"Guinea Pig Trials" Utilizing GPT: A Novel Smart Agent-Based Modeling Approach for Studying Firm Competition and Collusion

TL;DR

The paper tackles how firm competition and collusion emerge when firms can communicate, using a novel Smart Agent-Based Modeling (SABM) framework that embeds GPT-4 powered agents in a Bertrand duopoly with differentiable goods. It contrasts a base non-communicative setting with a communication-enabled variant, highlighting that tacit collusion arises around the Bertrand price without communication and shifts toward cartel-like pricing with communication, albeit with fuzzier convergence. The authors introduce memory- and planning-based behavior for agents, implement a structured prompting scheme, and evaluate variants including product differentiation, personas, and asymmetric costs to demonstrate robustness and insights into how communication and agent heterogeneity affect collusion dynamics. The results suggest SABM is a cost-effective, controllable method for exploring complex strategic interactions in economics and related fields, with practical implications for understanding how information exchange and agent heterogeneity shape price formation and anticompetitive outcomes.

Abstract

Firm competition and collusion involve complex dynamics, particularly when considering communication among firms. Such issues can be modeled as problems of complex systems, traditionally approached through experiments involving human subjects or agent-based modeling methods. We propose an innovative framework called Smart Agent-Based Modeling (SABM), wherein smart agents, supported by GPT-4 technologies, represent firms, and interact with one another. We conducted a controlled experiment to study firm price competition and collusion behaviors under various conditions. SABM is more cost-effective and flexible compared to conducting experiments with human subjects. Smart agents possess an extensive knowledge base for decision-making and exhibit human-like strategic abilities, surpassing traditional ABM agents. Furthermore, smart agents can simulate human conversation and be personalized, making them ideal for studying complex situations involving communication. Our results demonstrate that, in the absence of communication, smart agents consistently reach tacit collusion, leading to prices converging at levels higher than the Bertrand equilibrium price but lower than monopoly or cartel prices. When communication is allowed, smart agents achieve a higher-level collusion with prices close to cartel prices. Collusion forms more quickly with communication, while price convergence is smoother without it. These results indicate that communication enhances trust between firms, encouraging frequent small price deviations to explore opportunities for a higher-level win-win situation and reducing the likelihood of triggering a price war. We also assigned different personas to firms to analyze behavioral differences and tested variant models under diverse market structures. The findings showcase the effectiveness and robustness of SABM and provide intriguing insights into competition and collusion.

Figures (7)

• Figure 1: Base model without communication.
• Figure 2: Alternative model with communication allowed.
• Figure 3: Sample conversations output by the agents. Contents related to collusion are underlined. Firms Ed and Gill refer to Firms 1 and 2, respectively.
• Figure 4: Speed of collusion formation (first 200 rounds of Figure \ref{['fig:comm-price-all']}).
• Figure 5: Different levels of product differentiation.