Are Women’s and Men’s Pay Increase Trajectories Different in Nonprofit and For-Profit Human Services Organizations?

Abstract

Using the Survey of Income and Program Participation panel data, this study compares women’s and men’s pay increase trajectories and patterns of job mobility in the nonprofit and for-profit sectors. As recent studies suggested that industry-specific rather than economy-wide analysis is more appropriate in nonprofit/for-profit comparison, this study only focuses on the human services field. The results indicate that: (1) there was a selection in workers’ choices regarding staying or changing sector of employment; (2) nonprofit workers who chose to move to the for-profit sector tended to be those worse off in the nonprofit sector, whereas for-profit workers who chose to move to the nonprofit sector tended to be those better off in the for-profit sector, and both of the mover groups gained by moving; (3) in both nonprofits and for-profits, men enjoyed a steeper pay increase curve by occupying more managerial positions, having higher levels of education, and working longer hours; (4) on average, there was no statistically significant difference in pay increase trajectories between workers who chose to stay in their sector of employment. Consistent with previous studies, the regression analyses further support the argument that, within human services, there is not necessarily a sectoral difference in the gender pay gap—in other words, nonprofits are not necessarily more equitable than for-profits.

1. Introduction

As the U.S. economy becomes increasingly service-oriented, the country relies on both nonprofit and for-profit organizations to provide human services. The human services field serves the most vulnerable populations, including children, women, sexual and gender minorities, older adults, and people who are experiencing poverty and/or houselessness (Hasenfeld 2010; Salamon 2012). Existing surveys show that the human services workforce is highly gendered, with about 80% of the workers being women, and that human services jobs are generally low-paying (Faulk et al. 2013; Leete 2006; Zhao 2020). These patterns hold for both nonprofit and for-profit human services organizations.

The co-existence of nonprofit and for-profit organizations in the economy and in the human services field spurred an interest in comparing the pay and gender-based pay disparities between the two sectors (i.e., nonprofit, and for-profit). Based on economy-wide data (i.e., including all industries in the economy), studies have shown that the gender pay gap was smaller for nonprofits than for the for-profit sector (Hirsch et al. 2018; Leete 2000, 2006; Preston 1990, 1994; Preston and Sacks 2010). This implies that the nonprofit sector might have been a more equitable employer for female workers. However, recent studies have suggested that an industry-specific analysis is more appropriate (Faulk et al. 2013; Zhao 2020), considering that nonprofit organizations are concentrated in the human services field (which includes about 10 industries such as education, health services, social services, and residential care), whereas the for-profit sector is primarily made up of traditional business organizations in hundreds of business industries—with only a small proportion of organizations in human services (Ruggles et al. 2022). This creates an apple-to-orange comparison problem and could generate misleading results.

To avoid this problem, some more recent studies conducted industry-specific analyses. For example, using multiple-level modeling, Faulk et al. (2013) examined the industry effect on the nonprofit/for-profit gender wage gap comparison. They argued that the smaller gender pay gap observed in the nonprofit sector seems to be merely a side effect of an overall nonprofit pay compression; in other words, it is because the overall pay level of nonprofit workers is lower which leaves less room for gender pay disparity. Similarly, King and Lewis’ (2017) analyses of registered nurses found little evidence to support the hypothesis that nonprofit hospitals have a smaller gender pay gap than for-profit hospitals. Additionally, Zhao (2020) did not find a significant nonprofit/for-profit difference in the gender pay gap in their analysis of 10 human services industries that house the majority of nonprofit organizations. These studies suggest that, when comparing nonprofit vs. for-profit wage patterns, industry-specific analyses are more appropriate.

Furthermore, existing analyses are mostly based on cross-sectional data rather than longitudinal data; even though many studies have examined gender pay disparity among nonprofit executives (Finley et al. 2022; Lee and Lee 2021), no known empirical study has examined women’s and men’s pay increase trajectories over time in the two sectors—economy-wide or human services specific. We do not know if, once a worker is employed in the nonprofit sector, how their pay increase trajectory differs from their for-profit counterpart. What also remain unknown are patterns of job mobility, especially when moving between the nonprofit and for-profit sectors, among human services workers. These issues are closely related to employees’ career advancement opportunities (e.g., promotions) and job mobility in a specific sector.

To address this gap in the literature, this empirical study uses the Survey of Income and Program Participation (SIPP) 2008 Public Use Panel data (2008–2013) from the U.S. Census Bureau to track women’s and men’s job mobility and pay increase trajectories in the nonprofit and for-profit human services organizations. Because workers change jobs as well as sectors of employment (e.g., nonprofit, for-profit), this study also analyzes pay changes for four groups of workers: nonprofit (NP) stayers, workers that stayed in NP throughout the period of study; for-profit (FP) stayers, workers that stayed in FP; NP-FP movers, those who moved to FP from NP; and FP-NP movers, those who moved from NP to FP. It is found that, in both nonprofit and for-profit human services organizations, men were much more likely to occupy managerial positions than women and managers enjoyed a much steeper pay increase trajectory than non-managers. This pattern holds for both nonprofit and for-profit organizations, without a significant sectoral difference in female/male pay increase trajectories. This finding is consistent with previous industry-specific analyses (King and Lewis 2017; Zhao 2020) that, in the human services field where nonprofits and for-profits have a more balanced market share, there is no significant behavioral difference in terms of gender pay disparity between the two sectors. It thus further challenges the myth that the nonprofit sector is more gender equitable than the for-profit sector. The analysis also reveals a different between-sector job mobility pattern which will be detailed in the findings and discussions sections.

The following sections present theories and hypotheses, methods, findings, discussions, and limitations of this study1.

1.1. Theories

The literature points to individual determinants of pay levels at a certain point of employment, such as demographic characteristics like gender, education, and work experiences (as reviewed in Blau and Kahn 2017). Additional factors that explain gender differences in earnings growth include structural barriers to women’s career advancement. Extensive studies have shown that occupational segregation by gender and organizational situations that limit women’s opportunities to advance to managerial and leadership positions explain a large proportion of the persistent gender pay gap in general (Acker 1992; Blau et al. 2009; Blau and Kahn 2017; Blau and Winkler 2021). The remainder of this section reviews these theories in detail, with a focus on female-intensive workplaces.

1.2. Gender and Career Advancement in Female-Intensive Fields

Human services are known for their “female-dominated” workforce (Faulk et al. 2013; Preston 1990; Zhao 2020). Nevertheless, women’s numerical overrepresentation in the field does not necessarily translate into having power or control (McPhail 2004; Odendahl and O’Neill 1994). In fact, men have dominated leadership positions in the nonprofit sector (Hallock 2002); when women did have an opportunity to lead an organization, they were much more likely to be in smaller organizations, which tend to pay managers much less compared to bigger organizations (Lee and Lee 2021). This pattern is found to be a significant contributor to the gender pay gap in nonprofit executive compensation. In critique of the traditional term, “female-dominated”, McPhail (2004) termed this kind of workforce situation as “female-majority, male-dominated”.

A number of theories, including task differentiation, sex-based ascription, male tokenism (which results in a “glass escalator” effect for men), the effect of resource dependency on the “old boy club”, homophily—the labor supply pool’s sex composition, have explained women’s (lack of) opportunities for management positions in female-intensive fields.

1.2.1. Task Differentiation and Gender Segregation in the Workplace

According to Reskin (1988, p. 58), the major cause of gender wage disparity is “the segregation of women and men into different kinds of work.” This separation facilitates unequal value for men’s work when compared to women’s work and implies that this differential treatment is appropriate (Blau and Kahn 2017; Blau and Winkler 2021; Padavic and Reskin 2002). Women tend to be overrepresented in lower-paying jobs, including childcare, teaching, social work, and other helping professions. Sociologists have suggested that the low pay of female-intensive jobs results from a devaluation of caring work (Cohen and Huffman 2003; Reskin 1988). Because these jobs are often performed by women and the skills and talents required are often assumed to be natural attributes of women, it is assumed that they need no special compensation (Odendahl 1996; Odendahl and O’Neill 1994). Folbre and Smith (2017) termed this deficit as the “care wage penalty”. Preston (1990) pointed out that the relatively low wages and high level of representation of women in the nonprofit sector reflect an “occupational crowding” effect. The presence of institutional barriers and prejudices against women in male-dominated occupations could crowd women into female-intensive occupations, such as teaching, caretaking, and clerical jobs.

Task differentiation and occupational gender segregation is prevalent in human services organizations. Frontline service workers tend to be women, whereas at the managerial level, positions—especially executive positions—tend to be held by men (Dale 2017; Hallock 2002; Lee and Lee 2021; McPhail 2004; Odendahl 1996). This sharp distinction between women’s jobs and men’s jobs in human services represents the “sex labeling” and “sex stereotypes” of occupations and jobs (Padavic and Reskin 2002). By equating management with work requiring masculinity and “tough-mindedness,” male managers legitimize their role as leaders and differentiate themselves from the low-paying helping jobs that women have (Reskin and McBrier 2000).

1.2.2. Sex-Based Ascription in Managerial Employment

Ascription refers to the situation in which a status, position, or opportunity is distributed in part based on an ascribed characteristic (Kemper 1974). Historically, sex-based ascription has been the default in managerial employment and was practiced in a covert, even unconscious, way (Acker 1992; Kanter 1977; Powell 2018). For example, sex stereotypes are pervasive in the workplace and job assignment decisions are influenced by jobs’ sex labels; employment decision-makers often act on their biases and use sex as a proxy for productivity or employment cost. This ascription propensity favors men for management positions. To minimize selection costs, organizations may resist the creation of new and formal structures to select managerial personnel, and resort to gender ascription, instead. According to Salanick and Pfeffer (1978), when it is not clear who the best choice is, decision-makers tend to minimize risk by opting for persons who resemble themselves. Consequently, when women are under-represented in management, employers are likely to be more familiar with male managers and hence see a male candidate as less risky. Despite the progress in abandoning the ascription in personnel practice, men’s continued monopoly on management jobs indicates that organizational inertia exists, and sex-based ascription still influences women’s opportunities for advancement to managerial positions.

Another factor that might facilitate sex-based ascription is group power and in-group preference. Powerful groups have the tendency to institutionalize their privileges (Acker 1989; Blau and Winkler 2021). These patriarchal impulses could lead male organizational leaders to prohibit women from occupying positions that lead to their authority over men. In addition, people’s unconscious propensity to favor members of their own group can have the same ascriptive effect. This ascription is most likely to exist in top executive-level positions.

1.2.3. Tokenism and Men’s “Glass Escalator” Effect in Female-Intensive Jobs

Women often experience great difficulty entering male-dominated occupations, where they often face prevalent discrimination in the hiring process, and from supervisors, colleagues, and even outsiders (as reviewed in Blau and Winkler 2021). However, studies on men’s experiences in “female-dominated” occupations reveal the opposite: there is a preference for hiring men in these occupations. Men at higher levels receive preferential treatment, which closes off advancement opportunities for women with the same qualifications (Brandford and Brandford-Stevenson 2021; Hultin 2003; Simpson 2004; Williams 1992; Wingfield 2009). Williams (1992) termed this the “glass escalator” effect that men experience in traditionally female professions. The researchers pointed out that the tokenism effect and men’s master status are primary causes for this differential treatment of men in traditionally female occupations.

According to Kanter (1977), tokenism refers to a situation in which persons of one social type are in an extreme minority—less than 15 percent of the group. This proportional rarity significantly affects group interactions. For example, it leads to a heightened visibility of the token group members; their behavior is noticed more than that of the numerically dominant group members, and therefore, they receive more attention (Kanter 1977). Researchers studying men’s token experience in female-intensive occupations noticed that the effect of tokenism may differ for female and male tokens. The social status of the token’s group (e.g., male vs. female)—not their numerical rarity—is the crucial factor that determines whether the token encounters a “glass ceiling” or a “glass escalator” (Heikes 1991; Simpson 2004; Williams 1992). Therefore, men take the gender privilege with them when they enter predominately female occupations, and that translates into an advantage. This effect is especially salient for white men, compared to men of color (Brandford and Brandford-Stevenson 2021; Price-Glynn and Rakovski 2012; Wingfield 2009).

1.2.4. The Effect of Resource-Dependency and the “Old Boy Club”

According to resource-dependency theory, organizations need to align themselves with their environment (Aldrich and Pfeffer 1976; Pfeffer and Salancik 1978), and therefore, their staffing, especially the top executive appointments, tend to resemble the structure of their organizational context (Tharenou 1997). As government funding and private or corporate donations make up a substantial proportion of the revenue of human services organizations, and those who control resources in these institutions tend to be male (Stelter 2002), theoretically, human services organizations thus would prefer males over females for top executive positions to fit into the “old boy club”.

1.2.5. Homophily: The Gender Composition of Labor Supply

The gender composition of the pools from which organizations recruit managers reflects the readily available labor supply. Therefore, it is likely that the gender makeup of the nonmanagerial workforce affects that of its managerial workers (Reskin and McBrier 2000). Additionally, organizations have the tendency to match subordinates and supervisors of the same sex, which also could make the gender composition of the managerial workforce mirror that of the nonmanagerial workforce (Blum et al. 1994; Goodman et al. 2003; Kanter 1977). Given that human services industries are numerically dominated by female workers (Zhao 2020), it is reasonable to assume that the demand for female managers would be greater than that of male-dominated occupations. In other words, women would have more opportunities to work in managerial positions in human services organizations.

Hypotheses: Gender Differences in Earnings Growth in Human Services

In summation, women’s career advancement opportunities in a sector (i.e., nonprofit, or for-profit) of a specific human services industry are determined by the combination of the above forces. Given that all factors, except the sex composition of labor supply, favor men over women for managerial positions, men are expected to have better chances to advance to managerial positions, especially top executive positions, and thus have a steeper earnings growth curve. Nevertheless, as both the nonprofit and for-profit human services are highly gendered with a female-majority workforce (Ruggles et al. 2022; Zhao 2020), the effect of task differentiation, sex-based ascription, male tokenism, resource-dependency, and gender composition of the labor pool should be the same in each sector. Therefore, there should not be a significant nonprofit/for-profit difference in men’s and women’s career advancement opportunities and hence earnings growth within human services.

Hypothesis 1 (H1).

In the human services field, due to better career advancement opportunities, on average, men experience a steeper pay increase trajectory than women in both the nonprofit and for-profit sectors.

Hypothesis 2 (H2).

In the human services field, pay increase trajectories for nonprofit women and for-profit women are similar. This pattern also applies to men. In other words, the women/men pay increase trajectory difference in the nonprofit sector is similar to that of the for-profit sector.

2. Method

2.1. Data

The Survey of Income and Program Participation (SIPP) 2008 Public Use Panel data from the U.S. Census Bureau was used to answer the research questions. The SIPP survey design is a continuous series of national panels, with sample sizes ranging from approximately 14,000 to 52,000 interviewed households. The 2008 panel contains 16 waves of data for five years from 2008 through 2013 and has a large sample size of nonprofit workers. Specifically, respondents were surveyed every four months beginning in December 2008 until December 2013. Among the available SIPP panel datasets, the 2008 Panel covers the longest span of time. SIPP’s labor force section collects rich information about an individual’s work history from the beginning of the reference year through the interview months. Specifically, the data contains individual workers’ weeks of employment, the number of hours worked per week, and amount and type(s) of earnings, etc. It also collects business characteristics of the respondent’s job and employer, including industry, occupation, union status, number of employees (indicating the size of an organization), and incorporation status (e.g., nonprofit, for-profit). In addition to the advantage of being longitudinal, another advantage of SIPP data over the cross-sectional data used previously (Faulk et al. 2013; Zhao 2020) is that it contains employer characteristics. These are also key determinants of a worker’s wage and pay increase potential (Kalleberg and Van Buren 1996; Lee and Lee 2021; Villemez and Bridges 1988). For example, union status, whether a worker is unionized or not, is an important determinant of pay and potential pay disparity (Blau and Winkler 2021; Gould and McNicholas 2017).

This study uses the wave 1 and wave 13 data of SIPP 2008 panel, which covers a four-year time span. There are three reasons for choosing these two time points: (1) the attrition rate after wave 13 was too high (over 60%), whereas 64% of the wave 1 respondents remained in the panel through wave 13; (2) from wave 1 to wave 13, the panel lasted for four years, which is a sufficiently long period of time to track workers’ pay increase trajectories; (3) focusing on two points in time simplifies the analysis, so it does not need to account for change in employment status or sector of employment within the time span. Nevertheless, given that most American workers stay in their jobs for a long period of time (as reviewed in Borjas 2009), two time points should be able to capture most people’s job changes and should not introduce much error. This study restricts the sample to human services workers who worked in the nonprofit and for-profit sectors and were employed for all weeks in the reference month. As for industry restriction, this study adopts the Hasenfeld (1983) definition of human services—its primary function is to protect, maintain, or enhance the personal well-being of individuals and, following Zhao’s (2020) analysis on nonprofit/for-profit wage comparison, includes industries with at least 10% nonprofit workers and 10% for-profit workers. As a result, a similar set of human services industries to what were included in Zhao’s (2020) analysis is included (e.g., education, hospitals, health services, social services).

In tracking workers’ pay trajectories in two sectors of employment over a period, an inevitable issue is that some workers switch sectors. Therefore, the sample categorizes four groups of workers: (1) NP Stayers, who were employed in the nonprofit sector in both waves 1 and 13 (the sample size of this group is 752); (2) FP Stayers, who were employed in the for-profit sector in both wave 1 and 13 (sample size: 1434); (3) NP-FP Movers, who were employed in nonprofits in wave 1 but were employed in for-profits in wave 13 (sample size: 317); (4) FP-NP Movers, who moved from for-profits in wave 1 to nonprofits in wave 13 (sample size: 148). Longitudinal weights were used to account for the effect of attrition on the results. Because the Census Bureau had edited and imputed the Public Use data, missing data is not an issue, so the data was not re-imputed for this analysis.

2.2. Analytical Method

Because the Movers and the Stayers are inherently different from each other, this study analyzes the Stayer group and the Mover group separately.

First, for the NP and FP Stayers, ordinary least squares regression with robust standard errors is used to examine the difference of workers’ last-point hourly wage controlling the start-point wage and other baseline characteristics. Baseline individual characteristics that were controlled include full-time status, hours worked per week, occupation, industry, managerial position status, and educational level, as well as employer-side pay determinants such as union status, size of employer, and if the worker received employer-provided health insurance. Whether the worker has a child (or children) under 18 years old in the household is also controlled. These predictor variables are chosen to be included in reference to standardized gender pay gap wage analysis (Blau and Kahn 2017) and previous studies on nonprofit/for-profit differences in the gender pay gap (Faulk et al. 2013; Zhao 2020). This study controlled the start-point worker characteristics rather than that of the last point because changes in work characteristics during the time span are also influenced by the sector of employment. By controlling the start-point characteristics, the analysis captures the impacts of the sector of employment on these changes. For example, workers in one sector might be more likely to work part-time than in the other sector. Workers in different sectors also have different possibilities to advance to managerial positions, which are captured by the managerial position (being a manager or not) information in the data—SIPP data include occupational codes of workers which detail multiple managerial position titles. Both the SIPP data and the American Community Survey data (Ruggles et al. 2022) indicate that the percentage of nonprofit workers with a managerial title is much higher than that of for-profit. Including the managerial position variable helps adjust the estimate of the effect of one’s level of position on their wage.

Second, this study compares the baseline characteristics and four-year wage changes of the four groups: NP Stayers, FP Stayers, NP-FP Movers, and FP-NP Movers. This is to examine if there is a selection between the Stayer and Mover groups and how the selection functions.

2.3. Variables and Models

The OLS regression used the below model specification:

Ln (wage_ij−13) = a + β₁ *X_ij−1 + β₂ * female_i + β₃ * nonprofit_i + β₄ * female × nonprofit_i
+ β₅ * Ln (wage_ij−1) + ε_i

Ln (wage_ij−13) is the natural logarithm of a worker’s hourly wage at wave 13 (i.e., year four), and Ln (wage_ij−1) is the logarithm of their wage at wave 1 (i.e., year one). Hourly wage was calculated as monthly wage earning of the job divided by weeks worked and usual hours worked per week on this job. Coefficient β₂, β₃, and β₄ are the coefficients of interest. β₂ measures if there is a difference in four-year wage change between female FP Stayers and male FP Stayers; β₃ captures if the wage change differs between male FP Stayers and male NP Stayers; β₄ measures if the wage change differs between female NP Stayers and female FP Stayers. X_ij−1 is a vector of controls of baseline worker and employer characteristics, including a dummy for part-time work status (defined as working less than 35 h a week), being a manager or not, having children under 18 years old in the household or not, and if the worker had employer-provided health insurances, hours worked per week, estimated years of experience, estimated years of experience squared, and dummies for race, categorical educational level, size of the organization (measured by the number of employees), union status, detailed occupation codes, and detailed industry codes. The operationalization of these variables follows previous comparable analyses including Faulk et al. (2013), King and Lewis (2017), and Zhao (2020).

3. Findings

There are only negligible differences between the results of weighted analysis and those of unweighted analysis: the patterns generated by both analyses are the same. Therefore, only the results of the unweighted analyses are presented in the text (results of the weighted regression analyses are also presented in Table 5). As a reminder, the results presented here are all for human services workers only.

3.1. Worker Characteristics of the Stayers

As shown in

Table 1. Baseline worker characteristics of NP Stayers and FP Stayers (Part 1).

Worker Characteristics	Mean
	NP	FP	All
	n = 752	n = 1434	n = 2186
Baseline Hourly Wage	29.97	24.12	26.13
Female	26.82	21.25	23.16
Male	39.79	33.20	35.49
Female	0.76 1	0.76 1	0.76
Nonprofit Workers			0.34
Having Child in Household	0.35	0.41	0.39
Female	0.35 1	0.42 1	0.40
Male	0.33 1	0.39 1	0.36
Hours Worked Per Week	38.07	36.57	37.09
Female	36.77	35.70	36.07
Male	42.09	39.35	40.30
Part-Time Status	0.19 1	0.22 1	0.21
Female	0.23	0.24	0.24
Male	0.10	0.15	0.13
Experience (year)	25.01	22.64	23.46
Female	24.66 1	22.76 1	23.41 1
Male	26.09 1	22.26 1	23.59 1
Age (year)	45.22	41.89	43.03
Female	44.64	41.78 1	42.76 2
Male	47.03	42.23 1	43.90 2
Education
High School or Less	0.13	0.22	0.19
Some College	0.35	0.44	0.41
Bachelor’s Degree	0.29	0.21	0.24
Master’s Degree	0.14	0.09	0.11
Professional Degree	0.03	0.02	0.03
Doctoral Degree	0.05	0.02	0.03
Education–Female
High School or Less	0.14	0.23	0.20
Some College	0.38	0.47	0.44
Bachelor’s Degree	0.30	0.20	0.23
Master’s Degree	0.14	0.07	0.09
Professional Degree	0.03	0.01	0.02
Doctoral Degree	0.02	0.01	0.01
Education–Male
High School or Less	0.13	0.19	0.17
Some College	0.24	0.35	0.31
Bachelor’s Degree	0.27	0.22	0.24
Master’s Degree	0.17	0.12	0.14
Professional Degree	0.06	0.05	0.05
Doctoral Degree	0.13	0.07	0.09
Race
White	0.83	0.76	0.79
Black	0.11	0.16	0.14
Asian	0.02	0.03	0.02
Others	0.04	0.05	0.05
Race–Female
White	0.82	0.76	0.78
Black	0.11	0.17	0.15
Asian	0.03	0.03	0.03
Others	0.03	0.04	0.04
Race–Male
White	0.85	0.77	0.80
Black	0.08	0.13	0.11
Asian	0.01	0.02	0.02
Others	0.06	0.07	0.07

NP = Nonprofit; FP = For-profit. Data used: The Survey of Income and Program Participation (SIPP) 2008 Public Use Panel data from the U.S. Census Bureau. Sample restriction: Human services workers who were employed for all weeks in the reference month and stayed in their respective sector of employment (i.e., nonprofit, and for-profit) throughout the panel. Cells with a superscripted No. 1 means that the group differences in numbers are not statistically significant. Cells with a superscripted No. 2 means that the group differences in numbers are only statistically significant at the 0.1 confidence level. All other numbers are statistically significant at the 0.05 confidence level.

, the overall worker characteristics of human services workers, as well as the nonprofit/for-profit differences in worker characteristics shown by the SIPP data, are consistent with the results of Zhao’s (2020) analysis using the Census and the American Community Survey data. These datasets consistently show that the human services field had a highly gendered workforce, with female workers making up the majority of employees for all industries and in both the nonprofit and for-profit sectors. In other words, what was gendered is the human services field itself rather than a specific sector. The table also indicates that, in human services, the nonprofit and for-profit workforces were comparable but with slight differences. Specifically, a higher percentage of the nonprofit human services workforce was white (83% vs. 76%), and nonprofit workers were more educated, more experienced, and had a higher baseline hourly wage (29.97 vs. 24.12).

Additional descriptive analyses (see

Table 2. Baseline worker characteristics of NP Stayers and FP Stayers (Part 2).

Worker Characteristics	Mean
	NP	FP	All
	n = 752	n = 1434	n = 2186
Managers	0.13	0.09	0.10
Female	0.11	0.07	0.08
Male	0.20	0.14	0.14
Employer-Provided Health Insurances	0.88	0.77	0.81
Female	0.88 1	0.76 1	0.80 1
Male	0.87 1	0.80 1	0.83 1
Union Members	0.08 1	0.08 1	0.08
Female	0.08 1	0.07 1	0.07
Male	0.06 1	0.10 1	0.09
Size of Employer
Under 25 employees	0.19	0.27	0.24
25 to 99 employees	0.18	0.22	0.21
100+ employees	0.63	0.51	0.55
4-Year Wage Diff for Full-Time Workers	2.98 2	1.24 2	1.88
Female	2.71 1	1.14 1	1.71
Male	3.69 1	1.50 1	2.30
4-Year % Change of Wage for Full-Time Workers	0.22	0.19	0.20

Data used and sample restrictions are the same as Table 1.

) of the Stayers show that 88% of NP Stayers had employer-provided health insurance, whereas only 77% of FP Stayers were provided this benefit—even though there is no statistically significant gender difference in this benefit within either sector. A higher percentage of nonprofit human services workers were employed in larger organizations with over 100 employees (63% for nonprofits vs. 51% for for-profits). This means that nonprofit human services organizations tended to be larger than their for-profit counterparts. About 8% of human services workers of this sample were union members and there is no statistically significant difference in union status between nonprofit and for-profit human services workers.

Thirteen percent of nonprofit human services workers were in managerial positions, whereas only 9% of for-profit human services workers were in these kinds of positions. This indicates that nonprofit human services organizations had a different organizational structure with more employees with the title of “manager”. With respect to gender differences, in nonprofit human services organizations, 11% of female workers were managers, whereas 20% of male workers were managers; in for-profit human services organizations, 7% of female workers were managers, whereas 14% of male workers were managers. These differences are statistically significant at the 0.05 confidence level. Given that only 24% of nonprofit and for-profit workers were male (as shown in Table 1), the results indicate that men were much more likely to work in managerial positions than women in both sectors. This finding is consistent with the theories and previous empirical studies. Furthermore, in both nonprofits and for-profits, a significantly higher percentage of men (53%) had a bachelor’s degree or above than women (36%), and men worked longer hours (about 40) than women (about 36).

As shown in Table 1, The average start-point wage of NP Stayers was USD 29.97, compared to 24.12 for FP Stayers. Specifically, the average start-point wage of nonprofit men was 39.79; that of nonprofit women was 26.82; that of for-profit men was 33.2; and that of for-profit women was 21.25. These differences are all statistically significant at the 0.05 confidence level. The higher wage of nonprofit human services workers is likely due to their higher educational attainment, the greater number of years of experience, and the fact that they tended to work in larger organizations. For full-time workers, the average four-year difference in the hourly wage of NP Stayers was USD 2.98, compared to 1.24 for FP Stayers. According to the results presented in Table 2, the average percentage of the four-year difference in the hourly wage of NP Stayers was 22%, compared to 19% of FP Stayers. This means that, on average, nonprofit workers received a greater degree of wage increase than for-profit workers. Among the NP Stayers, women’s wages increased by about 22% whereas men’s wages increased by about 21%. For FP Stayers, women’s wages increased only by about 15%, but men’s wages increased by about 28%. This shows that, on average, the raw gender gap in wage increase was smaller in nonprofit than in for-profit human services organizations.

3.2. Worker Characteristics of the Movers

3.2.1. NP-FP Movers

To examine the possible selection in the moving behaviors of a proportion of human services workers, this section summarizes the baseline characteristics and wage changes of workers who moved between the nonprofit and for-profit sectors (see

Table 3. Baseline worker characteristics of the four groups.

	Mean
	NP Stayers	NP-FP Movers	FP Stayers	FP-NP Movers
	n = 752	n = 317	n = 1434	n = 148
Female	0.76	0.72	0.76	0.78
Hours Worked Per Week	38.07	36.51	36.57	36.78
Part-time	0.19	0.25	0.22	0.27
Experience (year)	25.01	19.39	22.64	21.92
Age (year)	45.22	39.05	41.89	41.33
Education
High School or Less	0.13	0.19	0.22	0.20
Some College	0.35	0.38	0.44	0.43
Bachelor’s Degree	0.29	0.26	0.21	0.22
Master’s Degree	0.14	0.09	0.09	0.05
Professional Degree	0.03	0.03	0.02	0.06
Doctoral Degree	0.05	0.05	0.02	0.03
Race
White	0.83	0.76	0.76	0.76
Black	0.11	0.15	0.16	0.18
Asian	0.02	0.04	0.03	0.01
Others	0.04	0.05	0.05	0.05
Managers	0.13	0.09	0.09	0.14
Female	0.11	0.08	0.07	0.09
Male	0.20	0.09	0.14	0.31
Employer-Provided Health Insurances	0.88	0.80	0.77	0.76
Union Members	0.08	0.07	0.08	0.09
Size of Employer
Under 25 employees	0.19	0.20	0.27	0.22
25 to 99 employees	0.18	0.21	0.22	0.16
100+ employees	0.63	0.60	0.51	0.62

Data used: The Survey of Income and Program Participation (SIPP) 2008 Public Use Panel data from the U.S. Census Bureau. Sample restriction: human services workers who worked in the nonprofit and for-profit sectors (including those who stayed in the same sector and those who switched sectors) and were employed for all weeks in the reference month.

).

Compared to those who stayed in the nonprofit sector, NP-FP Movers tended to be male, racial minorities, and part-time workers working in smaller organizations and non-managerial positions, and they tended to be less qualified and experienced and have lower levels of compensation. Specifically, 52% percent of the NP Stayers held a bachelor’s degree or higher, whereas only 43% of the workers who moved to for-profits had the same level of education. On average, NP Stayers had about 25 years of work experience, whereas the Movers had about 19 years of experience. The average start-point wage of NP-FP Movers was USD 23.3, which is significantly less than that of the NP Stayers (29.97). This pattern held for both women and men. All these differences are statistically significant at the 0.05 confidence level.

Furthermore, nonprofit human services organizations had a higher turnover at lower levels of the organizations but had less of a problem with staff retention at higher occupational levels. Specifically, the Mover–Stayer ratio of nonprofits is 317:752, whereas that of for-profits is 148:1434. In other words, although the nonprofit sector had a higher rate of worker loss, most of the turnover occurred in the lower levels of the organizations and affected workers with fewer qualifications.

3.2.2. FP-NP Movers

The profile of for-profit workers that moved to the nonprofit sector differs sharply. Compared to people who stayed in the for-profit sector, FP-NP Movers had a relatively higher level of educational attainment, and they tended to be employed in larger organizations and managerial positions. Specifically, 62% of the Movers, compared to 51% of the Stayers, were employed in organizations with over 100 employees. A total of 14% of the Movers were in managerial positions, whereas only 9% of the Stayers were at that occupational level. On the other hand, there were few differences in age, work experience, or racial composition between the Movers and the Stayers. In addition, the Movers were slightly less likely to have employer-provided health insurance. There was no statistically significant difference in the percentage of union members across all four groups—the proportion of union members was all very low across the board. In summary, the FP-NP Movers tended to be those who were relatively better off in the for-profit sector with higher wages, higher educational attainment, and higher occupational levels at larger organizations.

3.2.3. Wage Changes of the Four Groups

As shown in

Table 4. Baseline wages and wage changes of the four groups.

	Mean (USD)
	NP Stayers	NP-FP Movers	FP Stayers	FP-NP Movers
	n = 752	n = 317	n = 1434	n = 148
Baseline Hourly Wage	29.97	23.30	24.12	24.76
Female	26.82	21.94	21.25	20.07
Male	39.79	26.83	33.20	41.78
4-Year Wage Difference for Full-TimeWorkers	2.98	2.66	1.24	3.95
Female	2.71	2.44	1.14	1.85
Male	3.69	3.27	1.50	11.47
4-Year % Change of Wage for Full-TimeWorkers	0.10	0.11	0.05	0.16
Female	0.10	0.11	0.05	0.09
Male	0.09	0.12	0.05	0.27

Data used: The Survey of Income and Program Participation (SIPP) 2008 Public Use Panel data from the U.S. Census Bureau. Sample restriction: human services workers who worked in the nonprofit and for-profit sectors (including those who stayed in the same sector and those who switched sectors) and were employed for all weeks in the reference month.

, comparing the wage levels of the four groups, the average baseline wage was the lowest for NP-FP Movers, the highest for NP Stayers, and in the middle for FP-NP Movers and FP Stayers. With respect to the average four-year wage difference for full-time workers, FP-NP Movers had the highest wage increase (USD 3.95 increase and 16% percent increase), whereas the FP Stayers had the lowest wage increase (USD 1.24 increase and 5% percent increase). On average, the wage of male FP-NP Movers increased by USD 11.47 (27% percentage change), which is significantly higher than male NP Stayers (3.69, 9%), male NP-FP Movers (3.27, 12%), and male FP Stayers (1.50, 5%). As for women, on average, both the FP-NP Movers and the NP-FP Movers gained higher wages (9% and 11%, respectively) than the Stayer groups—NP Stayers’ wage increased 10% whereas FP Stayers’ increased only 5%. To summarize, regardless of gender or the original sector of employment, those who moved to another sector all had higher percentage wage increases compared to those who stayed.

The baseline wages and wage changes of the four groups prove the existence of selection in moving or staying behaviors. In general, nonprofit workers with high wages chose to stay, whereas those with unsatisfactory wages moved to the for-profit sector. Because the NP-FP Movers had the lowest baseline wages and the FP Stayers had higher baseline wages, it seems that for-profits paid this group of workers better than nonprofits. Furthermore, the NP Stayers and the FP-NP Movers were better off at the baseline, and they also enjoyed a high(er) wage increase over the four-year time span. The FP-NP Movers were the biggest winners with an average of a 16% increase in wages within four years.

3.3. Regression Results

Table 5. Regression results of the weighted and unweighted analyses.

	Model 1	Model 2	Model 3	Model 4
VARIABLES	Unweighted Analysis		Weighted Analysis
Female	−0.035	−0.037	−0.037	−0.040
	−0.041	−0.040	−0.041	−0.039
Nonprofit	0.026	0.023	0.046	0.040
	−0.067	−0.066	−0.064	−0.062
Nonprofit*Female	−0.023	−0.024	−0.044	−0.043
	−0.071	−0.070	−0.064	−0.064
Managers	0.379 **	0.375 **	0.506 ***	0.500 ***
	−0.159	−0.159	−0.190	−0.191
Having Children in Household	−0.008	−0.008	−0.001	−0.001
	−0.030	−0.030	−0.029	−0.029
Part-Time Status	0.119 ***	0.122 ***	0.122 ***	0.126 ***
	−0.041	−0.042	−0.040	−0.040
Education
High School or less
Some College	0.113 **	0.115 **	0.146 ***	0.147 ***
	−0.045	−0.045	−0.050	−0.050
Bachelor’s degree	0.304 ***	0.304 ***	0.354 ***	0.354 ***
	−0.060	−0.060	−0.062	−0.062
Master’s degree	0.365 ***	0.370 ***	0.388 ***	0.392 ***
	−0.064	−0.063	−0.068	−0.066
Professional degree	0.473 ***	0.471 ***	0.577 ***	0.574 ***
	−0.161	−0.164	−0.184	−0.188
Doctoral Degree	0.267 **	0.271 **	0.305 **	0.304 **
	−0.131	−0.129	−0.149	−0.145
HoursWorked Per Week	0.011 ***	0.011 ***	0.010 ***	0.010 ***
	−0.002	−0.002	−0.002	−0.002
Log HourlyWage at Baseline	0.324 ***	0.317 ***	0.316 ***	0.308 ***
	−0.045	−0.045	−0.038	−0.038
Years of Experience	0.005	0.005	0.006	0.006
	−0.005	−0.005	−0.006	−0.006
Years of Experience-Square	0.000	0.000	0.000	0.000
	0.000	0.000	0.000	0.000
Union Status		0.015		−0.018
		−0.043		−0.042
Size of Employer
Less than 25 employees
25 to 99 employees		0.014		0.010
		−0.034		−0.043
100+ employees		0.041		0.037
		−0.041		−0.051
Employer-Provided Health Insurances		0.049		0.06202 *
		−0.033		−0.033
Constant	1.211 ***	1.217 ***	1.353 ***	1.348 ***
	−0.192	−0.193	−0.235	−0.236
Observations	2107	2107	2107	2107
R-squared	0.563	0.564	0.580	0.581

Robust standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1. Data used: The Survey of Income and Program Participation (SIPP) 2008 Public Use Panel data from the U.S. Census Bureau. Sample restriction: human services workers who were employed for all weeks in the reference month and stayed in their respective sector of employment (i.e., nonprofit, and for-profit) throughout the panel. Analytical model: ordinary least squares regression with robust standard errors examining the difference of workers’ last-point hourly wage controlling the start-point wage and other baseline characteristics. Dependent variable: the natural logarithm of a worker’s hourly wage at the endpoint. The table presents the results of both weighted and unweighted analyses on the pooled data of NP Stayers and FP Stayers. Models 1 and 3 only controlled individual worker characteristics, whereas Models 2 and 4 added employer-side pay determinants.

presents the results of both weighted and unweighted analyses on the pooled data of NP Stayers and FP Stayers. Models 1 and 3 only controlled individual worker characteristics, whereas Models 2 and 4 added employer-side pay determinants. All the covariates were at the baseline time point.

The regression results show that being a female is associated with an about −3% penalty in terms of wage increase, even though the coefficients are not statistically significant. Both the coefficients of nonprofit status and the interaction term of nonprofit status and female are statistically insignificant at the 0.1 confidence level. Therefore, we are 90% confident to say that, in the U.S. human services field, after controlling for observed worker characteristics, worker baseline wages, and employer-size pay determinants, the pay increases for female and male nonprofit and for-profit workers who chose to stay in one sector were statistically indifferent. In other words, although women and men in nonprofits and for-profits started with different wage levels, if they chose to stay in the same sector, with adjustment of the effects of levels of positions and other pay determinants such as educational attainments, their pay increase trajectories would have no significant difference from their FP/NP counterparts within a four-year period; all four groups’ pay increase trajectories are statistically indifferent.

Furthermore, the results show that a 1% increase in the baseline wage is associated with a 0.32% increase in the endpoint wage. Compared to nonmanagers, on average, workers with a managerial title enjoyed a much sharper wage increase (about 38% in the unweighted analyses and 50% in the weighted analyses) across the four-year time span. Similarly, compared to workers with lower educational levels, workers with higher levels of education had much steeper wage increase trajectories. All levels of educational attainment are significantly associated with wage increases, meaning that people with higher levels of education experienced higher wage increases, regardless of the baseline wages. People with professional degrees had the sharpest wage increase: compared to people who did not have a bachelor’s degree, on average, the end-point wage of people with a professional degree was 47% higher, holding other worker characteristics and baseline wage constant. These results are consistent with the labor economics literature (Borjas 2009), which shows that the rewards of education have been growing in the era of skill-biased technical change. Furthermore, the more hours worked per week, the higher the end-point wage was for a worker, holding all other covariates constant. This means that the workplace rewards managerial positions, education, and long work hours.

In summary, Hypothesis 2 (i.e., in the human services field, there is no significant difference in women’s pay increase trajectories between the nonprofit and for-profit sectors) is supported. This hypothesis speaks to the nonprofit/for-profit sectoral differences in pay increase trajectories, and as argued in the literature review and indicated by previous empirical studies, this finding is expected given the similarities in labor force characteristics and the nature of work output of nonprofit and for-profit human services organizations. Hypothesis 1 (i.e., men experience a steeper pay increase trajectory than women in both the nonprofit and for-profit sectors), however, is not directly supported, meaning that there is no statistically significant gender difference in pay growth curve in both sectors after controlling for the effects of main pay determinants. However, it is noteworthy that in both nonprofit and for-profit human services organizations, men were much more likely to have higher levels of education, occupy managerial positions, and work significantly longer hours. Not surprisingly, these characteristics are associated with a much steeper pay increase curve. Therefore, we can say that in both nonprofit and for-profit human services organizations, men did experience a steeper pay increase trajectory due to their greater probability of being in managerial positions, having higher levels of education, and working longer hours in the workplace.

4. Conclusions and Discussion

Using wage and pay determinants information of the U.S. nonprofit and for-profit human services workers from two time points across a four-year time span, and by examining workers’ behaviors in changing or staying in their sector of employment, this study finds that: (1) there was a selection in workers’ choices regarding staying or changing sector of employment; (2) nonprofit workers who chose to move to the for-profit sector tended to be those worse off in the nonprofit sector, whereas for-profit workers who chose to move to the nonprofit sector usually were those better off in the for-profit sector, and both mover groups gained by moving; (3) after controlling for key pay determinants, on average, there was no statistically significant difference in pay increase trajectories between workers who chose to stay in their sector of employment; this pattern holds for both women and men. However, in both nonprofits and for-profits, men did enjoy a steeper pay increase curve by occupying more managerial positions, having higher levels of education, and working longer hours.

The high proportion of male managers in the FP-NP Movers group indicates a need for their skills in the nonprofit sector. This is understandable in the neoliberal era of managerialism (Abramovitz and Zelnick 2015): with the growing commercialization and professionalization of the nonprofit sector, for-profit management experience is increasingly valued. This trend, therefore, opens opportunities for for-profit managers to work in the nonprofit sector. This finding is consistent with the trend that a growing proportion of nonprofit organizations prefer to have managers with business degrees (Hoefer et al. 2013). Future research should examine if for-profit human services pay lower-level employees better than nonprofits and if nonprofits pay better for managerial positions. In addition, the sharp difference in the Mover-to-Stayer ratio between nonprofits (317: 752) and for-profits (148: 1434) may also indicate that it is easier for nonprofit workers to move to for-profit human services organizations, but it is more difficult for for-profit workers to enter nonprofit organizations; or, for-profit workers are less interested in moving to the non-profit sector. One possible explanation is that nonprofit human services organizations emphasize intrinsic motivation (Leete 2000; Zhao 2020), so they are less likely to welcome frontline workers that had worked for for-profits because the latter are less likely to have a track record proving their commitment to public services.

There are several possible explanations for the lack of gender and sectoral differences in pay increase trajectories. First, the female/male difference in pay increase trajectories is absorbed by the effects of managerial positions, educational levels, and working hours. After controlling for their impacts, the direct gender effect is negligible. In other words, it is the fewer opportunities to advance to managerial positions, acquire higher educational attainments, and their difficulty in meeting the demand of the long-hour work culture that hinder women’s pay increase potential. Second, previous analyses (Leete 2006; Faulk et al. 2013; Zhao 2020) on gender and sectoral differences in pay between the nonprofit and for-profit sectors only considered individual-level pay determinants (due to constraints on data availability); after controlling for organizational pay determinants such as the size of the organization, union status, and fringe benefits (e.g., health insurance), it is likely that gender differences in pay increase over a short period of time are negligible. In fact, studies on gender pay disparity among nonprofit executives have found that the size of the organization explains a substantial proportion of the female/male executive pay gap (Finley et al. 2022; Lee and Lee 2021). Third, workers’ end-point wages were mostly determined by their starting-point wages so the variances in end-point wages were primarily explained by the start-point wages, leaving limited variances to show group differences. Fourth, because the Stayers of each sector are a selective group, it is likely that they chose to stay because there would be no gain by moving to the other sector. Therefore, it is reasonable that the results show no sectoral difference in workers’ wage increases. Lastly, it is also possible that nonprofit and for-profit workers had no difference in wage increase trajectories: they started with different wage levels but enjoyed the same rate of wage increase over time. This is consistent with the labor market competition hypothesis (Leete 2006; Preston 1988; Steinberg 1990; Zhao 2020): when workers do similar work (work in the same industry and/or occupation) and have similar qualifications, labor market competition would eliminate pay differences across sectors of employment (i.e., nonprofit, or for-profit).

To summarize, the results have several practical and research implications. First, the job-moving patterns revealed by the longitudinal analysis indicate that a significant proportion of nonprofit human services workers, especially at the lower rank of organizations, chose to leave the sector to join the for-profit sector. Nonprofit human services organizations should pay close attention to staff retention at the frontline level, where the pay is notoriously low. Second, the moving of for-profit managers to nonprofits substantiates the need for the field to reflect on what kind of managers (e.g., training, experiences) better suit the needs of organizational development and client populations, and what kind of leadership pipeline (e.g., hire from outside or promote and train managers within) works better. Third, consistent with Zhao (2020), the regression analyses further support the argument that, within human services, there is not necessarily a sectoral difference in the gender pay gap—in other words, nonprofits are not necessarily more equitable than for-profits. Lastly, the field should pay due attention to women’s disadvantages in the workplace and take necessary organizational measures to achieve gender equity (e.g., opportunities for managerial positions and pay increases) in the workplace.

Limitations

This study has several limitations due to data availability and the complexity of the issue in question—tracking four groups of workers’ labor market behaviors and changes in pay. Specifically, first, the analyses use self-reported data, which is not as precise as administrative data. Second, without detailed job change information, this study cannot differentiate wage increases due to internal job changes (job changes within the organization) from those due to external job changes (change employers). It also cannot disentangle pay increases due to career advancement (i.e., job promotion) from those resulting purely from job mobility (voluntary change of jobs for various reasons or nonvoluntary change of jobs such as being laid off). Third, to simplify an already-complicated analysis, this study chose to only use the information of the start and end points of time rather than more detailed time points throughout the panel. Fourth, due to the small sample size, this study did not conduct multivariate analysis for the Movers group and only provided basic descriptive statistics. This limits this study’s ability to tell a nuanced story of the pay change of this group of workers.