Last Friday Arnold Kling at Econlog and Russell Roberts at Cafehayek commented on Larry Bartels book Unequal Democracy and I made some comments
about what they were saying.
Bartels used family income data in his book and both Arnold and Russell ignored this point and went into long discussion about why household income data is so distorted that the comparisons in Unequal Democracy were meaningless. Both Arnold and Russell are correct that the rapid growth in single women households over recent decades is a significant factor behind recent changes in poverty and this growth in single mothers and divorced women does impact the data. But GMU libertarians are not the only ones that know the household data is distorted by the growth in single women households. The Census Bureau where the data originates and Bartels are well aware of these problems. That is why Census also publishes data on family income and why Bartel used family income data in his book.
I suspect neither Arnold nor Russell have ever looked at the data to see how the growth of single women families has impacted the data. Rather they have heard someone use these criticisms of the data and just started used it because it agreed with their priors. But the data is readily available at Census.
First, what has happened to family income growth.
As the chart shows there was a sharp slowdown in family income growth sometime in the late 1970s. The trend growth rate of real family income slowed from 2.8% from 1950 to 1980 to only
0.8% from 1975 to 2005. The more recent growth rate was only 30% of the old trend and if the old trend growth had continued family income would have been 177% higher in 2005.
So what happened to married couples real income.
Married couples — including both families where the wife worked and where the wife stayed home — experienced a very similar slowdown in real income growth from a trend growth rate of 2.8% from 1950 to 1980 to only a 1.1% trend after 1975– the new trend is 40% of the old one If the old trend for married couples had continued their real income would have been 172% larger in 2005.
Now Russell Roberts argues that the growth in single women households distorted the data so severely that one could not make meaningful comparisons. But that does not look right to me.
It is very easy to see that the growth of single person household had a modest impact on real family income growth. Essentially, it is responsible for the fact that since 1975 real family income growth was 0.8% and married couples real income growth rate was 1.1%. This 0.3 percentage point impact from the growth in single person families clearly does not look like such a sever distortion that comparisons are meaningless. Moreover, the point that total family income would have been 177% higher and married couples income would have been 172% higher if the old trends continued does not look like such a sever distortion that no one can make meaningful comparisons. A five percentage point difference over 30 years does not look like a massive distortion that makes comparisons meaningless.
I’m not sure what else to do with this information. I’m sure that both Arnold and Russell heard this criticism at some right wing think tanks or blog and just picked it up without ever thinking about it. But too many people at right wing think tanks are paid to put out research that supports a point of view that one can not trust their research conclusions. The growth in single person households sounds dramatic and it could be very easy to accept someone quoting this data when they conclude that the income data is unrealistic. But as they say down home, if you lay down with dogs you get fleas.
What I’m curious about is what happens to the sharp 18 year old undergraduate at GMU that recognizes that their professors are dissembling and does not buy into their ideological distortions. I sure hope these professors do not grade on the basis of their students ideological purity.