This reminds me of the study by Philip Tetlock, where they found that 'experts' were **only slightly better than random guessing and the more confident ones tended to be worse.
I immediately think of Paul Krugman