We evaluate a comprehensive reform of Norwegian early retirement institutions in 2011 through the lens of a parsimonious random utility choice model. The reform radically changed work incentives and/or pension access-age for some (but not all) workers. We find that improved work incentives...