Bayesian Estimation of Economic Simulation Models Using Neural Networks

Donovan Platt

doi:10.1007/s10614-021-10095-9

Abstract

Recent advances in computing power and the potential to make more realistic assumptions due to increased flexibility have led to the increased prevalence of simulation models in economics. While models of this class, and particularly agent-based models, are able to replicate a number of empirically-observed stylised facts not easily recovered by more traditional alternatives, such models remain notoriously difficult to estimate due to their lack of tractable likelihood functions. While the estimation literature continues to grow, existing attempts have approached the problem primarily from a frequentist perspective, with the Bayesian estimation literature remaining comparatively less developed. For this reason, we introduce a widely-applicable Bayesian estimation protocol that makes use of deep neural networks to construct an approximation to the likelihood, which we then benchmark against a prominent alternative from the existing literature. Overall, we find that our proposed methodology consistently results in more accurate estimates in a variety of settings, including the estimation of financial heterogeneous agent models and the identification of changes in dynamics occurring in models incorporating structural breaks.

Highlights

IntroductionTo some extent, seen the emergence of a paradigm shift in how economic models are constructed
Introduction and Literature ReviewRecent years have, to some extent, seen the emergence of a paradigm shift in how economic models are constructed
A need to facilitate mathematical tractability and limited computational resources have led to a dependence on strong assumptions,1 many of which are inconsistent with the heterogeneity and nonlinearity that characterise real economic systems (Geanakoplos and Farmer 2008; Farmer and Foley 2009; Fagiolo and Roventini 2017)

Summary

Introduction

To some extent, seen the emergence of a paradigm shift in how economic models are constructed. A need to facilitate mathematical tractability and limited computational resources have led to a dependence on strong assumptions, many of which are inconsistent with the heterogeneity and nonlinearity that characterise real economic systems (Geanakoplos and Farmer 2008; Farmer and Foley 2009; Fagiolo and Roventini 2017). The Great Recession of the late 2000s and the perceived failings of traditional approaches, those built on general equilibrium theory, would lead to the birth of a growing community arguing that the adoption of new paradigms harnessing contemporary advances in computing power could lead to richer and more robust insights (Farmer and Foley 2009; Fagiolo and Roventini 2017). The extent to which this has been achieved in practice, remains open for debate (Hamill and Gilbert 2016)

Objectives

Methods

Results

Conclusion