From Economic Agents to Agentic Economies: A Systems Blueprint for Economic World Models

Economic World Models (EWMs) (Cong, 2025) are generative economic models that simulate how economies evolve from within by modeling heterogeneous agents, their beliefs and actions, and the market and institutional mechanisms through which their interactions produce aggregate outcomes. This paper develops an implementation roadmap for building economic world models as generative engines in which heterogeneous agents act, interact, adapt, and co-evolve with markets and institutions, thereby producing economic dynamics from the inside. We organize EWM systems into a six-level capability ladder, from fixed rulebased agent worlds to adaptive and LLM-based agent worlds, self-evolving agents, evolving institutional worlds, and sim-to-real economic twins aligned with real observations. A systematic literature survey across these levels reveals that existing work remains concentrated in lower-level agent and simulation environments, while systems with self-evolving agents, endogenous institutions, persistent empirical alignment, and validated economic mechanisms remain rare.

Introduction. Economics has long advanced through a tradition of “observe and explain (or predict).” Economists observe aggregate variables, collect micro-level evidence, estimate empirical relationships or causal effects, build theoretical and quantitative models, and ask what caused what (Haavelmo, 1944; Friedman et al., 1953; Heckman, 2001). This approach, albeit undeniably powerful, leaves open a deeper question of generative explanation. Deriving an equilibrium, fitting parameters, or forecasting a turning point does not by itself explain an economic phenomenon unless the outcome can emerge from the model inside. This idea motivates the central premise of this paper: to understand an economy, we should not only observe its outcomes, but build a world simulator which generates such outcomes.

Discussion / Conclusion. and Open Challenges This paper develops a CS/AI systems perspective on Economic World Models (EWMs), building on the economic framework of Cong (2025). We study EWMs as generative and interactive environments that model how economic states emerge from the decisions, interactions, and adaptations of heterogeneous agents under market mechanisms, institutional rules, and real-world constraints. EWMs emphasize the coupling between agent behavior and world dynamics: agents act based on incentives, information, beliefs, and constraints, while their actions are aggregated through economic mechanisms into prices, allocations, risks, institutions, and future states. Turning this implementation perspective into faithful, evolving, and reality-aligned systems remains a substantial challenge. We view EWM implementation not as a finished technical recipe, but as a research agenda. Several open challenges are central to this agenda. Behavioral realism. EWMs require agents whose behavior resembles real economic actors. Agents should not merely produce plausible narratives or rational choices; they should reflect how real agents perceive information, form beliefs, respond to incentives, face constraints, and adapt over time.