September 14, 2013
Brad Plummer has a good set of charts showing how different segments of the population have fared in the downturn. I have two minor quibbles with the selection. First, to show the decline in the labor share of output, chart 5 shows the labor share of GDP over the last three decades. This is slightly misleading. The depreciation share of GDP has risen by roughly two percentage points over this period, which means that if the division of wages and profits had stayed constant, the chart would still show a declining share of wages in GDP.
Folks should get in the habit if using net domestic product as the denominator. No one eats depreciation, if we want to look at distribution we should focus on net output, not gross output.
My other quibble is the use of the Sentier Research data for median income. This is a relatively new series which many journalists are turning to as a measure of family income. Unlike almost all the other widely used data sources, this one is not available for free. It might be worth paying for Sentier’s data if there were some valued added, but there really isn’t.
The Census produces annual data on family income. This is what people should look to for movements in median income over time. Sentier produces their data on a monthly basis, which may seem like a big gain, but it isn’t. The monthly movements (it’s actually a 3-month moving average) are dominated by noise. We are not really finding out what is happening with family income month to month, we are just picking up the impact of statistical quirks and erratic seasonal adjustment factors.
This is a case where free really is better. If you want to know what is going on with family income, stick with the Census Bureau.
Comments