Diffusion of ideas in networks with endogenous search

I study the diffusion of technology when the decision to search for productivity-enhancing technologies depends on the network of interactions between agents. Agents have the option to engage in costly learning from their first-degree connections. The more productive an agent's connections, the higher the gains from learning. Hence, the network affects the reservation productivity at which agents choose to learn and, therefore, affects aggregate productivity. I find that the denser the network, the higher total factor productivity and the lower inequality on average. This effect is due to less dispersed gains from learning across nodes when network density increases. Despite the increase in productivity, the share of agents that learn in equilibrium is not affected by network density.