Further Consideration of the Middle Class: An Increase in Single-Parent Families

Russ Roberts was kind enough to respond to my email asking him how he analyzed changing income distributions over the past several decades. I was pretty sure I had remembered him writing some things about it, and he pointed me to those things. Over the course of a couple posts I’ll review them.

This one draws from a 2009 post at Cafe Hayek. Here Roberts is responding to the following chart that was posted at the Freakonomics blog.

The chart could use some more labels, but basically it shows the annual percentage growth rate, in constant dollars, of mean* family income by quintile, and on the right, for the top 5% of families, for the 1947-1973 period and the 1973-2005 period.  The claim is that this shows that all the benefits of growth have been going to the wealthy, and  not to the middle and lower classes.

I think the data do demonstrate the stagnation in household income that we’ve seen in other data.  But Roberts argues the chart is meaningless because it’s too abstract.  Taken as abstract groups, the lowest quintile’s income has increased much more slowly in recent years than the lowest quintile’s income in previous years, but that doesn’t tell us how actual individuals have done because it’s not measuring the same groups of people.

Specifically, Roberts notes something that supports an idea I’ve had rolling around in my head (and which I mentioned, partly, in my discussion at Dispatches), which is that some of the stagnation of family income is surely attributable to an increase in single-family households (contra one commenter’s claim that income stagnation had happened despite an increase in two family incomes as women moved into the workforce).  I noted that there were more single-income households because people were marrying later, but I missed the more obvious point Roberts made:

The number of families increased dramatically simply because of divorce. There was also an increase in the number of families headed by single women with children. The quintile breaks-points changed, not because the economy was growing or shrinking but simply because of changes in the types of families.

True enough. What would median household income growth have looked like if households in the 1980s and ’90s contained fewer broken homes and single-mothers? Certainly, as Roberts notes, the quintile break points would be different.

I doubt that explains away all the stagnation in median household income growth rates, but it’s certainly part of the story, and a part that’s not attributable to some kind of economic malaise or nefarious socio-political scheme to shift wealth upward.


About J@m3z Aitch

J@m3z Aitch is a two-bit college professor who'd rather be canoeing.
This entry was posted in Uncategorized. Bookmark the permalink.

7 Responses to Further Consideration of the Middle Class: An Increase in Single-Parent Families

  1. D. C. Sessions says:

    OT: Although the stagnation in income (and as you know I prefer, wealth) for most of the US population worries me, I would argue that the larger threat to us politically is the rapid loss of intergenerational social mobility.

    Most analyses focus on upward social mobility, but the trend towards a permanent upper class is also worrisome.

  2. James Hanley says:


    I agree that loss of intergenerational mobility worrisome. But do we actually know it’s happening? Another item that Russ Roberts gave me, which I’ll post about soon, casts some doubt on that. And also, it’s hard to really judge that without having the perspective of several generations to look at, right? Not that you’re actually saying it quite this way, so I don’t mean this as a jab at you, but to say “right now intergenerational mobility is being lost” may be a non sequitur.

  3. D. C. Sessions says:

    And also, it’s hard to really judge that without having the perspective of several generations to look at, right?

    Well, of course we do have several generations to look at. Whether the time baseline is long enough to distinguish between noise and trend is another matter, but our mathematical ability to pull signal out of noise is better than most people appreciate.

    Just watch out for the null hypothesis.

  4. DensityDuck says:

    So, to some extent, the issue is the same as you see in a housing market where “average home price” is going up even though “total number of sales” is falling; what’s happening is that nobody’s buying the lower-price houses while all the high-end properties are still selling at the same rate. Therefore the “average home price” goes up, even though the actual sales prices are the same as they always were.

    The apparent effect is due to a change in the demographics of the population you’re studying. That is to say, you’re not looking at enough indicators.

    One of the things I learned in grad school is that no statistical conclusion can be based on the change of one parameter alone; any statistical study must be based on the movement of at least two variables.

  5. James Hanley says:


    So are you saying we do have multi-generational evidence that inter-generational mobility has declined? Could you provide a link?

  6. D. C. Sessions says:

    Nope. Pure methodological quibble.

    We don’t need to track mobility for the next 60 years to decide that something has happened in the last 20 (although doing so will narrow the error bars.)

    Further deponent sayeth naught.

  7. James Hanley says:

    Nope. Pure methodological quibble.


Comments are closed.