The Distribution of Wages and Salaries in 2012

For a while now, folks have been passing around the latest average wage index release from the Social Secuity Administration. It is meant to be the wages and salaries reported by employers through W-2 filings for their workers. I was curious to see what the 2012 Census data had to say on this question. So I went into the microdata and pulled out the distribution of personal wages and salary earnings (WSAL-VAL) among persons that had greater than zero wages and salary earnings. This was the result.

The result is somewhat similar to the SSA's figures. The 90th percentile line is exactly the same for instance. On the bottom, the Census' figures are significantly higher than the SSA's though. The SSA's data comes from administrative filings (not a survey) and is much more complete. So you would generally imagine it has better quality data here. However, it might also be that the two sources are measuring slightly different things.

Whichever source you use, there is a reason we don't measure poverty and such on the personal level. Focusing purely on personal income ignores that people who are living in larger families tend to pool income and resources, which affects their labor market behavior. For instance, suppose spouse one has an income of $50,000 a year and the family intends that spouse two primarily stay home and do the household and care labor of the family. However, they would like somewhat more than $50,000 a year and so spouse two gets a part-time job that makes $7,000 per year for a fairly small amount of labor. Looking at personal income (especially without looking at hours worked) could cause you to miconceive of this person as very impoverished and exploited and such when they aren't really.

None of this is to say personal income is not an important thing to look at for some purposes. But to the extent that people are interested in how people are actually living and what kind of income they actually have access to, family income is generally the way to go, which is why most income and poverty statistics use it instead of personal income.