Trust game (investment)
In a one-shot two-player trust game (Berg, Dickhaut, & McCabe, 1995) both players get an equal endowment (Z, say). Then, one of them (the sender or trustor i) gets the opportunity to send an amount (say, T) out of her or his endowment to the other player (the responder or trustee j). The amount sent is tripled by the experimenter (like a return on an investment) and the responder can return any part (R, say) of this tripled amount. Whereas homo economicus is predicted to return nothing and, thus, no transfers would take place by a similarly motivated individual, experimental results show that on average transfers equal about 50% of the endowment: while back-transfers are more or less equal to the transfer: (Berg et al., 1995, Glaeser et al., 2000, Camerer, 2003). Note that the efficient transfer equals the whole endowment.
For the application of the iATM model similar assumptions are used as before, except that now little money is assumed to be expected by the responder: . Consequently: . Another difference relates to utility. As receiving a transfer is less (negatively) emotional than the anger-arousing context of a power-to take game or ultimatum game, more room for deliberation by the responder may be expected (Fredrickson & Branigan, 2005). To allow for a more flexible (smoother) trade-off between the trustee’s utility and trustor’s utility, a simple loglinear instead of linear extended utility function is assumed: . Maximization of leads to , in line with what is observed.7
3.3. Suggestive neurobiological evidence
The affective tie mechanism, as a mechanism of affective bonding, raises the question how it relates to recent neurobiological findings regarding maternal care (attachment) and the bonding of mammals (Insel and Young, 2001, Numan, 2015, Numan, 2016, Numan and Young, 2016, Feldman, 2016, Feldman, 2017). Before going into this question, first, the evidence-based neural network model of maternal care proposed by Numan, 2015, Numan, 2016 is summarized (see also Numan & Young, 2016). This maternal care system is suggested to provide the neural foundation of human bonding more generally, like friendships; (Numan, 2015, Numan, 2020). Then, it will be indicated how existing neural evidence related to the ATM model (Bault et al., 2015) fits into this neural network model. Finally, additional evidence supportive of the more general uncertainty related iATM model will be put forward.
