Is there any value in self-publishing a book as an academic?

I would recommend that you ask yourself a question: why do I want this book "officially" published? You could, after all, just make the material free online via arXiv or as a webpage or an archival technical report of various types depending on its contents and your situation.

• Prestige of having a published book? A pay-to-publish press will generally give you negative credibility, since there are so many publishers that will publish your book without making you pay (including lots of dodgy low-quality ones that will publish pretty much anything). Pay-to-publish smells like desperation or resume-padding.
• Money from sales/royalties? If they thought it was going to make money, they would be paying you, not the other way around.
• Higher visibility? A pay-to-publish press will generally not be any good at promoting your work. They're not going to invest any real money in promotion if they aren't expecting to make significant money from sales.

Serious and reputable academic publishing houses are always looking for good books to publish. The right ones for your field will generally have booths at the major conferences in the field, with representatives that you can talk to.

The value of a pay-to-publish book is approximately the same as any other pay-to-publish material.

As described in this answer, the only way for someone to judge a non-peer-reviewed publications is to

Read them, see what they're worth.

So, if this book attracts a large number of readers in your academic community and they are impressed, it will enhance your reputation. This is not likely, unless you are already well-known in your field.

If people read the book and think it's bad, it will detract from your reputation.

If very few people bother to read it (which is the most likely case), it will confer (at best) zero benefit to you and will be a waste of money.

It may also confer negative reputation benefit. People who don't read the book may assume that the book is bad, and that if it was really good, you wouldn't have had to pay-to-publish.

For further reading, there's a relevant article over at InsideHigherEd.

Virtually none.

The main difference between academic/university presses and trade presses (including vanity presses and self-publishing) are that academic presses engage in a process of peer review before selecting which material to publish.

That is, academic presses will send your ms out to peer academics who then provide reader reports that will be used by the academic presses' board of directors/editors (usually academics themselves) to make the decision to publish the book or not -- usually after revisions indicated by said external readers.

This puts a book from an academic press in the category of peer-reviewed publications.

Note that there are some second and lower tier colleges don't care where a scholar's book comes from. If you suspect you are in one of these colleges -- ask your mentors and senior colleagues what you should do.

p.s. Once you have a certain degree of fame, you can publish with trade presses. Stephen Hawking really doesn't need the CV boost that might have happened if A Brief History had been published by U-Cambridge Press rather than Bantam Books. And if you are in fields that view books with disdain (say... physics) then it wouldn't really matter where or even if you published a book.