Public versus Private Information

Using the model by Morris and Shin (2002), we distinguish between how people perceive a state and how they act upon it. We show than even for perceptions, where the coordination motive plays no role, improving the quality of public information does not always reduce the forecasting error. The reason why this happens is because better information is always more relevant information. But if improvements in information attract more attention than they deserve, the overall effect may be detrimental to the accuracy of perceptions. Increases in private information quality, on the other hand, are always beneficial to the decisions. Moreover, the assumption of no private information error on average implies that agents would do better collectively if there was no public information signal.