Divergent Paths of US Counties: Multiscale Drivers and Spatial Dynamics of Population Growth and Decline (2010−2020)

Abstract

Population change across US counties has taken sharply different paths, with some areas experiencing steady growth while others face persistent decline. These trends are not random; they reflect deeper shifts in the economy and uneven regional development. This study contributes to demographic and regional research by treating population growth and decline as two distinct demographic regimes, each influenced by its own set of factors. We examine which factors are associated with county-level population change in the United States from 2010 to 2020, using data from 3130 counties with complete records across all explanatory variables, compiled from the American Community Survey, the Bureau of Economic Analysis and fiscal records. To capture both overall patterns and local variations, we combine standard regression methods (OLS, spatial lag and spatial error models) with Multiscale Geographically Weighted Regression (MGWR). Counties are explicitly divided into growing (N = 1548) and shrinking (N = 1582) groups, with models estimated separately for each. The findings show a moderate but consistent asymmetry between growth and decline. Declining counties tend to cluster together and are associated with consistent factors such as aging populations, weak labour markets, reliance on agriculture and, to a more limited extent, housing tenure. Growth, on the other hand, is more spatially variable and localized. It is associated with an agricultural employment structure operating at a broad scale, while natural increase and tax burden exhibit highly localized associations, and housing tenure shows a broader but uneven association. MGWR provides a substantially better fit than global models in both cases (R² = 0.546 for growing counties and 0.595 for shrinking counties, compared with 0.043 and 0.150 under standard OLS), highlighting that population change is a complex, multiscale process. Understanding these spatial patterns is crucial for developing policies and planning strategies that are sensitive to local conditions.