Big data on divorce in China

I recently read (and highly recommend) HE Xin's "‘No Malicious Incidents’: The Concern for Stability in China’s Divorce Law Practice," Social & Legal Studies, 2017, pp. 1-23. In it, he argues that (at least in the court he observed) the "law" of divorce is completely overridden by the perceived need to maintain stability, which in practice means never to give a verdict that will upset anyone who is credibly threatening violence. (Obviously, this gives grey hairs to judges in cases where both sides credibly threaten violence if the ruling doesn't go their way.)

As always, people will debate whether this phenomenon is marginal and trivial or instead says something fundamental about what the institutions of what is conventionally called the Chinese legal system are really all about. But this post is not about that. Instead, I want to use it to introduce an interesting source I just ran across: some big data analyses of Chinese divorce cases from Jan. 1, 2014 to Sept. 30, 2016.

First, the data confirms what He has told us elsewhere: it's much harder to get a contested divorce in China than you might think, given the surface permissiveness of the law. (After all, if one party insists that mutual affection has indeed broken down and no longer exists, on what basis could a court say that that's not true?) In fully 63% of cases, the court did not grant the petition for divorce. To that 63% of adjudicated cases, we should also add those cases where the court talked the petitioner into withdrawing the request for divorce.

To me the most interesting data - and the one that made me think of He's article - was on the educational level of couples involved in divorce disputes.

The graphic shows the educational level of one party on the Y axis and the other party on the X axis. The six categories from bottom to top and left to right are (1) illiterate/semi-illiterate, (2) primary school, (3) middle school, (4) high school, (5) technical/vocational/junior college, and (6) undergraduate and above. 

Note that the data aren't normalized to reflect the fact that there simply aren't a lot of people with a high educational level relative to those with a lower level, so this tells us nothing about whether people of a given educational level are more or less likely to get divorced. But the absolute numbers do tell us something interesting: that the practice of courts (after reading He's articles, I'm hesitant to call it "law") in contested divorces is driven heavily by people at the low end of the educational spectrum. Look at the size of the dots in the lower left section. This is a subject that might merit further study.