Filling the Gaps: Vacancy Chains and Agent-Based Models

This paper introduces agent-based model (ABMs) for studying vacancy-based mobility, such as finding a new apartment or getting a promotion. This approach outperforms current, canonical analytic vacancy models by: 1) modelling off-equilibrium cases; 2) simulating multiple, overlapping, exogenous shocks of varying durations and intensities; 3) instantiating dynamic systems with an inherent tendency towards stability, ex. prototypical bureaucracies; 4) increasing realism by relaxing standard Markovian assumptions; and 5) providing confidence bands for point estimates. Applying ABM vacancy models to an empirical example of occupational mobility reveals a previously undocumented cohort mechanism, the purge boomerang. These novel models also further the idiom of ABMs by explicitly coding vacancies as agents, thereby introducing a formal representation of absence, akin to the zero symbol in arithmetic.