Identifying income heterogeneity determinants using the method of moments quantile regression

Prior studies have typically concentrated on poverty status to determine anti‐poverty measures; however, this approach cannot sufficiently detect income heterogeneity. This study employs quantile regression for panel data to investigate the Korean Labour and Income Panel Study 2003–2020. Moreover, it adopts both household‐ and community‐level variables and separates demographic groups as working‐age and older adults, considering Korea's severe old‐age poverty. The findings indicate that household‐level characteristics, such as householder's gender, physical health, and employment status, present heterogeneous effects across the income distribution. Second, low‐income households are more vulnerable to regional economic and labour market downturns than high‐income neighbours. Lastly, although the National Pension, a backbone of the public pension system, provides limited supports for retirees because it was introduced much later than other countries, it assists low‐income old adults more effectively. Therefore, this study suggests more tailored redistribution measures, considering heterogeneous effects of household‐ and community‐level environments, and a further expansion of the National Pension to mitigate old‐age poverty.