Dynamically Aggregating Diverse Information

An agent has access to multiple information sources, each modeled as a Brownian motion whose drift provides about different component of an unknown Gaussian state. Information is acquired continuously—where the chooses both which sources sample from, and also how allocate attention across them—until endogenously chosen time, at point decision taken. We demonstrate conditions on agent's prior belief under it possible exactly characterize optimal acquisition strategy. then apply this characterization derive new results regarding: (1) endogenous for binary choice, (2) dynamic consequences manipulation, (3) strategic provision by biased news sources.