INTRODUCTION

Computational methods have become increasingly important in the analysis of dynamic general equilibrium problems. These methods are being used, for example, to study the incidence of tax and monetary policies in dynamic models of growth, commodity storage inv arious models of agricultural commodity markets, and price formationindyn amic models of asset markets. Many early computational methods relied primarily on intuitive economic tatonnement stories, and produced moderately successful algorithms. Even when these methods worked, they were usually slow. Furthermore, as we know from general equilibrium theory, tatonnement methods may not converge even with good initial guesses. Inthe past decade, the computational literature has made more use of formal mathematical tools from numerical analysis and perturbation theory. This use has resulted inmore powerful algorithms that canattack increasingly complex problems. These developments are particularly important when we try to solve models with several agents. This essay reviews the key ideas used in recent work, gives some examples of their advantages, and indicates the likely directions future work will take.

It is particularly appropriate that the 2000 World Congress of the Econometric Society include a survey of recent computational literature, because computational methodology is inherently an important part of what is broadly called “econometrics.” Ragnar Frisch, in his editorial in the initial issue of Econometrica, defined econometrics as the “unification of the theoretical-quantitative and the empirical-quantitative approach to economic problems.”