A doctorate in some areas of the humanities is worse than useless, because it opts you out of most of the jobs that can be gotten through the resume-and-interview process. You're overqualified for everything except laughably scarce tenure-track teaching positions.
You don't have to look at it that way, though. A PhD in a quantitative social science, or maybe a media studies program would be in a good position to start founding startups...
I just finished vol. 1 of Niall Ferguson's history of the rothschild banking family last night - and reading this makes me wonder how much great historical research we're missing out on because the economics of the historian profession is so bleak. There's got to be a better way to support people who want to work on these things than the present system - I personally know several people who wanted to be historians (who knows if they would have been any good) but ended up becoming IP and Tax lawyers instead. Sure, only a tiny fraction of historical writers will actually write things that I would bother to read, but stories like the above surely drive talented people away from these fields. (although I have to admit it's hard to imagine ferguson becoming an IP lawyer, or starting a web service company)
Cool. I'm guessing you liked the book? I picked up a copy just yesterday, on the off chance that it would tie in with some of my other historical interests. Does it mention the Fuggers?
At any rate, there is a better way to support the people who do work like this, although the mechanism is peripheral and counterintuitive. A parallel example might be to say that stock markets are important innovations in the field of public health, because they fund corporations that can aggregate and compartmentalize the risk of new drug discovery.
So without beating any farther around the bush: prediction markets (like Intrade) could monetize huge chunks of information that are currently only interesting to theoreticians, if they were capitalized at the level of today's stock or commodities markets.
If speculators were laying billion-dollar bets on, say, whether the US would invade Iraq in 2002 [1], the opinions of prominent historians and political scientists who study the Gulf region would be much more explicitly in demand. Such people would still have to publish to get prominent (so that the benefits of an open academic environment would still accrue to the public), and there would be additional pressure to produce falsifiable predictions instead of... whatever it is that gets produced now.
Here's a short intro for people who don't like math, signed by a veritable Who's Who of economics and public policy:
[1] It ought to be obvious why arbitrageurs would want to do this. I picked an old and inflammatory example to emphasize how many current political disagreements are at root questions of fact, about which people ought to be able to agree eventually...
