Yes, the closed cases has been steadily dropping but remains well above 2% - however, we expect that number to be higher than the actual mortality rate while the infection is still spreading/growing, as recovering from the disease takes longer than dying from it.

What I find more concerning is that the current ratio of deaths to cases is 3%, which I would expect to be an underestimate. Hopefully this is due to under-reporting of the mild cases, but it does imply either a major under-reporting of active cases or a significantly higher mortality rate than has been estimated so far (most likely the former).

I would take that number with a grain of salt too, as there is a lot of asymmetry with how and when that data is reported. There may be a large number of "recovering" cases that are not completely closed yet but have essentially zero chance of death in the future. The definition of recovered seems open to interpretation while death is not, and so the deaths would have a more immediate reporting period, as well as a shorter average interval of time since infection started.