Unemployment figures as published by the Ministry of Human Resources and Social Security (MOHRSS) on a quarterly basis is interesting not least as an example of how far data can diverge from reality. This is perhaps because the figure is so politically sensitive and is best illustrated by the figure’s relative imperviousness to external shocks.
At the end of the 1990s, as China reformed its State-Owned enterprises, independent reports estimated unemployment had surged to almost one fifth of the workforce, close to fifty million people. However figures collated by MOHRSS barely fluctuated, recording an unemployment rate around 4%.
A similar situation was witnessed in 2009 in the aftermath of the Financial Crisis, as swaths of factories closed and millions of migrant workers packed their bags and headed home. Independent reports indicated nearly 17% of China’s 275m strong migrant force was unemployed. Again, official figures barely fluctuated from the politically acceptable 4%.
So what is going wrong?
There are essentially two reasons for this inaccuracy. The first is that unemployment figures only measures urban residents receiving unemployment benefits from local governments. As registered urban residents account for less than half of China’s workforce and exclude China’s 275m strong migrant workers this leads to the first significant distortion.
The second problem arises from the number of ways in which local governments can bury unwanted claimants in bureaucracy to reduce the cost of welfare payments. Given the politically sensitive nature of unemployment, local governments have many reasons to perpetuate this situation.
Potential counterbalances to this figure are PMI employment sub-indices. These are far more sensitive indicators and represent a fairly reliable indicator of whether firms are increasing or decreasing their workforces.
Another indicator that has proven itself slightly more reliable than the MOHRSS figure is the national census figure, which regularly returns higher figures for unemployment – sometimes twice as high during periods of transition.