Information-consistent learning and shifts in long-run productivity

►We present an information-consistent learning framework within a DSGE macro model. ►Previous work assumes agents know more when optimizing than when learning. ►Earlier findings rely on inconsistent assumptions and not the presence of learning.