[EF4-5] An Agent-Based Understanding of Green Growth.

This preprint is situated within the context of the MATH+ project EF4-5. We delve into the dynamics of endogenous technical progress, often modeled through “learning by doing.” Specifically, we assume that technical progress evolves proportionally to the capital stock.

In the well-known Ramsey growth model, optimal growth paths are derived based on a representative agent who maximizes intertemporal utility. However, in a decentralized economy with numerous agents, a critical assumption prevails: actors tend to disregard their individual contributions to technical progress. Consequently, this external effect of investment—leading to technical progress—is not fully internalized, resulting in suboptimal outcomes.

Our contribution lies in an agent-based model that extends the standard Ramsey growth framework. Within this model, agents adopt different investment strategies. Some mimic the representative agent, while others explicitly consider their impact on technical progress. In the latter scenario, agents must form expectations about their peers’ investments.

Key findings from our investigation include:

1. Reproducibility: Standard economic results hold even in systems of homogeneous agents.
2. Capital Distribution: Unequal capital distributions tend to converge toward equality, fostering higher growth.
3. Qualitative Impact: Considering one’s contribution to technical progress does not fundamentally alter the overall picture.
4. Expectations vs. Reality: Economic performance remains robust, irrespective of how well agents’ expectations align with true investment values.

In summary, our study sheds light on the intricate interplay between investment decisions, agent heterogeneity, and economic outcomes in the context of economic growth for a one sector, closed economy.