Vinod's Blog
Random musings from a libertarian, tech geek...
Wednesday, February 27, 2008 - 09:30 AM Permanent link for Income Statistics
Income Statistics

One of the things that Thomas Sowell points out in this interview is how politicians are able to pick & choose different income stats in order to make different cases.   In particular, given election season, he notes that whenever anyone starts talking about "median household income" they're usually trying to make income statistics sound bad.

So, I did some digging to find the stats myself and the divergence between "household income" and "Real GDP Per Capita" is interesting to say the least -

[GDP data comes from here;  Median Household Income data is from here]

Why the delta?   Well, the key words are "median household". 

On the "household" side, demographics are changing dramatically.  For example, young single working folks (women in particular!) are far more likely to live in their own households now vs. in the past.  That combined with smaller families and more homeowners in general means that a given basket of income is spread over a larger number of homes.  Thus, and somewhat perversely, more flexibility in living status brings down this statistic significantly.

Similarly, the point in the distribution that's represented by "median" (rather than the "mean") is also changing dramatically.   Broadly speaking, the mean fluctuates less with changes in the distribution curve vs. the median.  For example, new immigrants (illegal or legal), fresh college grads, and the like populate the low end of the curve and enter the labor pool at a relative income level that hasn't changed much in the past few decades.  Today's grad students live at about the same income level as 20 yrs ago, AND there are a lot more of them, AND their incomes due to grad degrees will be far greater.  Additionally, phenomena like "assortive mating" as high income women seek out high income men skews the distribution vs. the old skool when men were more likely to marry housewives.

Wikipedia notes these effects from 1969 to 1996 and there's every reason to believe they continue to be the case in the dataset I highlighted above -

In 1969, more than 40% of all households consisted of a married couple with children. By 1996 only a rough quarter of US households consisted of married couples with children. As a result of these changing household demographics, median household income rose relatively slow despite an ever increasing female labor force and a considerable increase in the percentage of college graduates.

Permanent link for Income Statistics   Comments [ ] :: Main :: Archives