ABSTRACTOverconfidence is often viewed as encouraging entrepreneurs and CEOs to follow risky strategies such as entering new markets, engaging in innovation, or pursuing mergers and acquisitions. While such undertakings can generate excess returns and profits, overconfidence is frequently offered as an explanation for why so many business ventures fail. However, we show that in a setting where the decision maker does not know whether success is possible, theoretically, overconfidence can also have a discouraging effect, causing one to give up too soon. This counterintuitive result is driven by effectively misattributing one's own failure to an elevated assessment of the chance that success is not possible. In a controlled laboratory experiment, we find general support for the theoretical predictions, although empirically, participants are overly reluctant to engage in repeated innovation attempts.