ABSTRACT
Although economics is becoming increasing computational, economists are still ill at ease with simulation methods. We review central questions in the philosophy of simulation: what distinguishes simulations from other models and experiments? In what sense can simulations produce novel results? What implications does the epistemic opacity of simulations have for their explanatory value? We then explore these questions in the context of the Monte Carlo methods in econometrics, dynamic stochastic general equilibrium models, and the emerging field of agent-based macroeconomics, and we argue that the methodological peculiarities present in all of these cases provide us with interesting lessons about the way in which the role of theory is conceived in economics.
