Mitchell, Karlyn and Pearce, Douglas (2015): Direct Evidence on Sticky Information from the Revision Behavior of Professional Forecasters.