*2.3. Hypotheses*

Although many papers deal with trust and reciprocity [23], only a few analyze trust and reciprocity in a laboratory setting examining income inequality [24–27]. In our experimental finitely-repeated MTG, endogenous income inequality may emerge and may a ffect trust and reciprocity levels when players are informed of their partner's accumulated earnings. Specifically, we expect that trustors trust less and trustees reciprocate less when they are aware that their partner is thus far earning more. We formulate our two first hypotheses aimed to test for the e ffect that having information about the earnings of one's partner has on actual trust and reciprocity.

**Hypothesis 1 (H1).** *On average, the trustor sends lower amounts to the trustee when knowing that the trustee's accumulated earnings in the previous period were higher than their own.*

**Hypothesis 2 (H2).** *On average, the trustee returns higher amounts to the trustor when knowing that the trustor's accumulated earnings in the previous period were lower than their own.*

Even if a greater number of experimental studies find gender di fferences on trust in favor of males [28], some contradictory findings remain. For example, males send a higher portion of their endowment in experiments where the sample is heterogeneous in age. In addition, some authors attribute lower levels of female trust to their higher risk aversion [29]. However, when the characteristics of the participants are more homogenous, as in our case, the results tend to show higher levels of trust in females. This inspires our third and fourth hypotheses:

**Hypothesis 3 (H3).** *Compared to male trustors, female trustors send, on average, lower amounts.*

**Hypothesis 4 (H4).** *Compared to male trustees, female trustees return, on average, higher amounts.*
