Agree with Stephen. A simple set of hypotheses and a solid argument. Obviously I'm left curious as to why different recessions have displayed different changes in correlation - or if it's random.