Kevin Drum takes a courageous and rare stand on the internet, arguing that yes, downloading millions of files from JSTOR while evading attempts by both JSTOR, and the owner of the network you're using, to stop you, is, well, pretty close to stealing
[stepping onto soapbox]
Forget the analogy with stealing. The online pricing model at JSTOR and the major academic journals is totally inappropriate. The basic business model is to charge academic institutions a ton of money to give students and faculty "free" access to content. In order to protect that business model, those of us who are not academics get charged ridiculous amounts--I've seen $35 to download one journal article, for crying out loud.
In my opinion, there should be two sources of revenue for repositories of online research (by a repository, I mean ideally a database of just about every journal article and working paper on every subject; you do not have to achieve that level of coverage, but I want to make it clear that I am talking about more than just a single publisher's journals) :
1. Annual user fees from frequent users of, say, $20 per year. You are a frequent user if you download 10 or more papers a year. Up to that point, your downloads are free.
2. Donations. If the donors who give to higher education would divert just a small fraction of their funds to support an online repository, that would be plenty.
I am not wedded to this solution. What I want for online research pricing is "Don't be evil." JSTOR, the NBER, and the journals are evil. The way they treat intellectual property gives the slogan "property is theft" some credibility.
[stepping down from soapbox]