An Analysis on Regional Differences in the Food Consumption Patterns of Chinese Urban Households

This study addresses the difference of the regional food consumption patterns in the Chinese urban households using the panel data 'per capita annual consumption expenditure of urban households by region(2004-2009)'. The methodology of the paper is cluster analysis and regression model with constant slope coefficients and intercept that varies over the regions and time by panel data, which combine time series and cross-sectional data. The results are as follows;It is stated in this paper that 31 regions of China through 6 years are divided into five clusters having different food consumption patterns respectively, following to the observation of variations in pseudo-F and CCC as the number of clusters grows. In addition, it is detected that there exists a nearly consistent tendency that food consumption pattern of each regions varies as time progresses.
This present study analyzes on the differences of sectoral food consumption patterns by clusters using the linear regression models with constant slope coefficients and intercept that varies over the regions and time with panel data. In addition, Lagrange multipliers of Breusch and Pagan on the models of 5 food sectors give us strong evidence for the varying-intercept regression models, and the values of Hausman statistic suggest that the error components model is more appropriate than the dummy variable model to analyze the regional consumption patterns of 5 food sectors in China. And, this paper shows that income elasticities of some clusters distinctly differ from the others in several food sectors.
In conclusion, the research of this paper presents some logical reasons why it is very desirable that provinces in China are separated to some groups, so that the systematic grouping of food consumption patterns will improve the accuracy of the analysis in demand for food. Besides, these results of the study may suggest several points of interest for China-related food policy makers, planners and traders.