Actually it's not. You can have a hundred cranks making predictions about the economy and out of that hundred a few of them will appear to be accurate entirely through chance. A good track record in this case consists of not trying to make predictions from sets of data that are often basically random walks. Think of him as a doctor who advises a patient that to be healthy they have to avoid smoking and junk food and get regular exercise, rather than a witch doctor who can predict the day you'll have a heart attack. Basically, the question is does he act like a scientist or like a woo-woo? And to my knowledge, in his professional life, he has always acted like a scientist.