Diffusing workers in a multiplex world

A common question on our work on labour flow networks was whether certain types of workers would navigate specific types of networks, for example, as a function of their occupation. To address this, we developed a generalisation of the random walkers labour flow model to multiplex networks.

The study of labor mobility across firms is crucial to understand economic performance, unemployment, skills reallocation and other aspects that shape the economic life of nations. Modeling labor flows between firms has been a challenge due to the complexity arising from the distributed and heterogeneous nature of labor flows. In this paper, we introduce a discrete-time model of labor flowing on a multi-layered network (i.e. a multiplex graph). By introducing multiple layers, the model accounts for different mobility patters (e.g. industries, geographies, occupations, etc.), which is important to understand the reallocation of human capital, skills and knowledge. We apply the model to UK empirical micro-data and find that our measure of regional preferences for low versus high skilled workers vary significantly from a single to a multi-layer representation of the world.