The socially optimal level of altruism

It is already recognized by some specific models in the existing literature that altruism may have socially counterproductive effects. Economic theory also shows that self- interest often produces efficient outcomes. This thesis explores the relationship between altruistic preferences, punishment systems and the cultural evolution of morality. The central argument is that altruism has detrimental effects on the efficacy of punishment and the resultant incentives of agents to co-operate with socially efficient equilibria, and that the use of punishment can have a negative effect on the evolution of altruism. The sequential punishment model is presented - akin to an infinitely-repeated stage game, but sufficiently simple to allow determinate optimal punishment paths to be derived - and the impact of different levels of altruism fully analysed. It is shown that high levels of altruistic motivation - close to but slightly less than full altruism - cause the socially efficient equilibrium to break down. Although the model is only a highly stylized representation of social interaction, the key effects that drive these results should appear in many more specific examples. In summary, these are the temptation effect (more altruistic individuals are less tempted to do harm to others), the willingness effect (more altruistic individuals are less willing to inflict punishment), and the severity effect (punishments, such as a fine where the revenue is redistributed, are less severe for more altruistic individuals, because they place a higher value on the contribution of the revenue to the welfare of others). By embedding a simplified version of the sequential punishment model in a simulated indirect evolution framework, it is also established that the use of punishment can weaken the group selection mechanism, resulting in a lower level of altruism evolving. The normative consequences of this are shown to be ambiguous.EThOS - Electronic Theses Online ServiceGBUnited Kingdo