Abstract
This paper presents a model in which homogeneous rational agents choose between two competing technologies. Agents observe a private signal and a sample of other agents' previous choices. The signal has both an idiosyncratic and an aggregate component of uncertainty. I derive the optimal decision rule when each agent observes the decision of exactly two agents. Due to aggregate uncertainty, aggregate behavior does not necessarily reflect the true state of nature. Nonetheless, agents still find others' choices a good source of information, and they base their decisions partly on the behavior of others. Consequently, bad choices can be perpetuated in this environment: I show that aggregate uncertainty can lead to agents herding on the inferior technology with positive probability. I also present examples in which herding occurs for arbitrarily large sample sizes.