The analysts or institutions that make economic predictions may have other objectives apart from minimizing prediction errors. This article analyzes whether the people who make forecasts behave strategically, with the aim, for example, of obtaining publicity, increasing their income, achieving greater prestige or influencing the decisions of the economic agents. It first studies whether the experts formulate predictions that are radical or close to the consensus, giving rise to the so-called herding behaviour. Secondly, is considers whether reputation and age influence the degree of conservatism of the predictions. Lastly, it discusses to what extent the forecast specialists revise their early estimates on obtaining a higher volume of information. The strategic behaviour models considered are tested by using a US economy forecasting survey, the Livingston Survey, and another for the Spanish economy conducted by FUNCAS.

Key words: consensus, rationality, survey of forecasters.
JEL classification: C53, E37.