Non-coresident family as a driver of migration change in a crisis: The case of the COVID-19 pandemic

TL;DR

The paper investigates how non-coresident family, particularly parental ties, shaped inter-city migration in the COVID-19 crisis using Spectus GPS relocation data and IPUMS/ACS microdata. It combines micro- and macro-level analyses to show that people moved closer to family and that cities with more parental-family availability experienced larger net in-migration post-pandemic, after accounting for density and cost-of-living. A continuous-difference-in-differences model with a parental-availability proxy $v_i$ demonstrates a positive, significant influence of kin proximity on migration dynamics, quantified by $\ln \left[ \frac{y_i(\theta=1)}{y_i(\theta=0)} \right] = \beta v_i + \gamma + \sum_a \rho_a C_{ai}$. Together, the findings contribute to the demography-disaster nexus by highlighting kinship structures as a substantive driver of large-scale socioeconomic phenomena and offering implications for urban planning and policy in crisis contexts.

Abstract

Changes in U.S. migration during the COVID-19 pandemic show that many moved to less populated cities from larger cities, deviating from previous trends. In this study, building on prior work in the literature showing that the abundance of family ties is inversely related to population size, we analyze these migration changes with a focus on the crucial, yet overlooked factor of extended family. Employing two large-scale data sets, census microdata and mobile phone GPS relocation data, we show a collection of empirical results that paints a picture of migration change affected by kin. Namely, we find that people migrated closer to family at higher rates after the COVID-19 pandemic started. Moreover, even controlling for factors such as population density and cost of living, we find that changes in net in-migration tended to be larger and positive in cities with larger proportions of people who can be parents to adult children (our proxy for parental family availability, which is also inversely related to population size). Our study advances the demography-disaster nexus and amplifies ongoing literature highlighting the role of broader kinship systems in large-scale socioeconomic phenomena.

Figures (3)

• Figure 1: Changes in relocation patterns after the COVID-19 pandemic started in terms of city population size and parental family availability. Panel (a) shows the changes in the probability $z$ to relocate between (binned) city sizes before and during COVID-19 and suggests that movers from large cities were much more likely to relocate to small cities after the pandemic started than during the 2019 baseline period (red region in the bottom right corner). Panel (b) shows a binned scatterplot of our parental family availability proxy $v$ in relation to log-population (blue, left vertical axis) and net in-migration changes (grey, right vertical axis) after COVID-19 shock. The dots represent the mean vertical axis values given the horizontal axis bins (with 50 discrete, equidistant bins along the $v$-axis in total); error bars represent the 95% CI of the means. The grey line and blue curves are fitted regression lines for the means (linear and order-2 polynomial, respectively) with the shaded regions corresponding to the 95% CI of the regression estimates. The grey vertical axis is $\log \left[ y(\theta=1) / y(\theta=0) \right]$, which measures how much more (or less) of an attractor the cities in each bin became after the pandemic started (larger positive values indicate that on average the cities saw larger inflow per outflow after the pandemic started). Panel (c) shows a time-series of the average inflow per outflow of cities (log scale) grouped by quintiles of parental family availability $v$ (shaded regions correspond to the 95% CI of the means), suggesting proportionally high increases in net in-migration to cities in the high $v$-quintiles after the COVID-19 shock in April 2020 compared to the corresponding time in the prior year.
• Figure 2: Results for the micro-level analysis of movers in the IPUMS data, showing the rates for individual movers to (a) move into their parents' households ($\lambda_1$, Type 1 movers in the microdata) and (b) move back to their place of birth (POB) from elsewhere but not joining their parents' household ($\lambda_2$, Type 2 movers). Panel (c) shows $\lambda_3$, the rates for Type 3 moves, i.e., family household units moving back to their native place from elsewhere.
• Figure 3: A heatmap showing normalized density of our parental availability proxy variable $v$ conditioned on the general family availability $\phi$ obtained from mcleod2023originsmcleod2023origins (see Sections \ref{['sec:results-large-scale']} for discussion and methodology).