Mark Thoma points to a research paper by Lawrence Mishel and Heidi Shierholz titled "The sad but true story of wages in America". The study points to divergence between increase in productivity and increase in wages.
First we must be clear with which productivity we are talking about. Increasing labour productivity must be complimented with increasing wages. However, higher productivity without any change in labour productivity need not be. Is this possible? Of course it is.
At the risk of oversimplification, imagine a simple bread bakery. Its output is supported by various types of labour like bakers, sales staff, cleaning crew etc. Now imagine such a bakery buys an automatic bread making machine. All you do is put the dough in from one side and press a button. On the other side you get the most fabulous breads. In the new setup, the output supports sales staff, cleaning crew and a low-cost attendant. Now this machine needs maintenance which is on contract and there has been corresponding employment at the bread-machine manufacturing factory. But you can imagine the number of people supported by volume of bread produced is far lower. These people together take less percentage of sales price of bread but since their number is smaller they earn more. In a way, this is higher order of economies of scale in action. I agree that the knowledge provided by the bread-machine makers is high and deserves the monetary returns it gets. You can imagine this process running over and over again till gains accumulate with those that provide distinct knowledge. For example, the consulting baker at the bread-machine making factory would earn multiple times average baker.
So I would like to see change in capital intensity of production during the same time. An alternate, thus, would be that technology is resulting in polarization of incomes. On one side, it increases the scale that can be achieved by distinctive knowledge provider and hence scales up her returns / income, on the other it reduces the quality of labour required for the process and hence lowering the expected incomes of those still employed.
This change is structural shift from effort-based society to a knowledge-based society. Naturally, those without college degree, in other words ability to do knowledge work, have stagnant pay. In knowledge based society, the quality of effort is reduced and hence pay required to carry out the job reduces.
To understand real productivity and incomes, we need new concepts such as "wage content of a job" or its reverse "job content of a wage". The former would refer to wage change of specified standard set of jobs, sort of a job basket similar to consumption basket for inflation. The later would refer to the value of work that we can get done at a particular wage. If we measure this over time we will get better understanding of changes taking place in the economy. At the end, the job-profile, income profile and knowledge profile of a country should match and if it does, we can say capitalism is working fine.
I discuss some of these concepts in my book "Subverting Capitalism and Democracy".
My book "Subverting Capitalism & Democracy" is available on Amazon and Kindle.