Hey, maybe you can help me with this: chance encounters, geographic proximity, and innovative collaboration

Abstract

We know that proximity can influence collaborations, but it is unclear how proximity affects collaborator choice. Collaborator choice is important because it determines what knowledge is recombined, shaping the very nature of an innovation. In my dissertation, I consider both the context and mechanisms of proximity to explore the effect of proximity on collaborator choice. My first essay explores the context of proximity between knowledge workers’ residences, a setting that excludes some potential proximity mechanisms and modestly generalizes to between-firm spillovers. I show residential proximity predicts collaboration, as prior literature shows within a workplace. Notably, I show that only same-gender pairs seem influenced by residential proximity. This finding is potentially counterintuitive because, here, proximity leads to less diverse collaborator choices. I suggest homophily prohibits between-gender interactions outside an office from being substantive enough to increase awareness and influence collaborator choice. In the rest of my dissertation, I focus on the proximity mechanism of chance encounters as theoretically distinct from proximity. I predict and show that chance encounters should generally increase the variance of collaborator choice and diversity of knowledge in innovation. I also show that chance encounters do not seem to increase productivity. Instead, when chance encounters are less available, people turn to their prior contacts, limiting recombination potential. To empirically invoke variance in chance encounters instead of proximity, I develop and validate a measure using the flu to proxy for fewer chance encounters, borrowing from epidemiology findings that show when people fear a disease, they can engage in social avoidance behavior with proximate people.