Methodological Barriers to Model Transfer and Their Mitigation: The Case of Agent-Based Models in Macroeconomics

The presentation addressed the methodological challenges associated with transferring models across scientific domains, with a focus on the case of agent-based models in macroeconomics. It was argued that such transfers often encounter obstacles due to divergent methodological standards between source and target disciplines. The concept of methodological embedding — defined as the minimization of the semantic and/or syntactic distance between a transferred model’s construction assumptions and the methodological commitments of the target field — was proposed as a potential strategy to mitigate these challenges. However, it was noted that this strategy may sometimes compromise the epistemic function of models, giving rise to a dilemma: either the embedded model is dismissed as epistemically ineffective, or the non-embedded model is rejected for lacking methodological adequacy.