The Changing Distribution of Worker Earnings

On Friday, the Congressional Budget Office released a new study examining how worker earnings changed from 1979 through 2007. The report is full of important facts about the evolution of earnings throughout the earnings distribution and, in particular, among the highest earners.

For example, the following chart illustrates how the earnings of men and women (age 25-54) have changed at different points in the earnings distribution:

The chart confirms two well-known findings: men, on average, earn more than women, and high-earners have seen the largest earnings gains in recent decades. Other takeaways include:

• In real terms (i.e., adjusting for inflation), men at the 10^th and 50^th percentiles (of the male earnings distribution) in 2007 earned about the same as similarly situated men back in 1979.
• Women’s earnings have been growing faster than men’s. Women at the 10^th and 50^th percentile (of the female earnings distribution), for example, had higher earnings in 2007 than their counterparts in 1979.
• Women at the 90^th percentile in 1979 earned a bit less than the median man. In 2007, however, a woman at the 90^th percentile earned 66% more than the median man.

The following chart illustrates how earnings have evolved among the top 10% of men and the top 10% of women from 1989 to 2007:

For both men and women, earnings at the 95^th percentile have increased more than those at the 90^th, and earnings at the 99^th percentile have increased more than those at the 95^th. In short, the biggest earnings gains have occurred at the very top.

To calculate these types of figures, analysts have to make numerous decisions about what exactly they want to measure (or, given data limitations, what they can measure). CBO does a nice job of spelling those decisions out, noting, for example, the following limitations:

First, this paper examines annual earnings, not hourly wages. People’s annual earnings are determined both by their earnings per hour and by how many hours they work each year. Second, people’s annual earnings are not necessarily the entirety of their compensation. In focusing on earnings for its analysis, CBO did not take into account other forms of compensation, such as defined-benefit retirement plans, employers’ contributions to 401(k) plans, or employment-based health insurance. (Comprehensive data on compensation for individuals are not generally available over long periods.) Third, annual earnings do not necessarily represent all of the resources available to workers. Many people receive income from unemployment compensation, child support, or other sources over the course of a year. Also, many people share resources with family members, and some people have savings on which they can rely. Fourth, CBO excluded people who had no earnings over the course of a year and those who had earnings only from self-employment. People who had both earnings and income from self-employment were included in the analysis, but their income from self-employment was not considered.