I don't know that the graph tells this story.

It could be that the very slight decline in spending for the lower quintiles is from getting help that isn't counted (free preschool for example) or from products being more affordable (measuring money is not equivalent to measuring the value gotten) or from more people being able to avoid childcare expenses by being home.

It could be that the major increase in spending for the top quintile is wasted. Maybe the kid wants to own a horse. All sorts electronic things eat up money while delivering little value.