MICRO THEORY: Dr Min Weicheng (Shanghai Jiao Tong University)

Going for the Familiar: Long-Run Selection of Information Sources

This paper studies information acquisition under two-dimensional uncertainty in a social-learning framework. A society of short-lived agents sequentially learns about a persistent fundamental by choosing between a familiar source with known quality and an unfamiliar source whose quality is higher or lower. Under simple conditions, a novel phenomenon emerges: society almost surely settles on the familiar source, even when the unfamiliar source is of high quality and uncertainty about its quality is arbitrarily small. The mechanism is that signals from the unfamiliar source create correlation between the fundamental and its quality, which may reduce the informativeness of its future signals. Beyond the asymptotic result, the paper identifies environments in which settling occurs after only limited use of the unfamiliar source while large uncertainty about the fundamental remains. In these environments, the result extends to a single forward-looking agent who internalizes the value of experimentation. The findings show how uncertainty about information quality can shape information acquisition. As an implication, a high-quality research method, diagnostic technology, or analyst may lose attention while retaining a strong reputation, because past observations reduce the informational value of its future signals.