Wouldn't this be a much worse metric? It would have to net out to zero change on average by the definition of percentiles. If we take abs change to look at both downward and upward mobility, the measure wouldn't tell us where most downward mobility happened, up and down could all happen within the bottom 25% and none in the top 75% and this metric would say we are highly social mobile if there was a lot of movement there.
Wouldn't this be a much worse metric? It would have to net out to zero change on average by the definition of percentiles. If we take abs change to look at both downward and upward mobility, the measure wouldn't tell us where most downward mobility happened, up and down could all happen within the bottom 25% and none in the top 75% and this metric would say we are highly social mobile if there was a lot of movement there.