Determinants of Gender Differences in Change in Pay among Job-Switching Executives

Abstract

The authors investigate what determines differences in change in pay between men and women executives who move to new employers. Using proprietary data of 2,034 executive placements from a global search firm, the authors observe narrower pay differences between men and women after job moves. The unconditional gap shrinks from 21.5% in the prior employer to 15% in the new employer. After controlling for typical explanatory factors, the residual gap falls by almost 30%, from 8.5% at the prior employer to 6.1% in the new placement. This change reflects a relative increase in performance-based compensation for women and a lower level of unexplained pay inequality generally in external placements. Controlling for individual fixed effects, observed women have higher pay raises than do men. Finally, the authors find suggestive evidence that pay differences may also be moderated by differences in the supply and demand for women executives.

Keywords

executive pay executive labor market gender pay gap human capital external recruitment

Much of what we know about gender pay differences comes from studies of internal labor markets, although external markets—facilitated by search firms that help to identify external candidates—are becoming increasingly influential. Whereas search firms initially specialized in identifying candidates for only high-paying executive positions, today they are involved in placements across a variety of ranks, functions, and industries.

External job market research shows that placements can result in pay increases that are considerably larger than those for internal career tracks. But among job-switching managers, it is unclear whether or how such gains differ for women and men. Are gender differences evident in raises or in the factors that determine changes in pay? Our study addresses this question using compensation data for a large sample of executives changing jobs on the external labor market.

Our data include pre- and post- job change compensation, as well as a common set of explanatory variables, allowing us to employ comparative pay decomposition methods to examine the relation between external placements and wage structure. Wage structure, which reflects the prices set for various labor-market skills, measured and unmeasured, and the rents received for employment in particular sectors of the economy, has been shown to be an important factor in explaining gender pay differences over time and between countries. In addition, the structure of our data allows us to employ individual fixed-effect specifications to control for time-invariant unobserved individual attributes, a recurring confounding effect in the literature on gender pay differences.

Prior studies of the relation between mobility and the gender gap compare managers who use external job switches to advance their careers with managers who are promoted internally. Results suggest that men receive higher relative raises from external mobility than do women. It is unclear, however, whether these effects reflect differences in the qualifications of men and women who choose to apply for external positions, or discriminatory hiring and compensation practices. This phenomenon needs to be further unpacked. For example, do disadvantages for women originate in exclusion from consideration, in unfair hiring decisions, or in different compensation practices? Prior work indicates that those women who make it to final rounds of consideration for senior roles stand a stronger chance of being hired than men. This study investigates the next step in the process—do pay raises for placed women executives differ from those of comparable men? To examine this question, we explore the determinants of pay change among job-switching executives and how the determinants differ by gender.

Our data also allow us to explore the changing role of unobserved factors on the residual gender pay gap, together with the shifting weights of explanatory factors, which have not been studied for a population of executives switching positions. Partitioning the explanations of changes in pay in this way provides structure to study potential channels that attenuate, reproduce, or extend gender pay differences.

Related Literature

Executive Pay Gap Studies

The general pay gap has diminished since the 1980s (Blau and Kahn 1992, 1994, 2006, 2007, 2017; Loprest 1992). The ratio of average pay for women to average pay for men rose from 62% in 1980 to 74% in 1989, but has increased less rapidly since, to 77% in 1998 and 79% in 2010 (Blau and Kahn 2017). For high earners, the segment that includes executives, progress in closing the gap has been particularly slow in recent years (Kassenboehmer and Sinning 2014; Blau and Kahn 2017).

Given the influence and visibility of executives, the gender pay gap among this segment is of particular interest. Putting aside the use of stock, the average cash compensation of the Forbes 800 CEOs increased from $700,000 in 1970 to more than $2.2 million in 2000 (Murphy and Zábojník 2004). Women have typically not benefited from this steep pay rise, however, as they are under-represented in top management. In 2019, women made up only 27% of executive/senior-level officers at S&P 500 companies, 21% of board seats, and 11% of the top-five earners of the company.¹ Their exclusion from high-paying jobs contributes to the overall pay gap.

But what is the pay gap for women who become senior managers? Evidence on this question is mixed, as studies differ in how they have decomposed the gender gap into various explanatory factors and residual unobserved factors. Those that control for observables such as company size, job position, age, and seniority found that women are paid less than men, with estimates ranging from 5% to 16% (Bertrand and Hallock 2001; Bell 2005; Muñoz-Bullón 2010; Selody 2010; Elkinawy and Stater 2011; and Albanesi, Olivetti, and Prados 2015).

By contrast, Gayle, Golan, and Miller (2012) found that, controlling for executive rank and background, women executives actually earn more than men and are promoted faster. Leslie, Manchester, and Dahm (2017) argued that women qualified to fill senior leadership positions are paid a premium for their diversity value. Adoption of diversity initiatives (Bartels, Nadler, Kufahl, and Pyatt 2013; Mayer, McCluney, Sonday, and Cameron 2015) and the smaller talent pools for women senior executives (Betrand and Hallock 2001) may, therefore, work to narrow pay disadvantages for women.

Differences in the pay gap may also be partially driven by variation in how pay is measured. Bertrand and Hallock (2001) used salary, bonus, and stock option grants. Gayle et al. (2012) used salary, bonus, options, stock grants, and pensions. However, even studies that use consistent definitions of compensation reach varying conclusions. Focusing exclusively on salary, Elkinawy and Stater (2011) found that women are paid less than men, whereas Muñoz-Bullón (2010) found that gender differences in salary are largely explained by occupational title effects.

That employer attributes, including size and industry, influence the residual gender pay gap indicates the important role of occupational segregation. Jobs differ by seniority and function, and controlling for such job-specific classifications can affect the residual gap. Contextual details of the person, employer, and job matter greatly in determining where differences in pay originate. They show that gendered pay differences need to be unpacked to identify the channels through which differences may arise.

Studies of the Gender Gap and Mobility among Managers

Gender pay differences have largely been studied in the context of internal labor markets (Quintana-García and Elvira 2017); pay differences that occur in external placements have received less attention. Yet external markets are growing in importance (Bidwell, Briscoe, Fernandez-Mateo, and Sterling 2013). From the 1970s to the early 2000s, external CEO hires more than doubled from 15% to 33% (Murphy and Zábojník 2007); by 2010, 54% of positions with an annual salary of $150,000 or more involved external search firms (Hamori and Koyuncu 2011).

Switching employers has been linked to larger pay raises than have internal moves (Pfeffer and Baron 1988; Sonnenfeld and Peiperl 1988; Sonnenfeld, Peiperl, and Kotter 1988); however, such gains may differ by gender. Survey-based studies indicate that job-switching women managers are disadvantaged relative to their male counterparts (Brett and Stroh 1997; Dreher and Cox 2000; Lam and Dreher 2004; Dreher, Lee, and Clerkin 2011).² Comparing cash compensation differences between movers and stayers, these studies found that the relative raises for movers are greater for men than for women. Analysis of publicly traded technology companies also finds that mobility disadvantages women, particularly in variable compensation (Quintana-García and Elvira 2017). Such results suggest the disparities found in internal labor markets may also hold or even be magnified in external markets.

Fernandez and Abraham (2011) cautioned that external labor markets are governed by different factors than are internal markets. For example, search firms, which shortlist candidates for hiring firms, play an important role in external markets but not in internal markets. Botelho and Abraham (2017) reported that shortlisting practices of search firms can disadvantage women candidates, particularly when search costs for advisors are high. Although access to networks of placement intermediaries may be a barrier to job-switching women (Dreher et al. 2011), prior studies that condition on pools find no evidence that women are disadvantaged relative to comparable men (Fernandez and Abraham 2011; Fernandez and Campero 2017). Similar findings were reported in studies of biopharmaceuticals (Fernandez and Abraham 2011), retail banking firms (Fernandez and Abraham 2010), and C-level³ and board of director positions (Fernandez-Mateo and Fernandez 2013). Although these findings do not rule out discrimination, since explanatory variables such as rank, industry, and occupation may be affected by discrimination (Blau and Kahn 2017), they further refine the understanding of the origins of pay differences. Among externally placed executives, such candidate pool effects have been shown to explain much of the disparity in hiring decisions (Fernandez-Mateo and King 2011; Fernandez-Mateo and Fernandez 2016); conditioning on pools, hiring decisions do not show clear evidence of disadvantages for women.

These studies primarily focus on how gender bias affects candidate pools and hence final placement outcomes. However, it is also important to consider how women who achieve placements are rewarded (Kahn 2014). In this article, we study externally placed executives, to investigate whether women, who clear initial hurdles to high-paying jobs, receive raises that are different from similarly qualified men.

Methods

Data

We examine a population of job switchers for whom we can observe prior and placed compensation outcomes. This approach allows us to measure the contemporaneous change in compensation at the time of the switch for a given individual. Our data come from a top-five global executive search firm. The sample comprises 2,034 executive placements across multiple functions, industries, seniority levels, and geographies from 2004 to 2011. For each placement, we have data on geography, position title, industry, company, and cash compensation levels from both prior and placed positions. Individual attribute data include gender and education. To these data, we add details on individual career histories published on LinkedIn.

Data from search intermediaries provide the additional benefit of capturing job moves among executives below the C-suite. Prior work examining gender compensation differences among executives has relied mostly on mandatory compensation disclosures for the five highest-paid executives in publicly traded companies. Within this elite population, the representation of women is quite low; in studies examining the top-compensated executives, women held fewer than 8% of such positions (Bertrand and Hallock 2001; Demerjian, Lev, and McVay 2012; Gayle et al. 2012; Guthrie, Sokolowsky, and Wan 2012; Quintana-García and Elvira 2017). Additionally, focus on such senior positions introduces a strong selection bias; individuals at these levels have demonstrated exceptional abilities, regardless of gender (Gayle et al. 2012). Not surprisingly, perhaps, prior work warns against generalizing gender pay gaps from patterns created by such an elite group (Bertrand and Hallock 2001; Gayle et al. 2012). Because our data include lower-level placements, the representation of women is notably higher (18%) and potentially more representative of where most women executives work.

Our data also allow us to examine how job rank and job function affect compensation, which cannot be explored using data for only the five highest-paid executives. Job rank in a reporting hierarchy is a measure of professional attainment and is strongly correlated to pay levels. Functional role (such as finance, human resources, IT, marketing, and operations) also captures differences in work tasks that can affect compensation through occupational segregation (Anker 1997; Blackburn, Browne, Brooks, and Jarman 2002; Smith, Smith, and Verner 2013). For such functional specialization, human-capital investment decisions are often made at the beginning of a career and can have persistent effects on career paths and consequent compensation outcomes (Bielby and Baron 1984, 1986a, 1986b).

We also use data from the Bureau of Labor Statistics on the prevalence of female managers by industry. Data on the prevalence of women by function is collected from 2014 Capital IQ⁴ data on the top-20 officers of Fortune 500 firms.

Measures

We present summary statistics for the dependent and independent variables described below, by gender and for the full sample, in Table 1.

Table 1.

Summary Statistics by Gender for Sample of 2,034 Executives Accepting Placements in New Firms during the Period 2004 to 2011

				Male			Female
Variable	Obs.	Mean	S.D.	Obs.	Mean	S.D.	Obs.	Mean	S.D.	t-statistic
Comp – placed (′000)	2,034	330.9	260.0	1,670	338.1	249.8	364	298.0	300.6	2.37*
Comp – prior (′000)	2,034	293.9	388.0	1,670	306.1	422.0	364	238.2	143.4	5.31*
Total.experience	2,034	18.8	6.4	1,670	19.1	6.3	364	18.1	6.6	2.60*
Highest.deg = pre-Bachelors	2,034	0.01	0.07	1,670	0.01	0.07	364	0.01	0.07	0.02
Highest.deg = Bachelors	2,034	0.36	0.48	1,670	0.36	0.48	364	0.36	0.48	0.15
Highest.deg = Postgrad	2,034	0.63	0.48	1,670	0.63	0.48	364	0.63	0.48	0.15
Ivy.league	2,034	0.05	0.22	1,670	0.05	0.22	364	0.04	0.20	0.93
MBA	2,034	0.12	0.33	1,670	0.12	0.33	364	0.12	0.33	0.04
Risk.comp – prior	2,034	0.19	0.17	1,670	0.19	0.18	364	0.15	0.15	4.00*
Risk.comp – placed	2,034	0.21	0.15	1,670	0.22	0.15	364	0.20	0.15	2.00*
Job.rank – placed	2,034	0.48	0.50	1,670	0.46	0.50	364	0.60	0.49	5.07*
Job.rank – prior	2,034	0.58	0.50	1,670	0.56	0.50	364	0.65	0.48	3.01*
Firm.size – placed ≤500	2,034	0.27	0.44	1,670	0.27	0.44	364	0.26	0.44	0.48
Firm.size – placed = 501 to 1,000	2,034	0.09	0.28	1,670	0.09	0.29	364	0.06	0.24	1.91
Firm.size – placed = 1,001 to 5,000	2,034	0.21	0.41	1,670	0.22	0.41	364	0.16	0.37	2.51*
Firm.size – placed = 5,001 to 10,000	2,034	0.14	0.34	1,670	0.13	0.34	364	0.16	0.36	1.14
Firm.size – placed ≥10,001	2,034	0.30	0.46	1,670	0.29	0.45	364	0.36	0.48	2.58*
Firm.size – prior ≤500	2,034	0.19	0.39	1,670	0.19	0.39	364	0.20	0.40	0.61
Firm.size – prior = 501 to 1,000	2,034	0.06	0.24	1,670	0.07	0.25	364	0.05	0.22	1.08
Firm.size – prior = 1,001 to 5,000	2,034	0.19	0.39	1,670	0.19	0.39	364	0.20	0.40	0.67
Firm.size – prior = 5,001 to 10,000	2,034	0.14	0.34	1,670	0.14	0.35	364	0.11	0.32	1.51
Firm.size – prior ≥10,001	2,034	0.42	0.49	1,670	0.42	0.49	364	0.43	0.50	0.45
Public – placed	2,034	0.53	0.50	1,670	0.53	0.50	364	0.57	0.50	1.61
Public – prior	2,034	0.60	0.49	1,670	0.61	0.49	364	0.59	0.49	0.72

Mean for men and women significantly different at the p < 0.05 level.

Dependent Variable

Our primary outcome variable is salary plus performance bonus compensation (comp). We do not have data on stock or option awards. Mean total compensation for prior and placed jobs are approximately $294,000 and $331,000, respectively. Average compensation in the prior job setting is $306,000 for men and $238,000 for women, implying a gender pay difference of 22%; post-placement, mean compensation increases to $338,000 for men and $298,000 for women, indicating that the unconditional gender gap is 12% at the time of placement. Consistent with prior research, our empirical tests transform compensation levels by taking the natural log to diminish the effect of outliers.

Independent Variable

The key independent variable, female, takes the value 1 when the executive is female, 0 when male. Of the placed executives, 18% are women. Each executive appears twice in the data, once for his or her prior position and again for his or her placed position. In individual fixed-effects specifications, we focus on the effect of gender on compensation in the context of job moves. The primary variable of interest is then the interaction of female with the binary variable, placed, which takes the value of 1 for jobs representing a placement position, 0 if a prior position.

Control Variables

Human-capital theory links compensation with distinctive capabilities gained through experience and investments in education. We control for years of professional experience using the variable total.experience, defined as total number of work years. The average executive has almost 19 years of professional experience, with men having more experience than women (19.1 years for men versus 18.1 for women).

Three variables capture investments in education. Highest degree attained (highest.deg) is an indicator variable that captures whether an executive’s highest formal educational level was (1) pre-bachelors, (2) bachelors, or (3) postgraduate for those earning the equivalent of a master’s degree or higher, including all post-baccalaureate professional schools. We also construct binary variables that indicate whether the executive attended an Ivy League school (ivy.league) and if he or she has an MBA degree (mba), since these variables have been connected to perceived executive performance (Bertrand and Schoar 2003; Miller, Xu, and Mehrotra 2015). We observe no significant difference in the educational attainment of men and women executives in the sample: 63% have a postgraduate degree, 5% have attended an Ivy League school, and 12% have an MBA.

We control for share of cash compensation that is performance-based bonus (risk.comp) by dividing target performance bonus by the total cash compensation (i.e., salary and bonus). On average, bonuses for women are 15% of total compensation prior to external placement, and 20% after, compared to 19% and 22% for men.

At the job level, we include controls for job type based on job title classification. As executives attain higher reporting levels, the complexity of their tasks increases, leading to higher compensation levels (Agarwal 1981). Additionally, the scope of their influence and responsibility increases. Executive job titles are coded by keywords and classified by categories including “manager,”“director,”“vice president,” and so on. We coarsen the classification into a binary division of senior versus non-senior rank, with those at C-suite, president, and vice-president levels considered senior (job.rank = 0), and other positions non-senior (job.rank = 1). Prior to placement, nearly half the placements are for positions in which executives are not C-level, presidents, or vice-presidents, indicated by the mean job.rank value of 0.58. As we expected, post-placement, mean job.rank drops to 0.48, indicating that most of the sample candidates accepted promotions to move. Both before and after the move, men on average occupy higher-ranked positions than do women.

We also use job function (job.funct) to capture effects from occupational differences. The variable job.funct refers to the standard classification of departments by business function. Executive job titles are coded by function into categories that include finance, human resources, IT, marketing, operations, and so on. Function names and their breakdown of representation in the data are reported in Table 2, along with aggregate data for all women occupying a top 20 position at S&P 500 companies, as reported by Capital IQ in 2014. Both before and after the job switch, the percentage of sample women executives in each function is quite similar to those for women more broadly. The two exceptions are R&D, where the sample percentage is almost twice the aggregate (17% pre- and 19% post-switch, versus 9%), and legal (with a post-switch percentage of women of 43% versus 26% for the S&P 500).

Table 2.

Percentage of Women by Job Function for Sample of 2,034 Executives Accepting Placements in New Firms during the Period 2004 to 2011

	Aggregate % women S&P 500^a	% women indicator	Prior			Placed
	Aggregate % women S&P 500^a	% women indicator	Count	% oftotal	% women	Count	% of total	% women
Human resources	48	High	191	9	43	191	9	44
Marketing	28	High	213	10	27	209	10	27
Legal	26	High	47	2	32	53	3	43
Corporate	26	High	176	9	22	149	7	22
IT	18	High	119	6	18	138	7	16
Finance	17	Low	354	17	15	383	19	14
Sales	17	Low	218	11	10	210	10	9
Other	17	Low	150	7	15	76	4	17
Operations	10	Low	183	9	8	215	11	9
General management	10	Low	337	17	9	362	18	9
R&D	9	Low	46	2	17	48	2	19
Total			2,034	100	18	2,034	100	18

The aggregate frequency of women in management roles by job function is from Capital IQ data for women listed among the top 20 positions in their companies for Fortune 500 companies in 2014.

We include a set of contextual controls at the firm, individual, and job level. At the firm level, we control for the firm’s two-digit North American Industry Classification System (NAICS) industry (industry). Table 3 shows the percentage of women managers by industry for the placed sample and for the US economy using data from the Bureau of Labor Statistics (BLS) reported in 2000. On average, the percentage of women by industry is significantly lower for the sample executives (18% both before and after the job switch) than that reported by the BLS (39%). One plausible explanation for this difference is that our sample disproportionately represents larger companies (with fewer senior women) than do aggregate data covered by the BLS (which include many small companies founded by women). The correlation between the two is quite low (37% for pre- and 29% for post-switch data), however, suggesting that the women in our sample also come from different industries than typical women managers. Adjusting for this scale difference, industries that are particularly underrepresented in our sample include agriculture and real estate; those that are over-represented include wholesale (pre-switch), diversified (pre-switch), arts and entertainment (pre-switch), and education (both pre- and post-switch). Despite these differences, the frequency of men and women across industries remains remarkably stable in the pre- and post-switch periods.

Table 3.

Industry Frequency for Sample of 2,034 Executives Accepting Placements in New Firms during the Period 2004 to 2011

			Prior			Placed
Industry	Aggregate % women managers ^a	% women indicator	Count	% of total	% women	Count	% of total	% women
Health care	66	High	24	1.2	25.0	39	1.9	30.8
Financial	53	High	227	11.2	19.4	244	12.0	20.9
Education	53	High	22	1.1	45.5	29	1.4	20.7
Other services	45	High	100	4.9	16.0	112	5.5	18.8
Arts and entertainment	43	High	16	0.8	31.2	9	0.4	66.7
Accommodation and food services	43	High	34	1.7	23.5	44	2.2	25.0
Public administration	41	High	4	0.2	25.0	2	0.1	50.0
Professional services	40	High	172	8.5	23.3	142	7.0	19.0
Information	40	High	205	10.1	18.5	198	9.7	14.1
Agriculture	39	High	3	0.1	0.0	3	0.1	0.0
Real estate	39	High	15	0.7	0.0	14	0.7	7.1
Mining and extraction	39	High	24	1.2	16.7	33	1.6	18.2
Diversified	39	High	3	0.1	33.3	5	0.2	20.0
Retail	38	High	117	5.8	24.8	103	5.1	23.3
Transport and warehousing	26	Low	52	2.6	15.4	57	2.8	22.8
Wholesale	24	Low	15	0.7	33.3	8	0.4	12.5
Manufacturing	22	Low	964	47.4	15.4	949	46.7	15.8
Construction	12	Low	37	1.8	2.7	43	2.1	11.6
Total			2,034	100.0	17.9	2,034	100.0	17.9

The frequency of women in management roles by industry is for the US economy using data from the Government Accountability Office (GAO) and Bureau of Labor Statistics reported in 2000.

The control for firm size (firm.size) is the number of employees, measured as a categorical variable reflecting size classifications of less than 500; 501 to 1,000; 1,001 to 5,000; 5,001 to 10,000; and greater than 10,000. The distribution of sample executives by firm size is reported in Table 1. The executives systematically move from firms in the largest size category (42% pre and 30% post) to those in the smallest (19% to 27%). A similar pattern holds across genders.

To control for whether the company is publicly traded, public is an indicator variable that takes the value 1 for public firms, and 0 otherwise. Summary data from Table 1 show a modest switch among sample executives from public (60% before to 53% after) to private firms. This pattern is more pronounced for men than for women.

Because the data are global, we control for the region (region) of the job placement, grouping job placements into Africa, Asia, Europe, North America, South America, and Australia/New Zealand. Table 4 reports the distribution of placements by region. Blau and Kahn (1992, 2003) reported that international differences in the gender pay gap are to a significant extent explained by cross-country differences in overall wage structure, that is by the prices paid for various skills and qualifications in a given market, beyond differences in the skills and qualifications of men and women in those markets. Countries that reward measured and unmeasured labor market skills more (and punish their absence more harshly) have larger gender pay differences. Although these findings pertain to wages over the country populations broadly rather than the executive labor market studied in this article, they underscore the importance of accounting for wage structure, which we do in our analysis. Not surprisingly, perhaps, given their populations and executive labor market mobility, the majority of sample placements are in North America and Europe. The pre- and post-placement regional frequency of sample executives is virtually identical, however, for the full sample and for male and female executives. Finally, to control for macroeconomic trends, we include dummies to capture year effects (year), from 2004 to 2011.

Table 4.

Frequency of Men and Women Executives by Region for Sample of 2,034 Executives Accepting Placements in New Firms during the Period 2004 to 2011

	Prior			Placed
Region of placement	Count	Male frequency (%)	Female frequency (%)	Count	Male frequency (%)	Female frequency (%)
Africa	12	1	1	11	0	1
Asia	177	8	12	178	8	12
Europe	748	37	35	750	38	35
North America	934	46	46	927	46	46
South America	113	6	6	112	5	6
Australia/New Zealand	50	3	1	56	3	2
Total	2,034	100	100	2,034	100	100

We report correlations between the variables in Table 5. A consistent positive correlation is evident between the same variables in pre- and post-placed periods. Therefore, given the challenge of presenting the full correlation matrix for all pre- and post-placed variables, we present the matrix for the pre-placed period only. Consistent with prior research, we find a significant predictable correlation between compensation and experience, job rank, gender, and location in North America (Abowd and Kaplan 1999; Bertrand and Hallock 2001; Gayle et al. 2012; Quintana-García and Elvira 2017).

Table 5.

Bivariate Correlations Using Pre-placed Data for Sample of 2,034 Placed Executives

	Variable	1	2	3	4	5	6	7	8	9	10	11	12	13
1	Log(comp)	1.00
2	Female	−0.11**	1.00
3	Bachelors	−0.03	0.00	1.00
4	Postgrad	0.03	0.00	−0.99**	1.00
5	Ivy	0.09**	−0.02	−0.12**	0.12**	1.00
6	MBA	0.03	0.00	−0.19**	0.19**	0.24**	1.00
7	Total.experience	0.22**	−0.06*	0.05	−0.05	0.04	−0.03	1.00
8	Job.rank	−0.37**	0.06*	0.00	0.00	−0.06*	0.00	−0.22**	1.00
9	Risk.comp	0.41**	−0.08**	−0.04	0.04	0.03	0.00	0.07*	−0.17**	1.00
10	Public	0.01	−0.02	−0.02	0.03	0.01	0.00	0.00	0.03	0.05	1.00
11	Firm.size $ ≤ $ 500	0.01	0.01	−0.01	0.00	0.04	0.00	0.04	−0.07*	−0.02	−0.27**	1.00
12	Firm.size $ > $ 10,000	−0.07*	0.01	−0.03	0.03	−0.03	0.02	−0.03	0.11**	0.01	0.38**	−0.42**	1.00
13	Region: Europe	−0.08**	−0.02	−0.04	0.03	−0.13**	−0.01	−0.20**	0.27**	−0.10**	−0.05	0.03	−0.06	1.00
14	Region: North America	0.20**	−0.01	0.03	−0.03	0.17**	−0.03	0.25**	−0.39**	0.17**	0.06*	0.00	0.03	−0.70**

p < 0.01; *p < 0.05.

Empirical Methods

We estimate a variety of model specifications of the change in gender pay gap following external placement. The first, ordinary least squares (OLS), is estimated separately for the period before and after placement controlling for other factors related to pay, such as experience, education, position, and geography:

\begin{matrix} \log (com p_{it}) = β_{0} + β_{1} femal e_{i} + β_{2} total . experienc e_{i} + β_{3} highest . degre e_{i} + β_{4} mb a_{i} \\ + β_{5} ivy . leagu e_{i} + β_{6} risk . com p_{i} + β_{7} job . ran k_{i} + β_{8} job . func t_{i} + β_{9} firm . siz e_{i} \\ + β_{10} publi c_{i} + β_{11} industr y_{i} + β_{12} regio n_{i} + β_{13} yea r_{i} + ε_{i .} \end{matrix}

(1)

We then estimate Oaxaca-Blinder (Blinder 1973; Oaxaca 1973) and Juhn, Murphy, and Pierce (1991) (hereafter JMP) decompositions. For these two approaches, regressions using Equation (2) below are run separately for men and women before and after placement:

\begin{matrix} \log (com p_{it}) = β_{0} + β_{1} total . experienc e_{i} + β_{2} highest . degre e_{i} + β_{3} mb a_{i} \\ + β_{4} ivy . leagu e_{i} + β_{5} risk . com p_{i} + β_{6} job . ran k_{i} + β_{7} job . func t_{i} + β_{8} firm . siz e_{i} \\ + β_{9} publi c_{i} + β_{10} industr y_{i} + β_{11} regio n_{i} + β_{12} yea r_{i} + ε_{i .} \end{matrix}

(2)

The Oaxaca-Blinder decomposition analyzes how the observed explanatory factors have changed and how the unexplained portion of the pay gap has changed before and after placement. The JMP decomposition, which has been used to study differences in the gender pay gap over different time periods (Blau and Kahn 1994, 1997) and between different countries (Blau and Kahn 1992, 1996), is used here to study how pay differences change following a job move. We conduct JMP decomposition of the gender gap for the “human capital specification,” containing education and work experience measures (highest.deg, mba, ivy.league, and total.experience). We then add controls for firm attributes (firm.size, public, industry, region) and call this the “human capital and firm specification.” Next, we report a specification that also includes job attributes (job.rank and job.funct). Finally, we include risk.comp to estimate the “full specification.”⁵

To help control for unobserved ability differences among the sample executives, we then estimate an individual fixed-effects model that pools observations before and after placement:

\begin{matrix} \log (com p_{it}) = β_{1} (femal e_{i} \times place d_{t}) + β_{2} place d_{t} + β_{3} risk . com p_{it} \\ + β_{4} job . ran k_{it} + β_{5} job . func t_{it} + β_{6} firm . siz e_{it} + β_{7} publi c_{it} \\ + β_{8} industr y_{it} + β_{9} regio n_{it} + γ_{i} + ε_{it .} \end{matrix}

(3)

The parameter (γ_i), which represents the individual fixed effects, absorbs time-invariant factors such as female, total.experience, highest.deg, ivy.league, and mba (Allison 2009). Consequently, there is no estimated coefficient for female. The coefficient on the interaction term, female×placed, captures the effect of gender on the pay raise at the time of the job switch. Standard errors for fixed-effects regressions are clustered at the individual level.

The concern that unobserved differences drive much of the estimated gender gap (Hensvik 2014) is particularly relevant for studies that compare movers and stayers as two distinct populations, since managers with the opportunity to move may have other unobserved attributes that are correlated with higher pay and gender. Having drawn our sample from the database of a global executive search firm, we condition on one unobserved attribute identified in the literature, gender differences in access to placement channels (Dreher et al. 2011). Our sample comprises women with access to a global executive search firm, though, who might be drawn from a more select talent pool than the sample of men (Gayle et al. 2012). Individual fixed-effect specifications help us address this concern.

Our study is limited to the cash components of compensation—salary and target performance bonus. These pay components are argued to be more comparable across industries and levels (Brett and Stroh 1997), and make it more comparable with prior work (Brett and Stroh 1997; Dreher and Cox 2000; Lam and Dreher 2004; Dreher et al. 2011; and Leslie et al. 2017 all use cash compensation as measured outcomes of pay). However, the lack of visibility to other components of pay, including long-term incentive plans (LTIPs), stock, and options, limits the completeness of our consideration of pay differences. This is particularly concerning given the growing share of pay for such components (Edmans, Gabaix, and Jenter 2017).

Empirical studies suggest that gender differences in pay are higher in such variable components compared to salaries (Muñoz-Bullón 2010; Carter, Franco, and Gine 2017; Quintana-García and Elvira 2017). These studies suggest the role of gender differences in risk preferences (Gneezy, Niederle, and Rustichini 2003; Gneezy and Rustichini 2004; Niederle and Vesterlund 2007; Croson and Gneezy 2009) may explain these outcomes. To the extent that women are more focused on stable income, they may negotiate more forcefully on base salary (Muñoz-Bullón 2010). Thus, a focus on cash compensation may indicate only a partial story of progress on pay parity.

To examine this issue, albeit imperfectly, we separate the components of cash compensation into salary and target performance bonus. Bonuses are the component of cash compensation at risk.⁶ If risk preferences guide selective interest in compensation by component, we may observe the changes in cash compensation follow a similar pattern—that is, that the gap closes more on salaries compared to bonuses. This pattern would suggest that women are gaining on fixed salary, but might be losing ground on variable pay, implying that the exclusion of stock and option compensation from our analysis is likely to lead us to overstate the degree to which women receive higher raises than do men.

Results

OLS Models for Prior and Placed Employer Pay Gaps

We report separate OLS regression models for prior and placed positions in Tables 6 and 7. Model 1 of Table 6 reports the unconditional gender pay gap, with female as the only explanatory variable. The gender pay gap is 21.5% (e^0.195– 1). Including the year dummies in the base model (model 2) leaves the gap unchanged at 21.7%. Model 3’s addition of controls for human capital narrows the gap only marginally to 18.5%, reflecting the similar education and work experience qualifications of men and women in our sample. The human capital and employer specification (model 4), which includes the control variables firm.size, public, industry, and region, also leaves the gender gap largely unchanged (18.4%). The most dramatic reduction in the gender gap, to 11.7%, is for model 5, which includes job-specific controls, job.rank and job.funct. Finally, the introduction of risk.comp drops the gender gap to 8.5% (model 6). These results suggest that gender differences in job positions along with the level of performance-based risk in compensation contribute much to explaining pay differences in the prior employer.

Table 6.

Gender Compensation Gap at Prior Positions for 2,034 Executives Placed via the External Labor Market during the Period 2004 to 2011

	(1)	(2)	(3)	(4)	(5)	(6)
	Dependent variable = log(comp)
	No controls	Base	Human capital	Human capital & Employer	Human capital & Employer & Job	Human capital & Employer & Job & Risk.comp
Constant	12.406**	12.324	11.487**	11.217**	11.761**	11.811**
	(0.072)	(.)	(0.293)	(0.273)	(0.380)	(0.408)
Female	−0.195**	−0.196**	−0.170**	−0.169**	−0.111**	−0.082**
	(0.046)	(0.045)	(0.041)	(0.030)	(0.032)	(0.026)
Total.experience			0.023**	0.020**	0.015**	0.014**
			(0.003)	(0.003)	(0.003)	(0.003)
Risk.comp	No	No	No	No	No	Yes
Highest.deg	No	No	Yes	Yes	Yes	Yes
MBA	No	No	Yes	Yes	Yes	Yes
Ivy.league	No	No	Yes	Yes	Yes	Yes
Job.rank	No	No	No	No	Yes	Yes
Job.funct	No	No	No	No	Yes	Yes
Firm.size	No	No	No	Yes	Yes	Yes
Public	No	No	No	Yes	Yes	Yes
Industry	No	No	No	Yes	Yes	Yes
Region	No	No	No	Yes	Yes	Yes
Year	No	Yes	Yes	Yes	Yes	Yes
Observations	2,034	2,034	2,034	2,034	2,034	2,034
Adjusted R-squared	0.012	0.014	0.068	0.130	0.247	0.344

Notes: Standard errors in parentheses.

p < 0.01; **p < 0.05.

Table 7.

Gender Compensation Gap at Placed Positions for 2,034 Executives Placed via the External Labor Market during the Period 2004 to 2011

	(1)	(2)	(3)	(4)	(5)	(6)
	Dependent variable = log(comp)
	No controls	Base	Human capital	Human capital & Employer	Human capital & Employer & Job	Human capital & Employer & Job & Risk.comp
Constant	12.587**	12.480**	11.751**	11.687**	11.887**	12.242**
	(0.051)	(0.059)	(0.203)	(0.261)	(0.227)	(0.179)
Female	−0.140**	−0.143**	−0.120**	−0.132**	−0.061*	−0.059*
	(0.028)	(0.027)	(0.026)	(0.027)	(0.030)	(0.022)
Total.experience			0.019**	0.015**	0.011**	0.010**
			(0.003)	(0.003)	(0.002)	(0.002)
Risk.comp	No	No	No	No	No	Yes
Highest.deg	No	No	Yes	Yes	Yes	Yes
MBA	No	No	Yes	Yes	Yes	Yes
Ivy.league	No	No	Yes	Yes	Yes	Yes
Job.rank	No	No	No	No	Yes	Yes
Job.funct	No	No	No	No	Yes	Yes
Firm.size	No	No	No	Yes	Yes	Yes
Public	No	No	No	Yes	Yes	Yes
Industry	No	No	No	Yes	Yes	Yes
Region	No	No	No	Yes	Yes	Yes
Year	No	Yes	Yes	Yes	Yes	Yes
Observations	2,034	2,034	2,034	2,034	2,034	2,034
Adjusted R-squared	0.010	0.014	0.086	0.181	0.289	0.463

Notes: Standard errors in parentheses.

p < 0.01; **p < 0.05.

Table 7 reports analogous estimates of the gender gap for the same population of executives at the placement firms. Model 1 reports the unconditional gender gap of 15%, 30% less than for model 1 of Table 6. The gap is unchanged (15.3%) when year dummies are added to the base model (model 2). Model 3, the human capital specification that includes education and work experience variables, lowers the gap to 12.7%, comparable to model 4 (14.1%), which includes firm-specific controls. Similar to Table 6, the inclusion of job-specific controls has the greatest impact on the gender gap, closing the unexplained pay difference to 6.3% (model 5). Finally, the introduction of risk.comp, in model 6, lowers the gap modestly to 6.1%, suggesting that risk plays a limited role in driving pay differences among new employers.

Oaxaca-Blinder Decomposition of Prior and Placed Employer Pay Gaps

Table 8 reports the Oaxaca-Blinder decomposition for the prior and placed employer. The decomposition is from the viewpoint of women. Group differences in the predictors are weighted by the coefficients of women to determine the endowment effect. The endowment effect captures the expected change in log earnings for women if these women had predictor levels of men. For the coefficient effect, the difference in coefficients between men and women are weighted by the predictor levels of women, implying that the coefficient effects represent expected changes in log earnings for women if they had the coefficients of men, given their current predictor levels.

Table 8.

Oaxaca-Blinder Decomposition of Gender Compensation Gap for 2,034 Executives Placed via the External Labor Market during the Period 2004 to 2011

	Prior period			Placed period
	Coefficient	Std. Error	p-value	Coefficient	Std. Error	p-value
Overall
Men	12.406	0.017	0.000	12.587	0.013	0.000
Women	12.212	0.035	0.000	12.447	0.028	0.000
Difference	0.195	0.039	0.000	0.140	0.030	0.000
Endowments	0.124	0.038	0.001	0.097	0.030	0.001
Coefficients	0.088	0.034	0.011	0.062	0.026	0.017
Interaction	−0.017	0.034	0.615	−0.018	0.025	0.469
Endowments
Total.experience	0.010	0.006	0.084	0.009	0.005	0.062
MBA	0.000	0.003	0.971	0.000	0.001	0.971
Ivy	0.000	0.002	0.942	0.000	0.001	0.898
Highest.deg	0.000	0.002	0.951	0.000	0.001	0.945
Risk.comp	0.076	0.020	0.000	0.025	0.013	0.052
Job.funct	0.015	0.020	0.468	0.037	0.017	0.025
Job.rank	0.025	0.010	0.013	0.030	0.010	0.002
Firm.size	0.000	0.006	0.965	0.008	0.006	0.193
Public	−0.001	0.002	0.679	−0.001	0.003	0.564
Industry	−0.012	0.018	0.506	−0.012	0.010	0.240
Region	0.006	0.009	0.514	0.007	0.006	0.285
Year	0.005	0.006	0.456	−0.006	0.006	0.269
Coefficients
Total.experience	0.076	0.092	0.412	0.028	0.069	0.689
MBA	−0.013	0.013	0.291	−0.003	0.010	0.764
Ivy	0.007	0.007	0.339	0.004	0.005	0.439
Highest.deg	−0.458	0.419	0.275	0.120	0.314	0.704
Risk.comp	−0.133	0.033	0.000	0.002	0.035	0.950
Job.funct	0.081	0.095	0.392	−0.047	0.077	0.538
Job.rank	−0.049	0.050	0.326	−0.013	0.035	0.705
Firm.size	−0.142	0.154	0.355	0.272	0.144	0.059
Public	0.044	0.045	0.330	0.002	0.031	0.958
Industry	0.261	0.621	0.674	−0.294	0.469	0.530
Region	−0.034	0.374	0.927	0.332	0.278	0.232
Year	−0.039	0.584	0.946	−0.221	0.386	0.567
Constant	0.487	1.027	0.635	−0.118	0.766	0.877
Interaction
Total.experience	0.004	0.005	0.434	0.002	0.004	0.692
MBA	0.000	0.002	0.971	0.000	0.000	0.972
Ivy	0.002	0.003	0.500	0.001	0.002	0.548
Highest.deg	0.000	0.002	0.919	0.000	0.001	0.997
Risk.comp	−0.032	0.011	0.004	0.000	0.003	0.950
Job.funct	0.000	0.023	0.984	−0.008	0.018	0.658
Job.rank	0.006	0.007	0.350	0.003	0.008	0.706
Firm.size	−0.001	0.006	0.885	−0.012	0.007	0.100
Public	0.002	0.003	0.563	0.000	0.003	0.959
Industry	0.003	0.019	0.877	−0.010	0.012	0.420
Region	0.006	0.007	0.441	0.002	0.005	0.713
Year	−0.006	0.007	0.385	0.004	0.007	0.584

Table 8 shows that gender pay differences are smaller in the placed job versus the prior job. For the prior position, 0.124 log points of the raw pay gap is explained by endowment effects, primarily job controls (job.funct and job.rank) and risk.comp, and 0.088 log points by differences in coefficients. By contrast, at the placed position, 0.097 log points of the raw pay gap is explained by endowment effects (again, primarily by job.funct, job.rank, and risk.comp) and 0.062 by differences in coefficients. This outcome suggests relatively similar patterns of explanatory factors between prior and placed jobs. Further analysis for the placement period shows that had women been in the same function and rank as men, the raw gender gap would have been lowered by 0.067 log points. The gap would have widened by 0.012 log points if women worked in the same industry, indicating that women are located in higher-paying industries. Finally, the endowment effect of risk.comp, the amount of compensation that is performance based, drops in influence from 0.076 to 0.025 log points, suggesting that at-risk pay is less of a gender discriminator in the placed position.

JMP Decomposition of the Change in Pay Gap

We apply the decomposition method suggested by JMP (Juhn, Murphy, and Pierce 1991) to analyze how the change in gap can be attributed to changes in the explanatory factors. In these decompositions, the male group is the group that determines reference coefficients and reference residual distributions. Table 9 reports the results. Panel A summarizes the gender gap and the changes in pay percentiles. For the observed job change, the unconditional pay gap is smaller by 0.054 log points, and the mean sample woman manager’s pay would rise from the 40th percentile in the pay distribution to the 41st.

Table 9.

JMP Gender Compensation Gap for 2,034 Executives Placed via the External Labor Market during the Period 2004 to 2011

Panel A: Overall gap, before and after placement
	Period		Differential	Mean female percentile
	Prior		0.194	0.401
	Placed		0.140	0.414
Panel B: Analysis of gap over job switch
			(1)	(2)	(3)
			Human capital & Employer	Human capital & Employer & Job	Human capital & Employer & Job & Risk.comp
B.1. Descriptive statistics
	Mean female residual:
	Prior		0.169	0.111	0.082
	Placed		0.132	0.061	0.059
	Standard deviation of residuals for males:
	Prior		0.629	0.581	0.549
	Placed		0.459	0.424	0.369
		Difference:	0.17	0.157	0.180
B.2. Decomposition of change in differential
	Observed variables		−0.002	0.026	−0.028
	Observed prices		−0.021	−0.034	−0.019
Decomposition of observed factors			−0.023	−0.008	−0.047
	Gap		0.006	−0.032	0.042
	Unobserved prices		−0.037	−0.014	−0.049
Decomposition of unobserved factors			−0.031	−0.046	−0.007

Notes: JMP: Juhn, Murphy, and Pierce (1991).

Table 9, panel B.1 shows the gender gap after controlling for a variety of explanatory factors. Controlling for human capital and firm attributes (model 1), the gender gap drops from 0.169 log points at the prior employer to 0.132 log points at the placement employer. When we add job-specific variables of job.rank and job.funct (model 2), the gap is lowered from 0.111 log points to 0.061 log points. Finally, when we include risk.comp (model 3), it falls from 0.082 log points to 0.059 log points. These findings confirm that the pay gap is consistently lower in the placed setting than in the prior setting, and that a sizable share of the gender gap is explained by differences in job-specific attributes and the level of performance-based pay.

Panel B.2 decomposes changes in the gender gap for various specifications. Controls for human capital, employer attributes, and job attributes (see model 2) explain a relatively small share of the change in the gender difference in pay (−0.008 log points). Decomposing this into volume and price effects reveals offsetting effects, however. The volume effect (0.026 log points) shows that women did not narrow the pay gap by moving to more attractive male-dominated jobs or employers, whereas the price effect (–0.034 log points) indicates a decline in returns to these positions for men, which did narrow the gap. This drop in returns to observable attributes for men (relative to women) is consistent with Leslie et al. (2017), who argued that among individuals demonstrating the potential to fill senior leadership positions in organizations, women enjoy higher returns because they both possess desired qualifications and help organizations achieve diversity goals.

By contrast, the gap and unobserved price effects in model 2 worked to close the gender gap. Controlling for human capital, employer attributes, and job attributes, placement enabled women to move up the distribution of pay relative to men, reflected in a −0.032 log point change in the pay differences. In addition, a reduction in male residual inequality narrowed the gap. Had women maintained their percentile positions from the prior employer, the reduction in pay disparity as measured by the male residual inequality would have helped close differences in pay by an additional 0.014 log points. This finding suggests that unobserved factors contributed substantially to closing the pay gap.

Model 3 of panel B.2 adds the risk.comp control for the JMP decomposition results. The total effect (−0.047 log points) indicates that risk.comp is a substantial explanatory factor. Of the 0.054 log points change in raw gender gap, 87% (0.047 log points) is explained by observables. Approximately 60% of this (0.028 log points) is driven by changing endowment effects, reflecting women assuming a greater share of at-risk compensation. The remaining 40% (0.019 log points) is driven by diminished returns to observables for males.

Unobserved factors account for only a −0.007 log point share of change in pay difference. This shift from model 2 suggests that changes in the level of risk-based compensation are important in explaining the change in the gap, as inclusion of risk.comp materially shifts the explanatory weight of observable factors from low to high. The further decomposition of unobserved factors into gap effects versus unobserved price effects indicates the two are offsetting. The gap effect of 0.042 log points implies that, after controlling for observables, women have actually lost ground to men in their relative position in the wage distribution. This is offset, however, by the price effect on unobserved factors (−0.049 log points), coming from a shift in residual male pay inequality.

Comparing results from model 2 and model 3, the explanatory contribution of performance-based compensation appears to account for a substantial portion of the change in the pay differences. In model 2, most of the change in the gender gap is explained by changes in the unobserved component. By contrast, the inclusion of risk.comp in model 3 indicates that much of the gap is explained by observed factors and that a change in at-risk compensation assumed by women in placed positions contributes substantially to their higher raises.

Across all models, the wage structure, or “price effects,” are first-order factors. The sum of observed and unobserved price effects shows how the gap has changed as a result of changing prices. The change in standard deviation of male residuals from the wage function in Table 9, panel B.1 shows that in each specification the dispersion of residual pay for men shrinks considerably.

These results suggest that the residual pay differences narrowed following the job switch, because sample women managers took on more performance risk in their pay, and not because they moved to jobs and employers with characteristics that commanded higher returns. In addition, diminished returns to observed attributes for men helped narrow the gap. On average, women actually dropped in percentile ranks compared with men, which would have unraveled these gains. However, this effect was offset by the price effect on unobservables, which showed a reduction in male pay inequality between periods.

Thus, these JMP decompositions support two explanations for the observed narrowing of pay differences—that women took on more at-risk performance-based pay and benefited from lower general pay disparity among executive placements.

Individual Fixed-Effect Models

One explanation for our earlier findings is that placed women in our sample are systematically higher quality than our placed men, a form of sample-selection bias. This could explain the willingness of sample women to accept higher at-risk pay, the lower returns for sample men on observables, and the smaller gender pay gap among the placed employers. To further examine how such unobservable factors affected changes in the gender gap following external placements, we estimate the pay model using time-invariant fixed effects. Fixed effects capture personal attributes, such as education, gender, and experience, that do not vary, although employers could value them at different levels of importance.

Table 10 reports estimates of the fixed-effects models. The fixed effect absorbs any personal characteristics of the sample managers, including education, gender, and experience. The impact of the job change on the pay of women relative to men is captured by the estimate for female×placed. Model 1, which contains no time-varying controls, reports that the change in gender gap is 0.054 and statistically significant only at the p < 0.10 level, just beyond the bounds of the p < 0.05 level. It implies that, controlling for individual fixed effects, women receive an additional raise of 5.5% (e^0.054– 1) beyond their male counterparts. This effect changes only modestly when the model includes time-varying employer attributes (estimate of 0.059 in model 2) and job attributes (estimate of 0.062 in model 3). Model 4 introduces the risk.comp control and lowers the female×placed estimate to 0.04, which is insignificant. This attenuation is consistent with the findings of the JMP decomposition analysis and suggests that the risky component of compensation contributes materially toward narrowing gender pay differences. This finding is reinforced by results for salary and for bonus, both reported in Table 11. The coefficients on female×placed are not significant for models 1 through 3 in which the dependent variable is log(salary). For models 4 through 6, in which the dependent variable is log(bonus), the female×placed estimates are large (0.570 to 0.618) and significant at the p < 0.10 level.

Table 10.

Change in Gender Compensation Gap for Executives Taking New Positions via the External Labor Market; Fixed-Effects Regressions

	(1)	(2)	(3)	(4)
	Dependent variable = log(comp)
	Human capital	Human capital & Employer	Human capital & Employer & Job	Human capital & Employer & Job & Risk.comp
Female × Placed	0.054	0.059*	0.062*	0.040
	(0.028)	(0.028)	(0.029)	(0.025)
Placed	0.181**	0.186**	0.182**	0.160**
	(0.012)	(0.013)	(0.013)	(0.012)
Risk.comp				0.917**
				(0.102)
Job.rank	No	No	Yes	Yes
Job.funct	No	No	Yes	Yes
Industry	No	Yes	Yes	Yes
Firm.size	No	Yes	Yes	Yes
Public	No	Yes	Yes	Yes
Region	No	Yes	Yes	Yes
Observations	4,068	4,068	4,068	4,068
R-squared	0.132	0.160	0.171	0.279

Notes: Standard errors in parentheses.

p < 0.01; **p < 0.05.

Table 11.

Change in Gender Salary Gap and Bonus Gap for Executives Taking New Positions via the External Labor Market; Fixed-Effects Regressions

	(1)	(2)	(3)	(4)	(5)	(6)
	Dependent variable = log(salary)			Dependent variable = log(bonus)
	Human capital	Human capital & Employer	Human capital & Employer & Job	Human capital	Human capital & Employer	Human capital & Employer & Job
Female × Placed	0.010	0.013	0.009	0.570	0.618	0.608
	(0.026)	(0.026)	(0.026)	(0.345)	(0.345)	(0.347)
Placed	0.163**	0.167**	0.170**	1.587**	1.637**	1.580**
	(0.014)	(0.016)	(0.016)	(0.150)	(0.154)	(0.158)
Job.rank	No	No	Yes	No	No	Yes
Job.funct	No	No	Yes	No	No	Yes
Industry	No	Yes	Yes	No	Yes	Yes
Firm.size	No	Yes	Yes	No	Yes	Yes
Public	No	Yes	Yes	No	Yes	Yes
Region	No	Yes	Yes	No	Yes	Yes
Observations	4,068	4,068	4,068	4,068	4,068	4,068
R-squared	0.081	0.100	0.109	0.072	0.086	0.096

Notes: Standard errors in parentheses.

p < 0.01; **p < 0.05.

Overall, the fixed-effect tests reinforce the JMP decomposition findings. On average, the women in our sample receive higher at-risk bonus pay raises than do their male counterparts, reducing the gender differences in pay. We cannot rule out that this reflects quality differences between placed men and women, however, and that new employers use bonus awards more aggressively than did the prior employers to reward such quality differences.

Investigating Moderators of Findings

To the extent that prices are a function of supply and demand, contextual differences in labor market conditions could affect pay raises by gender. When the supply of women managers is low (high), or the demand of women managers is high (low), expected pay raises could be higher (lower). Supply and demand forces are difficult to isolate, however, given the many competing explanations for conditions such as scarcity and pay differences. As a result, we are unable to causally identify the effect of variation in supply or demand on gender differences in changes in pay. Instead, we use the fixed-effect model to investigate the impact of a series of contextual moderator factors that are likely to be associated with supply and demand effects.

Placement Seniority

More-senior placements are expected to have a greater percentage pay increase for women than for senior men, thereby reducing the gender pay gap because such placements are higher profile. Many organizations seek to augment diversity (Bartels et al. 2013), yet women are not proportionately represented among senior management ranks. Leslie et al. (2017) argued that women who can fill senior management roles are more valuable to organizations for the purposes of advancing diversity goals than are men of similar qualifications, warranting higher pay. Such hires create more visible examples of support for gender balance, since the positions are more prestigious and carry greater opportunity for impact. We, therefore, construct the variable sr, which takes the value 1 for senior placements by candidates currently in a senior position, and 0 otherwise.

Table 12 reports estimates of the gender pay gap for senior (sr = 1) and less-senior positions (sr = 0). Model 1 estimates show that women with current senior titles who are placed in senior positions receive a statistically reliable 0.139 log point additional increase in pay beyond the 0.086 log point average pay increase for comparable males. By contrast, women not moving between senior positions (model 2) do not receive any differential pay increases other than the 0.237 log point increase earned by comparable male placements. A Wald chi-square test indicates that the difference in the gender pay raise for senior and non-senior job changes is statistically reliable (p < 0.05). These findings continue to hold when we include risk.comp in the fixed-effect models (see Table 13).

Table 12.

Comparison of Gender Compensation Effects for 2,034 Placed Executives Taking New Positions via the External Labor Market by Seniority and by Public vs. Private Company Status Using Fixed-Effect Regressions

	(1)	(2)	(3)	(4)
	Dependent variable = log(comp)
	Moves between senior positions	No moves between senior positions	Moves between public companies	No moves between public companies
Female × Placed	0.139**	0.016	0.115*	0.028
	(0.042)	(0.037)	(0.049)	(0.032)
Placed	0.086**	0.237**	0.217**	0.157**
	(0.014)	(0.017)	(0.022)	(0.015)
Job.rank	No	No	Yes	Yes
Job.funct	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes
Firm.size	Yes	Yes	Yes	Yes
Public	Yes	Yes	No	No
Region	Yes	Yes	Yes	Yes
Observations	1,368	2,700	1,554	2,514
R-squared	0.182	0.201	0.229	0.157

Notes: Difference in the female×placed coefficients between model 1 and model 2 are significant at the p < 0.05 level. Difference in the female×placed coefficients between model 3 and model 4 are significant at the p < 0.10 level. Standard errors in parentheses.

p < 0.01; **p < 0.05.

Table 13.

Change in Gender Compensation Gap for Executives Taking New Positions via the External Labor Market, including Control for Percentage of Compensation as Performance Bonus; Fixed-Effects Regressions

	(1)	(2)
	Dependent variable = log(comp)
	Moves between senior positions	No moves between senior positions
Female × Placed	0.094**	0.002
	(0.036)	(0.033)
Risk.comp	0.853**	0.953**
	(0.118)	(0.139)
Placed	0.068**	0.212**
	(0.013)	(0.017)
Job.rank	No	No
Job.funct	Yes	Yes
Industry	Yes	Yes
Firm.size	Yes	Yes
Public	Yes	Yes
Region	Yes	Yes
Observations	1,368	2,700
R-squared	0.366	0.292

Notes: Standard errors in parentheses.

p < 0.01; **p < 0.05.

Placement Industry

Placements in industries with a relatively low percentage of women are expected to pay relatively more to women than to men, both because of women’s relative scarcity and because they are likely to demand a higher premium to compensate for incremental career risk and personal disruptions from a move. The greater career risk for women in industries and functions with fewer women stems from the lack of role models and networks that have been argued to be important to professional attainment (Lyness and Thompson 2000). To estimate this hypothesis, we construct the variable fem.industry, which takes the value 1 for companies in industries with a higher percentage of women in management positions than the industry median (26%) reported by the 2000 BLS study of 2-digit NAIC industries (see Table 3), and 0 otherwise.

As reported in model 1 of Table 14, women executives who take positions in industries with relatively few women (fem.industry = 0) receive an economically and statistically significant 0.132 log point additional increase in pay beyond the 0.208 log point increase for comparable men. By contrast, women do not receive any differential pay growth apart from the 0.158 log point increase earned by comparable men when they take new positions in industries with a relatively high representation of women (see model 2 for fem.industry = 1). A Wald chi-square test, however, indicates that the difference in female×placed estimates for industries with high and low frequencies of women is not significant at traditional levels.

Table 14.

Comparison of Gender Compensation Effects for 2,034 Placed Executives Taking New Positions via the External Labor Market by Industry and by Function, Using Fixed-Effect Regressions

	(1)	(2)	(3)	(4)
	Dependent variable = log(comp)
	Male-dominated industry	Less male-dominated industry	Male-dominated function	Less male-dominated function
Female × Placed	0.132*	0.037	0.134*	0.031
	(0.061)	(0.030)	(0.057)	(0.033)
Placed	0.208**	0.158**	0.183**	0.168**
	(0.019)	(0.017)	(0.017)	(0.020)
Job.rank	Yes	Yes	Yes	Yes
Job.funct	Yes	Yes	No	No
Industry	No	No	Yes	Yes
Firm.size	Yes	Yes	Yes	Yes
Public	Yes	Yes	Yes	Yes
Region	Yes	Yes	Yes	Yes
Observations	1,718	2,350	2,272	1,796
R-squared	0.201	0.148	0.181	0.165

Notes: Standard errors in parentheses.

p < 0.01; **p < 0.05.

Placement Function

Placements for functional positions with a low percentage of women are expected to pay more for women than for comparable men, once again, because of their scarcity and because women candidates are likely to demand a higher premium to compensate for incremental career risk and personal disruption. To test this effect, we use the variable fem.function, which takes the value of 1 for functions with a higher percentage of women managers than the median function (based on aggregate data for the top 20 officers of the Fortune 500 firms, as reported by Capital IQ in 2014), and 0 otherwise (see Table 2).

Model 3 of Table 14 shows that women hired on the external labor market in functions with relatively few women (fem.function = 0) earn an additional 0.134 log point pay increase beyond the 0.183 log point increase for men placed in these same functions. By contrast, in functions with a higher frequency of women (model 4, fem.function = 1), women do not earn materially more from changing positions than the 0.168 log point pay increase earned by comparable men. A Wald chi-square test, however, indicates that the difference in female×placed estimates for functions with high and low frequencies of women is not significant at traditional levels.

Visibility of a Publicly Traded Company

Public companies face greater scrutiny not only for financial performance but also for their conduct. Given stronger pressure to support gender-based parity (Tinsley, Wade, Main, and O’Reilly 2017), public companies may be more willing to pay women more, since such hires position them for better balance in representation. To examine this question, we construct the variable public, which takes the value of 1 for job moves taking place between public companies, 0 otherwise.

Turning back to Table 12, we report estimates of the gender pay gap for job switches occurring between publicly traded companies (public = 1) and private (public = 0). Model 3 estimates show that women who move jobs between publicly traded companies receive a statistically reliable 0.115 log point additional increase in pay beyond the 0.217 log point average pay increase for comparable males. Women whose moves include privately held companies, however, do not receive a differential pay increase from the 0.157 log point increase earned by comparable male placements (model 4). A Wald chi-square test indicates that the difference in the gender pay raise occurring between publicly traded and private companies is statistically reliable, but only at the p < 0.10 level.

In summary, supply and demand factors are difficult to identify precisely. The low supply of women in certain industries and functions, for example, could be a response to weaker demand and/or higher gender discrimination. By contrast, environments with a greater demand for women managers may have attracted more women (Ali, Kulik, and Metz 2011; Elkinawy and Stater 2011; Leslie et al. 2017; Quintana-García and Elvira 2017). We do not attempt to estimate the effect of variation in supply or demand on the gender gap and, therefore, interpret the findings with caution.

Perhaps it is not surprising that the most consistent and reliable evidence of higher raises for women from these tests is for senior placements, where supply and demand effects are easier to separately identify. Prior research shows that, as diversity is increasingly valued, demand effects for senior placements favor women. Controlling for pool size, senior women are actually more likely than senior men to be hired (Fernandez and Abraham 2010, 2011; Gayle et al. 2012). The immediate supply of available women candidates is limited, however, because of high attrition rates (Gayle et al. 2012) and the lengthy work experience required.

Robustness Checks on Regional Differences

Table 8 does not indicate that regional differences account for substantial gender differences in pay. We do not find this surprising since only 5.5% of sample managers switch between regions. Table 4 shows that the frequency of placements by region are comparable for men and women. The only exception is for Asia, where women placements are 12%, higher than that of men at 8%. In unreported individual fixed-effect regressions (available upon request from the authors), Asian placements have a negative effect on pay raises for women, though this effect is not precisely estimated. Because of limitations of the sample, we cannot precisely characterize the impact of regional differences on gender differences in changes in pay.

Discussion and Conclusion

We find that following an external job move, women receive higher raises than do their male counterparts, thereby reducing pay inequality. Two factors underlie this change. First, residual pay inequality was smaller at the new employer relative to the old. And second, increased at-risk performance-based pay for women in the new positions contributed to narrowing pay differences between men and women. However, women continued to earn lower performance-based pay than men after the move. The results of fixed-effects models reinforce these findings—the smaller gender pay differences remained after controlling for time-invariant individual effects but were attenuated when we controlled for performance-based compensation.

In addition, we find suggestive evidence that the narrowing of the gender pay gap post job change may be related to factors associated with supply and demand. Notably, in circumstances with lower supply of or higher demand for women candidates, women appear to have measurably higher raises than do men, even accounting for compensation structure.

We recognize a number of limitations to our study. First, we do not observe any offer that the sample executives may have received to remain with their original employers, or for that matter any other external offers received. As a result, we assume that the pay increase for the year of job switch is the difference between new and old pay. If the incumbent employer comes close to matching the new employer’s offer, our measured reduction in pay differences is overstated. We argue, however, it is likely that most of any pay increase offered by the old employer would not have been made had the manager not received an outside offer, suggesting that our estimate is relevant for the sample managers.

Second, our sample is subject to two potential sources of selection bias. Similar to the studies of the top earners of public companies, our results may be influenced by the selectivity of women in the sample, which reflects those who gain access to these placement networks. Our fixed-effect models suggest that such unobserved individual attributes do not explain the results. But we cannot rule out that this form of selection exists, and new employers differentially recognize and value higher-quality women who opt to move. Another form of selection bias can arise if men are willing to move for smaller pay increases than are women. This situation could arise if women face greater costs for moving attributable to family disruptions. Alternatively, women may be more likely than men to consider factors other than pay in their decision. Groysberg (2008) found that although men focus largely on compensation, women consider a broader range of criteria, including fit, values, and managerial styles of potential colleagues. Incumbent employers might well anticipate this, and price discriminate against their women employees, offering less opportunity for promotions and raises. Greater selectivity on the part of women may also result in better subsequent job matches, yielding higher pay. All these factors could explain why we observe that external moves generate higher pay for women than for men.

Third, as recognized in earlier studies, identifying explanations for differences in pay does not eliminate the possibility that it is the result of discrimination since explanatory variables such as rank, industry, and occupation may also be affected by discrimination. Fourth, our data only capture salary and bonus, so we are not able to take stock and other elements of long-term incentive compensation into account. Fifth, our sample has disproportionate representation of women from manufacturing, raising questions about selection bias and generalizability.

Finally, although our work follows in the steps of earlier studies that seek to understand the complex processes that have served to perpetrate the gender gap, our study is descriptive rather than theory testing. Nonetheless, we believe that our findings inform theory on the role of differences between internal and external labor markets on hiring decisions. External hiring decisions are subject to greater information deficits since hirers lack information from repeated observation of internal candidates in various settings (Bidwell 2011; Quintana-García and Elvira 2017). This lack of information creates opportunity for more discretion (Arvey and Campion 1982; Dipboye 1992; Judge, Higgins, and Cable 2000; Posthuma, Morgeson, and Campion 2002). Conversely, Bidwell (2011) pointed out that external hirers tend to place greater emphasis on formal markers of ability, leading to decisions that are more objective and less vulnerable to bias. Greater formality in employment decision-making has been connected to a reduction in bias generally, and gender bias in particular (Pfeffer and Cohen 1984; Baron, Davis-Blake, and Bielby 1986; Sutton, Dobbin, Meyer, and Scott 1994; Reskin 2000; Elvira and Graham 2002). If external hiring decisions weigh hiring criteria differently, in a way that favors more objective criteria, they may help attenuate gender bias and/or reduce overall variance in unexplained pay differences. Our findings on price effects are consistent with this hypothesis. Nonetheless, we do not formally test this theory, creating an opportunity for further study of how differences in internal and external labor markets affect pay mechanisms and gender pay disparities.

Footnotes

Acknowledgements

We are grateful for the comments and suggestions provided by Robin Ely and for the extensive research support from Robin Abrahams. The authors are equal contributors to this article and are listed in alphabetical order.

The project was supported by financial assistance from the Department of Faculty Research and Development of the Harvard Business School. It did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Data for this study have been provided by a large global executive placement firm on the condition of anonymity. Such firms can be approached directly for requests of similar data, which they may provide at their discretion.

Additional results are available from the corresponding author at eric.lin@westpoint.edu.

ORCID iD

Eric Lin

1

See (accessed July 2019).

2

Dreher and Cox (2000) and Dreher et al. (2011) focused on comparisons between white men and minorities and women, whereas Brett and Stroh (1997) and focused exclusively on gender-based differences.

3

“C-level” and “C-suite” refer to those high-ranking executives that hold the top positions of an organization, by function. The “C” signifies the “chief” prefix common to the job titles, as in chief executive officer (CEO), chief operating officer (COO), chief financial officer (CFO), chief information officer (CIO), and so on.

4

Capital IQ is a database produced by Standard & Poor’s that can be used to find detailed company financial information for US and international public and private companies and investment firms.

5

These stepwise decompositions are modeled on the approach conducted in Blau and Kahn (2006) and . In these studies, the authors partitioned the decompositions into “human capital specification” (race, education, and work experience) and “full specification” (the inclusion of all other explanatory controls in addition to those in the human capital specification).

6

Given that the link between manager accomplishments and bonuses is more direct than the link between manager accomplishments and general stock prices, prior research has argued that performance bonuses have a more potent incentive effect than stock-based compensation (Murphy 2013; ).

References

Abowd

John M.

Kaplan

David S

. 1999. Executive compensation: Six questions that need answering. Journal of Economic Perspectives13(4): 145–68.

Agarwal

Naresh C.

1981. Determinants of executive compensation. Industrial Relations 20(1): 36–46.

Albanesi

Stefania

Olivetti

Claudia

Jose Prados

María

. 2015. Gender and dynamic agency: Theory and evidence on the compensation of top executives. In Polachek

Solomon W.

Tatsiramos

Konstantinos

Zimmermann

Klaus F.

(Eds.), Research in Labor Economics: Gender in the Labor Market, Vol. 42, pp. 1–60. Bingley, UK: Emerald Group Publishing.

Ali

Muhammad

Kulik

Carol T

Metz

Isabel

. 2011. The gender diversity–performance relationship in services and manufacturing organizations. International Journal of Human Resource Management 22(7): 1464–85.

Allison

Paul D.

2009. Fixed Effects Regression Models. Thousand Oaks, CA: Sage.

Anker

Richard

. 1997. Theories of occupational segregation by sex: An overview. International Labour Review 136(3): 315–39.

Arvey

Richard D.

Campion

James E

. 1982. The employment interview: A summary and review of recent research. Personnel Psychology 35(2): 281–322.

Baron

James N.

Davis-Blake

Alison

Bielby

William T

. 1986. The structure of opportunity: How promotion ladders vary within and among organizations. Administrative Science Quarterly 31(2): 248–73.

Bartels

Lynn K.

Nadler

Joel T.

Kufahl

Katie

Pyatt

Jodie

. 2013. Fifty years after the Civil Rights Act: Diversity-management practices in the field. Industrial Organizational Psychology: Perspectives on Science and Practice 6(4): 450–57.

10.

Bell

Linda

. 2005. Women-led firms and the gender gap in top executive jobs. IZA Discussion Paper No. 1689. Bonn, Germany: Institute for the Study of Labor.

11.

Bertrand

Marianne

Hallock

Kevin F

. 2001. The gender gap in top corporate jobs. Industrial and Labor Relations Review 55(1): 3–21.

12.

Bertrand

Marianne

Schoar

Antoinette

. 2003. Managing with style: The effect of managers on firm policies. Quarterly Journal of Economics 118(4): 1169–208.

13.

Bidwell

Matthew J.

2011. Paying more to get less: The effects of external hiring versus internal mobility. Administrative Science Quarterly 56(3): 369–407.

14.

Bidwell

Matthew J.

Briscoe

Forrest

Fernandez-Mateo

Isabel

Sterling

Adina

. 2013. The employment relationship and inequality: How and why changes in employment practices are reshaping rewards in organizations. Academy of Management Annals 7(1): 61–121.

15.

Bielby

William T.

Baron

James N

. 1984. A woman’s place is with other women: Sex segregation within organizations. In Barbara F. Reskin (Ed.), Sex Segregation in the Workplace: Trends, Explanations, Remedies, pp. 27–55. Washington, DC: National Academies Press.

16.

Bielby

William T.

Baron

James N

. 1986a. Men and women at work: Sex segregation and statistical discrimination. American Journal of Sociology 91(4): 759–99.

17.

Bielby

William T.

Baron

James N

. 1986b. Sex segregation within occupations. American Economic Review 76(2): 43–47.

18.

Blackburn

Robert M.

Browne

Jude

Brooks

Bradley

Jarman

Jennifer

. 2002. Explaining gender segregation. British Journal of Sociology 53(4): 513–36.

19.

Blau

Francine D.

Kahn

Lawrence M

. 1992. The gender earnings gap: Learning from international comparisons. American Economic Review 82(2): 533–38.

20.

Blau

Francine D.

Kahn

Lawrence M

. 1994. Rising wage inequality and the US gender gap. American Economic Review 84(2): 23–28.

21.

Blau

Francine D.

Kahn

Lawrence M

. 1996. Wage structure and gender earnings differentials: An international comparison. Economica 63(250): S29–S62.

22.

Blau

Francine D.

Kahn

Lawrence M

. 1997. Swimming upstream: Trends in the gender wage differential in 1980s. Journal of Labor Economics 15(1): 1–42.

23.

Blau

Francine D.

Kahn

Lawrence M

. 2003. Understanding international differences in the gender pay gap. Journal of Labor Economics 21(1): 106–44.

24.

Blau

Francine D.

Kahn

Lawrence M

. 2006. The US gender pay gap in the 1990s: Slowing convergence. Industrial and Labor Relations Review 60(1): 45–66.

25.

Blau

Francine D.

Kahn

Lawrence M

. 2007. The gender pay gap: Have women gone as far as they can? Academy of Management Perspectives 21(1): 7–23.

26.

Blau

Francine D.

Kahn

Lawrence M

. 2017. The gender wage gap: Extent, trends, and explanations. Journal of Economic Literature 55(3): 789–865.

27.

Blinder

Alan S.

1973. Wage discrimination: Reduced form and structural estimates. Journal of Human Resources 8(4): 436–55.

28.

Botelho

Tristan L.

Abraham

Mabel

. 2017. Pursuing quality: How search costs and uncertainty magnify gender-based double standards in a multistage evaluation process. Administrative Science Quarterly 62(4): 698–730.

29.

Brett

Jeanne M.

Stroh

Linda K

. 1997. Jumping ship: Who benefits from an external labor market career strategy? Journal of Applied Psychology 82(3): 331–41.

30.

Carter

Mary Ellen

Franco

Francesca

Gine

Mireia

. 2017. Executive gender pay gaps: The roles of female risk aversion and board representation. Contemporary Accounting Research 34(2): 1232–64.

31.

Croson

Rachel

Gneezy

Uri

. 2009. Gender differences in preferences. Journal of Economic Literature 47(2): 448–74.

32.

Demerjian

Peter

Lev

Baruch

McVay

Sarah

. 2012. Quantifying managerial ability: A new measure and validity tests. Management Science 58(7): 1229–48.

33.

Dipboye

Robert L.

1992. Selection Interviews: Process Perspectives. Cincinnati, OH: Thomson South-Western.

34.

Dreher

George F.

Cox

Taylor H

. 2000. Labor market mobility and cash compensation: The moderating effects of race and gender. Academy of Management Journal 43(5): 890–900.

35.

Dreher

George F.

Lee

Jeong-Yeon

Clerkin

Thomas A

. 2011. Mobility and cash compensation: The moderating effects of gender, race, and executive search firms. Journal of Management 37(3): 651–81.

36.

Edmans

Alex

Gabaix

Xavier

Jenter

Dirk

. 2017. Executive compensation: A survey of theory and evidence. European Corporate Governance Institute (ECGI) – Finance Working Paper No. 514/2016. Accessed at https://ssrn.com/abstract=2992287.

37.

Elkinawy

Susan

Stater

Mark

. 2011. Gender differences in executive compensation: Variation with board gender composition and time. Journal of Economics and Business 63(1): 23–45.

38.

Elvira

Marta M.

Graham

Mary E

. 2002. Not just a formality: Pay system formalization and sex-related earnings effects. Organization Science 13(6): 601–17.

39.

Fernandez

Roberto M.

Abraham

Mabel

. 2010. From metaphors to mechanisms: Gender sorting in (to) an organizational hierarchy. MIT Sloan Research Paper No. 4779-10. Accessed at https://ssrn.com/abstract=1589012.

40.

Fernandez

Roberto M.

Abraham

Mabel

. 2011. Glass ceilings and glass doors? Internal and external hiring in an organizational hierarchy. MIT Sloan Research Paper No. 4895-11. Accessed at https://ssrn.com/abstract=1804896.

41.

Fernandez

Roberto M.

Campero

Santiago

. 2017. Gender sorting and the glass ceiling in high-tech firms. ILR Review 70(1): 73–104.

42.

Fernandez-Mateo

Isabel

Fernandez

Roberto M

. 2013. Women on top: Executive search and the gender gap in allocation to top corporate jobs. American Sociological Association Annual Meeting, New York.

43.

Fernandez-Mateo

Isabel

Fernandez

Roberto M

. 2016. Bending the pipeline? Executive search and gender inequality in hiring for top management jobs. Management Science 62(12): 3636–55.

44.

Fernandez-Mateo

Isabel

King

Zella

. 2011. Anticipatory sorting and gender segregation in temporary employment. Management Science 57(6): 989–1008.

45.

Gayle

George-Levi

Golan

Limor

Miller

Robert A

. 2012. Gender differences in executive compensation and job mobility. Journal of Labor Economics 30(4): 829–72.

46.

Gneezy

Uri

Niederle

Muriel

Rustichini

Aldo

. 2003. Performance in competitive environments: Gender differences. Quarterly Journal of Economics 118(3): 1049–74.

47.

Gneezy

Uri

Rustichini

Aldo

. 2004. Gender and competition at a young age. American Economic Review 94(2): 377–81.

48.

Groysberg

Boris

. 2008. How star women build portable skills. Harvard Business Review 86(2): 74–82.

49.

Guthrie

Katherine

Sokolowsky

Jan

Wan

Kam-Ming

. 2012. CEO compensation and board structure revisited. Journal of Finance 67(3): 1149–68.

50.

Hamori

Monika

Koyuncu

Burak

. 2011. Career advancement in large organizations in Europe and the United States: Do international assignments add value? International Journal of Human Resource Management 22(4): 843–62.

51.

Hensvik

Lena E.

2014. Manager impartiality: Worker-firm matching and the gender wage gap. ILR Review 67(2): 395–421.

52.

Judge

Timothy A.

Higgins

Chad A.

Cable

Daniel M.

2000. The employment interview: A review of recent research and recommendations for future research. Human Resource Management Review 10(4): 383–406.

53.

Juhn

Chinhui

Murphy

Kevin M.

Pierce

Brooks

. 1991. Accounting for the slowdown in black-white wage convergence. In Kosters

Marvin H.

(Ed.), Workers and Their Wages: Changing Patterns in the United States, pp. 107–43. Washington, DC: American Enterprise Institute Press.

54.

Kahn

Lawrence M.

2014. New evidence on gender and the labor market: A symposium. ILR Review 67(2): 283–86.

55.

Kassenboehmer

Sonja C.

Sinning

Mathias G

. 2014. Distributional changes in the gender wage gap. ILR Review 67(2): 335–61.

56.

Lam

Simon S.

Dreher

George F

. 2004. Gender, extra-firm mobility, and compensation attainment in the United States and Hong Kong. Journal of Organizational Behavior 25(7): 791–805.

57.

Leslie

Lisa M.

Manchester

Colleen

Dahm

Patricia

. 2017. Why and when does the gender gap reverse? Diversity goals and the pay premium for high potential women. Academy of Management Journal 60(2): 402–32.

58.

Loprest

Pamela J.

1992. Gender differences in wage growth and job mobility. American Economic Review 82(2): 526–32.

59.

Lyness

Karen S.

Thompson

Donne E

. 2000. Climbing the corporate ladder: Do female and male executives follow the same route? Journal of Applied Psychology 85(1): 86–101.

60.

Mayer

David M.

McCluney

Courtney

Sonday

Laura

Cameron

Lindsey

. 2015. Giving managers the business: The effectiveness of business and moral case arguments for increasing organizational diversity. Academy of Management Annual Meeting, Vancouver.

61.

Miller

Danny

Xiaowei

Mehrotra

Vikas

. 2015. When is human capital a valuable resource? The performance effects of Ivy League selection among celebrated CEOs. Strategic Management Journal 36(6): 930–44.

62.

Muñoz-Bullón

Fernando

. 2010. Gender-compensation differences among high-level executives in the United States. Industrial Relations 49(3): 346–70.

63.

Murphy

Kevin J.

2013. Executive compensation: Where we are, and how we got there. In Constantinides

George M.

Harris

Milton

Stulz

René M.

(Eds.), Handbook of the Economics of Finance, Vol. 2, pp. 211–56. Oxford, UK: North Holland/Elsevier.

64.

Murphy

Kevin J.

Zábojník

Ján

. 2004. CEO pay and appointments: A market-based explanation for recent trends. American Economic Review 94(2): 192–96.

65.

Murphy

Kevin J.

Zábojník

Ján

. 2007. Managerial capital and the market for CEOs. Accessed at https://ssrn.com/abstract=984376.

66.

Niederle

Muriel

Vesterlund

Lise

. 2007. Do women shy away from competition? Do men compete too much? Quarterly Journal of Economics 122(3): 1067–101.

67.

Oaxaca

Ronald

. 1973. Male-female wage differentials in urban labor markets. International Economic Review 14(3): 693–709.

68.

Pfeffer

Jeffrey

Baron

James N

. 1988. Taking the workers back out: Recent trends in the structuring of employment. In Staw

Barry M.

Cummings

Larry L.

(Eds.), Research in Organizational Behavior, pp. 257–303. Greenwich, CT: JAI Press.

69.

Pfeffer

Jeffrey

Cohen

Yinon

. 1984. Determinants of internal labor markets in organizations. Administrative Science Quarterly 29(4): 550–72.

70.

Posthuma

Richard A.

Morgeson

Frederick P.

Campion

Michael A

. 2002. Beyond employment interview validity: A comprehensive narrative review of recent research and trends over time. Personnel Psychology 55(1): 1–81.

71.

Quintana-García

Cristina

Elvira

Marta M

. 2017. The effect of the external labor market on the gender pay gap among executives. ILR Review 70(1): 132–59.

72.

Reskin

Barbara F.

2000. The proximate causes of employment discrimination. Contemporary Sociology 29(2): 319–28.

73.

Selody

Karen

. 2010. Is the risk worth the reward for top female executives? University of California Berkeley Working Paper. Berkeley: University of California.

74.

Smith

Nina

Smith

Valdemar

Verner

Mette

. 2013. Why are so few females promoted into CEO and vice president positions? Danish empirical evidence, 1997–2007. ILR Review 66(2): 380–408.

75.

Sonnenfeld

Jeffrey A.

Peiperl

Maury A

. 1988. Staffing policy as a strategic response: A typology of career systems. Academy of Management Review 13(4): 588–600.

76.

Sonnenfeld

Jeffrey A.

Peiperl

Maury A.

Kotter

John P

. 1988. Strategic determinants of managerial labor markets: A career systems view. Human Resource Management 27(4): 369–88.

77.

Sutton

John R.

Dobbin

Frank

Meyer

John W.

Scott

W. Richard

. 1994. The legalization of the workplace. American Journal of Sociology 99(4): 944–71.

78.

Tinsley

Catherine H.

Wade

James B.

Main

Brian G. M.

O’Reilly

Charles A.

III.

2017. Gender diversity on U.S. corporate boards: Are we running in place? ILR Review 70(1): 160–89.