If two countries both report a 5-point shift, how do you know the shift is actually the same size if the points are counted differently?

Perhaps the actual change was bigger or smaller but the approach to measurement highlighted the change more or less strongly. There's no statistical validity to be had.