Apart from income inequality, income polarisation is another important indicator of residents’ income gap. The measurement of income polarisation from traditional household surveys is often inaccurate, however, due to the under-coverage of top incomes. In order to correct for this bias, this paper combines the Chinese Household Income Project (CHIP) database with the Top Incomes in China (TIC) database. We examine the levels, trends, and heterogeneity of income polarisation for Chinese residents in 2002, 2013 and 2018, and conduct different simulation analyses of tax policies. Our conclusions are as follows. First, the estimates based on the CHIP data show that income polarisation levels increase from 2002 to 2013 but decline slightly from 2013 to 2018. After including the TIC data, income polarisation presents a persistent upwards trend over the period. The results are supported by robustness checks, which indicate the significance of top incomes for measuring income polarisation. Second, after combining household survey (HS) data with top income data, the degree of alienation keeps rising from 2002 to 2018, which indicates an expanding income gap between different resident groups. Third, heterogeneity analyses confirm higher polarisation levels for females, people with college degrees and people living in rural areas compared to males, those without college degrees and those living in urban areas, respectively. Lastly, simulation analyses support the benefits of personal income tax in decreasing income polarisation. Tax penalties for magnates, famous actors, and digital economy workers have similar effects, and higher tax rates are associated with lower income polarisation levels. This paper not only facilitates understanding of China’s true levels of income polarisation, but also provides evidence for the reform of the income distribution system
