Should journal editors know the identity of authors during the review process?

A junior philosopher writes:

Why is it that, typically, journal editors know the identity of the authors whose manuscripts they make decisions about?

I know that some journal editors process manuscripts blind, but many do not.  Given that referees are always blind, one might wonder whether it matters whether editors are.  But as we all know, editors often make important decisions not dictated soley by referee reports.  First, editors sometimes decide to reject papers before sending them to referees.  Second, editors decide which referees to choose, and how many referees reports to gather.  Often, different referees have different personalities that may be known to editors (for example, some may be harsher than others).  Third, the editor must decide what to do with they receive conflicting recommendations from referees (which, I hear, is not
uncommon).  At each of these stages it is possible that that the editor's decision could be influenced, either consciously or subconsiously, by factors that should be irrelevant, such as the prestige of the author's institution, the author's race or gender, the editor's perception of the quality of other papers by the author, whether the editor knows the author personally, etc.

That is the argument in favor of blinding editors.  What are the arguments against?  I can't think of many, but perhaps that's just because I'm not very familiar with the review process.  I would love to hear what others think on this matter.  I can see that it would complicate things administratively: there would have to be a way for editors to communicate their decisions to the authors without knowing who they are.  But this doesn't seem insurmountable.

Thoughts from readers?