AbstractWith a continuous decline in the cost of manipulating data and a continuous increase in the richness of data banks, policymakers have increasing opportunities to build and apply so‐called microsimulation models—models that attempt to simulate the behavior of the individuals in a large population under a specified program. The efforts of the Department of Labor to use a model in evaluating proposed changes in the unemployment insurance system point up both the power and the weaknesses of such models. Any user who applies these models without attempting to understand which of their strengths and weaknesses are most important for analyzing the problem at hand is asking for trouble. Easy to use or not, these models are not user‐fri