We use linked longitudinal data on employers and employees to estimate how the 2003-2005 Hartz reforms affected the wages of displaced German workers after they returned to work. We also present a simple new method to decompose the wage effects into components attributable to selection on...