A couple of replies. Russ both says

Does Tyler really believe that the median male worker of today has the same standard of living today as in 1969? My original challenge to Tyler was to give me his interpretation of data points like the one he invokes here. It is certainly consistent with his story but it’s too consistent. Per capita GDP has increased a lot since 1969. Did none of it go to the median male worker? Besides problems with inflation measurement in the wage data, there is also a problem with benefits. By the way, the income data leaves out a great deal as well.

First, unless we are getting into issues between the Implicit GDP deflator and the Consumer Price Index, then inflation bias is not going to explain the deviation of real per capita GDP from real median income. Both are deflated figures.

Also, I think it helps here to move away from the median, that way we can avoid discussions about filtering down.

A core question is whether or not increases in GDP are being reflected in market wages and if so for whom. Below I graph per capita GDP versus mean male and female income.


You can see that even mean growth for men lags far behind GDP per capita. Moreover there were two broad periods of stagnation from 1969 to 1993 and from 1998 to now. That can’t be explained merely by recession. We have only one period of clear growth.

Now, lets look at indices, to get a better look at relative change.


Check that out. Female income tracks GDP per capita almost dead-on. These two series are not derived from the same underlying data. So, the fact that female mean income tracks real GDP is notable.

This makes it harder to say that this is simply something about the difference between per capita GDP and real income.

Now lets dispense with the difference between male and female. Easy to come by data only goes back to 1974 here but we still see the same phase pattern we saw in the male data if less pronounced.


Now, lets do GDP per hour worked. Here we can see a pick up in GDP per hours worked in the 90s as well though unlike the boost in mean incomes it persists through the dot-com recession.


I think the take away is that there are genuine phases, or put another way, the 90s really were different. Something was going in the economy from 1969 to at least 2011 with the exception of a brief period in the 90s.

I don’t think it can be explained away by appeals to inflation bias.

Second, Russ says,

I bring up the distorting effect of divorce on the data not because I think it redeems the last 40 years. That’s not what we’re talking about. We’re talking about whether the economy is broken. If the average person does not share in economic growth of the magnitude we’ve had in the last 40 years then we should have a revolution not a modest increase in taxes on the rich or a hike in the minimum wage.

I don’t know if we should go around having revolutions regardless of what the median income data say.

However, more to the point: the economy in and of itself cannot be broken. The economy simply is.

We might have certain feelings about the economy but “the economy is broken” is not fundamentally a statement about the economy it is a statement about our emotions.

How people feel is important, but it is not at issue here and on a more basic level I don’t know if its really relevant what the income data say when it comes to that. If people are sad or upset then they are sad or upset. This needs to be addressed.

Analyzing the income data, however, is not likely to make many sad people happy or many happy people sad.

As for the deviation of real per capita GDP from mean income I think we have some plausible candidates for the discrepancy – health care seems to be the big driver, but we’ll have to look at that evidence in a later post.