Symmetry and dissipation as the basic mechanism of social mobility, explaining distance scaling of migration patterns

TL;DR

This paper introduces a fundamental physics-inspired mechanism based on symmetry and dissipation to explain and predict the distance scaling of social migration flows, moving beyond empirical models.

Contribution

It proposes a new theoretical framework for social mobility that accounts for distance scaling and predicts the scaling exponent, grounded in physical analogies.

Findings

The model explains the value of the scaling exponent as 2.

It accounts for deviations and saturation in migration flows.

Verified with residential migration data from Australian cities.

Abstract

Models of social mobility inspired by the Newton's law of gravity have been used for several decades to describe migrations of people, goods, and information. Despite an eminent reference and widespread use, these models lack the background theory, being often viewed as a collection of empirical recipes which rely on adjustable parameters. Here we propose a tractable and fundamental mechanism of social mobility, which explains distance scaling of migration flows and predicts the value of the scaling exponent. The mechanism reveals two key aspects framing social flows, which have direct analogy in physics: symmetry and dissipation. In particular, we identify the conditions for the social gravity scaling, when the power law exponent equals 2, and explain deviations from this behaviour, including saturation transitions. The resulting flow distribution is determined by the spatial structure of the underlying social network, rather than by distance explicitly. The theory is verified for residential migration in suburb networks of major Australian cities with diverse structure, population and size. The mechanism is directly translatable to other social contexts, such as flows of goods or information, and provides a universal understanding of how dynamic patterns of social mobility emerge from structural network properties.