Why Harvard Might Not Be So Great (Or, one reason median wage gains have slowed since 2000)

Imagine that someone had come up with a vaccine or health supplement or drug that they claimed had incredible health benefits. They shared data that showed how people taking this were more fit, higher performing, and far healthier than the average American. Looking at the results, you would have to be impressed.

You ask to take it as well. At that point you're told that you can't just take it. You have to qualify. And when you learn the qualification criteria, you're blown away. They only let Olympic athletes take their product. And you suddenly suspect that their incredible results have more to do with their screening than the product itself.

Higher education does at least two things. One, it educates. Two, it is a signal to potential employers.

Harvard has about 22,000 students, roughly 17,000 from the US. Last year, about 3.6 million American kids graduated from high school. If Harvard's incoming class of freshmen is one-fifth of its total enrollment, slightly less than 0.1% of American high school graduates go to Harvard.

Is Harvard a fabulous place to get an education? I think it must be but we won't ever know how much better it is than other schools because Harvard will never run a randomized trial of simply letting any kids in from any school or neighborhood. It is possible that the list of America's top universities would change if we collected data from a decade of randomized trials rather than data testing the health benefits of a supplement only given to Olympic athletes.

The American economy has made the transition from an industrial to an information economy. (Which simply means that new jobs and wealth are not being created by folks standing at factory lines manipulating things but instead by folks sitting at computers manipulating the symbols of things.) Once upon a time, white collar workers were a subset of our workforce. They were the managers and analysts who constituted only a small portion of the many factory workers, the men (it was always men) who could wear a tie without worrying about it getting caught in machinery. Now, knowledge workers outnumber factory workers about 4 to 1. What this means is that an education system designed to signal to the market, "These graduates are the best and brightest and you should hire them for your key management and analyst positions" is less helpful. The point of education is not to identify the top 5%. 

The point is to make the median graduate more valuable. 

How have we done at improving the productivity of folks at the median since the turn of the century? Not as well. Between 2000 and 2015, real median personal income was unchanged. In the previous 15 years, it rose 28%. In the last 15 years of the 20th century, we could raise median wages by making factory workers more productive and moving a higher percentage of new workers from factories to cubicles. Productivity gains in factories continued as they had for nearly a century; the shift from lower productivity in factories to higher productivity behind a computer accounted for the rest of the gains. 

Now, though, gains in productivity have to come from making the median knowledge worker more productive. We still don't know as much about that.

As knowledge work becomes more common - that is, as a greater portion of our workforce moves into the information economy - signaling to markets which graduates are our best and brightest will be less important and our ability to make all of these graduates more productive becomes far more important. The latter will rely on a different set of strategies.

[to be continued]