institutionalized procedure that facilitates sanction enforcement. This study specifically investigates the latter dimension, examining how the mere existence of a punishment institution influences the formation of normative expectations. In other words, we examine whether behaviors such as extortion, when unpunishable and unmonitorable, are perceived as more socially appropriate. Using a neutrally framed harassment bribery game experiment, we manipulate the presence of a third-party punisher alongside with a Krupka–Weber norm-elicitation procedure to analyze shifts in participants’ normative perceptions. Our findings reveal no significant impact of the punishment institution on normative expectations. While a weak consensus emerges regarding the inappropriateness of extortion behavior, substantial disagreement persists concerning the merits of punishing extortionists. Policy implications of these results are discussed.
Greenhouse gases in the atmosphere pose a risk to human life and ecosystems. The planet’s temperature is gradually increasing, causing changes in weather patterns, rising sea levels, and extreme weather events. In 2015, countries agreed on the 2030 Agenda for Sustainable Development, which includes 17 goals, one of which is climate action. As an EU member, Italy has committed to reducing GHG emissions by 55
This paper examines social enterprises (SEs) through the lens of Amartya Sen’s notion of commitment. The hybrid nature of SEs combines social objectives with profit maximisation, often privileging the social dimension over the financial one. For this reason, SEs appear paradoxical within a neoclassical framework. The paper brings together the open definitional debate surrounding the nature of SEs and an interpretation grounded in the capability approach. Its aim is to demonstrate that the paradoxes identified within the neoclassical framework tend to dissolve once SEs are understood as capabilities-oriented firms. This methodological perspective makes it possible to illuminate how SEs take decisions, with particular reference to Sen’s concept of commitment. It also enables a clearer analytical distinction between SEs and neoclassical firms, while preserving the complexity of motives and objectives that characterises both, thereby helping to define the boundaries between these two distinct types of firm.
This study examines the relationship between online social interactions, sentiment dissemination, and stock market returns using Reddit data. We find that a small number of active users significantly influence others by disseminating sentiment within their networks. Active users have a more pronounced influence on less-active users when they share similar beliefs and during periods of increased uncertainty. Moreover, we show that prior-day abnormal network sentiment positively affects future stock market returns. Finally, we evaluate a network-based market timing strategy that effectively reduces drawdowns and highlights the practical implications of online interactions.
As data trading markets expand, risks such as data breaches, misuse, and the circulation of falsified information increasingly threaten trust and security. This study constructs a tripartite evolutionary game model involving the government regulators, data providers, and blockchain service providers. Prospect theory is embedded into both the payoff structure and risk perception process to capture behavioral biases. The model compares static and compliance-contingent regulatory incentive mechanisms, with MATLAB-based simulations used to examine system stability and evolutionary trajectories under varying parameter combinations. Results show that: (1) Under static incentive schemes, excessively high penalties or rewards may dampen compliance incentives, suggesting that stricter regulation does not necessarily yield higher compliance; (2) compliance-contingent incentives, which link reward–penalty intensity to compliance levels, are more effective in driving the system toward high-compliance and high-security equilibria under lower regulatory pressure; and (3) cognitive biases such as overconfidence and loss aversion distort strategic decisions and weaken the effectiveness of traditional incentive tools. By integrating prospect theory into blockchain-based data trading governance for the first time, this paper identifies the effective boundary of punishment intensity and highlights the disruptive role of risk perception biases in incentive design. The findings offer theoretical and practical insights for optimizing differentiated regulatory strategies and platform security governance, contributing to the development of a safer and more sustainable data trading ecosystem.
Incentive structures that support desired outcomes are crucial for building an egalitarian system. However, existing theories explicitly accounting for the social interactions that shape these structures are scarce. Thus, this study examines three types of political economies in which individual decisions depend on the actions of others and their consequences. Using simulations and mean-field solutions, it analyzes labor supply, occupational choice, and production decisions within these systems to explore the role of price signals, redistributive schemes and social interactions in achieving desirable equilibrium outcomes.
Employer branding involves strategies to create a positive corporate image, attracting and retaining high-quality workers. However, the impact of these policies on employees remains unclear in existing literature. In this study, drawing from econophysics literature—particularly the Maxwell–Boltzmann distribution—we use an evolutionary game theory model to investigate the population exposure to these strategies. Through agent-based methods, we analyze two-player populations seeking an optimal equilibrium, exploring the influence of wage offers and employee consumption levels. Additionally, we consider external sponsors, like relatives or universal income providers, who can subsidize wages. Our findings indicate the significant role of external sponsors in game dynamics, prompting their consideration in human resources management.
The experimental literature in economics largely acknowledges the role of social capital in social payoff enhancing behavior and/or voluntary provision of public goods. However, social capital is often measured using different proxies that, in principle, may have different impacts on contributions. In this article, we unpack social capital in three dimensions (i.e., school, family, and friends) and we analyze which dimension matters for contributions in economic experiments. We collect data from an ultimatum game, a trust game, and a public good game carried out with children in primary and secondary schools in Sicily, Italy. To measure social capital, we also collect detailed information on social attitudes related to different contexts. We find that, as expected, social capital matters for contributions. More interestingly, we find that the only component that affects contributions is social capital at school. Our findings shed a light on how the school environment may be extremely important for the formation of social capital, especially when this is linked to behavior within the same environment.
Standard behavioral frameworks often overlook how social network topology shapes the coordination of inflation expectations. To address this, I build a hybrid agent-based model that integrates “narrative-rooted” heuristics within a standard New Keynesian structural scaffold. In this framework, beliefs evolve through dual channels: performance-based selection (heuristic switching) and social diffusion (DeGroot learning). Simulations across canonical topologies reveal that seeding a target-based narrative in high-centrality nodes compresses forecast dispersion and accelerates convergence. However, a structural trade-off emerges: while performance evaluations dampen distorting narratives, strong social persuasion reduces volatility but simultaneously decouples expectations from economic fundamentals. This occurs because agents prioritize peer alignment over objective data, highlighting the double-edged nature of social networks: they stabilize expectations when credible narratives diffuse through hubs, but sever the link between beliefs and fundamentals when conformity overrides economic signals.
This experimental study introduces a threshold inequality into a classic threshold public goods game to understand coordinating behavior under asymmetric conditions. We use a novel design with “advantaged” players who have a lower threshold than a “disadvantaged” player, reflecting real-world public good scenarios. The threshold inequality disrupts the efficient coordination we observe in the control sessions leading to welfare losses relative to a control group without a disadvantaged player. Threshold inequality creates opportunities to help a disadvantaged player while enabling advantaged players to potentially free ride. We find evidence that threshold inequalities make the Pareto efficient provisioning of public goods more challenging, leading to suboptimal outcomes.
