Convergence and mobility of rural household income in China : New evidence from a transitional dynamics approach

Purpose The purpose of this paper is to study convergence and income mobility of China’s rural households. Design/methodology/approach The data of rural household income per capita are employed to compute the transitional dynamics in the rural sector. The analyses are conducted at two spatial levels, namely, the national and provincial levels. Ergodic distributions are computed to provide a forecast of future income distributions, whereas Mobility Probability Plots are constructed to offer detailed information on the transitional dynamics. Findings The income distributions are found to have considerable persistence. Another finding is that most of the households (except the extremely low-income households) have a tendency of moving downwards in the income distribution though they are more likely to remain in the same levels of relative income because of their high persistence. Convergence to a unimodal income distribution is possible in the long run, however, the households will converge to a value which is far below China’s per capita gross domestic product. Research limitations/implications Since a lot of the rural households would congregate to the lower part of the income distribution if the transitional dynamics remain unchanged, therefore, it calls for government intervention. Practical implications More resources should be diverted to the rural sector. Social implications The finding also shows that the provinces have very different transitional dynamics even if they are situated in the same economic zone. Thus, the government should formulate province-specific development polices so as to promote greater equality. Originality/value Given that no recent research has been conducted on convergence and transitional dynamics of rural household income. Therefore, this paper attempts to fill the gap in the literature by investigating the pattern and future development of rural household income in China through the use of stochastic kernel approach.