An adapted version of ”the trust game with revenge” is applied to a Swedish setting. Senders – the first-movers - can keep an endowment of SEK 30, or give fractions or all to an unknown receiver. Donations are multiplied by five before reaching receivers, who may, or may not, send back part or the entire received amount. Half of the receivers are given information that the sender has the opportunity to exact revenge, while the remaining are not given this information. Results differ from Fehr and Gächter (2000) J. Econ. Perspect. 14, 159-181, in the sense that the share of endowments sent in the first stage is around two thirds, compared to less than one third in Fehr and Gächter. Furthermore, they find a very strong effect of punishment while we find almost no effect. An efficiency frontier is defined and results show that that only 25 % of the outcomes reach this frontier due to lack of trust. If senders were confident that receivers would return at least 20 % of the donated amount, it is optimal to donate the whole endowment. Only about one-fifth returned of the receivers returned less than this, so for the most part the lack of trust is unwarranted.
