“Last-Place Aversion”: Evidence and Redistributive Implications