Traditional open source licenses like GPL create a hard-to-cross chasm between it and traditional economics; essentially you can trade code for code, but cannot trade code for anything else (with the traditional medium of that trade being $).
Imagine if some highly successful OSS product "expanded" GPL to include a monetary contribution clause (i.e if you don't want copyleft then just pay X -- instead of contributing code the licensor has the option to contribute money). That money could trivially be used to hire programmers and therefore be converted back into code... or used to pay contributers.
Of course this opens up a huge can of worms, like who gets paid how much (I think that it was very wise for OSS and GPL to steer clear of these issues during its incubation period). As many people have posted, a "trust network" can only be part of the equation... I think that a more quantitative algorithm could be created that captures contributions in code, docs, and community. But the big problem with any quantitative algorithm is that people could change their style to trick the algorithm into thinking they contributed more; for example, to fake out SLOC counts its pretty easy to deliberately write large amounts of code to do small jobs. Enter the trust network; perhaps it can be used to catch and regulate abusers.
I've expanded on the idea in my blog here: http://effluviaofascatteredmind.blogspot.com/2009/03/thoughts-on-gpl-open-company-concept.html. I think I've already exceeded what most slashdot readers really want to read :-).