No one thinks markets are perfect, and EMH never says this. The proof that markets are efficient is that it is so improbable one can generate alpha — something you, like most EMH critics, concede. But the implications do not seem obvious. That you were able to find one person in 2004 and turn his measured warning into something that would have drastically reversed the regulatory emphasis on weakening underwriting standards is classic hindsight wisdom.
Thanks to Chris Dillow for the pointer.
The wider point is that though markets are very far from perfect (and we have been shown many examples of that over the last couple of years), we should not be comparing markets against perfection but against the alternative. The alternative is some command (probably political of firm-orientated) structure. The Guardian looks here at local government planning.