The AHA Flap

I am trying to wrap my head around the American Historical Association's controversial statement that Ph.D. dissertations should be unavailable in digital form for six years. Here is the core of it:

At the same time, however, an increasing number of university presses are reluctant to offer a publishing contract to newly minted PhDs whose dissertations have been freely available via online sources.  Presumably, online readers will become familiar with an author’s particular argument, methodology, and archival sources, and will feel no need to buy the book once it is available.  As a result, students who must post their dissertations online immediately after they receive their degree can find themselves at a serious disadvantage in their effort to get their first book published; it is not unusual for an early-career historian to spend five or six years revising a dissertation and preparing the manuscript for submission to a press for consideration. 

I wish there was some evidence from publishers to this effect. A dissertation is very different from a book, as all scholars know it is much more raw and it lacks the gravitas of a press, thus making it much less likely to be sought out and cited (is there data on dissertation citation?). This article in The Atlantic suggests the evidence does not bear that out. On the flip side, via Raul Pacheco-Vega, here is the Oxford University Press site saying "We will not usually consider for publication any book held in its entirety or in
significant part in an institutional or commercial electronic depository." Is that the norm? The AHA doesn't say. Also, the "presumably" is a bit weird. Is there any evidence for it?
I have no direct involvement in this, though it interests me because I've always been close to historians (in addition to friends here at UNC Charlotte, I had historians both on my M.A. and Ph.D. committees) and some of my work is political history and bookish. So I feel I can relate to a certain extent, though I must say this flap really highlights how different history and political science are.
This may well just reflect an ongoing debate within the discipline about the primacy of publishing a book for tenure. If a book (or what really becomes THE BOOK for an assistant professor) is the only way to get tenure, then there is an incentive to keep everything about your book manuscript secret until it gets published, assuming that presses will ignore you if anyone can look at your dissertation. Unfortunately, this means that knowledge takes second place to process. It's a conscious effort to get as few people as possible to know who you are and what you do.
For junior scholars, this makes perversely good sense because they need that job. However, hiding knowledge cannot be good for any discipline (though here is an historian's take to the effect that if you wrote the dissertation, you should have the right to bury it in a hole if you want). I would hope there would be some other way of providing a viable path to tenure. THE BOOK is deeply entrenched, but need that be permanent?