From BLS.gov's data on educational attainment, employment, & weekly wage/salary for the dozen years 2000 - 2011:

For wage/salary, & employment data of full time employees over age 25 (by educational attainment) it's clear that wage/salary growth didn't keep pace with productivity growth.... so employee's income's don't keep pace with productivity gains.  Only 2'/3's of the total productivity gains were applied to wage/salary growth in the composite.  What's more important though is that virtually the entire gain in wages/salaries were applied to employees with a Bachelor's or higher degree, while those with a high school degree or lower took a major hit in both numbers of employed and in the wage/salary losses (in real terms). 

Overall real wage/salary growth over the 12 year period was 2.8%... which occurred by decreasing composite wage/salary paid out for the employees with lower educational levels, while increasing the composite wage/salary outlay for those with Bachelor's and higher levels of education.

Over the same period productivity grew by 4.4%... so only 2/3's (65%) of the increase in productivity were recognized in the composite outlay for wages/salaries.  Significantly though, the 2/3's of productivity gain recognized as a change in wage/salary were recognized only in employees with a Bachelors or higher degree.  .... effectively those gains were at the expense of employees with less than a Bachelor's degree. 

This is merely the continuation of the means by which median family income has continually and increasingly diverged from productivity beginning in the late 1970's.  This also shifts the distribution of income from the middle class. 

The change in employment for the period shows that the annual rate of loss was in those with high school diploma and less (-1.12%/year), with all the gains in those with a Bachelor's degree or better (+1.78%/year)... .while the overall change in employment (full time, 25 years and older) was only 0.22%/year for the most recent dozen years.

The weighted average weekly wage/salary in 2011 $ for each educational attainment group is shown in the table below.  The composite wage/salary outlay for those with a Bachelors and better educational attainment is over $41.7 billion while for those with a high school diploma or less is $20.1 billion.   This constitutes a 108% greater composite income of employees with the higher educational attainment.   This doesn't even take those workers under 25 or those of any age who aren't working full time!... so the divergence is even greater than indicated. 

* in thousands

In contrast the 2000 numbers show the higher educational bracket's composite wage/salary income was $34.8 billion, while those with the lower educational levels was $24.0 billion, constituting a 45% greater income for those with the higher educational levels. Thus, over the last dozen years the divergence of incomes has increased by a factor of 2.4x (= 108% / 45%).... which amounts to a rate of divergence increase of ~ 10% / year. 

This isn't just because the higher educational group gained $7 billion (out of $35 billion or +20%) ), but also because the lower group actually lost $4 billion (out of $20 billion, or -20%).  In other words the differential grew by $11 billion (or by ~ 9

The upshot is that the divergence of middle family incomes with productivity gains over time, and over the most recent dozen years as well, appears to be due to the productivity gains occurring by higher level educated employees creating advances in productive methods which displace the labor provided by the lower level employees.  The effect is that the remaining lower educated employees are paid increasingly less in composite (in real terms) because their availability increases n the market place.. those displaced become available for lower paying employment... while a greater number of the higher level educated employees are highered and paid at 2-3 times that of the lower levels, though they also have lost income in real terms over the dozen years.  The net is that employment overall barely changes (+0.2%/year), but productivity increases and total real outlay of wages/salaries increases at only half the rate of the productivity gains. 

There's an economic dynamic at play here ....  as long as the more highly educated employees can effect productivity gains at roughly twice the rate (or more) of overall wage/salary growth, then it's economically beneficial to capital owners profits as well as to the employment of the more highly educated (though not to their real income gains overall... i.e. only the top decile those with advanced degrees had a real income gain over that past 12 years... all the rest saw real income's decline).  

With advances in technology increasing at an increasing rate then it will become increasingly possible to create greater and greater productivity gains with relatively fewer and fewer of the higher educated hired to effect the productivity gains.  This reminds me of the period in time when technology company's were training their offshore engineering counterparts how to do the jobs they were doing domestically.... effectively training their replacements.  

The excess of gains in productivity over employment costs go to the capital owners as profits so that increasing rates of productivity gains go increasingly to capital owners and relatively less to the higher educated echelon of employees over the long run.  This continually increases the rate of supply of now lower educated employees... whether or not they become better educated with time, with simultaneous reduction in demand for their employment at higher wages... leading to an excess of supply of labor available for lower skill labor hence reductions in real incomes of that proportion of our workforce.  

You can see this in the income quartiles in the following chart which shows the percent changes in incomes by each educational group over the last dozen years. The greatest loss in weighted average incomes occurred in the less than High School educated group, but the high school + group (some college) wasn't far behind, and the Bachelor's group was next in line in loss of real incomes. 

All this goes to show & say is that the divergence of median family incomes, and those of most employees continues to diverge from productivity gains and consequently GDP in real terms.  This is an unsustainable economic model without some means of redistributing economic gains to the broader base.of the society.  In any event it appears that gov't policies other than those that restrict unions or effective union formation aren't directly responsible for the increasing divergence of median family incomes with productivity gains and GDP growth rates since the late 1970's.  What's causing it is the capitalist system with increasing technological improvements that can be translated to productivity gains by the more highly educated at the direct economic expense of the larger & broader labor supply.  The effect has been going on increasingly for the last 30+ years... with no end in yet in sight  Consolidations and offshoring are just peripheral aspects which exacerbate and perhaps accelerate the problem, but they are not a direct cause. 

 Addendum:

To  illustrate a new reality in the economics of educational attainment and wages, consider the following scenario. 

Assume you're a brilliant person having just graduated High School with straight A's and taken all advanced placement courses.  Assume this lets you be able to attend a major university on a full tuition & books scholarship and that if you did this you can graduate with an advanced (6 yr) degree.  Further, assume that with that 6 year advanced degree you can start working at the median wage level (the median wage for persons with an advanced degree).

But, by attending the university for 6 years you are forgoing 6 years of wages in the meantime.

Now for some choices.  Assume that on graduating high-school, because of your brilliance and straight A's in advanced placement courses you could start right away and for the 1st year be paid at the lower 10 percentile of wages (for high school graduates with no college).  That you can earn the 2nd Quartile wage during the 2nd year of employment, over the next two years advance to earn the median wage for high school graduates, then in the 5th year earn the 3rd quartile wage, and in the 6th year (and beyond) earn the upper 10% of high school graduates wage.

If you attend the university for the 6 years you could otherwise be earning decent money and moving up the wage scale, your forgoing the above earnings.  

With the above assumptions how many years of work does it take after finishing the 6 year degree to earn back what you would have earned over the same 6 years by not attending the university.  Answer: 3.4 years. 

In other words, by spending 6 years at the university you need to spend the next 3.4 years working just to break even. 

OK, you say... great, a 6 year investment in education to get an advanced degree get's me to break even in just 3.4 more years.

So 9.4 years elapse from high school graduation to break even by investing in the 6 year advanced degree education.

But during the same 3.4 years you would have been earning the upper 10% wage level of high school graduates, compared to earning the median wage for those with an advanced degree.  So after these elapsed 9.4 years how much more would you have earned by attending the university than by starting work right out of high school?  

Answer: You would have actually earned 47% LESS (not more!) than had you worked directly out of high school for those 9.4 years.

OK, so how much longer (after the 9.4 years) would you have to work at the median wage for an advanced degree wage earner to equal the earnings by working at the upper 10% level wage for those with just a high school degree?   

Answer: You can't make up the difference in your entire working career!!!! even if you worked for another 40 years you would still have earned 4% LESS than had you not gone to university and graduated with an advanced (6 year) degree. 

In other words if you're among the smart ones in high school, and even if it costs you (or your parents) nothing to attend university for 6 years to graduate with an advanced degree, you could earn more by applying your high school and smarts to working your way up the wage ladder competing with high school graduates than you can be attending college and competing with college graduates with advanced degrees.  

If you add in the costs of attending a university to get the 6 year degree it's even worse of course.

BTW, the situation was better in real terms in 2000 than it is in 2011 in this regard.  And it was even far better than that earlier still.  The reason is that as educational levels of attainment have continually increased, the differential in wage levels as decreased .... across the board.  For example the differential between a upper 10% High School graduate wage and the median wage for an advanced degree was 6.7% (relative to the high school upper 10% wage level) in 2000, while it's been reduced to a differential of just 4.7% by 2011. 

This isn't to say that the differentials will continue to erode (though there's no reason why they wouldn't that I can fathom... and they have been continuously as greater proportions of the work force have higher educational attainment levels), nor that the satisfaction of work in a field of choice with an advanced degree doesn't have other (non-financial) benefits... but in terms of who makes more money over time it's not the college grad with an advanced degree ... it's the smart high school kid that is able to differentiate himself from other high school grads and work their way up the wage ladder rapidly in competing with other high school grads of a far lower capability on average.   BTW, this assume both work only 40 hrs a week.... my experience and that of my peers and current college degreed workers is that they work far more than a 40 hour week on average.... 50 hrs / week is probably a closer estimate. 

Just thought it was relevant to looking a bit into how the future of income with educational attainment will tend to play out.