Expectations of lifetime earnings are key factors in many individuals’ decisions, from weighing education options to deciding on career paths. However, lifetime earnings differ widely across individuals, and uncovering the factors that explain these differences can be challenging. Senior Economist José Mustre-del-Río and Assistant Economist Emily Pollard used a unique data set combining administrative and survey data to assess variation in lifetime earnings. Their findings were published in the Bank’s Economic Review in March 2019.

What makes this study unique?

Measuring lifetime earnings—as well as identifying what explains differences in lifetime earnings—can be challenging. First, measuring lifetime earnings requires data on entire lifetimes and cannot be proxied by earnings at a point in time. For example, medical doctors may temporarily have low earnings while in residency but will likely see their earnings rise thereafter. Similarly, individuals raising young children may temporarily work fewer hours but may work more as their children age. Second, examining which individual-level factors help explain earnings differences requires detailed demographic data.

Our study dealt with these challenges by using data from the U.S. Census Bureau’s Survey of Income and Program Participation Synthetic Beta (SSB). These data take respondents from the bureau’s Survey of Income and Program Participation (SIPP) and match them with Social Security Administration (SSA) and Internal Revenue Service (IRS) Form W-2 earnings records. These records allowed us to construct entire earnings histories for a large sample of individuals. Additionally, because the data are based on a sample of individuals surveyed in the SIPP, they include a host of demographic characteristics (such as race, education, and marital and parental status) that typically are not available in administrative data.

Which kinds of workers were included in this assessment?

To ensure that we had a good picture of earnings over a person’s entire career, we included only individuals who were age 18-25 in 1978, the start of our sample. We then followed these individuals through 2011—the last year for which data were available—when they were age 51-58. Additionally, we restricted our sample to people with a high degree of labor market attachment. In particular, we included only people with at least 17 years of positive earnings—meaning that they worked for pay during at least half of the 34-year sample.

Which factors help explain lifetime earnings?

Our results suggest that observable characteristics such as gender, race, age, education and labor market experience explain a little more than half of lifetime earnings differences. However, among these characteristics, labor market experience—the fact that some individuals systematically work more years than others—accounts for roughly 40 percent of the difference in earnings. In contrast, standard demographic characteristics such as gender, race or education alone explain no more than 15 percent of differences in lifetime earnings. Thus, cumulative labor market experience appears to be crucial in explaining lifetime earnings differences across individuals.