Assortative Matching in Marriages

The analysis measures local assortative matching patterns at the province level, a level above the prefecture, using marriages formed after 2005 and lasted till the census time in 2010. Allowing assortative matching patterns to vary by region alleviates the endogeneity problem that matching patterns adapt to local marriage market opportunities. The decision to use the province instead of the prefecture as the unit is a compromise between using information about marriages from recent years and proximate geographical areas. Some prefectures do not have enough observations of recent marriages to infer robust matching patterns and variations in matching patterns across prefectures within a province tend to be small. For each of the 30 provinces, I calculate the empirical distribution of the age gap and the education gap between the husband and the wife. The distributions show similar curves with moderate variations across provinces (see Online Resource 1).