Long Work Hours, Part-Time Work, and Trends in the Gender Gap in Pay, the Motherhood Wage Penalty, and the Fatherhood Wage Premium
==================================================================================================================================

* Kim A. Weeden
* Youngjoo Cha
* Mauricio Bucca

## Abstract

We assess how changes in the social organization and compensation of work hours over the last three decades are associated with changes in wage differentials among mothers, fathers, childless women, and childless men. We find that large differences between gender and parental status groups in long work hours (fifty or more per week), coupled with sharply rising hourly wages for long work hours, contributed to rising gender gaps in wages (especially among parents), motherhood wage penalties, and fatherhood wage premiums. Changes in the representation of these groups in part-time work, by contrast, is associated with a decline in the gender gap in wages among parents and in the motherhood wage penalty, but an increase in the fatherhood wage premium. These findings offer important clues into why gender and family wage differentials still persist.

*   gender inequality
*   family wage gap
*   gender wage gap
*   motherhood wage penalty
*   fatherhood wage premium
*   work hours
*   long work hours
*   overwork

After converging relatively rapidly in the 1970s and 1980s, the gender gap in hourly wages shrank only modestly over the next thirty-five years. Today, median weekly wages of full-time women are 83 percent of the median weekly wages of men, an increase of just three percentage points since 2004 (BLS 2015). This stalled convergence in the gender gap in wages has led to a large and vibrant research literature that seeks to understand why change has been so slow (England 2010).

One of the key empirical insights of this literature is that the gender gap in wages at the aggregate level is perpetuated by persistent gender differences in individual labor market behaviors: whether men and women work for pay, the occupations and industries in which they work, and the number of hours per week they work. These gender differences emerge in the context of structural changes in the distribution of jobs with particular attributes (such as expected work hours) and in the wages associated with these attributes, resulting in complex and offsetting effects on the gender gap in wages.1 For example, Youngjoo Cha and Kim Weeden (2014) show that the diffusion of long work hours in the United States in the 1990s and 2000s, coupled with the persistent gender gap in long work hours and rising hourly compensation for long work hours, was associated with an increase in the gender gap in wages after adjusting for other wage-relevant attributes. These trends largely offset wage-equalizing shifts in women's educational attainment.

A second empirical insight is that much of what appears to be a gender wage gap is better understood as a gender-specific family gap in pay or, as they are known in the economic and sociological literatures, the *motherhood wage penalty* and *fatherhood wage premium*: mothers earn less than observationally similar childless women (Pal and Waldfogel, this volume; Waldfogel 1998; Budig and England 2001; Avellar and Smock 2003; Gangl and Ziefle 2009; Staff and Mortimer 2012; Cooke 2014; Kahn, Garcia-Manglano, and Bianchi 2014), and fathers earn more than observationally similar childless men (Pal and Waldfogel, this volume; Waldfogel 1998; Lundberg and Rose 2000; Glauber 2008; Killewald 2012). As in the broader gender literature, the research on family wage differentials emphasizes wage-relevant labor market behaviors, including work hours, of parents compared with childless adults. Curiously, however, explicit efforts to tie trends in the distribution of work hours between mothers and childless women, or between fathers and childless men, to trends in family wage gaps are relatively few and far between. Those that do typically emphasize shifts in the distribution of parents and childless adults across part-time and full-time work but ignore long work hours (see, for example, Waldfogel 1998; Pal and Waldfogel, this volume; Buchmann and McDaniel, this volume).

We bring together these two streams of research by describing the empirical relationship between trends in work hours in the United States, the gender wage gap among parents and among childless adults, and the family wage gap among women and among men. The first set of comparisons is motivated by our expectation that changes in work hours and in the hourly pay of different work hours had a stronger association with the gender gap in wages among parents than among childless adults. The second set is motivated by our expectation that growth or decline in family wage gaps result from the interplay of changes in the distribution of work hours among parents and childless adults and structural changes in the hourly pay associated with part-time work, full-time work, and long work hours.

We assess these expectations using nationally representative labor force data from the Current Population Survey (CPS). First, we briefly situate our analyses in the broader literatures on the sources of the stagnation in the gender and family gap in wages.

## WORK HOURS AND THE GENDER WAGE GAP

In accounting for the slow convergence in the gender gap in wages over the last two decades, gender scholars point to widespread cultural beliefs about the existence of deeply rooted and often biologically based differences in men's and women's traits and skills (Charles and Grusky 2004; Ridgeway 2011; Cotter, Hermsen, and Vanneman 2011). These cultural beliefs affect men's and women's labor market decisions, employers’ hiring decisions, and institutional configurations such as the availability of policies that would support workers who wish to combine paid and unpaid labor. Because the underlying cultural beliefs are slow to change, so the argument goes, gender differences in the labor market behaviors that generate a gender gap in wages are also slow to change.

One implication of this argument is that the proximate sources of the gender wage gap, and of stagnation in the gender wage gap, are likely to be especially pronounced among parents. After all, a core feature of gender essentialism is the belief that women naturally excel at nurturance, personal service, and childrearing (Charles and Grusky 2004, 19; Ridgeway and Correll 2004). This stereotype makes it likely that more mothers than fathers or childless adults will devote a greater share of their time to childrearing, more difficult for mothers to avoid social sanctions if they do not curtail their paid work hours during their childrearing years, and less likely that employers will hire mothers for jobs that require stereotypically male-typed skills or work hours (Ridgeway and Correll 2004).

This prediction is supported, albeit indirectly, by the empirical literature on the proximate sources of the gender gap in wages. Take, for example, housework, which affects the time and energy that is available for paid labor. The housework time gaps between mothers and fathers are larger than the gaps among all adults (Bianchi et al. 2012), and though a growing percentage of Americans state a preference for egalitarian divisions of household labor (Gerson 2009; Pedulla and Thébaud 2015), parental gaps in housework have stalled. Or, consider discrimination, which affects wage gaps directly but also creates unequal opportunities for workers to enter jobs with different levels of pay and work hours. Although several audit or other quasi-experimental studies show evidence of gender discrimination in high status occupations, others show that such discrimination is limited to mothers (Correll, Benard, and Paik 2007).

The conflict between work and nonwork roles may also be greater for mothers than it is for fathers or childless adults. This is especially true in workplaces and occupations where employers expect workers to put in long hours, and where workers derive social status from working long work hours (Jacobs and Gerson 2004; Sharone 2004; Reid 2015). In these settings, the ideal worker is someone who is available to clients and supervisors at all hours of the day or night, is able to travel or relocate for work, and prioritizes career success over family or leisure (Williams 2000). This image of an ideal worker is hard to reconcile with the stereotype of the ideal mother, a mother who is available to her family at all hours of the day or night, is able to travel or relocate to support her children's enrichment activities, and prioritizes family over career success. To be sure, fathers, childless men, and childless women in “greedy organizations” (Coser 1967) or occupations may also experience work-life conflicts, but they are often more able to “pass” as ideal workers, in part because employers interpret their behavior more favorably than identical behavior by mothers (Reid 2015; Correll, Benard, and Paik 2007; Williams 2003). As a result, the gender gap in long work hours among parents is likely to be greater among parents than childless adults and, given its roots in gender essentialist beliefs, also more resistant to change.

Organizational scholars have also noted that a growing share of jobs are organized as part-time work, contingent work, temporary work, contract work, and other “nonstandard employment relations” that weaken any expectation, tacit or otherwise, that employees will work a regular, forty-hour work week (Kalleberg 2011). The growth of these nonstandard employment relations, like the emergent culture of overwork, has different implications for gender gaps in work hours among parents than among childless adults. Historically, the gender gap in part-time work has been much greater for parents than for childless adults: mothers have long been more likely to work part-time than childless women, presumably because part-time work is easier to combine with the time demands of childrearing; fathers, by contrast, have been much less likely to work part time than childless men, perhaps because of cultural expectations surrounding male breadwinning (Townsend 2002). However, the growth in part-time work has largely been in involuntary part-time work (Kalleberg 2000), and it has been accompanied by a decline in the gender gap in part-time work, especially among parents.

Gender gaps in work hours, and changes in them, are only relevant to the gender gap in hourly pay if different work hours are associated with different levels of hourly pay. In this regard, part-time work often pays lower hourly wages than full-time work that is comparable in terms of tasks and skill requirements (Kalleberg 2000, 2011). Claudia Goldin (2014) argues that pay is nonlinear with respect to hours in many occupations, because hours are worth more if they are worked continuously in long blocks and if they are timed to overlap with the hours of colleagues. Moreover, the relative pay of different work hours is changing: for example, Cha and Weeden find that the hourly pay of workers who put in fifty or more hours per week has increased dramatically relative to full-time workers with similar observed attributes (2014; see also Kuhn and Lozano 2008).

These wage disparities for different work hours, when coupled with gender gaps in work hours and trends in those hours that vary by parental status, have potentially important implications for gender gaps in wages. First, in a world in which mothers are more likely to work part time and less likely to work long hours than other groups, and in which fathers are more likely to work long hours and less likely to work part time, gender gaps in wages will be greater among parents than among childless adults merely because of differences in the relative pay of different work hours. Second, and related, changes in the relative pay for part time and long work hours will have stronger associations with trends in the gender gap in wages among parents than among childless adults.

## WORK HOURS AND THE FAMILY WAGE GAP

As we noted in our introductory comments, we also wish to flip our comparisons around to assess how trends in within-gender family wage differentials—the motherhood wage penalty and the fatherhood wage premium—are associated with trends in work hours and in the relative pay of different work hours. Research estimates that the motherhood wage penalty is between 6 percent and 15 percent per child, the higher estimates coming from models that correct for differential selection of mothers and childless women into paid labor (Gangl and Ziefle 2009). Approximately one-third of the education-adjusted motherhood wage penalty disappears in models that also adjust for work experience; much of the rest is associated with the employment situation of mothers after childbirth, especially their greater likelihood of entering part-time work. Similarly, much of the fatherhood wage premium is associated with work hour behaviors: when men become fathers, they increase their paid work hours by an estimated forty hours per child per year, and their hourly pay increases by about 4 percent per child (Lundberg and Rose 2000; Glauber 2008).2

Unlike the literature on the gender gap in wages, the literature on family wage gaps has been less concerned with trends. Indeed, much of it relies on longitudinal data from a single cohort, using within-person variation in parental status to estimate the “causal” effect of having a child on wages. There are important exceptions. For example, Markus Gangl and Andrea Zeifle show that the motherhood wage penalty was smaller for a cohort of women born in the late 1950s than for a cohort born in the early 1960s (2009). Sarah Avellar and Pamela Smock find no change in the motherhood wage penalty between a cohort of women born between 1944 and 1954 and another born between 1980 and 1984 (2003). However, Pal and Waldfogel show a decline in the motherhood wage penalty, from 6 to 1 percent, between 1993 and 2013 (this volume). Similarly, Shelly Lundberg and Elina Rose find a decline in the fatherhood wage premium between men born between 1943 and 1950 and those born between 1951 and 1974 (2000).

Our goal is not to replicate these studies’ focus on identifying, to the extent possible with observational data, the causal effect of parental status on wages. Rather, it is to describe the empirical relationship between trends in family wage gaps, changes in the distribution of work hours between mothers and childless women and between fathers and childless men, and changes in the wage returns to different work hours. We also take a more encompassing view of work hours than much of the family wage gap literature, focusing on long work hours as well as part-time hours.

## DATA, VARIABLES, AND METHODS

We first describe, for each of our four gender and parental status groups, trends in: wage and salary employment; work hours, conditional on employment; and the hourly wages associated with different levels of work hours, providing both unadjusted mean wages and mean wages after adjusting for education, experience, and other standard wage predictors. We then use a technique developed by Chinhui Juhn, Kevin Murphy, and Brooks Pierce (1991) to decompose trends in the gender gap in wages within parental groups, and trends in the family gap in wages within genders, into changes that are generated by shifts in the share of the relevant groups who work part time, full time, or long work hours (the *quantity effect*), and changes that are generated by shifts in the relative wages of different work hours (the *price effect*).

### Data

The data for these analyses come from the May (1969–1984) and Merged Outgoing Rotation Groups, or MORG (1979–2014) of the Current Population Survey files compiled and distributed by the National Bureau of Economic Research.3 Our base analytic sample is limited to noninstitutionalized civilian workers ages eighteen to sixty-four years. Except for our analysis of trends in wage and salary employment as a share of the total population, we restrict the sample to currently employed wage and salary workers with nonmissing values on the parental status questions (see following section). The multivariate analyses of wages further restrict the sample to workers with nonmissing and nonzero hours with valid wage information. Finally, the Juhn, Murphy, and Pierce (JMP) decomposition analyses are further restricted to CPS years 1984 and 2014 or, for the period-specific decompositions, 1984, 1993, 2004, and 2014. The sample sizes differ across analyses, and we present them in the notes to the tables and figures. All analyses use the BLS-provided sampling weights.

### Variables

The outcome variable in our analyses is hourly wages. We estimate the hourly wages of workers who are not typically paid by the hour by dividing weekly wages by the number of hours usually worked per week. We exclude workers whose hourly wages fall below $1 per hour or more than $100 per hour in 1979 U.S. dollars, and we multiply wages that are top-coded by the BLS to ensure confidentiality by 1.4 (Card and DiNardo 2002). We adjust nominal wages for inflation using the Bureau of Economic Analysis's Personal Consumption Expenditures Deflator, and express all wages in 2014 dollars. In the multivariate and decomposition analyses, we take the natural log of hourly wages (measured in pennies); in the bivariate analyses, we keep wages in the natural metric to ease interpretation.

We measure parental status with a binary variable that we constructed from a CPS variable that indicates the number of children under the age of eighteen who are related to the household head (up through the 1988 CPS) or in the primary family (the 1989–2014 CPS) and who reside in the sampled household.4 The CPS files are constructed so that all members of the household (May) or primary family (MORG), including children, receive the same value on the parental status variable. We identify parents and childless adults by limiting the May and 1984–1988 MORG sample to heads of households, including heads of single-person households, and their spouses. In the 1989 through 2014 surveys, data on children were collected for primary families rather than households. For these surveys, we limit the sample to adult heads of primary families, their spouses, and respondents who are not in primary families, where the latter includes childless adults who live alone or in nonfamily groups (such as roommates or boarders).

Two other complexities with the parental status measure need to be kept in mind when interpreting the results. First, the variables necessary to identify parents are missing from the 1982 and 1983 surveys (May and MORG), the 1994 through 1998 MORG surveys, and the 1999 MORG surveys collected before November. Second, we cannot identify parents of children who do not reside in the household or who are over the age of eighteen.5 In supplementary analyses, we reestimate our models on a sample of men and women of childrearing years (age eighteen to forty-five). We found similar patterns as in the full sample, but with greater differences between childless adults and parents in the association between long work hours and wage differentials.

Our measure of work hours is based on a CPS variable that asks respondents how many hours they usually work, referring to the main job or, in the May surveys, “this job.” In the MORG series, usual hours are edited and missing cases imputed by the BLS. Beginning with the 2000 survey, the BLS added a category for “hours vary,” which constitutes about 3 percent of wage and salary workers. Rather than exclude these cases, we assume that their hours worked last week are a reasonable proxy for usual work hours. This proxy will overstate usual hours at the main job for workers with more than one job, but understate usual hours for workers who are not working in the reference week because of illness, vacation, holidays, strikes, or temporary layoffs.6 In the May supplements, the usual hours variable is not available until 1973, so we begin our descriptive analysis of trends with this year. If usual hours are missing, we assume hours worked at the main job, a variable available only for the 5 percent of the sample who are dual job holders and only prior to 1981, are a valid proxy. If this variable is also missing, we exclude the case. Of the many specifications we tested, this provides the closest match to distribution of work hours in the MORG data for the years between 1979 and 1984, when both May and MORG files are available. Even so, we recommend caution in comparing across the May and MORG series.

We convert work hours into a set of five dummy variables using standard cut points in the work-family literature and in administrative publications: one to twenty hours, twenty-one to thirty-four hours, thirty-five to forty hours, forty-one to forty-nine hours, and fifty hours or more. We use this five-category measure in our bivariate analyses, but for ease of presentation aggregate to a three-category measure—part time (one to thirty-four hours), full time (thirty-five to forty-nine hours), and long work hours (fifty hours or more)—in our multivariate analyses.7 Sensitivity checks fit to data pooled across parental status show that other cut points generate very similar results (see Cha and Weeden 2014).

Our multivariate wage equations adjust for standard predictors of wages: race (non-Hispanic white, non-Hispanic African American, Hispanic, other race), age and its square, education (less than high school, high school, some college, college graduate, advanced degree), marital status (currently married, unmarried), potential years of work experience (age in years minus schooling in years minus 6) and its square, region, metropolitan residence, and public-sector employment. The JMP decompositions of the family wage gap include an identical set of covariates, but the decompositions of the gender wage gap exclude marital status because the assumption that the association of marital status with wages is equal for men and women is unsustainable. Tables A1 and A2 present descriptive statistics for the covariates in our multivariate analyses by gender and parental status for 1984 and 2014.

Our wage equations do not adjust for occupation, union membership, employer tenure, or actual work experience. Employer tenure and actual work experience are not available in the CPS, and union membership is not consistently available. However, analyses comparing CPS with other data sources show that including union membership, employer tenure, and actual work experience does not appreciably alter the relationship between work hours and wages (Cha and Weeden 2014). Occupation is available in the CPS, but adjusting for it may understate the true associations between work hours and wages, and hence the magnitude of family wage differentials, if one assumes that occupation is partly endogenous to parental status (Gangl and Ziefle 2009). As a sensitivity check, we reestimated our multivariate and decomposition models using data from which we first purged all possible association between wages and occupations by fitting fixed effects of occupations. Where these results differ from our main results, we report them. Finally, we also tested models that exclude potential work experience, on the argument that experience, too, is endogenous to parental status; this specification yielded nearly identical work hour coefficients as those that we present here.

### Methods

Most of our analysis rests on simple bivariate trends or on standard ordinary least squares (OLS) regression, which we assume is familiar to most readers. The JMP decomposition method, however, warrants some explanation. It begins with a standard wage equation fit to data from one demographic group (for example, men or childless adults), and assumes that the observed associations between the covariates and wages for this group also hold for the other demographic group (such as women or parents) in the absence of discrimination.8 Formally,![Formula][1]</img>  

where *y*it is the log of wages for individual *i* in year *t*; x is a vector of independent variables; b is a vector of regression coefficients; σ is the residual standard deviation of the baseline group's wages for year *t*; and θ is a standardized residual with a mean of zero and variance of 1 for each year. We provide these regression coefficients in tables A3 and A4.

The change in the between-group wage gap between two time points, denoted by 0 and 1, can be decomposed into four components:

![Formula][2]</img> ![Formula][3]</img> ![Formula][4]</img> ![Formula][5]</img> 
In these equations, Δ denotes the average male-female (or parent-childless adult) difference in the variable it precedes. The *observed x effect*, equation (2), also known as the quantity change effect, is the portion of the change in the gender (or family) wage gap between the two time points that is associated with changes in the quantity of each of the observed predictors (such as experience or education) in x. The *observed price effect*, equation (3), is the portion of the change in the gender (or family) wage gap that is associated with changes in the net wage returns to each observed predictor. Equations (4) and (5) estimate the contribution of price and quantity changes in unobserved variables on the changes in the gender or family wage gaps and are not central to our discussion.

## RESULTS

### Trends in Wage and Salary Employment

Changes in work hours, conditional on being employed, take place against the backdrop of changes in participation in paid and unpaid labor. In figure 1, we graph trends in the percentage of each of the four parental status and gender groups (relative to the total population) who are employed as wage and salary workers as of the reference week. Figures A1 and A2 provide analogous trends by race.

![Figure 1.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F1.medium.gif)

[Figure 1.](http://www.rsfjournal.org/content/2/4/71/F1)

**Figure 1.** 
Trends in Percentage of Men and Women in Wage and Salary Employment

Trends in wage employment show the now-familiar story of partial convergence across parental status and gender groups, a convergence driven both by the declining share of fathers and childless men who are employed for pay and by the rising share of mothers and childless women who are employed for pay. The decline in men's wage employment is evident throughout the forty-five years in the CPS: childless men, for example, decreased their wage employment from 76 percent in 1969 to 66 percent in 2014. The growth of women's wage employment was concentrated in the 1970s and 1980s, when it increased from 36 to 59 percent (mothers) and from 50 to 64 percent (childless women). For both groups of women, the share in wage employment declined slightly between 1999 and 2014 (see also Byker, this volume).

### Trends in Work Hours

Figures 2 through 5 present, for each gender and parental status group, the percentage of wage earners in each work hour category: at least fifty hours per week (the top and darkest shaded area); the two full-time categories (the next two areas from the top, corresponding to forty-one to forty-nine hours and thirty-five to forty hours, respectively); and the two part-time categories (the two lightest shaded areas, corresponding to twenty-one to thirty-four hours and, at the very bottom, one to twenty hours). Figures A3 through A6 provide analogous trends by race.

![Figure 2.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F2.medium.gif)

[Figure 2.](http://www.rsfjournal.org/content/2/4/71/F2)

**Figure 2.** 
Percentage of Workers in Work Hour Categories, Childless Men

The gender gap in long work hours between childless men and childless women (figures 2 and 3, respectively), and between fathers and mothers (figures 4 and 5), is both substantial and persistent. In the early 1970s, when the CPS began collecting information on usual work hours, 15 percent of childless men and 4 percent of childless women worked fifty or more hours per week. By 1999, the peak of long work hours in the United States, 21 percent of childless men and 10 percent of childless women worked them. The share of these groups declined slightly through the 2000s, dropped precipitously in the aftermath of the Great Recession, and began to rise again during the economic recovery. As of 2014, 18 percent of childless men and 9 percent of childless women worked at least fifty hours per week, still below peak levels but up 1 to 2 percentage points since the recession. Figures 2 and 3 show that childless men and childless women's share of long work hours moved in tandem, meaning that the gender gap in long work hours among childless adults remains essentially unchanged.

![Figure 3.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F3.medium.gif)

[Figure 3.](http://www.rsfjournal.org/content/2/4/71/F3)

**Figure 3.** 
Percentage of Workers in Work Hour Categories, Childless Women

![Figure 4.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F4.medium.gif)

[Figure 4.](http://www.rsfjournal.org/content/2/4/71/F4)

**Figure 4.** 
Percentage of Workers in Work Hour Categories, Fathers

![Figure 5.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F5.medium.gif)

[Figure 5.](http://www.rsfjournal.org/content/2/4/71/F5)

**Figure 5.** 
Percentage of Workers in Work Hour Categories, Mothers

Among parents, the gender gap in long work hours was just as persistent (compare figures 4 and 5). In 1973, 20 percent of fathers worked at least fifty hours per week, versus only 3 percent of mothers. These percentages rose to a late-1990s peak of 24 percent of fathers and 6 percent of mothers. The recent recession had less of an impact on long work hours among parents than it did among childless adults: the percentage of fathers declined by only 2 percentage points between 2007 and 2009, and the percentage of mothers by less than 1 point. By 2014, the percentage of mothers rebounded to its prerecession peak of 6 percent, and the percentage of fathers rose to 20 percent. Despite these fluctuations, the gap between fathers and mothers was nearly the same in 2014 as it was in 1984, though it expanded slightly in the middle years.

We also note that the gender gap in long work hours among parents greatly exceeds the gap among childless adults. In 2014, for example, the ratio of fathers to mothers who work long hours was more than three to one, compared to a two to one ratio among childless adults. This comparatively large gender gap in long work hours among parents is driven both by the smaller percentage of mothers than childless women who work long hours, and by the larger percentage of fathers than childless men who work long hours. This implies, of course, that the within-gender gaps in long work hours between parents and childless adults have a different “sign” for men and women. Moreover, within-gender gaps in long work hours widened between the beginning of the 1980s and the late 1990s, and although they narrowed in the subsequent years, they did not return to earlier levels. As we will discuss below, this opens the door for a quantity change effect of long work hours on the family wage gaps.

Trends in part-time work are indicated by changes in the bottom area (one to twenty hours per week) and the second area from the bottom (twenty-one to thirty-four hours per week) in figures 2 though 5. With the exception of the sharp decline in very low work hours among mothers in the late 1970s, the trends in the two categories of part-time work are consistent; for the sake of brevity, we discuss them as one category.

Like the gender gap in long work hours, the gender gap in part-time work is greater among parents than among childless adults. As shown in figures 2 through 5, the percentage of part-time workers is the highest among mothers, who are followed in turn by childless women, childless men, and fathers. This pattern is evident in all years, but strongest in 1970s and 1980s, in large part because the percentages of men and women who work part time converged in more recent years. This recent convergence was driven by the well-known decline in the percentage of mothers who work part time (from 30 percent in 1973 to 25 percent in 2014), as well as by an increase in the percentage of part-time fathers (from 3 percent in 1973 to 5 percent in 2014) and especially of part-time childless men (from 5 percent in the early 1970s to 10 percent in 2014). All groups show an uptick in part-time work during the Great Recession, but their postrecession experiences differed: childless women quickly returned to prerecession levels and held steady, mothers returned to prerecession levels and continued to decline, but neither childless men nor fathers had returned to prerecession levels of part-time work by 2014.

### Trend in Mean Wages by Work Hours, Parental Status, and Gender

The association between trends in work hours and trends in gender and family wage gaps depends, of course, on the wages associated with different work hours. Figures 6 through 8 present the mean unadjusted wages associated with long work hours (figure 6), full-time work thirty-five to forty-nine hours per week, figure 7), and part-time work (one to thirty-four hours per week, figure 8). Figures 6 through 8, as well as subsequent figures that include wage information, begin with 1984, the first year of the MORG series in which the measures of parental status is available.

![Figure 6.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F6.medium.gif)

[Figure 6.](http://www.rsfjournal.org/content/2/4/71/F6)

**Figure 6.** 
Unadjusted Mean Hourly Wages, Long Hours

![Figure 7.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F7.medium.gif)

[Figure 7.](http://www.rsfjournal.org/content/2/4/71/F7)

**Figure 7.** 
Unadjusted Mean Hourly Wages, Full-Time

![Figure 8.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F8.medium.gif)

[Figure 8.](http://www.rsfjournal.org/content/2/4/71/F8)

**Figure 8.** 
Unadjusted Mean Hourly Wages, Part-Time

The first noteworthy result of figures 6, 7, and 8 is that the unadjusted hourly wages for men and women who work long hours rose faster than the hourly wages of full-time workers of the same gender and parental status. For example, fathers who worked long hours earned, on average, $22 per hour (in 2014 dollars) in 1984 and $33 per hour in 2014, a 50 percent increase (figure 6). By comparison, full-time fathers also earned $22 per hour in 1984, but their mean hourly wages had increased to only $26 per hour by 2014 (figure 7). The mean wages of mothers and childless women who worked long hours also increase more rapidly than those of full-time women: mothers who worked long hours saw their mean hourly wages nearly double (from $16 in 1984 to $30 in 2014), compared with a 50 percent increase (from $14 to $21) for full-time mothers (compare figures 6 and 7). By the early 2010s, the hourly wage gaps between mothers and childless women who worked long hours, on one hand, and childless men, on the other, had largely disappeared.

The hourly wage growth for part-time work was much less substantial, and though gender and parental status groups varied somewhat, for no group did growth match that of workers who worked full time or long hours (figure 8). A trend in the wages of part-time men is hard to discern, in part because so few fathers work part time. Part-time mothers and childless women saw their mean wages rise, but only modestly in absolute and percentage terms. Specifically, part-time mothers’ mean wages grew by $5 per hour, from $12 to $17, between 1984 and 2014, an increase of 42 percent. Part-time, childless women's mean wages grew by only $3 per hour, from $12 to $15, or 25 percent.

The overall story of unadjusted wage growth, then, is some convergence across gender and parental status groups, driven by the comparatively weaker wage growth for childless men than for mothers or childless women. Even so, the trend lines across gender and parental status groups are by and large parallel, and gender and family gaps in mean unadjusted wages remain substantial in 2014.

These cross-group differences in unadjusted wages graphed in figures 6 through 8 could, of course, merely reflect compositional changes if, for example, the proportion of workers with a college degree grew faster among workers who put in long hours than among those who work full time. We can gain some leverage on this by regressing (logged) hourly wages on work hours and predictors of wages (see also tables A1 and A2). The coefficients pertaining to long work hours are presented in figure 9, and those pertaining to part-time work hours in figure 10; all work hour coefficients are large enough multiples of their standard errors to be significant at conventional levels (α = 0.05). These coefficients are interpreted as the proportional wage increase (values greater than unity) or decrease (values less than unity) associated with long work hours relative to full-time workers, after adjusting for other predictors.

![Figure 9.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F9.medium.gif)

[Figure 9.](http://www.rsfjournal.org/content/2/4/71/F9)

**Figure 9.** 
Adjusted Mean Hourly Wages, Long Hours

![Figure 10.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F10.medium.gif)

[Figure 10.](http://www.rsfjournal.org/content/2/4/71/F10)

**Figure 10.** 
Adjusted Mean Hourly Wages, Part-Time

For the most part, work hour differences in trends in adjusted wages show a similar pattern as trends in unadjusted wages (compare figures 9 and 10 with figures 6 through 8). The adjusted hourly wages of workers who work fifty or more hours per week rose more dramatically than those of full-time workers of the same gender and parental status. Notably, however, figure 9 shows that the mean adjusted wages of workers who put in long hours fell well short of those of full-time workers in the first half of our data series, implying a “wage penalty” for long work hours. The magnitude of this wage penalty declined steadily between 1984 and the early 1990s, and by 2000 it had become a wage premium (coefficients exceeded unity). The wage premium grew until the early 2010s, after which it leveled off somewhat (see the trend line for fathers). Even so, by 2014, workers who worked fifty or more hours per week earned between 4 and 7 percent more than their full-time counterparts, even adjusting for education and other wage-relevant characteristics.

The second key result in figure 9 is that the trend in the group-specific wage premium for long work hours did not differ substantially by gender or parental status: among mothers, for example, it was comparable to the long-hour wage premium among fathers. Note that this does not imply that mothers and fathers receive identical wages for long work hours (see figures 6, 7, and 8), because the relevant comparison is within groups.

Figure 10, the analogous graph for part-time wages, shows that the adjusted mean wages of part-time workers fell short of those of full-time workers in all years and for all gender and parental status groups. The size of this part-time wage penalty varies by gender and parental status. It is largest among childless men, such that part-time childless men earn approximately 67 percent of the wages of full-time childless men adjusting for other attributes, and smallest among mothers, such that part-time mothers earn about 80 percent of the wages of full-time mothers.

Between 1984 and the mid-2000s, the negative wage differential between part-time work and full-time work did not change appreciably for any parental status or gender group, but afterward it declined sharply. The wage gap between part-time and full-time mothers, for example, grew by about 4 percentage points between 2007 (0.84) and 2014 (0.80). Taken together, figures 9 and 10 imply that up until the late 2000s, growth in the relative wages associated with long work hours was the dominant trend. In the last decade, the decline in the relative wages associated with part-time work was also pronounced. The JMP decompositions, which we turn to next, tease out these relationships between changes in the relative wages of work hours and changes in gender and parental status wage gaps.

### Decomposition of Trends in Gender Wage Gaps

The JMP decompositions relevant to trends in the gender wage gaps within parental status groups are presented in table 1. Between 1984 and 2014, the gender gap in wages among parents decreased by 0.19 log points, and the gender gap in wages among childless adults decreased by 0.20 log points (see table 1 entries for total change in gender wage gap). Changes in observed factors account for about 25 percent of the change in the gender wage gap for childless adults (that is, 0.049/0.193 = 0.25) and 17 percent of the change for parents (0.033/0.197 = 0.17).

View this table:
[Table 1.](http://www.rsfjournal.org/content/2/4/71/T1)

**Table 1.** 
Selected Coefficients from JMP Decomposition of Gender Wage Gap

For both groups, changes in long work hours are associated with widening the gender wage gap, adjusting for other observed factors, but this association is stronger among parents than childless adults. Among childless adults, rising wage returns to long work hours are associated with an increase in the gender gap in wages by 0.016 log points, about 9 percent of the size of the total change in the gender wage gap. Among parents, this price effect of long work hours is nearly twice as large: 0.029 log points, comparable to 15 percent of the total change in the gender wage gap.9 Put differently, in a hypothetical world in which the hourly wages associated with long work hours remained at 1984 levels, the gender wage gap among childless adults would be about 9 percent lower than we observe today, and the gender wage gap among parents would be 15 percent lower. To put this in context, the effect of changes in the wage returns to different educational degrees is about 10 percent of the total change for childless adults, and 5 percent of the total change for parents (for all decomposition coefficients, see tables A5 and A6).10

By contrast, the quantity change effect of long work hours is quite small for both parents and nonparents. For childless adults, changes in the gender gap in long work hours are associated with increase in the gender wage differential by a trivial 1 percent of the total change. For parents, the estimated quantity change is essentially null (0.2 percent). This is anticipated by figures 2 through 5, which show no difference between 1984 and 2014 in the gender gap in long work hours.

By taking 1984 and 2014 as the starting and ending points, the preceding decomposition results may gloss over differences in the timing of price and quantity effects of long work hours. Columns 2 through 4 of table 1 present results from three JMP decompositions that use data from 1984, 1993, 2004, and 2014 to estimate wage trends across three time periods. These period-specific results show that the contribution of rising prices for long work hours to the expansion of the gender gap in wages is particularly pronounced in the first two decades of our data. This finding is consistent with figures 9 and 10, which show that the steeper wage growth of long work hours relative to full-time work leveled off around 2010. Very little of the change in the gender gap in wages is associated with changes in the shares of the parents relative to childless adults who work long hours (the quantity change effect) in any of the time periods.

In contrast to long work hours, changes in part-time work were associated with declines in the gender gaps in wages for parents and childless adults, primarily through nontrivial changes in the shares of these groups who work part time (see table 1). Between 1984 and 2014, the reduction of the part-time work hour gap between childless men and childless women was associated with an estimated 0.013 log point reduction in the gender gap in wages, or 7 percent of the total decline. Among parents, the decline in the gender gap in part-time work was associated with a 0.024 log point decline in the gender gap in wages, or 12 percent of the total change.11 The quantity effects of part-time work in the metric of log points are relatively unaffected by eliminating all occupation-wage association from the data. However, in terms of percentage of the total change in the wage gap, purging the occupation-wage association reduces the quantity changes by more than half, to 3 percent for childless adults and 5 percent for parents.

Changes in the adjusted hourly wages of part-time work, by contrast, had very little impact on gender wage gaps for either parents or childless adults across the 1984 to 2014 period. The period from 2004 to 2014 is a possible exception: in these ten years, the declining hourly wages of part-time work relative to full-time work was associated with an expansion of the gender wage gap among parents, and among childless adults, that largely offset the contraction of the wage gap attributable to convergence in the shares of men and women who worked part time. However, the price change effect in this period is greatly attenuated after we purge all occupation-wage associations from the data. The more robust story is that convergence in the shares of part-time workers among fathers and childless men, and mothers and childless women, contributed to convergence in the within-group gender wage gaps.

### Decomposition of Trends in Family Wage Gaps

Our final set of results quantifies the association between changes in work hours and within-gender wage differentials between parents and childless adults. Table 2 provides the relevant estimates for wage differentials among women and men.12 In both cases, the within-gender wage gap is calculated by subtracting parents’ wages from childless adults’ wages: it is positive for women (because childless women earn more, on average, than mothers) and negative for men (because childless men earn less, on average, than fathers). A positive coefficient for women means that the covariate is associated with an increase in the motherhood wage penalty; a positive coefficient for men means that the listed covariate is associated with a decline in the fatherhood wage premium.

View this table:
[Table 2.](http://www.rsfjournal.org/content/2/4/71/T2)

**Table 2.** 
Selected Coefficients from JMP Decomposition of Family Wage Gap

The first rows in each panel of table 2 provide estimates of the total changes in the family wage gaps for women and men, respectively. The wage gap between mothers and childless women declined by 0.041 log points, about 4 percentage points, between 1984 and 2014. By contrast, the wage gap between fathers and childless men increased: in 1984, fathers earned 6 percent higher wages than observationally similar childless men, but by 2014 they earned 11 percent higher wages. The increase in the fatherhood wage premium is 0.037 log points, again about 4 percentage points.

The growth in the hourly pay of those who work long hours was associated with an increase in the family wage gaps for both men and women. Between 1984 and 2014, rising hourly wages for long work hours was associated with an expansion of the motherhood wage penalty by an estimated 0.004 log points, about 9 percent of the total change in motherhood wage penalty. In percentage terms, this price change effect appears to be larger in 1993 and 2004 than in either of the two flanking decades (see columns 2–4, table 2). Rising pay for long work hours also increased the wage gap between fathers and childless men, but by an even greater magnitude: 0.006 log points, or 18 percent of the total change, between 1984 and 2014. Changes in the relative share of mothers and childless women, and of fathers and childless men, who work long hours were not associated with changes in the motherhood wage penalty or fatherhood wage premium, respectively, perhaps because these quantity changes were so modest (see also figures 2 through 5).

Just as in the gender wage decompositions, the association between trends in part-time work and trends in family wage gaps is driven primarily by changing shares of each group in part-time work (quantity change effects), not by changing hourly wages for part-time work (price change effects). The comparatively rapid decline in the share of mothers who work part-time compressed the motherhood wage penalty by 0.012 log points, or 31 percent of the total change between 1984 and 2014. During the same period, the comparatively rapid increase in part-time work among childless men increased the fatherhood wage premium by about 0.004 log points, or 11 percent of the total change (see table 2, column 1, panel B).13 For both genders, price change effects of part-time work were modest in the 1984 to 2014 period. Column 4 in table 2 shows some indication of price change effects in the 2004 to 2014 period, as indicated by the positive price change coefficient for women (0.003 log points) and negative price change coefficient for men (–0.003 log points). The within-gender comparisons thus show that price changes in long work hours were more influential for changes in the fatherhood wage premium than the motherhood wage penalty, whereas quantity changes in part-time work were more influential for changes in the motherhood wage penalty than in the fatherhood wage premium.

## CONCLUSION

One of the puzzles of contemporary gender scholarship is why gender and family wage gaps have been so persistent. In this article, we have focused on one proximate source of between-group wage inequalities, namely work hours. Our analysis is inspired by the argument that cultural beliefs about men and women's “natural” traits lead to persistent gender and parental status gaps in wage-relevant behaviors and outcomes. In this context, it is unsurprising that even though the share of workers in all gender and parental status groups who worked fifty or more hours per week rose in the 1990s through the mid-2000s, the gap in long work hours between “mothers and others” remained the most extreme. Gender and parental status gaps in part-time work hours have been less resistant to change, although a far greater share of mothers still work part time than childless women, fathers, or childless men.

“Sticky” gender and parental status gaps in work hours, coupled with changes in how different levels of work hours are compensated, have a strong association with trends in the gender gap in wages, in the motherhood wage penalty, and in the fatherhood wage premium. We have shown, first, that the rise (and later, partial retreat) in the share of Americans who work long hours had the net effect of increasing the gender wage gap (especially among parents) the motherhood wage penalty, and the fatherhood wage premium. Second, the association between trends in work hours and trends in wage gaps emerges because the mean hourly wages of workers who work long hours grew markedly compared to the wages of other, observationally similar workers who “only” work full time. This wage growth was gender and parental-status neutral in the sense that all long-hour workers saw their relative wages grow. It was not neutral in its consequences for aggregate levels of inequality. Because of the distributions of long work hours, as a group mothers benefited the least, and fathers the most, from the steep increase in the wage premium for working fifty or more hours per week. Our results suggest that the gender gap in human-capital adjusted wages among parents would be 15 percent lower if the hourly wages of long work hours had not increased. Similarly, the gender gap in wages among childless adults would be 8 percent lower, the motherhood wage penalty would be 9 percent lower, and the fatherhood wage premium would be 18 percent lower than what we in fact observe between 1984 and 2014.

We have also shown that changes in the part-time work were associated with declining gender gaps in wages for parents and for childless adults, a decline in the motherhood wage penalty, and an increase in the fatherhood wage premium. However, these associations are driven by changes in composition, not in the relative wages of part-time work: specifically, a growing share of childless men and a declining share of women, particularly mothers, who work part time. These shifts in the composition of part-time work were associated with a decline of 12 percent in the wage gap between mothers and fathers, 6 percent between childless women and men, and 30 percent between mothers and childless women. They were associated with an 11 percent increase in the wage gap between fathers and childless men over the thirty years of our study, albeit off a lower baseline gap than the other comparisons.

Across the entire labor force, changes in the shares of part-time workers from each demographic group and changes in the relative wages of different work hours had offsetting effects on between-group inequalities in human-capital adjusted wages. The convergence in part-time work hours across parental status and gender groups over the thirty years of the CPS data suggest that greater equality in work hours is possible, and with it a further reduction in the gender gap in wages. However, much of this convergence may be at the expense of men and women who would prefer full-time work but cannot find it.

Convergence in long work hours, too, would go far to reduce wage gaps. Such convergence could, logically, occur by reducing the share of fathers and childless men who work long hours or by increasing the share of mothers and childless women who do. Given evidence that many workers put in long work hours less out of preference than out of the fear that they will incur career penalties if they do not (Clarkberg and Moen 2001; see also Reid 2015), and given the association between long work hours and productivity may be more illusory than real, the preferred solution (from the standpoint of maximizing happiness) would seem to be to reduce the career sanctions for workers who avoid long hours. Our results also suggest, however, that gender and family wage gaps are affected by changes in the workplace and in workplace policies that affect disparities in the wages associated with different work hours. To the extent that part-time work is disproportionately minimum wage work, raising the minimum wage is likely to decrease the motherhood wage penalty. At the same time, policy changes that benefit workers who put in long hours, such as raising the salary threshold for overtime pay, may have the unanticipated consequence of exacerbating gender and family wage gaps.

## Acknowledgments

We thank Martha Bailey, Sheldon Danziger, Tom DiPrete, and the anonymous reviewers for their helpful comments on an earlier version of this paper.

## APPENDIX

View this table:
[Table A1.](http://www.rsfjournal.org/content/2/4/71/T3)

**Table A1.** 
Means and Standard Deviations for Variables in Regression and Decomposition Analyses, 1984

View this table:
[Table A2.](http://www.rsfjournal.org/content/2/4/71/T4)

**Table A2.** 
Means and Standard Deviations for Variables in Regression and Decomposition Analyses, 2014

View this table:
[Table A3.](http://www.rsfjournal.org/content/2/4/71/T5)

**Table A3.** 
Regression Coefficients in JMP Decompositions of the Gender Gap in Wages, Childless Adults

View this table:
[Table A4.](http://www.rsfjournal.org/content/2/4/71/T6)

**Table A4.** 
Regression Coefficients in JMP Decompositions of the Gender Gap in Wages, Parents

View this table:
[Table A5.](http://www.rsfjournal.org/content/2/4/71/T7)

**Table A5.** 
Coefficients from JMP Decomposition of Gender Gap in (Logged) Wages, Childless Adults

View this table:
[Table A6.](http://www.rsfjournal.org/content/2/4/71/T8)

**Table A6.** 
Coefficients from JMP Decomposition of Gender Gap in (Logged) Wages, Parents

View this table:
[Table A7.](http://www.rsfjournal.org/content/2/4/71/T9)

**Table A7.** 
Coefficients from JMP Decomposition of Family Gap in (Logged) Wages, Women

View this table:
[Table A8.](http://www.rsfjournal.org/content/2/4/71/T10)

**Table A8.** 
Coefficients from JMP Decomposition of Family Gap in (Logged) Wages, Men

![Figure A1.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F11.medium.gif)

[Figure A1.](http://www.rsfjournal.org/content/2/4/71/F11)

**Figure A1.** 
Participation in Wage and Salary Employment, Men

![Figure A2.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F12.medium.gif)

[Figure A2.](http://www.rsfjournal.org/content/2/4/71/F12)

**Figure A2.** 
Participation in Wage and Salary Employment, Women

![Figure A3.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F13.medium.gif)

[Figure A3.](http://www.rsfjournal.org/content/2/4/71/F13)

**Figure A3.** 
Work Hours Among Wage and Salary Employees, Non-Hispanic White Men

![Figure A4.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F14.medium.gif)

[Figure A4.](http://www.rsfjournal.org/content/2/4/71/F14)

**Figure A4.** 
Work Hours Among Wage and Salary Employees, Non-Hispanic White Women

![Figure A5.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F15.medium.gif)

[Figure A5.](http://www.rsfjournal.org/content/2/4/71/F15)

**Figure A5.** 
Work Hours Among Wage and Salary Employees, Non-Hispanic Black Men

![Figure A6.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/2/4/71/F16.medium.gif)

[Figure A6.](http://www.rsfjournal.org/content/2/4/71/F16)

**Figure A6.** 
Work Hours Among Wage and Salary Employees, Non-Hispanic Black Women

## FOOTNOTES

*   1. We realize that one cannot safely interpret the results of wage equation models based on observational data as causal, except under the implausible assumption that the observed covariates capture all wage-relevant differences between mothers and fathers, for example, or between mothers and childless women. We use the term *effect* sparingly. The exception is in our discussion of our decomposition models, where *price change effect* and *quantity change effect* are standard terminology (see data and methods).

*   2. Shelly Lundberg and Elina Rose (2000) report a nonlinear association between children and men's work hours and wages: the first child is associated with a large increase in hours and wages; the second, a smaller increase; and subsequent children, no increase. Evidence of nonlinearities in the motherhood wage penalty is mixed (see, for example, Gangl and Ziefle 2009). We focus on the average family wage differential for parents and childless adults.

*   3. We prefer the May-MORG to the March series because the latter has far fewer cases, reports annual income (which, for job changers, may not have been earned from the occupation or work hours in the reference week), and does not report usual hours at the main job.

*   4. For ninety-six cases in the 2010 through 2014 surveys, the CPS variable indicating the number of children was logically inconsistent with a variable indicating the ages of the children. We exclude these cases from our analysis.

*   5. Alexandra Killewald (2012) shows that fathers of nonresident and nonbiological children, unlike other fathers, do not experience a wage premium. One might anticipate an attenuated motherhood wage penalty for nonresident and nonbiological mothers.

*   6. We also imputed usual hours for the hours-vary cases using multivariate imputation, but because the results using this measure were nearly identical, we fall back on the simpler proxy.

*   7. We do not differentiate between voluntary and involuntary part-time hours. Our prior work showed that the two forms of part-time work have similar associations with wages (Cha and Weeden 2014).

*   8. We begin with a wage equation for men (in the analysis of gender wage gaps) and childless adults (in the analyses of family wage gaps). Results from analyses that use wage equations for women and parents, respectively, have a similar pattern. We report key results from this specification in the footnotes; full results are available from the first author.

*   9. If we use the wage equation for women as the base for the Juhn, Murphy, and Pierce decomposition, we find a price change effect of long work hours 0.019 log points (10 percent) among parents and 0.011 log points (6 percent) among childless adults. If we purge the data of all occupation-wage associations, we find a smaller price change effect: 7 percent for parents versus 14 percent in the main results; 6 percent for childless adults versus 8 percent in the main results.

*   10. The education price effects are calculated by summing the price change coefficients of the education categories and dividing by the total change (x100).

*   11. In decompositions that use the women's wage equation as the base, these percentages are 5 percent (childless adults) and 8 percent (parents).

*   12. Tables A7 and A8 present the full set of coefficients for the JMP decompositions of the family wage gap.

*   13. If we use the women's wage equation as the base for the decomposition, the quantity effect of part-time work decreases to 25 percent of the total for women, 10 percent for men.

*   Copyright © 2016 by Russell Sage Foundation. All rights reserved. Printed in the United States of America. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher. Reproduction by the United States Government in whole or in part is permitted for any purpose. We thank Martha Bailey, Sheldon Danziger, Tom DiPrete, and the anonymous reviewers for their helpful comments on an earlier version of this paper. Direct correspondence to: Kim Weeden at kw74{at}cornell.edu, 323 Uris Hall, Ithaca, NY 14853; Youngjoo Cha at cha5{at}indiana.edu; and Mauricio Bucca at mebucca{at}gmail.com. A one-click Stata replication package for the analyses in this paper is available at [www.kimweeden.com/manuscripts](http://www.kimweeden.com/manuscripts).

Open Access Policy: *RSF: The Russell Sage Foundation Journal of the Social Sciences* is an open access journal. This article is published under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.

## REFERENCES

1.  Avellar, Sarah, and Pamela Smock. 2003. “Has the Price of Motherhood Declined over Time? A Cross-Cohort Comparison of the Motherhood Wage Penalty.” Journal of Marriage and the Family 65(3): 597–607.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1111/j.1741-3737.2003.00597.x&link_type=DOI) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000184833200008&link_type=ISI) 

2.  Bianchi, Suzanne M., Liana C. Sayer, , Melissa A. Milkie, , and John P. Robinson. 2012. “Housework: Who Did, Does or Will Do It, and How Much Does It Matter?” Social Forces 91(1): 55–63.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1093/sf/sos120&link_type=DOI) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000307837000009&link_type=ISI) 

3.  Buchmann, Claudia, and Anne McDaniel. 2016. “Motherhood and the Wages of Women in Professional Occupations.” RSF: The Russell Sage Journal of the Social Sciences 2(4). doi: 10.7758/RSF.2016.2.4.05.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.7758/RSF.2016.2.4.05&link_type=DOI) 

4.  Budig, Michelle J., and Paula England. 2001. “The Wage Penalty for Motherhood.” American Sociological Review 66(2): 204–25.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.2307/2657415&link_type=DOI) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000168194200003&link_type=ISI) 

5.  Bureau of Labor Statistics (BLS). 2015. “Highlights of Women's Earnings in 2014.” Report no. 1058. Washington: U.S. Department of Labor. Accessed June 27, 2016. [http://www.bls.gov/opub/reports/womens-earnings/archive/highlights-of-womens-earnings-in-2014.pdf](http://www.bls.gov/opub/reports/womens-earnings/archive/highlights-of-womens-earnings-in-2014.pdf).
    
    

6.  Byker, Tanya. 2016. “The Opt-Out Continuation: Education, Work, and Motherhood from 1984 to 2012.” RSF: The Russell Sage Journal of the Social Sciences 2(4). doi: 10.7758/RSF.2016.2.4.02.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.7758/RSF.2016.2.4.02&link_type=DOI) 

7.  Card, David, and John E. DiNardo. 2002. “Skill-Biased Technological Change and Rising Wage Inequality: Some Problems and Puzzles.” Journal of Labor Economics 20(4): 733–83.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1086/342055&link_type=DOI) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000179222700002&link_type=ISI) 

8.  Cha, Youngjoo, and Kim A. Weeden. 2014. “Overwork and the Slow Convergence in the Gender Gap in Wages.” American Sociological Review 79(3): 457–84.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1177/0003122414528936&link_type=DOI) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000340424400005&link_type=ISI) 

9.  Charles, Maria, and David B. Grusky. 2004. *Occupational Ghettos: The Worldwide Segregation of Women and Men*. Stanford, Calif.: Stanford University Press.
    
    

10. Clarkberg, Marin, and Phyllis Moen. 2001. “Understanding the Time-Squeeze: Married Couples’ Preferred and Actual Work-Hour Strategies.” American Behavioral Scientist 44(7): 1115–35.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1177/0002764201044007005&link_type=DOI) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000168886200004&link_type=ISI) 

11. Cooke, Lynn Prince. 2014. “Gendered Parenthood Penalties and Premiums across the Earnings Distribution in Australia, the United Kingdom, and the United States.” European Sociological Review 30(3): 360–72.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1093/esr/jcu044&link_type=DOI) 

12. Correll, Shelley J., Stephen Benard, , and In Paik. 2007. “Getting a Job: Is There a Motherhood Wage Penalty?” American Journal of Sociology 112(5): 1297–338.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1086/511799&link_type=DOI) 

13. Coser, Lewis A. 1967. “Greedy Organizations.” European Journal of Sociology 8(2): 196–215.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1017/S000397560000151X&link_type=DOI) 

14. Cotter, David, Joan M. Hermsen, , and Reeve Vanneman. 2011. “The End of the Gender Revolution? Gender Role Attitudes from 1977 to 2008.” American Journal of Sociology 117(1): 259–89.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1086/658853&link_type=DOI) 
    
    [PubMed](http://www.rsfjournal.org/lookup/external-ref?access_num=22003521&link_type=MED&atom=%2Frsfjss%2F2%2F4%2F71.atom) 

15. England, Paula. 2010. “The Gender Revolution: Uneven and Stalled.” Gender & Society 24(2): 149–66.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1177/0891243210361475&link_type=DOI) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000275756800001&link_type=ISI) 

16. Gangl, Markus, and Andrea Ziefle. 2009. “Motherhood, Labor Force Behavior, and Woman's Careers: An Empirical Assessment of the Wage Penalty for Motherhood in Britain, Germany, and the United States.” Demography 46(2): 341–69.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1353/dem.0.0056&link_type=DOI) 
    
    [PubMed](http://www.rsfjournal.org/lookup/external-ref?access_num=21305397&link_type=MED&atom=%2Frsfjss%2F2%2F4%2F71.atom) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000268057900007&link_type=ISI) 

17. Gerson, Kathleen. 2009. *The Unfinished Revolution: How a New Generation Is Reshaping Family, Work, and Gender in America*. Oxford: Oxford University Press.
    
    

18. Glauber, Rebecca. 2008. “Race and Gender in Families and at Work: The Fatherhood Wage Premium.” Gender and Society 22(1): 8–30.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1177/0891243207311593&link_type=DOI) 

19. Goldin, Claudia. 2014. “A Grand Gender Convergence: Its Last Chapter.” American Economic Review. 104(4): 1091–119.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1257/aer.104.4.1091&link_type=DOI) 

20. Jacobs, Jerry A., and Kathleen Gerson. 2004. *The Time Divide: Work, Family, and Gender Inequality*. Cambridge, Mass.: Harvard University Press.
    
    

21. Juhn, Chinhui, Kevin M. Murphy, and Brooks Pierce. 1991. “Accounting for the Slowdown in Black-White Wage Convergence.” In *Workers and Their Wages*, edited by Marvin H. Kosters. Washington, D.C.: AEI Press.
    
    

22. Kahn, Joan R., Javier Garcia-Manglano, , and Suzanne M. Bianchi. 2014. “The Motherhood Penalty at Midlife: Long-Term Effects of Children on Women's Careers.” Journal of Marriage and the Family 76(1): 56–72.
    
    

23. Kalleberg, Arne L. 2000. “Nonstandard Employment Relations: Part-Time, Temporary, and Contract Work.” Annual Review of Sociology 26: 341–65.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1146/annurev.soc.26.1.341&link_type=DOI) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000089631300016&link_type=ISI) 

24. Kalleberg, Arne L. 2011. *Good Jobs, Bad Jobs: The Rise of Polarized and Precarious Employment Systems in the United States, 1970s to 2000s*. New York: Russell Sage Foundation.
    
    

25. Killewald, Alexandra. 2012. “A Reconsideration of the Fatherhood Wage Premium: Marriage, Coresidence, Biology, and Father's Wages.” American Sociological Review 78(1): 96–116.
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000314239900005&link_type=ISI) 

26. Kuhn, Peter J., and Fernando A. Lozano. 2008. “The Expanding Workweek? Understanding Trends in Long Work Hours Among U.S. Men, 1979–2006.” Journal of Labor Economics 26(2): 311–43.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1086/533618&link_type=DOI) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000255217500004&link_type=ISI) 

27. Lundberg, Shelly J., and Eliana Rose. 2000. “Parenthood and the Earnings of Married Men and Women.” Journal of Labour Economics 7(6): 689–710.
    
    

28. National Bureau of Economic Research (NBER). 1969–1984. “CPS May Supplement.” Accessed April 11, 2016. [http://www.nber.org/data/may.html](http://www.nber.org/data/may.html).
    
    

29. National Bureau of Economic Research (NBER). 1979–2014. “CPS Merged Outgoing Rotation Groups.” Accessed April 11, 2016. [http://www.nber.org/data/morg.html](http://www.nber.org/data/morg.html).
    
    

30. Pal, Ipshita, and Jane Waldfogel. 2016. “The Family Gap in Pay: New Evidence for 1967 to 2013.” RSF: The Russell Sage Journal of the Social Sciences 2(4). doi: 10.7758/RSF.2016.2.4.04.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.7758/RSF.2016.2.4.04&link_type=DOI) 

31. Pedulla, David S., and Sarah J. Thébaud. 2015. “Can We Finish the Revolution? Gender, Work-Family Ideals, and Institutional Constraint.” American Sociological Review 80(1): 116–39.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1177/0003122414564008&link_type=DOI) 
    
    [PubMed](http://www.rsfjournal.org/lookup/external-ref?access_num=26365994&link_type=MED&atom=%2Frsfjss%2F2%2F4%2F71.atom) 

32. Reid, Erin. 2015. “Embracing, Passing, Revealing, and the Ideal Worker Image: How People Navigate Expected and Experienced Professional Identities.” Organization Science 26(4): 997–1017.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1287/orsc.2015.0975&link_type=DOI) 

33. Ridgeway, Cecilia L. 2011. *Framed by Gender: How Gender Inequality Persists in the Modern World*. Oxford: Oxford University Press.
    
    

34. Ridgeway, Cecilia L., and Shelley J. Correll. 2004. “Motherhood as a Status Characteristic.” Journal of Social Issues 60(4): 683–700.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1111/j.0022-4537.2004.00380.x&link_type=DOI) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000225373200003&link_type=ISI) 

35. Sharone, Ofer. 2004. “Engineering Overwork: Bell-Curve Management at a High-Tech Firm.” In *Fighting for Time: Shifting Boundaries of Work and Social Life*, edited by Cynthia Fuchs Epstein and Arne L. Kalleberg. New York: Russell Sage Foundation.
    
    

36. Staff, Jeremy, and Jeylan T Mortimer. 2012. “Explaining the Motherhood Wage Penalty During the Occupational Career.” Demography 49(1): 1–21.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1007/s13524-011-0068-6&link_type=DOI) 
    
    [PubMed](http://www.rsfjournal.org/lookup/external-ref?access_num=22037996&link_type=MED&atom=%2Frsfjss%2F2%2F4%2F71.atom) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000300065100001&link_type=ISI) 

37. Townsend, Nicholas W. 2002. *The Package Deal: Marriage, Work, and Fatherhood in Men's Lives*. Philadelphia, Pa.: Temple University Press.
    
    

38. Waldfogel, Jane. 1998. “Understanding the ‘Family Gap’ in Pay for Women with Children.” The Journal of Economic Perspectives 12(1): 127–56.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1257/jep.12.2.127&link_type=DOI) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000073756400008&link_type=ISI) 

39. Williams, Joan C. 2000. *Unbending Gender: Why Family and Work Conflict and What to Do About It*. New York: Oxford University Press.
    
    

40. 1.  Williams,  Joan C, ed
    
    . 2003. “Litigating the Glass Ceiling and the Maternal Wall: Using Stereotyping and Cognitive Bias Evidence to Prove Gender Discrimination.” Employee Rights and Employment Policy Journal 7(2).

 [1]: /embed/graphic-1.gif
 [2]: /embed/graphic-2.gif
 [3]: /embed/graphic-3.gif
 [4]: /embed/graphic-4.gif
 [5]: /embed/graphic-5.gif