*This paper investigates mechanism design for allocating a good with a positive externality without monetary payments, in a setting where each agent knows not only their own valuation but also the valuations of other agents to whom they are connected in a network. The planner's goal is to allocate the good to the agent with the highest valuation. Given the externality — which increases with the valuation of the agent receiving the good — each agent prefers that the good be allocated to the agent with the highest valuation to maximize their own utility. The paper explores the concept of belief-free implementation as proposed by Bergemann and Morris (2009) and demonstrates that the planner can use this partial incentive alignment to accurately identify the agent with the highest valuation. This is achieved by soliciting information from each agent about their own valuation and the valuations of their neighbors, regardless of the agents' beliefs. The paper proves that the network must be such that there are at least 2 agents who are connected to everyone to achieve the belief-free implementation. However, by assuming that agents use only undominated strategies, the paper proposes a mechanism that ensures efficient allocation, irrespective of agents' beliefs, when the network includes at least one agent who is connected to all other agents.

**Men tend to select into leadership positions more frequently than women. We conduct a lab experiment to examine how confidence contributes to this phenomenon. Our design allows us to measure different dimensions of confidence and correlate them with willingness to lead in different leadership contexts. For both men and women, we find that confidence explains willingness to lead to a larger extent when the leader’s role is to advise followers than when it is to take decisions on their behalf. This suggests that interventions aimed at boosting women's confidence may be more effective in reducing the gender-leadership gap in environments of the former type.