Everyone know's where this one is going: the salaries for post-docs (and of course grad students) in the life-sciences are low, the ones in the so-called hard-sciences only marginally higher (I am including engineering in the hard sciences as there is a great overlap) and in social sciences the lowest. The annual survey of post-docs by the Scientific Research Society Sigma Xi (predictably) shows the same figures year after year (link): more than 50,000 people hold postdoctoral appointments in the United States. The majority are federally funded (69%). Most are doing research in the life or health sciences, but 22% are in the physical sciences or engineering, and 4% are working in the social sciences or humanities.
Now for the fun facts: the median salary for a worker with a bachelor's degree was $45,000, if you have a master's the figure jumps to $56,000 (an increase of ~24%) and if you have a doctoral degree or professional degree, the median is about $71,500 (a 59% increase over a bachelor's degree). That's all well and good, now where do these post-docs rank? Sensibly, one might say they can't expect to get as much money as in industry or the free market - fair to assume - so lets say they make more than a master's degree holder, i.e. more than $60,000. Far be it from it however, the number is $38,000 (a decrease of ~16% over a bachelor's degree holder).
The question that is open to debate is now what mechanisms are there to increase post-doc pay? There are no post-doc unions (imagine the horror at a faculty meeting if all post-docs would go on strike) nor lobbying groups (who would pay the lobbyists their six-figure salary?). Add to that the fact that the majority of the post-doctoral researchers in the US are actually not US citizens, nor permanent residents and you have a nice set up for a constantly rotating, cheap and extremely skilled workforce that is unable and possibly unwilling to campaign for change so as to not endanger their livelihood, careers etc... Simply put, as long as the majority of the funding for post-doc positions comes from the federal government, the salaries are going to remain low.
What is then to be done? I for one would welcome increased collaboration with industrial research departments: meet half-way, not all industry research is so far away from the lab and not all lab research is useless to industry. In an effort to increase the wages and thereby increases retention of the nations most intensely schooled workforce, I say increase collaboration with the industry, after all, that is where the money is.