Wage Inequality and the Rise in Labor Force Exit: The Case of US Prime-Age Men

Change in their relative earnings accounted for 44% of the growth in labor force exits among non-college men between 1980-2019, suggesting a decline in social status is a likely factor driving the decline in prime-age labor force participation. “This paper investigates whether prime-age non-college men are more inclined to leave the labor force when their expected earnings fall relative to the earnings of other workers in their labor market. The empirical model takes into account that a job not only provides economic security but also affirms a worker’s social status, which is tied to their position relative to their age range peers. According to [a regression analysis] estimate, a 10% growth in expected earnings has an associated 0.12 percentage point decrease in the exit rate. Contrarily, a 10% growth in reference earnings has an associated 0.13 percentage point increase in the exit rate, fully discounting the earnings effect. These coefficients offer suggestive evidence that non-college men’s labor market exit behavior is tied to the relative values of their earnings. Over the course of the study period, non-college men’s relative earnings declined 30% on average. Based on the estimates, this decline in relative earnings had an associated 49 percentage point increase in the exit rate, accounting for 44% of the total growth in the exit rate among non-college men over this period. In contrast, changes in real earnings alone account for only 18% of the total growth in exit rate“

Their bottom line

“…Over the course of the study period, non-college men’s relative earnings declined 0.30 log point (see Figure 2), which was the combined result of a 0.17- log-point decline in their real median earnings and a 0.23-log-point growth in the reference earnings. A back-of-the-envelope calculation based on the OLS estimates suggests that this decline is associated with a 0.49 percentage point increase in the exit rate, accounting for 44 percent of the net growth in non-college men’s labor force exit rate during this period. In contrast, changes in real occupation earnings alone account for only 18 percent of the total increase in the labor exit rate for non-college men, leaving a larger unexplained growth trend….”

The core of the paper, “…this paper investigates whether prime-age noncollege men are more inclined to leave the labor force when their expected earnings fall relative to the earnings of other workers in their labor market. The empirical model takes into account that a job not only provides economic security but also affirms a worker’s social status, which is tied to their position relative to their age range peers. The model uses a difference-in-differences framework and for identification relies on the variation in non-college men’s relative earnings across states and occupations over time. The sample consists of state-occupation-level panel data on labor force exit rates, occupation earnings, and job loss risk matched with information on the state-level earnings distribution and a set of state socioeconomic controls over the period 1980 through 2019….”

The Evidence

“…The results from the ordinary least squares (OLS) regression analysis show that labor force exit rates decline with a worker group’s expected earnings but increase with their reference earnings, defined as the average earnings in a state across all prime-age workers. According to the estimate, a 10 percent growth in expected earnings has an associated 0.12 percentage point decrease in the exit rate. Contrarily, a 10 percent growth in reference earnings has an associated 0.13 percentage point increase in the exit rate, fully discounting the earnings effect. These coefficients offer suggestive evidence that non-college men’s labor market exit behavior is tied to the relative values of their earnings. Over the course of the study period, non-college men’s relative earnings declined 30 percent on average. Based on the estimates, this decline in relative earnings had an associated 49 percentage point increase in the exit rate, accounting for 44 percent of the total growth in the exit rate among non-college men over this period. In contrast, changes in real earnings alone account for only 18 percent of the total growth in exit rate….”

“…The relationship between relative earnings and labor force exit rates varies by worker demographics. Across races and ethnicities, a correlation between labor force exit rate and relative earnings is identified among non-Hispanic White men but absent among non-Hispanic Black and Hispanic men. Across age groups, the magnitude of the correlation between exit rate and relative earnings declines with age, with the strongest correlation observed among men aged 25 to 34. The correlation between relative earnings and exit rate also weakens when the sample is stratified by a worker’s marital status. These findings suggest that a decline in social status among White men and marriage market sorting are potential channels through which relative earnings affect men’s labor force exit decisions. A supplementary analysis further reveals that the labor force exit rate of non-college men increases with both the level and the skewness of the reference group’s earnings distribution. Conditional on the state median earnings not increasing or decreasing, non-college men are more likely to leave the labor force when the top earners in a state make disproportionately more than the other workers, showing further evidence that earnings inequity may undermine non-college men’s work incentives…”

The Decline in Less Educated Men’s Labor Market Status

“…From 1980 to 2019, the US prime-age non-college men’s labor market return steadily declined relative to that of college-educated workers. During this period, noncollege men’s median weekly earnings fell 17 percent, while college-educated men experienced a 20 percent earnings gain, increasing the weekly earnings gap nearly 150 percent from $280 to $687. In addition, non-college men’s earnings premium over women’s quickly dissipated as gender equity improved. Over this period, the median earnings of women grew 32 percent, significantly outpacing the growth rate of men’s earnings. Most of the growth was from college-educated women, whose earnings increased a stellar 39 percent….Figure 1 plots how relative earnings have changed over time for US men and women by college education status….”

“…Compared with their 1980–1982 level, non-college men’s relative earnings have followed a consistent downward trend over the last 40 years, despite periodic real earnings gains. Specifically, from 1980 to 2019, the value of non-college prime-age men’s median earnings fell from 104 percent to only 69 percent of the average earnings across their prime-age peers. In contrast, relative earnings for college-educated men, college-educated women, and non-college women grew initially, from 1980 to 1993, and then saw a modest downward trend reflecting the accelerated earnings growth at the upper tail of the earnings distribution and the overall stagnation of earnings at the median during the later years of the sample period. Nevertheless, the decline rate was substantially smaller compared with the rate at which relative earnings for non-college men fell. Together, these patterns highlight the earnings dispersion across education levels as well as the increased skewness in the earnings distribution over the last four decades….”

Relative Earnings and Labor Force Exit Rate: An Inverse Relationship

“…Figure 2 plots the changes in college-educated and non-college men’s relative earnings (panel A) and labor force exit rate (panel B) over time from their 1980–1982 values. Compared with college-educated men, non-college men have experienced a much steeper fall in relative earnings as well as a more sizable rise in their labor force exit rate, showing the negative correlation between the two. This inverse relationship also appears across subgroups of non-college men….” image001.png “…Panel A of Figure 3 plots the changes in relative earnings for non-college men by their occupations’ skill-level requirements. Following the 2010 Standard Occupational Classification System (SOC) and the convention in this literature, high-skill occupations here include management, professional, and public safety occupations (SOC: 11–31, 331–333). Middle-skill occupations include sales and administrative support occupations (SOC: 41, 43); production, installation, and repair occupations (SOC: 49, 51); and transportation and construction occupations (SOC: 45, 47, 53). Low-skill occupations comprise various service occupations (SOC: 33–39).5 While non-college men in all three occupation groups experienced sizable declines in relative earnings, the decline was steeper for workers in the middle-skill occupations (–33 percent) than for those in the low-skill occupations (–26 percent) or the highskill occupations (–20 percent). Panel B of Figure 3 shows that non-college men in the middle-skill occupations also experienced the most rapid increase in labor force exit rate, followed by workers in the low-skill occupations and in the high-skill occupations. The decline in non-college men’s relative earnings during the 1980–2019 period also varied spatially. Because of state-level differences in industrial composition, the college wage premium rose at different speeds across states, and the extent of the decline in non-college men’s relative earnings ranged from 10 percent to 64 percent depending on where a worker lived….Together, these descriptive statistics offer suggestive evidence that the labor force exit rate increases when relative earnings fall. These patterns, however, are subject to the influence of other confounding factors and may be only coincidental….” image003.png

Pinghui Wu, “Wage Inequality and the Rise in Labor Force Exit: The Case of US Prime-Age Men,” Federal Reserve Bank Of Boston, December 2022, https://www.bostonfed.org/publications/research-department-working-paper/2022/wage-inequality-and-the-rise-in-labor-force-exit-the-case-of-us-prime-age-men.aspx