The ‘Endowment Effect’ and the Revision Costs Hypothesis