This is a BS saying because it's so easy to find counterexamples.

Let's try to use the number of child deaths, for example. By your saying, trying to do that is bad. Because governments probably will somehow resurrect dead children to look good at this metric. Right?

Look at what France, Japan, and USA consider childhood deaths. Back when I did medical software and wrote reports, France and Japan both lumped at lot of deaths post birth into 'death at birth' to lower their childhood death statistics.

No, but it is easy to see how reporting the death of children could be withheld - say for instance, if this metric had penalties associated with it.