Here is an interesting paper by Wenchao Jin with a good account of why Chinese women stop working at such an extraordinarily young age. I like that it is attentive not only to the details of public policy but also of social institutions:

In 2013, the employment rate among 55-64-year-old urban women in China stands at 27%, well below the rates seen in most other countries at all levels of development. The urban female 55-64 employment rate is around 50−55% in the UK, Thailand (which has similar GDP per capita to China) and the Philippines (which has lower GDP per capita). …

The first and foremost explanation is the low pension age for urban women. The biggest public pension schemes set the formal retirement age at 50 for female workers and 55 for female ”cadres” or managers, and 60 for men. However, there are different rules for people with special circumstances like disabilities and compliance is not perfect, so the age at which a woman becomes eligible for a public pension can be as early as 45 or as late as 60. According to the China Health and Retirement Longitudinal Study 2011, 90% of current female pensioners completed the retirement process by 55. …

The second explanation of low employment rate among older females is their adult children. First, (expected) financial transfers from children when one is old and frail reduce the need to accumulate a large stock of assets through working. About 60% of women in their 50s live with their children, and about 70% receive financial support from their non-coresident children in the last year. There is some evidence that the financial transfers from children respond to parental incomes. Thus, transfers from children have a wealth effect as well as an insurance effect, both of which have implications for parents’ decisions on labour supply and saving.

Moreover, demands from adult children for domestic services such as grandchild care mean less time is available for paid work in the market. In urban China, the majority of women have grandchildren before 60 and, conditional on having grandchildren, the majority spend time looking after their grandchildren for an average of more than 30 hours a week. Moreover, I find that urban female employment is significantly negatively correlated with having grandchildren (conditional on her own age, education and so on), and this correlation comes from those women with more educated children. This is consistent with the hypothesis that parents cooperate with their adult children when choosing between market and domestic labour supply and leisure.

Older women are thus both pushed out of the labor force by the official retirement age, and also pulled out by demands of family. At least the first of these factors is likely to change over time, as the Chinese government has confirmed it plans to gradually raise official retirement ages for men and women. (The normal retirement age for women in most OECD countries ranges from 60-67.) The model developed in the paper suggests the effects of this change would be extremely large:

Raising the female pension age from its current level (which varies across individuals and has a mean of 50.6) to 60 would increase the employment rate by 28 percentage points on average over age 50-59. The average age at which women leave the labour force would also increase from 52.9 in the baseline to 55.7.