Hi, I'm Estee.

Labour market mismatch can arise from workers having limited attention. This paper proposes a Generalized Partially Directed Search model, extending on existing literature by allowing inattentive workers to have diverse priors and heterogeneous attention costs. I show that mismatch can be inherited from bias in workers' default search strategies, and heterogeneous attention costs could contribute to greater variability in the equilibrium outcomes. I also explore equilibrium multiplicity that was not adequately accounted for in previous studies. Equilibria where workers adopt different application strategies may generate both higher market efficiency and lower monopsony power than when workers employ the same application strategies. This information-theoretic approach to model job search offers new policy insights on the basis of attention.

The growing adoption of large language models (LLMs) presents potential for deeper understanding of human behaviours within game theory frameworks. This paper examines strategic interactions among multiple types of LLM-based agents in a classical beauty contest game. LLM-based agents demonstrate varying depth of reasoning that fall within a range of level-0 to 1, which are lower than experimental results conducted with human subjects in previous literature, but they display similar convergence patterns towards Nash Equilibrium choice in repeated settings. Through simulations that vary the group composition of agent types, environments with lower strategic uncertainty enhance convergence for LLM-based agents, and environments with mixed strategic types accelerate convergence for all. Results with simulated agents not only convey insights on potential human behaviours in competitive settings, they also offer valuable understanding of strategic interactions among algorithms.

We study whether large language models (LLMs) can predict human strategic behavior from pre-play communication. Using three canonical laboratory games that vary in incentive alignment and communication structure, we provide LLMs and incentivized human forecasters with identical transcripts and ask them to predict players' subsequent actions. Using GPT-5 as our main model, we find that it consistently outperforms humans and achieves accuracy well above chance, especially when incentives are aligned and communication is bilateral. The performance gap arises almost entirely from correctly forecasting cooperative actions, while both humans and GPT-5 struggle to anticipate defection. These results suggest that strategic communication contains systematic information about future behavior that humans underutilize and that LLMs are able to exploit more effectively.

Workers' past application choices can act as heuristics for future decisions. By integrating learning theory into a search model, this work explores the role of experiences on workers' application choices. It provides an evolutionary perspective to labour market dynamics, and offers insights on equilibrium selection. I propose two market structures, where wages are unobservable and observable to workers, and model workers' application strategies over time using reinforcement learning and best response dynamics respectively. I show that in the presence of multiple equilibria, experience-based learning generally induces workers to coordinate on a more efficient, locally asymptotically stable equilibrium in which they apply with high probability to different firms, in both static and dynamic wage-setting environments. Learning models not only highlight potential mechanisms for equilibrium selection, they also suggest process-oriented policies to improve market efficiency.

Large language models (LLMs) have been extensively used as the backbones for general-purpose agents, and some economics literature suggests that LLMs are capable of playing various types of economics games. Following these works, to overcome the limitation of evaluating LLMs using static benchmarks, we propose to explore competitive games as an evaluation for LLMs to incorporate multi-players and dynamicise the environment. By varying the game history revealed to LLM-based players, we find that most LLMs are rational in that they play strategies that can increase their payoffs, but not as rational as indicated by Nash Equilibria. Moreover, when game history is available, certain types of LLMs, such as GPT-4, can converge faster to Nash Equilibrium strategies, suggesting a higher rationality level in comparison to other models. We provide an economics arena for the LLMs research community as a dynamic simulation to test rationality and strategic reasoning ability.

As the world shifts towards greater consumerism, there is an increasing tendency for individuals to distinguish themselves through the goods they purchase, which assert their taste and implicit social status. While the pursuit of social status has been well-explored, most studies focus on luxury consumption; limited research has investigated cultural goods (books, etiquette classes, etc.) as another signalling tool. This paper seeks to understand and provide a theoretical grounding for individuals' choices between status goods, and its implications for policy and social mobility. The choice to signal status through luxury or cultural goods is evaluated under two scenarios: when the wage-enhancing benefits of cultural goods are not revealed, and when they are. Individuals always prefer luxury goods in the first instance, but upon satisfying certain conditions, those endowed with high cultural and social capital may consume cultural goods instead. The paper proposes policy measures for a welfare-maximising social planner to drive greater consumption of cultural goods, which also has a positive impact on intergenerational mobility.