Abstract:
We experimentally test overconfidence in investment decisions by offering partic- ipants the possibility to substitute their own for alternative investment choices. Overall, 149 subjects participated in two experiments, one with just one risky as- set, the other with two risky assets. Overconfidence increases (i) with the absolute deviation from optimal choices, (ii) with task complexity, and (iii) decreases with uncertainty as indicated by the difference between willingness to pay and to accept.