http://www.newyorkfed.org/research/capital_markets/ycfaq.html Updated August 6th. July 2010 = 0.31 % http://www.newyorkfed.org/research/capital_markets/Prob_Rec.pdf
No offense, you post the best links here, but in the world I live in a 31% probability of anything based on correlation is noise. That would be clear if they published the scattergram of the data sets they used. If you saw it, you would say, "Hmmmm. There's almost something there. Maybe we are looking at the wrong independent varaible."