I was reading Simon Wren-Lewis's recent post, which took me to another one of his posts, which then linked to Noah Smith's May 2013 post on how poor DSGE models are at forecasting.