A Microfounded Herding Model and Its Estimation On German Survey Expectations

The paper considers the dynamic adjustments of an average opinion index that can be derived from a microfounded framework where the individual agents switch between two kinds of sentiment with certain transition probabilities. The index can thus represent a general business climate, i.e., expectations about the future course of the economy. This approach is empirically tested with the survey expectations published by the ZEW and ifo institute. The estimated coefficients make economic sense and are highly significant. In particular, besides effects from fundamental data like the output gap in the recent past, one can identify a strong herding mechanism within both panels, such that the agents do not just join the majority but, metaphorically speaking, follow each single motion of the crowd. In addition, the transition probabilities of the ZEW agents are found to be influenced by the ifo climate but not the other way around.