Why is it that although nearly all research published in academic journals is funded in one way or another by taxpayers, these journals are nearly all private, commercially run operations that charge high subscription fees and exclude the public from accessing them? And what a lot of money there is to be made charging extortionate fees to university libraries for access to knowledge.
What service exactly do they provide in return? Typesetting? Printing? Is that really so hard in this electronic age? Their content comes free (though some journals have the nerve to charge submission fees in one way or another). The editors are generally unpaid academics volunteering their research time (though often with generous expenses from the publisher and university support). The quality control system of peer-review is provided on a voluntary unpaid basis by expert academics in the field, who generally acquired that expertise in publicly funded universities and who are using their publicly funded research time to do it. In economic terms, academic publishers are rentiers: business who make their money from other people's work and investment simply by virtue of occupying a strategic position in a production system that comes with bottle-neck owning negotiating privileges.
All in all it looks like the big academic publishers have found a way to take the public's money and convert it into private profit without even providing a public service.
Once one recognises this outrageous situation the next issue is what to do about it. Here I think it's important to think about why publishing has been left to private (profit-making) companies in the first place, and that means thinking about why academics themselves don't want to do it. Academics want a comfortable research environment that allows them to get on with their work. They want a comfortable office, but don't want to be concerned with building maintenance; they want a comfortable salary, but don't want to think about department budgets. Likewise, they want a system for publicising their research without having to set up publishing institutions and worrying about how to set up sophisticated websites, cross-referencing systems, or type-setting; and they want to be able to quickly identify the value of an article among the thousands on any particular topic, without expending their precious time reading it. The only thing they care about is whether this system works well enough that they don't have to think about it.
There's no reason academics couldn't set up their own journals and do a better job. (I've actually been involved in setting one up - all it takes is a budget of a few hundred euros for a simple website and a lot of time.) But it's not where academics' comparative advantage lies. Every hour they spend on things like that or other administrative duties is an hour lost to their productive work. And so it's not surprising that, despite occasional outbreaks of academic run online journals like PloS, most academics are happy to keep outsourcing the whole business to whomever can promise to do a reasonable job and not bother them about it. And that they don't much care if it's run by private or public institutions (since the bills go to the university library).
The whole point of academic publishing is to separate wheat from chaff and so facilitate the price mechanism (impact factor) that allows the market for ideas to run smoothly. That means that it is a natural oligopoly, like credit rating agencies - islands of reputational reliability that channel the mass of information into a recognisable market order. The question of whether this essential piece of academic infrastructure should be run on private profit seeking or public good principles is of most interest not to academics themselves (who don't care what happens so long as their working isn't disrupted), but to the public in general and specific groups like libraries, journalists, and lay-academics. That means it is a political debate that reaches far beyond the universities.