This must be my day to be baffled. I don’t see how Mark Perry reaches the conclusion that he does:
America’s middle-class is disappearing, but into higher, not lower, income categories
based on the data that he presents. He presents four graphs, all drawn from Census Bureau data. The first illustrates average household income, median household income, and average household size. The second illustrates average and median household income per member. The third illustrates median family income for married couples, both spouses working. The fourth illustrates percent income shares for high middle and local income groups.
What leaps out at you from the graphs is that incomes for the highest group have grown substantially, the others not so much.
My first quibble is that I think his definition of “middle income” is peculiar. How does he arrive at the distribution he does? A “drunkard’s search”? It would seem to me that reasonable definitions would be median plus or minus one standard deviation (middle would be two-thirds, a sixth “high”) or median plus or minus standard deviations (95% middle, 2.5% high). Under his definition “high income” is about a twelfth, either too high or too low to be intuitive.
But my biggest gripe is that I just don’t see how the conclusion can be reached from the data. Wouldn’t the large proportion of middle and higher income jobs, particularly in the $80,000 to $120,000 range, that have gone to immigrants provide an alternative explanation? If that’s the case it’s not so much a “middle income disappearing into upper income” story as one of importing a new upper middle income group to replace the previous one.
$100k seems like a reasonable cutoff, assuming you use real dollars. I think the point of the piece is that more households earn over that threshold than they did in the past. His numbers look good at first pass.
Steve
Fallacy of aggregation
Additionally, $100K is an arbitrary cut-off.
This is basically what that series of Pew reports on the shrinking middle class states. The number of households btw/ 67% to 200% of median income has been declining since ’71. More of this change is due to an increasing number of households above 200% of median income than below 67%, but both have increased.
http://www.pewsocialtrends.org/2015/12/09/the-american-middle-class-is-losing-ground/
Which is another way of saying that $100K as the cut-off may be artful.
There are other ways of reaching the results. Imagine five widgets. One widget has a value of 1, three of them have a value of 10, and the fifth has a value of 50. Subtract 4 from each of the value 10 widgets, add one widget with a value of 8, and add 4 to the 50 widget. Mathematically, that model has all of the characteristics of Dr. Perry’s and, presumably, Dr. Perry would group the new 8 widget with the 50 widget rather than with the widgets that now have value 6 (because 8 is his cut-off).
When you state a proposition as Dr. Perry has, you must also demonstrate that your case is the only or best explanation for the observable facts. When you respond to someone else’s proposition you only need demonstrate that there are other explanations or that another explanation satisfies the facts better. I’ve done that.
My relatively simple model has other virtues. Dr. Perry’s proposition and case do not explain why so many people are unhappy. To do that you must add another factor, e.g. they’re ignorant ingrates, bamboozled by the unscrupulous, etc. That fails by Occam’s Razor. Not only does my model explain why people are unhappy (60% of the people are no better off or worse off than they were), it’s more robust than Dr. Perry’s case. By that I mean that the more widgets you add and the more you adjust their values, the more closely my model resembles the real world. Intuitive and robust are pretty strong advantages.
Steve, it’s understandable that you might find that Dr. Perry’s proposition and case make intuitive sense because your profession is one of the few that, over the period of the last 50 years, has moved from the middle income range to the upper income.