I think the mortality rate as reported is way too low. Even if we take China's bogus numbers at face value, the denominator is wrong. Pick whatever time you like between confirming a case and the patient dying. Let's say it is a week. In that case, you should take the cumulative number of diagnosed cases from a week against reported deaths today.
It’s possible the opposite is true. Mild cases may not be detected or reported, and that would lead to a lower mortality rate.