I thought the same thing, but there isn't enough data to suggest that it's wrong. I think that there is a portion "unemployed" a portion "changed careers" and

I thought the same thing, but there isn't enough data to suggest that it's wrong. I think that there is a portion "unemployed" a portion "changed careers" and a portion that are too small to individually plot. So that entire graph may only be like 75% of those that were originally at those companies.