How Persistent are Shocks to GDP? Evidence from the Forecasting Community.

Abstract. This paper uses forecast revisions to provide direct evidence on the persistence of shocks to GDP. The evidence from the forecasting record of professional, academic and government economists suggests that innovations in GDP exhibit far greater persistence than would be implied by simple trend stationary models. Instead, the persistence matches or exceeds the levels of persistence generated by low-order ARIMA models. Preliminary and incomplete. Suggestions and comments needed