When Lone Wolf Defectors Undermine the Power of the Opt-Out Default.

High levels of cooperation are a central feature of human society, and conditional cooperation has been proposed as one proximal mechanism to support this. The counterforce of free-riding can, however, undermine cooperation and as such a number of external mechanisms have been proposed to ameliorate the effects of free-riding. One such mechanism is setting cooperation as the default (i.e., an opt-out default). We posit, however, that in dynamic settings where people can observe and condition their actions on others' behaviour, 'lone wolf' defectors undermine initial cooperation encouraged by an opt-out default, while 'good shepherds' defeat the free-riding encouraged by an opt-in default. Thus, we examine the dynamic emergence of conditional cooperation under different default settings. Specifically, we develop a game theoretical model to analyse cooperation under defaults for cooperation (opt-out) and defection (opt-in). The model predicts that the 'lone wolf' effect is stronger than the 'good shepherd' effect, which - if anticipated by players - should strategically deter free-riding under opt-out and cooperation under opt-in. Our experimental games confirm the existence of both 'lone wolf' defectors and 'good shepherd' cooperators, and that the 'lone wolf'effect is stronger in the context of organ donation registration behaviour. We thus show a potential 'dark side' to conditional cooperation ('lone wolf effect') and draw implications for the adoption of an opt-out organ donation policy.