In this paper we come closer to this strategy as we test the hypothesis that a linear combination of selected past idiosyncratic shocks recorded by the equity price of a given firm helps track and forecast aggregate business cycle fluctuations. At this stage we like to anticipate, however, that, somewhat against the benefit achieved by pooling many forecasts suggested by this strand of literature, our conclusions are that pooling individual information does not typically represent a good alternative to a situation in which instead a small number of regressors (. a subset of the full information set whose composition changes over time) are selected according to.
