Massive Genealogies Distinguish Frontier from Steady-State Internal Migration

Abstract: Recent studies of human migration have focused on modern issues of international economics, politics, urbanization, or commuting. Here we make use of very large anonymized genealogies which offer quantitative metrics and models before census data became available. In European and North American data from 1400 to 1950 we find two distinct patterns of lifetime migration. The steady-state pattern shows a universal power-law distribution of migration distance; by its early appearance it cannot be dependent on post-industrial technology. The frontier pattern, in contrast, is not scale-free with its much longer average distances. All migration distances are well fit by a three parameter model; the temporal and geographic patterns of the fitted parameters give new insight to American internal expansion 1620-1950. Frontier migration is also highly directional and asymmetric; gravity models do not apply. The American frontier pattern arose from the colonial-era steady-state within a generation, plateaued for three generations, then returned to a more mobile steady-state, a sequence paralleled by the Steppe migrations that brought the Bronze Age to Neolithic Europe. The transient frontier pattern is enabled by large-scale technological or numeric imbalance and geographic opportunity; when these forces abate, a new steady-state begins.