Social Security System Differences and Population Mobility: Evidence from the Guangdong–Hong Kong–Macao Greater Bay Area

Background and Objectives: Population mobility plays a critical role in promoting regional economic integration, labor market efficiency, and balanced development. In the Guangdong–Hong Kong–Macao Greater Bay Area (GBA), one of China’s most dynamic economic regions, population movement has intensified alongside rapid urbanization and cross-border economic activity. However, institutional differences—particularly in social security systems—remain a significant barrier to the free and efficient movement of labor. Due to the “one country, two systems” framework and the coexistence of multiple administrative, legal, and welfare regimes, substantial disparities exist in social security coverage, benefit structures, contribution rates, and financing mechanisms across Guangdong, Hong Kong, and Macao. These differences generate institutional costs that may distort migration decisions and hinder optimal allocation of human resources. Against this backdrop, this study aims to examine how differences in social security systems within the GBA affect population mobility, both theoretically and empirically, and to identify the mechanisms through which these institutional costs influence migration patterns. Methodology: The study adopts a mixed theoretical–empirical approach. First, it develops a conceptual framework grounded in new institutional economics and migration cost–benefit theory, conceptualizing differences in social security systems as institutional costs that influence mobility decisions. These costs are categorized into three components: public costs (government fiscal expenditures on social security), enterprise costs (employer contributions), and individual costs (employee contributions). Second, the study employs panel data from 11 cities in the Guangdong–Hong Kong–Macao Greater Bay Area over the period 2018–2021, a timeframe selected based on key policy changes enabling cross-border participation in mainland social insurance systems. A two-way fixed effects model is estimated to control for both regional heterogeneity and time effects. Population mobility is measured using comparable indicators across jurisdictions, while social security cost variables are constructed based on harmonized financial and institutional data. The analysis further incorporates control variables including wage income, economic development level, unemployment rate, and fiscal expenditure. Robustness checks are conducted using alternative dependent variables, extended sample periods, and additional controls, including pandemic-related factors and structural characteristics of the labor force. Key Findings: The empirical results reveal three main findings. First, differences in social security systems significantly hinder population mobility in the GBA. Higher total social security costs—arising from institutional fragmentation, benefit non-portability, and transfer inefficiencies—are associated with lower levels of population inflows. This finding confirms the hypothesis that institutional costs act as barriers to labor mobility, even in economically integrated regions. Second, the effects of social security costs vary across responsible agents. Public costs and enterprise costs exert a statistically significant negative impact on population mobility, reflecting fiscal pressure, congestion effects in public services, and increased labor costs for firms. In contrast, individual costs exhibit a positive relationship with mobility, suggesting that higher personal contributions may signal better employment conditions and stronger social protection, thereby attracting labor. Third, the impact of social security costs is heterogeneous across city types. Non-core cities and those within the Pearl River Delta are more sensitive to increases in social security costs, while core cities and special administrative regions show weaker or statistically insignificant effects. Additional robustness tests confirm the stability of these findings and highlig