Abstract This study explores a nested representation of ethical, moral, social identity, motivated, opportunistic and reciprocal agent preferences to characterize screening contracts in a principal–agent model under adverse selection. This leads to a ranking of the type of social preferences that principals should seek in agents, based upon the information rents associated with each agent type. When moral hazard is introduced the ranking further depends upon the interaction of limited liability with self-selection. These results are interpreted in light of the 2010 Dodd-Frank Act and principal–agent experiments.