Um what's to stop the expert making a mistake and not realising that he has insufficient expertise?
The scenario is that Wikipedia editors have found an expert with a proven track-record in the specific subject area of the article, and solicited his opinion. In this scenario, the expert's track-record provides some evidence that the expert has the expertise -- we wouldn't be relying on his or her self-evaluation. Such evidence could never be conclusive, but I think it's a lot better than none at all.
-- Matt (User:Matt Crypto)
This is NOT a choice between one single person's academic accreditation vs. one single person's self-evaluation. It's a choice between the ''collective'' wisdom and joint assessment of a larger group of people** vs. one single person's academic accreditation.
** some of them likely will just so happen to hold academic degrees themselves, without that ever being an issue or getting mentioned
- Jens