Administration & Society
http://aas.sagepub.com/
Modeling Nonprofit Employment : Why Do So Many Lesbians and
Gay Men Work for Nonprofit Organizations?
Gregory B. Lewis
Administration & Society 2010 42: 720 originally published online 27 August
2010
DOI: 10.1177/0095399710377434
The online version of this article can be found at:
http://aas.sagepub.com/content/42/6/720
Published by:
http://www.sagepublications.com
Additional services and information for Administration & Society can be found at:
Email Alerts: http://aas.sagepub.com/cgi/alerts
Subscriptions: http://aas.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
Citations: http://aas.sagepub.com/content/42/6/720.refs.html
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
377434
AAS
Modeling Nonprofit
Employment: Why Do So
Many Lesbians and Gay
Men Work for Nonprofit
Organizations?
Administration & Society
42(6) 720­–748
© 2010 SAGE Publications
DOI: 10.1177/0095399710377434
http://aas.sagepub.com
Gregory B. Lewis1
Abstract
Why are people with same-sex partners more likely than married people
to work for nonprofit organizations (NPOs)? Analysis of 2000 Census data
suggests that smaller gay–straight pay disparities for men in the nonprofit
sector, occupational choices, and ability to afford nonprofit employment
explain some overrepresentation of partnered gay men but not of partnered
lesbians. Even after controlling for all these factors, people with samesex partners remain more likely than married people to work for NPOs,
suggesting that a strong desire to serve others is an important factor.
Keywords
sector choice, nonprofit employment, sexual orientation, pay disparities
People with same-sex partners are much more likely than married people to
work for nonprofit organizations (NPOs). As Americans generally see charitable institutions as performing more socially and morally responsible work
than private firms and governments, this overrepresentation of lesbians, gay
men, and bisexuals (LGBs) may surprise the majority of Americans who
1
Georgia State University, Atlanta
Corresponding Author:
Gregory B. Lewis, Andrew Young School of Policy Studies, Georgia State University, Atlanta,
GA 30302-3965
Email: glewis@gsu.edu
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
721
Lewis
believe that homosexuals share hardly any of their moral values. Brooks
(2006) suggests that volunteering and giving to NPOs distinguishes between
the charitable and the selfish, and several studies indicate that most nonprofit
employees donate a share of their potential earnings by accepting lower pay
than they could earn working for a for-profit firm. Does the striking concentration of gay men and lesbians in the nonprofit sector indicate a misperception of the benevolence of LGBs or of NPO employees?
Using a 5% sample of the 2000 Census, I first establish that people with
same-sex partners are much more likely to work for NPOs than those with
spouses or opposite-sex partners. I then test two nonaltruistic hypotheses to
explain the disparity. First, LGBs’ overrepresentation in NPOs could be a
side effect of their residential and occupational choices or of other characteristics. If NPOs are concentrated in West Coast or Northeastern cities where
LGBs tend to live or if NPOs disproportionately employ people in occupations where LGBs work, for instance, LGBs could be more likely than others
to work for NPOs without preferring nonprofit employment. Second, partnered gay men and lesbians might have no stronger commitment to service
but be better able to afford to choose the nonpecuniary benefits of the nonprofit sector, because they are more likely to have a well-paid partner to subsidize their endeavors, less likely to have children to support, or face smaller
wage penalties in a less-discriminatory nonprofit sector.
Separate regression analyses by sex, sector, and educational level confirm
earlier findings that most workers make a financial sacrifice by choosing
nonprofit employment and that men with male partners earn less than comparable married men. They also reveal that gay–straight pay disparities for men
are markedly smaller in the nonprofit sector. (Women with female partners
earn about the same as married women in both sectors.) Logit analyses by sex
and educational level confirm that gender, education, occupation, industry,
location, and the size of sectoral pay differences all influence whether one
chooses nonprofit employment, though the analyses do not replicate earlier
findings that racial and ethnic minorities are more likely than comparable
Whites to work for NPOs. Even controlling for all these factors, however,
people with same-sex partners are more likely than comparable married people to work for NPOs, suggesting that altruistic motives help explain the
overrepresentation.
LGBs, Altruism, and Nonprofit Employment
Most Americans disapprove of LGBs. In five surveys since 2000, 52% to
57% say “sexual relations between two adults of the same sex” are “always
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
722
Administration & Society 42(6)
wrong.” In 14 surveys since 2000, between 48% and 58% call homosexual
relations “morally wrong.” This moral disdain might just be for homosexual
sex and not for LGBs as human beings; however, “feeling thermometers”
that ask people to rate homosexuals and other groups on a scale from 0 to
100 based on the warmth of their feelings toward them show that “between
1984 and 2002, the American public moved from feelings that best can be
described as icy (a mean score of 30) to a temperature just a shade below
neutral (46)” (Egan & Sherrill, 2005). When the Washington Post (1998)
asked whether homosexuals “generally share most of your moral and ethical
values, some of your moral and ethical values, or hardly any,” only 6% said
most and another 29% said some; in contrast, 33% chose hardly any and 24%
volunteered none.
This made all the more surprising tentative findings that people with
same-sex partners are more likely than married and heterosexually partnered
people to work for NPOs (Badgett & King, 1997; Lewis, 2006), suggesting
higher levels of benevolence among LGBs. Nonprofit employees are widely
reputed to be more altruistic than those in the for-profit sector (Jeavons, 1992;
Mirvis, 1992; Mirvis & Hackett, 1983)—or at least to have stronger political
or religious commitments to social change (Onyx & Maclean, 1996; RoseAckerman, 1996). The opportunity to serve others is a major attraction of
work in the nonprofit sector, allowing many NPOs to attract volunteer labor
and to pay lower salaries than for-profit firms to employ comparable workers
(Frank, 1992; Preston 1989, 1990; Weisbrod, 1983). Nonprofit employees,
especially managers and professionals, appear to donate part of their labor by
accepting below-market pay (Frank, 1996; Hansman, 1980; Preston, 1989;
Rose-Ackerman, 1996).
NPOs can attract employees despite low wages partly because many people find nonprofit work more socially responsible, meaningful, and personally rewarding than for-profit or government employment (Frank, 1996;
Light, 2002; Mirvis, 1990; Mirvis & Hackett, 1983). As opportunities to help
others or benefit society increase job satisfaction even for people who do not
prioritize helping others in choosing a job (Lewis & Frank, 2002), this may
compensate for lower pay in the nonprofit sector (Preston, 1989, 1990). NPO
employees have higher job satisfaction than similar individuals in the forprofit sector (Benz, 2005). Public administration scholars have devoted much
effort to describing the impact of public service motivation on seeking or
keeping government jobs (Crewson, 1997; Karl & Sutton, 1998; Perry &
Wise, 1990; Rainey, 1982; Wittmer, 1991). Similarly, those who value service to others or prioritize meaningful work over high pay should be more
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
723
Lewis
likely to choose employment in the nonprofit than in the for-profit sector
(Light, 2002).
Several economists dispute the claim of low NPO pay, however, which
would make altruistic explanations for choosing nonprofit employment
unnecessary. Reanalyzing data used by Weisbrod (1983), Goddeeris (1988)
found no pay disadvantage for public-interest lawyers. Using 1990 Census
data, Leete (2001) concluded that the pay disadvantage to working in the
nonprofit sector is disappearing. Analyzing 1994-1998 Current Population
Survey data, Ruhm and Borkowski (2003) found that NPO pay is competitive. Although economists often find that government pays above-market
rates (e.g., Belman & Heywood, 1989; Gyourko & Tracy, 1988; Moulton,
1990; Smith, 1976—a conclusion widely rejected in the public administration community—they do not find that NPOs pay better than the for-profit
sector.
Why Do LGBs Choose Nonprofit Employment?
Before seeking altruistic explanations, we need to consider whether LGBs’
overrepresentation in the nonprofit sector is driven by their residential and
occupational choices or by other personal characteristics rather than by a
desire to work for NPOs. LGBs are much more likely than heterosexuals to
live in urban areas, on the West Coast, and in New England (D. A. Black,
Sanders, & Taylor, 2007; Black, Gates, Taylor, & Sanders, 2000, 2002; Gates
& Ost, 2004). If nonprofit jobs are more urban than for-profit jobs and concentrated in those areas, high numbers of LGBs in NPOs would naturally
result.
LGBs also have different occupational distributions than heterosexuals
(Badgett, 2001; Blandford, 2003). As occupations differ substantially
between the for-profit and nonprofit sectors (Ruhm & Borkowski, 2003),
overlapping distributions could lead to disproportionate numbers of LGBs
working for NPOs. Among other differences, both gay men and lesbians are
more likely than heterosexuals to choose occupations atypical for their gender (Badgett, 2001; Blandford, 2003). Black et al. (2007, p. 65) show that the
average man in a same-sex couple works in an occupation where 47% of the
workers are women; for the average heterosexually coupled man, only 39%
are women. The average woman with a female partner works in an occupation where 55% of her coworkers are women; for women with male partners,
the percentage is 60%.
The nonprofit workforce is predominantly female (Benz, 2005; Mirvis,
1990; Mirvis & Hackett, 1983; Onyx & Maclean, 1996; Preston, 1989),
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
724
Administration & Society 42(6)
largely, according to Preston (1990), because of the sector’s higher concentration of “traditionally female” jobs. Leete (2000, p. 440) notes that the average man in the nonprofit sector works in an occupation that is 79% female,
whereas in the for-profit sector he works in an occupation that is only 41%
female. Gay men’s higher propensity to work in more “female” occupations
could help explain their disproportionate nonprofit employment (though not
that of lesbians).
Ruhm and Borkowski (2003) also note that NPO employees “are concentrated in eight narrowly defined industries” (p. 1006). If LGBs are disproportionately drawn to those industries, they will be more likely to work for
NPOs. As people consider most of these industries to be more socially
responsible than others (Frank, 1996), however, choosing them could demonstrate an altruistic bent even in the absence of nonprofit employment.
Second, LGBs may have no stronger desire to serve others than heterosexuals do, but they may be more able to afford to choose nonpecuniary over
material rewards. Black et al. (2007) argued that the greater difficulty and
expense same-sex couples face in having or adopting children makes them
more likely to be childless, which, in turn, loosens their time and money constraints, especially for couples with two male earners. This can lead to greater
consumption of leisure activities; men in gay male couples work somewhat
fewer hours than men in heterosexual couples and are more likely to live in
expensive, “high-amenity” cities (Black et al., 2002, 2007).
It may also lead to more nonprofit employment. If socially responsible
work is morally rewarding but comes at a financial cost, those with more
financial resources (or fewer constraints) should be more likely to choose
socially responsible work (Frank, 1996). This could help explain why college
graduates are much more likely than high school graduates to work for NPOs
(Mirvis, 1990; Mirvis & Hackett, 1983; Preston, 1989, 1990): College graduates receiving below-market pay from NPOs still earn more than high school
graduates working for private firms. It may also help explain why women are
far more likely than men to work for NPOs: Wives’ pay typically comprises
a smaller share of household income than husbands’ pay, making a pay penalty for the wife less costly to the family. As lesbians are more likely than gay
men to have children and less likely to have a partner with a large paycheck,
however, affordability will explain less of the overrepresentation of lesbians
in NPOs.
Gay men may also be more able to afford nonprofit employment than married men if they face less wage discrimination in that sector. Several studies
find that gay men earn 15% to 30% less than comparably educated and experienced straight men, though lesbians may earn more than comparable straight
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
725
Lewis
women (Allegretto & Arthur, 2001; Badgett, 1995; Barrett, Pollack, & Tilden,
2002; Berg & Lien, 2002; Blandford, 2003; Black, Makar, Taylor, & Sanders,
2003; Carpenter, 2004, 2005; Carpenter & Gates, 2008; Clain & Leppel,
2001; Jepsen, 2007; Klawitter & Flatt, 1998). Experiments show that employers are less likely to offer job interviews to lesbian and gay job applicants
(Crow, Fok, & Hartman, 1998; Hebl, Foster, Mannix, & Dovidio, 2002;
Weichselbaumer, 2003), supporting an attribution of gay men’s lower pay to
discrimination.
Because NPOs appear to discriminate less against women and minorities
than for-profit firms do, they may discriminate less against LGBs as well.
Preston (1990) attributed much of women’s concentration in the nonprofit
sector to smaller gender pay disparities—the cost to working in the sector is
lower for women than men. Leete (2000) found much lower variation in pay
in NPOs than in for-profit firms, especially at managerial and professional
levels, and much lower pay disadvantages for women and minority men. (She
attributes the pattern to the higher importance of pay equity in organizations
that rely on intrinsically motivated workers.) If LGB–straight disparities are
also smaller in NPOs, LGBs pay a smaller wage penalty for nonprofit
employment and may be attracted to a less discriminatory workplace.
Alternatively, lesbians and gay men might have a special preference for
working in the nonprofit sector. Having experienced societal disdain, LGBs
probably identify more with out-groups generally than heterosexuals do.
LGBs are also much more liberal than heterosexuals (Egan, 2009; Lewis,
Rogers, & Sherrill, 2003). If liberalism and out-group identification predict a
public service orientation, then nonprofit employment may attract LGBs
more than heterosexuals. On the other hand, LGBs are also strikingly less
religious than others (Egan, 2009; Lewis et al., 2003), which would tend to
discourage nonprofit employment, especially in the many religious NPOs.
Data and Method
The 5-Percent Public Use Microdata Sample (PUMS) of the 2000 Census
provides detailed information on individuals in a random 5% sample of U.S.
households. In every household, the person who owns or rents the dwelling
is designated the householder, and others are identified by their relationships
to him or her. The Census lists a wide array of possible relationships (e.g.,
husband/wife, housemate, boarder), including “unmarried partner.” If the
householder and the unmarried partner are the same sex, I classify them as
members of a same-sex couple.1 Because the PUMS provides no indicator of
the sexual orientation of those without partners, and because coupled and
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
726
Administration & Society 42(6)
single individuals differ systematically, I follow the lead of most research on
LGBs using Census data and drop all people who do not live with a partner.
I restrict the sample to full-time, full-year workers (those who worked at
least 40 hr a week for at least 48 weeks during 1999) who were at least 18
years old. The Census long form asked individuals detailed questions about
their “chief job activity or business last week” and instructed those with multiple jobs to describe the one where they worked the most hours. After naming their employer, they indicated whether they worked for “a private-for-profit
company or business,” “a private not-for-profit, tax-exempt, or charitable
organization,” a “local government (city, county, etc.),” a “state government,”
the “federal government,” or for themselves or their family. Leete (2001)
noted that data processors checked responses on sector “for consistency with
answers to questions on employer name, location, industry, and occupation”
and “could use a directory of company names to identify the correct . . . legal
form of an organization” (p. 145). Leete (2001) and Ruhm and Borkowski
(2003) concluded that although some nonprofit employees misidentified their
sector, those errors result in under- rather than overstatement of differences
between the sectors.
To establish whether higher percentages of those with same-sex than different-sex partners work for NPOs, I calculate the percentage of each sex and
couple type who work in each sector. To test whether overrepresentation of
LGBs in NPOs is a side effect of other choices, I calculate what percentage
of each group would work in each sector if all people in a particular location
or occupation or industry had the same probability of working in a particular
sector. For instance, I created a location variable with 345 values, one for
each PMSA (Primary Metropolitan Statistical Area) and one each for the nonmetropolitan areas in each state, calculated the percentage of all workers in
each location who worked in each sector, assigned that probability to each
worker in that location,2 then averaged the probabilities by gender and partner type. I used the same process for detailed occupation and industry codes.
I next examine whether most people would earn more working for private
firms than for NPOs, by running a series of regression models with the natural
logarithm of 1999 earnings as the dependent variable. Following Leete (2001),
I drop government employees and the self-employed and run separate models
for nonprofit and for-profit employees, allowing the two sectors to reward different characteristics (including sexual orientation) differently. Because most
analyses find very different pay patterns for lesbians and gay men, and because
many independent variables also affect male and female heterosexuals differently, I split the sample by sex. Because the nonprofit workforce is very highly
educated, I also analyze college graduates and others separately.
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
727
Lewis
The key independent variables are two dummy variables coded 1 for people with same-sex or different-sex unmarried partners. The models include
fairly standard control variables. Education is measured in years for those
without college diplomas and as a set of dummy variables for college graduates (master’s degree, professional degree, and doctoral degree; those with a
bachelor’s as their highest degree are the reference group). Work experience
is estimated as “age – education – 6” and is entered in both linear and quadratic form to allow a curvilinear relationship. Weeks worked in 1999 and the
natural logarithm of hours worked in a typical week proxy for work effort and
344 dummy variables for the location (described above) proxy for labor market conditions. I include four dummy variables for race/ethnicity: Latino,
African American, Asian American, and Other or Mixed Race. Non-Hispanic
Whites are the reference group.3 Additional dummy variables indicate
whether the employee is a naturalized citizen, is not a citizen, has limited
English ability, or has a disability. I repeat all regressions, adding 25 dummy
variables for broad occupational category. If people choose their occupations
before finding their jobs, sectoral pay differences should control for these
choices; if employers place workers in occupations (e.g., manager), however,
they should not.
In Table 3, I convert regression coefficients into expected percentage differences in earnings by exponentiating all coefficients, subtracting 1, and
multiplying times 100. In Table 4, I estimate differences between people’s
expected earnings in the two sectors by calculating their expected pay twice,
once based on coefficients from a nonprofit regression and once from a forprofit regression, and subtracting the latter from the former.
Next I model employment in the nonprofit sector, using logit analysis on
a dummy dependent variable coded 1 for employees of NPOs. I run models
separately by sex and college graduation status. I include education, work
experience, race/ethnicity, citizenship, English ability, and disability as control variables, but do not add hours or weeks worked. I also include three
variables indicating the percentage of all people in that location, occupation,
or industry who work for NPOs.4
To capture how well an individual can “afford” NPO employment, I
include several variables. First, the smaller the expected nonprofit pay penalty from the previous regressions, the more likely nonprofit employment
should be. Second, the greater the household resources, the more likely nonprofit employment should be. The model includes a dummy variable coded 1
if the partner/spouse works at least 20 hr per week and the natural logarithm
of the partner’s earnings (coded 0 for those whose partners are not employed).
Third, children increase a household’s expenses, making nonprofit employment
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
728
Administration & Society 42(6)
less affordable. Dummy variables for children younger than age 18 years
and those younger than age 6 years in the household should have negative
coefficients.
To determine which of these factors best explain sectoral employment differences between LGBs and married people, I run many variations on each
logit model. To a bivariate model that only includes couple type, I add (separately) industry, occupation, location, education, work experience, race/ethnicity/citizenship, children, partner’s work status, or expected salary
differences to see how the Has Same-Sex Partner coefficient changes. If
LGB–straight differences on that independent variable contribute to the gross
employment differences of Model 1, then the coefficient will shrink when
that variable is added to the model. Because the order of entry could matter,
I also start from the “full” model and drop each variable (or set of variables)
individually to see how the Has Same-Sex Partner coefficient changes. This
time, if LGB–straight differences on that variable contribute to the gross
employment differences, the Has Same-Sex Partner coefficient should grow
when I fail to control for it.
Findings
Descriptive Statistics
The nonprofit sector employs a disproportionately female, college-educated,
and LGB workforce (Table 1). NPOs employed 5.9% of the full-time workforce in 1999 but 9.5% of partnered women and 11.9% of partnered college
graduates. Men with male partners were more than twice as likely as married
men to work for NPOs (10.3% vs. 4.6%), and women with female partners
were about one quarter more likely than married women to do so (13.7% vs.
10.6%).
Although nonprofit employment varies widely by location, residential differences cannot explain the large number of LGBs working for NPOs. If
LGBs and heterosexuals living in the same PMSA had identical probabilities
of working for NPOs, 6.6% to 6.8% each group would work for NPOs (Panel
3). Differences in occupation and industry matter far more. If LGBs and heterosexuals in the same occupation were equally likely to choose nonprofit
employment, men with male partners would be substantially more likely than
married men (7.9% vs. 5.0%) and women with female partners would be
slightly more likely than married women (10.4% vs. 9.9%) to work for NPOs
(Panel 4). Differences would be even larger if LGBs and heterosexuals in the
same industry had the same probability of working for NPOs (8.4% vs. 4.7%
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
729
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
Actual distribution
All employees
Men with male partner
Married men
Men with female partner
Women with female partner
Married women
Women with male partner
College graduates only
Men with male partner
Married men
Men with female partner
Women with female partner
Married women
Women with male partner
Expected distribution (all employees)
Based on location
Men with male partner
Married men
Men with female partner
Women with female partner
Married women
Women with male partner
6.7
7.4
7.0
12.4
14.0
8.7
65.4
68.8
75.4
49.0
53.6
65.5
76.6
76.1
76.1
75.9
76.1
76.2
15.5
9.4
6.6
21.7
16.8
14.0
6.7
6.6
6.7
6.8
6.6
6.8
7.3
7.3
7.3
7.3
7.2
7.3
5.2
7.0
5.0
8.9
8.3
5.1
4.4
5.1
5.1
5.1
5.2
5.1
7.6
7.6
6.3
13.1
11.3
8.2
5.0
4.5
3.0
8.3
6.6
4.3
(continued)
4.1
5.5
2.9
3.9
4.1
3.0
4.7
6.8
4.7
3.8
4.3
3.6
5.0
4.8
4.8
4.9
4.8
4.8
Local
State
Federal
government government government
75.4
78.3
86.3
65.1
70.4
80.1
For-profit
10.3
4.6
2.7
13.7
10.6
7.5
Nonprofit
Table 1. Actual and Expected Percentage Working in Each Sector
730
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
76.8
78.7
85.1
67.5
70.0
8.4
4.7
3.1
11.3
10.5
8.1
Women with male partner
78.6
76.0
78.5
83.2
67.7
70.9
76.7
For-profit
7.9
5.0
3.4
10.4
9.9
7.9
Nonprofit
Based on occupation
Men with male partner
Married men
Men with female partner
Women with female partner
Married women
Women with male partner
Based on industry
Men with male partner
Married men
Men with female partner
Women with female partner
Married women
Table 1. (continued)
5.6
5.5
6.7
5.3
8.6
8.7
6.0
7.0
6.0
8.7
8.1
6.1
4.5
5.5
4.5
3.3
8.0
6.6
5.5
4.4
3.6
7.7
6.6
5.3
3.2
4.6
5.1
3.7
5.4
4.6
4.1
3.7
5.4
3.2
4.6
4.2
Local
State
Federal
government government government
731
Lewis
for men, and 11.3% vs. 10.5% for women, Panel 5), though industry choice
may be too similar to sector choice to be an appropriate independent variable.
Still, actual LGB employment in NPOs is higher than that predicted by either
occupation or industry.
People with same-sex partners work for somewhat different industries
than married people within the nonprofit sector (Table 2). NPO employees
with same-sex partners are more than twice as likely as married NPO employees (6 times for men) to work for museums and art galleries; nearly twice as
likely to work for civic, social, and advocacy organizations; and about one
half more likely to work for universities. They are only half as likely to work
for elementary and secondary schools and only one fourth (men) to one half
(women) as likely to work for religious organizations.5 Coupled gay men are
also about one half more likely than married men to work for hospitals or
other health NPOs, though coupled lesbians are somewhat less likely than
married women to do so.
Earnings Differences
As in previous studies, men with male partners earn significantly less than
similar married men (Table 3). However, gay–straight pay disparities for
men are twice as large in the for-profit as in the nonprofit sector (17.9% vs.
8.6% for college graduates and 15.9% vs. 8.5% for those without college
diplomas).6 Narrower gay–straight pay gaps in the nonprofit sector echo
smaller gender and race pay disparities in the sector (Leete, 2000; Preston,
1990), though my findings are somewhat mixed on race. In contrast, women
with female partners earn about the same as comparable married women in
both sectors among college graduates and 5% more among women without
college diplomas, with or without controlling for occupation.
The vast majority of nonprofit employees would be expected to earn more
if they worked in for-profit firms, but expected earnings differences vary
widely across groups. Among people who worked for NPOs, 97% of male
college graduates, 92% of men without college diplomas, 85% of female college graduates, and 48% of women without college diplomas had higher
expected earnings in the for-profit sector. Pay penalties are higher for college
graduates than for others and for men than for women (Table 4). Average
expected earnings differences ranged from 32% for male college graduates to
0 for women without college diplomas. Racial differences were small for
women without college diplomas; for others, Whites typically paid a larger
penalty than minorities.
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
732
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
Male
partner
20.8
5.5
4.0
7.4
13.1
10.6
9.9
6.1
4.2
0.7
6.7
11.3
Hospitals
Religious organizations
Elementary and secondary schools
Family and child services
Colleges and universities
Other health
Civic, social, and advocacy organizations
Savings, insurance, and other financial services
Business and professional organizations
Labor unions
Museums and art galleries
Other
Table 2. Organizations That Employ Nonprofit Workers
18.0
21.3
8.8
5.7
10.0
4.8
5.5
3.0
2.4
2.2
1.1
17.3
Wife
Men with
18.5
2.7
6.2
12.3
9.0
4.3
9.0
2.9
2.6
2.8
2.1
27.5
Female
partner
24.7
2.4
6.6
18.0
9.7
10.4
10.9
1.8
3.0
0.6
1.5
10.7
31.1
5.6
12.8
13.6
6.0
8.1
6.4
4.6
2.1
0.5
0.9
8.4
27.1
1.2
6.0
19.2
5.0
9.5
10.1
5.1
2.7
0.7
1.6
11.9
25.2
11.8
10.7
10.5
7.7
6.8
6.2
3.9
2.3
1.3
1.1
12.6
Female
Male
All
partner Husband partner partnered
Women with
733
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
Model 1. No occupational controls
Has same-sex partner
Has different-sex partner
Latino
African American
Asian American
Other or mixed race
Model 2. Controlling for broad
occupational category
Has same-sex partner
Has different-sex partner
Observations
−8.5**
(2.84)
−11.0**
(9.69)
−14.3**
(12.37)
−17.9**
(19.19)
−17.8**
(8.30)
−13.2**
(7.53)
−9.9**
(3.54)
−10.8**
(10.12)
30,065
−15.9**
(21.91)
−13.5**
(88.69)
−14.9**
(75.43)
−19.0**
(103.61)
−21.2**
(57.12)
−13.2**
(39.00)
−14.7**
(20.83)
−13.3**
(83.60)
949,071
−16.2**
(14.32)
−14.0**
(28.32)
320,646
−17.9**
(15.40)
−16.8**
(33.31)
−20.0**
(33.71)
−23.4**
(41.57)
−8.7**
(15.64)
−18.6**
(21.02)
−13.9**
(6.56)
−12.2**
(8.12)
43,100
−8.6**
(3.56)
−9.4**
(5.55)
−11.8**
(6.86)
−16.3**
(12.27)
−11.2**
(7.02)
13.2**
(5.54)
4.4**
(5.73)
−3.6**
(18.65)
517,521
5.0**
(6.21)
−5.2**
(26.41)
−10.2**
(39.42)
−10.0**
(42.39)
−8.9**
(20.65)
−7.6**
(17.49)
5.3*
(2.23)
−3.1**
(4.89)
52,900
5.0
(1.92)
−4.5**
(6.38)
−7.5**
(8.77)
−9.1**
(13.02)
2.9
(1.70)
−5.8**
(4.58)
−1.3
(1.06)
−7.9**
(14.36)
136,843
−1.6
(1.25)
−8.2**
(16.25)
−14.5**
(19.75)
−11.4**
(18.05)
−2.9**
(4.17)
−11.5**
(10.43)
−1.7
(1.14)
−3.7**
(4.26)
41,924
−1.4
(0.83)
−3.7**
(3.91)
−7.4**
(5.71)
−3.9**
(3.75)
4.4**
(3.14)
−5.4**
(2.94)
College graduate
For-profit Nonprofit For-profit Nonprofit For-profit Nonprofit For-profit Nonprofit
No college diploma
College graduate
No college diploma
Women
Men
Table 3. Expected Salary Differences, by Sector, Sex, and Education
734
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
Male
without
diplomas
−2
−13
−6
−13
−4
−9
−7
−11
−12
Same-sex partner
Married
Different-sex partner
White
Asian
Black
Latino
Other
Total
−22
−33
−23
−34
−24
−22
−18
−24
−32
Male
college
graduates
−1
0
1
0
13
0
1
2
0
Female
without
diplomas
−19
−13
−12
−14
−2
−7
−6
−6
−13
Female
college
graduates
Expected percentage pay difference
Table 4. Expected Pay Differences Between Nonprofit and For-Profit Sectors
58
93
81
95
72
85
81
90
92
Male
without
diplomas
92
97
94
98
93
91
85
91
97
Male
college
graduates
51
48
46
49
7
49
44
36
48
Female
without
diplomas
93
84
83
88
54
76
68
67
85
Female
college
graduates
Percentage with negative difference
735
Lewis
Men with same-sex partners typically paid much smaller penalties than
married men. Among gay men without college diplomas, the expected pay
penalty was only 2%, and 42% were expected to earn more in the nonprofit
sector; in contrast, married men without diplomas paid a 13% penalty, on
average, and only 6% were expected to earn more in the nonprofit sector.
Among male college graduates, the expected pay penalty was much higher
for both those with male partners (22%) and those with wives (33%) and
nearly universal (92% and 97%, respectively), but again members of samesex couples faced less disadvantage in the nonprofit sector. In contrast,
female college graduates with female partners were more likely to pay a penalty (93% vs. 84%), and paid a larger one (19% vs. 13%), than married
women.
Choice of Sectors
In Table 5, Model 1 is a bivariate logit model for nonprofit employment, with
partner type as the only independent variable. The Has Same-Sex Partner
coefficients indicate that men with male partners are 3.4 percentage points
more likely than married men to work for NPOs if they do not have college
diplomas (6.5% vs. 3.1%) and 7.2 percentage points more likely to do so if
they are college graduates (19.2% vs. 12.0%). Women with female partners
are 1.6 points less likely than married women to work for NPOs among those
without college diplomas (8.0% vs. 9.6%) and 6.8 points more likely to do
so among college graduates (30.7% vs. 23.9%).
Changes in the Has Same-Sex Partner coefficients between Models 1 and
2 (the “full” model) suggest how much of the gross differences in nonprofit
employment can be accounted for by LGB–straight differences in industry,
occupation, location, pay gaps, partner’s pay, children, race/ethnicity, education, and work experience. Strikingly, adding the full set of control variables
only meaningfully changes the Has Same-Sex Partner coefficient for people
without college diplomas, shrinking it by about half for men (from .767 to
.421) but shifting it in the opposite direction for women (the significant
underrepresentation disappears, replaced by an insignificant overrepresentation). For college graduates, the Has Same-Sex Partner coefficient shrinks
only 10% for men, and it expands 15% for women. Holding the other variables at their means and comparing LGBs to married people of the same sex,
the predicted probability of working for an NPO is about 50% higher for
those with same-sex partners for men (1.7% vs. 1.1% for those with less than
a college diploma and 6.5% vs. 4.0% for college graduates) and for female
college graduates (18.4% vs. 13.2%).
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
736
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
No college diploma
Model 2
0.421**
(4.54)
−0.132**
(4.06)
0.943**
(8.42)
0.004
(0.08)
0.010*
(2.26)
−0.007
(0.32)
0.004
(0.16)
0.094**
(2.93)
−0.243**
(7.91)
−0.151*
(2.40)
Model 1
0.767**
(11.87)
−0.362**
(15.20)
Has same-sex partner
Has different-sex partner
Expected nonprofit earnings
disadvantage
Partner is employed at least
20 hr per week
Natural logarithm of
partner’s earnings
Household has children
younger than age 18 years
Household has children
younger than age 6 years
Latino
African American
Asian American
Model 1
0.509**
(6.61)
−0.181**
(3.75)
0.870**
(11.15)
−0.076*
(2.01)
0.029**
(7.18)
−0.015
(0.65)
−0.033
(1.30)
−0.107*
(2.07)
−0.283**
(5.86)
−0.110*
(2.45)
Model 2
College graduate
0.559**
(12.06)
−0.439**
(14.62)
Men
Table 5. Logit Model for Nonprofit Employment, by Sex and Education
−0.206**
(3.35)
−0.428**
(25.56)
Model 1
0.081
(1.00)
−0.038
(1.74)
0.494**
(5.13)
0.104*
(2.22)
−0.010*
(2.09)
0.009
(0.63)
0.018
(0.95)
−0.102**
(3.98)
−0.346**
(15.71)
−0.340**
(6.75)
Model 2
No college diploma
0.345**
(8.00)
−0.384**
(17.17)
(continued)
0.397**
(6.03)
−0.037
(1.12)
0.345**
(4.78)
0.066
(1.06)
−0.001
(0.18)
−0.011
(0.50)
−0.015
(0.58)
−0.254**
(5.54)
−0.418**
(11.14)
−0.223**
(5.23)
Model 2
College graduate
Model 1
Women
737
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
Other or mixed race
Has limited English ability
Naturalized citizen
Not a citizen
Disabled
Years of education
Master’s degree
Professional degree
Doctoral degree
Work experience
Model 1
Men
0.017**
(4.78)
0.152**
(2.79)
0.176**
(3.40)
−0.012
(0.28)
0.166**
(3.95)
0.053*
(2.40)
0.064**
(12.52)
Model 2
No college diploma
Table 5. (continued)
Model 1
0.439**
(17.61)
0.406**
(14.72)
0.560**
(16.17)
0.022**
(5.69)
−0.143
(1.87)
−0.299**
(2.65)
−0.356**
(8.24)
−0.171**
(3.95)
−0.153**
(4.48)
Model 2
College graduate
Model 1
0.030**
(12.05)
−0.010
(0.24)
−0.031
(0.57)
−0.200**
(5.95)
−0.184**
(4.71)
−0.096**
(5.34)
0.116**
(21.52)
Model 2
No college diploma
(continued)
−0.127
(1.88)
−0.252*
(2.45)
−0.348**
(8.72)
−0.343**
(7.66)
−0.334**
(10.49)
0.281**
(13.52)
0.138**
(4.60)
0.396**
(8.14)
0.026**
(7.84)
Model 2
College graduate
Model 1
Women
738
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
979,142
979,141
−0.000
(0.33)
0.024**
(35.19)
0.071**
(20.15)
0.080**
(223.41)
.0014
.4884
Note: Absolute values of z statistics are in parentheses.
*p < .05. **p < .01.
Observations
Work experience squared
Probability of nonprofit
employment in occupation
Probability of nonprofit
employment in metro area
Probability of nonprofit
employment in industry
McFadden’s pseudo-R2
Model 1
.0014
Model 1
363,747
−0.000
(1.23)
0.011**
(21.76)
0.089**
(23.88)
0.078**
(173.96)
.6163
Model 2
College graduate
363,747
Men
Model 2
No college diploma
Table 5. (continued)
570,423
.0021
Model 1
570,423
−0.000**
(5.51)
0.013**
(32.12)
0.088**
(34.21)
0.070**
(253.54)
.4213
Model 2
No college diploma
.0021
178,769
178,769
−0.000**
(2.74)
0.010**
(21.36)
0.100**
(29.50)
0.064**
(155.72)
.4673
Model 2
College graduate
Model 1
Women
739
Lewis
Industry and occupation are the most important predictors of nonprofit
employment, followed by location (based on standardized odds ratios; Long
& Freese, 2006, pp. 177-179). Although the McFadden pseudo-R2 values
range from .42 to .62 for the full models, they are only .02 to .08 when those
three variables are dropped. The percentage of people in one’s industry who
work for NPOs is by far the strongest predictor of whether one works for an
NPO, but industry choice does not explain the overrepresentation of LGBs in
NPOs. Only among less-educated men are those with same-sex partners disproportionately concentrated in industries with high nonprofit employment.
Adding the industry variable to the “bivariate” model shrinks the Has SameSex Partner coefficient for men without college diplomas from .767 to .565.
For male college graduates, adding or dropping the industry variable barely
affects the coefficient. For both female groups, controlling industry shifts the
coefficient in a positive direction, indicating that coupled lesbians are less
likely than married women to work in NPO-heavy industries.
The percentage of people in one’s occupation who work for NPOs is the
second-strongest predictor of NPO employment, and occupational choice
helps explain why gay men without college diplomas are overrepresented in
the sector. (The Has Same-Sex Partner coefficient drops from .767 to .427
when occupation is added to the bivariate model and rises from .421 to .505
when it is dropped from the full model.) For the other groups, adding or dropping the variable has little impact on the coefficient, or shifts it in the wrong
direction. Location affects sector nearly as much as occupation (it actually
changes the McFadden pseudo-R2 a bit more) but, as in Table 1, it does not
help explain why LGBs are more likely to work for NPOs. LGB–straight differences in nonprofit employment are slightly larger when location is controlled than when it is not (starting from either the bivariate or full model).
The ability to afford nonprofit employment makes some difference, especially for men. The smaller one’s expected pay penalty for nonprofit employment, the more likely one is to work for an NPO. Among male college
graduates, for instance, the average predicted probability of working for an
NPO was 16.3% for those whose expected pay was higher in the nonprofit
sector and only 11.8% for those whose expected pay was lower in NPOs.
Although the positive effect is highly significant for all four groups, it is not
large: a one-standard-deviation increase in relative pay only raises the probability of nonprofit employment by 0.3 to 1.5 percentage points (holding the
other variables at their means).
Having an employed partner increases a man’s probability of working for an
NPO, and the higher his partner’s earnings, the higher his probability of doing
so. For women, however, higher partner’s earnings lower the probability of an
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
740
Administration & Society 42(6)
NPO job for those without college diplomas and have no effect for college
graduates. Having children does not significantly affect the probability for
working for an NPO for any group.
Affordability helps explain why men with male partners are more likely
than married men to work for NPOs, but it does not help explain sectoral differences for women. For men, the Has Same-Sex Partner coefficient is about
20% lower in the full model than in the same model without relative pay,
partner’s work status and earnings, and the presence of children; the bivariate
coefficient also drops by .12 to .15 when those variables are added in the
men’s equations. For the women, however, these variables have virtually no
effect on the Has Same-Sex Partner coefficient.
Additional education increases the probability of nonprofit employment
for all groups, as does work experience. The somewhat higher educational
level of men with male partners than of married men explains a small amount
of the sectoral difference. Somewhat surprisingly, given previous research on
the diversity of the nonprofit sector, African Americans and Asian Americans
were significantly less likely than comparable Whites to work for NPOs in all
four groups, and Latinos were significantly less likely to do so in three groups.
Limitations
Because Census data only identify LGBs who live with their partners, I
dropped the unpartnered and compared people with same-sex partners to married people. This choice of comparison group is particularly important for
men, because an extensive empirical literature shows that husbands make
substantially more than apparently comparable single men, even if they have
female partners (Antonovics & Town, 2004; Cornwell & Rupert, 1997;
Ginther & Zavodny, 2001; Gray, 1997; Korenmann & Neumark, 1991;
Krashnisky, 2004; Loh, 1996; Stratton, 2002). Three potential explanations of
the marriage premium for men, all finding some support in the research, suggest different implications of comparing partnered gay men to married men.
First, if employers discriminate in favor of married men, comparing coupled gay men to demographically similar married men is fair. Second, if marriage makes men more productive (e.g., because their wives specialize in
household production [cooking, cleaning, child-rearing], allowing the husbands to focus on their careers), however, this comparison may overstate pay
discrimination. To test this possibility, I restricted the sample to men whose
wives or partners work full-time to eliminate the single-earner households
where specialization is most likely; the unexplained gay–straight pay differences dropped for both college graduates and others in both sectors, but the
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
741
Lewis
pattern of smaller disparities in the nonprofit sector persisted. Third, if more
productive men are more likely to marry (e.g., because women assess additional characteristics [not asked on the Census] that make men both more
desirable spouses and more productive workers), the comparison could also
be unfair, if gay men care less about productivity in choosing mates than
heterosexual women do. Selection into partnership may work differently for
LGBs than selection into marriage does for heterosexuals (Black et al. 2000;
Carpenter & Gates, 2008; Gates & Ost, 2004). However, wealthier and better-educated lesbians and gay men are more likely to have partners (Carpenter,
2003; Carpenter & Gates, 2008), suggesting that LGBs do consider productivity characteristics in coupling decisions, and the lower rate of coupling for
LGBs suggests that same-sex partnership is more selective than marriage.
This analysis also treats the ability to afford nonprofit employment as
exogenous. People without children and with well-paid partners or spouses
could be more likely to choose nonprofit employment, but couples where one
or both partners want to work for NPOs could choose to have both partners
work or not to have children. In the latter case, my findings would overstate
the impact of affordability. As those apparent effects were fairly weak, however, the bias appears minor.
This analysis also excluded the public sector, although Table 1 suggests
underrepresentation of gay men and college-educated lesbians in federal and
local government employment. Given a long history of explicit anti-LGB
discrimination in government employment (Johnson, 2004; Lewis, 1997), a
continuing ban on open LGBs in the military, and fairly weak protections
against discrimination in the federal service and most state and local governments today (Lewis & Pitts, in press), LGBs with high public service motivation may be drawn to nonprofit over public sector employment. That is,
altruistic impulses that might push heterosexuals toward public sector jobs
might direct LGBs to the nonprofit sector. The overrepresentation of LGBs in
NPOs swamps their underrepresentation in government, however, even after
controlling for the factors considered here (Lewis & Pitts, in press). Inclusion
of government employees might moderate but should not fundamentally
change the conclusions.
Conclusion
The nonprofit sector employed 7.2% of the U.S. workforce in 2004 (Salamon
& Sokolowski, 2006, p. 3), up from 6.4% in 1990, as nonprofit employment
has grown nearly twice as rapidly as the general economy for 15 years (Irons
& Bass, 2004, p. 3). Its ability to grow despite paying comparable workers
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
742
Administration & Society 42(6)
substantially less than the private, for-profit sector indicates that superior
opportunities to help others and benefit society, along with other nonpecuniary benefits, play a key role in attracting qualified workers. Pay is not irrelevant, as the size of the financial penalty they pay individuals affects whether
nonprofit employment is chosen. A one-standard-deviation increase in relative pay only raises the probability of nonprofit employment by 0.3 to 1.5
percentage points, however, and other signs that a worker can afford to pay a
penalty (having an employed partner with high earnings and not having children) also have a limited impact. Nonprofit sector pay appears to be high
enough to ensure a qualified workforce but low enough to attract the more
altruistic portions of the labor force.
Thus, the overrepresentation of partnered lesbians and gay men in the nonprofit sector suggests they have higher levels of altruism or commitment to
social change, unless alternative explanations can be found. Partnered gay
men’s overrepresentation owes something to smaller earnings penalties,
greater likelihood of having employed partners whose salaries compensate
for those penalties, and lower probability of raising children. Even with all
these factors (including occupation and industry) controlled, however, gay
men’s odds of nonprofit employment remain half to two thirds higher than
those of comparable married men. These nonaltruistic explanations do an
even worse job of explaining the overrepresentation of partnered lesbians:
They do not differ from married women in ways that would predict overrepresentation in the nonprofit sector, but among college graduates their odds of
nonprofit employment are also 50% higher.
This pattern casts doubt on a widespread belief that LGBs share few, if
any, of the ethical and moral values of heterosexual Americans (Washington
Post 1998). It also clashes strongly with popular stereotypes that LGBs lead
hedonistic, self-centered, superficial lives. Black et al. (2002, 2007) have
examined how the different economic incentives LGBs face affect their life
choices but have tended to focus on greater consumption of leisure activities
and higher probabilities of living in expensive, “high-amenity” cities.
Researchers may want to consider whether different incentives also lead to
greater opportunities for service in other ways.
Managers of nonprofit organizations need to consider the implications of
their disproportionately LGB workforces. Explicit prohibition of discrimination on the basis of sexual orientation and provision of domestic partner benefits may be especially important for NPOs in attracting and retaining
high-quality employees. NPOs may want to target more recruitment efforts at
the LGB community, for new employees as well as for volunteers and funds.
Diversity training may help heterosexual employees adapt to environments
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
743
Lewis
where LGBs are likely to open on the job and develop organizational cultures
that value sexual diversity. Such efforts may increase NPOs’ competitive
advantage over private firms and governments for certain types of employees, though they may also make workplaces more challenging for religious
conservatives.
Scholars should also consider LGBs in their analyses of the nonprofit and
public sectors. Such marked LGB–straight differences in choice of nonprofit
employment suggest that representation and pay issues may also be important in government. Nondiscrimination protections could have important
impacts on recruitment and retention of both LGB and heterosexual employees or on the delivery of services. LGB and straight managers might differ in
style or effectiveness. Research has been hindered by the absence of good
data: None of the federal surveys ask for sexual orientation, and personnel
records do not include the information. A new sexual orientation question in
the 2008 American National Election Studies led to 5% self-identifying as
LGB and only 1% refusing to answer the question. A similar question on the
Federal Human Capital Survey could provide scholars large enough samples
to study a wide variety of issues.
Acknowledgment
An earlier version of this article was presented at the annual meeting of the
Association for Research on Nonprofit Organizations and Voluntary Action,
November 15, 2007, Atlanta, Georgia. I am grateful to Lakshmi Pandey for substantial help in working with the Census data and to Chester Galloway for excellent
research assistance.
Declaration of Conflicting Interests
The author declared no conflicts of interests with respect to the authorship and/or
publication of this article.
Funding
This research was funded in part by the Williams Institute of the UCLA School of
Law.
Notes
1. When same-sex couples entered their marital status as “married,” the Census Bureau
changed their marital status and relationship codes, recoding them as “unmarried
partners.” Black et al. (2006) showed convincingly that most couples whose marital
and relationship status were “allocated” in this way had actually made an error in
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
744
2.
3.
4.
5.
6.
Administration & Society 42(6)
recording the sex of one of the spouses. Following their advice and the practice
of others (e.g., Carpenter & Gates, 2008; Jepsen, 2007), I have dropped everyone
whose sex, marital status, or relationship code was allocated.
In nonmetropolitan Alabama, for example, 3.63% worked for NPOs and 72.27%
worked for private firms; I therefore assigned each worker in that area a 3.63%
probability of nonprofit employment and a 72.27% probability of for-profit
employment.
Latino is coded 1 for all those who checked “Spanish/Hispanic/Latino.” The race
variables are coded 1 only if Latino is coded 0. African American is coded 1 for
those who checked only “Black, African American, or Negro” and Asian American is coded 1 for those who checked only one or more of the Asian options. Those
who checked “American Indian or Alaska Native” or multiple races (and were not
Latino) are coded 1 on Other or Mixed Race. The reference group is those who
only checked “White.”
These are calculated as the percentage of private sector employees (dropping the
self-employed and government workers) who worked for NPOs within each metropolitan area, detailed occupation, or detailed industry (separately). Using the full
set of dummy variables was not practical computationally. In one group (male college graduates), the far simpler probability variable contributed almost as strongly
to the McFadden’s pseudo-R2 and did not meaningfully change the Has Same-Sex
Partner coefficient.
Because coupled gay men are more than twice as likely as married men to work
for NPOs, they are actually one half (rather than one fourth) as likely as married
men to work for religious organizations. This also means that they are more than
12 times as likely as married men to work for museums and art galleries.
Occupational choice accounts for some of the difference: Controlling for 25 broad
occupational categories makes the gay–straight pay gaps more similar in the two
sectors.
References
Allegretto, S. A., & Arthur, M. M. (2001). An empirical analysis of homosexual/heterosexual male earnings differentials: Unmarried and unequal? Industrial and
Labor Relations Review, 54, 631-646.
Antonovics, K., & Town, R. (2004). Are all the good men married? Uncovering the
sources of the marital wage premium. American Economic Review, 94, 317-321.
Badgett, M. V. L. (1995). The wage effects of sexual orientation discrimination.
Industrial and Labor Relations Review, 48, 726-739.
Badgett, M. V. L. (2001). Money, myths, and change: The economic lives of lesbians
and gay men. Chicago, IL: University of Chicago Press.
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
745
Lewis
Badgett, M. V. L., & King, M. C. (1997). Lesbian and gay occupational strategies. In
A. Gluckman & B. Reed (Eds.), Homo economics: Capitalism, community, and
lesbian and gay life (pp. 73-86). New York, NY: Routledge.
Barrett, D. C., Pollack, L. M., & Tilden, M. L. (2002). Teenage sexual orientation,
adult openness, and status attainment in gay males. Sociological Perspectives,
45, 163-182.
Benz, M. (2005). Not for the profit, but for the satisfaction? Evidence on worker wellbeing in non-profit firms. Kyklos, 58, 155-176.
Belman, D., & Heywood, J. S. (1989). Government wage differentials: A sample
selection approach. Applied Economics, 21, 427-439.
Berg, N., & Lien, D. (2002). Measuring the effect of sexual orientation on income:
Evidence of discrimination? Contemporary Economic Policy, 20, 394-414.
Black, D. A., Gates, G., Sanders, S., & Taylor, L. (2007). The measurement of samesex unmarried partner couples in the 2000 U.S. census. California Center for Population Research, University of California–Los Angeles.
Black, D. A., Gates, G., Taylor, L. J., & Sanders, S. G. (2000). Demographics of the
gay and lesbian population in the United States: Evidence from available systematic data sources. Demography, 37, 139-154.
Black, D. A., Gates, G., Taylor, L. J., & Sanders, S. G. (2002). Why do gay men live
in San Francisco? Journal of Urban Economics, 51, 54-76.
Black, D. A., Makar, H. R., Taylor, L. J., & Sanders, S. G. (2003). The effects of sexual orientation on earnings. Industrial and Labor Relations Review, 56, 449-469.
Black, D. A., Sanders, S. G., & Taylor, L. J. (2007). The economics of lesbian and gay
families. Journal of Economic Perspectives, 21(2), 53-70.
Blandford, J. M. (2003). The nexus of sexual orientation and gender in the determination of earnings. Industrial and Labor Relations Review, 56, 622-642.
Brooks, A. C. (2006). Who really cares: The surprising truth about compassionate
conservatism: America’s charity divide—Who gives, who doesn’t, and why it matters. New York, NY: Basic Books.
Carpenter, C. (2003). Sexual orientation and body weight: Evidence from multiple
surveys. Gender Issues, 21(3), 60-74.
Carpenter, C. (2004). New evidence on gay and lesbian household incomes. Contemporary Economic Policy, 22, 78-94.
Carpenter, C. (2005). Self-reported sexual orientation and earnings: Evidence from
California. Industrial and Labor Relations Review, 58, 258-273.
Carpenter, C., & Gates, G. (2008). Gay and lesbian partnership: Evidence from California. Demography, 45, 573-590.
Clain, S. H., & Leppel, K. (2001). An investigation into sexual orientation discrimination as an explanation for wage differences. Applied Economics, 33, 37-47.
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
746
Administration & Society 42(6)
Cornwell, C., & Rupert, P. (1997). Unobservable individual effects, marriage and the
earnings of young men. Economic Inquiry, 21, 285-294.
Crewson, P. E. (1997). Public-service motivation: Building empirical evidence of
incidence and effect. Journal of Public Administration Research and Theory, 7,
499-518.
Crow, S. M., Fok, L. Y., & Hartman, S. J. (1998). Who is at greatest risk of workrelated discrimination—Women, Blacks, or homosexuals? Employee Responsibilities and Rights Journal, 11, 15-26.
Egan, P. J. (2009, March). Group cohesion without group mobilization: The case of
lesbians, gays, and bisexuals (Working Paper). Retrieved from http://politics.
as.nyu.edu/docs/IO/4819/cohesionwomobilization.pdf
Egan, P. J., & Sherrill, K. (2005). Neither an in-law nor an outlaw be: Trends in Americans’ attitudes toward gay people. Public Opinion Pros (February). Retrieved
from http://www.publicopinionpros.com/
Frank, R. H. (1996). What price the moral high ground? Southern Economic Journal,
63(1), 1-17.
Gates, G. J., & Ost, J. (2004). The gay and lesbian atlas. Washington, DC: Urban
Institute.
Ginther, D. K., & Zavodny, M. (2001). Is the marriage premium due to selection?
The effects of shotgun weddings on the return to marriage. Journal of Population
Economics, 14, 313-328.
Goddeeris, J. H. (1988). Compensating differentials and self-selection: An application
to lawyers. Journal of Political Economy, 96, 411-428.
Gray, J. (1997). The fall in men’s return to marriage: Declining productivity effects or
changing selection? Journal of Human Resources, 32, 481-504.
Gyourko, J., & Tracy, J. (1988). An analysis of public- and private-sector wages
allowing for endogenous choices of both government and union status. Journal of
Labor Economics, 6, 229-253.
Hansmann, H. B. (1980). The role of nonprofit enterprise. Yale Law Journal, 89, 835-901.
Hebl, M. R., Foster, J. B., Mannix, L. M., & Dovidio, J. F. (2002). Formal and interpersonal discrimination: A field study of bias toward homosexual applicants. Personality and Social Psychology Bulletin, 28, 815-825.
Irons, J. S., & Bass, G. (2004). Recent trends in nonprofit employment and earnings:
1990-2004 Tax and Budget Reports. Washington, DC: OMB Watch.
Jeavons, T. H. (1992). When the management is the message: Relating values to management practice in nonprofit organizations. Nonprofit Management and Leadership, 2, 403-417.
Jepsen, L. K. (2007). Comparing the earnings of cohabiting lesbians, cohabiting heterosexual women, and married women: Evidence from the 2000 Census. Industrial Relations, 46, 699-727.
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
747
Lewis
Johnson, D. K. (2004). The lavender scare: The Cold War persecution of gays and
lesbians in the federal government. Chicago, IL: University of Chicago Press.
Karl, K. A., & Sutton, C. L. (1998). Job values in today’s workforce: A comparison of
public and private sector employees. Public Personnel Management, 27, 515-516.
Klawitter, M. M., & Flatt, V. (1998). The effects of state and local antidiscrimination policies for sexual orientation. Journal of Policy Analysis and Management, 17, 658-686.
Korenmann, S., & Neumark, D. (1991). Does marriage really make men more productive? Journal of Human Resources, 26, 282-307.
Krashnisky, H. A. (2004). Do marital status and computer usage really change the
wage structure? Journal of Human Resources, 39, 774-791.
Leete, L. (2000). Wage equity and employee motivation in nonprofit and for-profit
organizations. Journal of Economic Behavior and Organization, 43, 423-446.
Leete, L. (2001). Whither the nonprofit wage differential? Estimates from the 1990
Census. Journal of Labor Economics, 19, 136-170.
Lewis, G. B. (1997). Lifting the ban on gays in the civil service: Federal policy toward
gay and lesbian employees since the Cold War. Public Administration Review, 57,
387-395.
Lewis, G. B. (2006). The demographics of Georgia I: lesbian and gay couples. (FRC
Brief 113): Fiscal Research Center, Andrew Young School of Policy Studies,
Georgia State University.
Lewis, G. B., & Frank, S. A. (2002). Who wants to work for the government? Public
Administration Review, 62, 395-404.
Lewis, G. B., & Pitts, D. W. (2010). Representation of lesbians and gay men in federal, state, and local bureaucracies. Journal of Public Administration Research
and Theory, in press.
Lewis, G. B., Rogers, M. A., & Sherrill, K. (2003, May). Sexual identity, sexual
behavior, and group socialization: Does gay sex turn people into liberal democrats? Paper presented at the annual meeting of the American Association for
Public Opinion Research, Nashville, TN.
Light, P. C. (2002). Pathways to nonprofit excellence. Washington, DC: Brookings
Institution.
Loh, E. S. (1996). Productivity differences and the marriage wage premium for White
males. Journal of Human Resources, 31, 568-589.
Long, J. S., & Freese, J. (2006). Regression models for categorical dependent variables using Stata (3rd ed.). College Station, TX: Stata Press.
Mirvis, P. H. (1992). The quality of employment in the nonprofit sector: An update
on employee attitudes in nonprofits versus business and government. Nonprofit
Management and Leadership, 3, 23-41.
Mirvis, P. H., & Hackett, E. J. (1983). Work and work force characteristics in the
nonprofit sector. Monthly Labor Review, 106(4), 3-12.
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010
748
Administration & Society 42(6)
Moulton, B. R. (1990). A reexamination of the federal–private wage differential in the
United States. Journal of Labor Economics, 8, 270-293.
Onyx, J., & Maclean, M. (1996). Careers in the third sector. Nonprofit Management
and Leadership, 6, 331-345.
Perry, J. L., & Wise, L. R. (1990). The motivational bases of public service. Public
Administration Review, 50, 367-373.
Preston, A. E. (1989). The nonprofit worker in a for-profit world. Journal of Labor
Economics, 7, 438-463.
Preston, A. E. (1990). Women in the white-collar nonprofit sector: The best option or
the only option? Review of Economics and Statistics, 72, 560-568.
Rainey, H. G. (1982). Reward preferences among public and private managers: In
search of the service ethic. The American Review of Public Administration, 16,
288-302.
Rose-Ackerman, S. (1996). Altruism, nonprofits, and economic theory. Journal of
Economic Literature, 34, 701-728.
Ruhm, C. J., & Borkowski, C. (2003). Compensation in the nonprofit sector. Journal
of Human Resources, 38, 992-1021.
Salamon, L. M., & Sokolowski, S. W. (2006). Employment in America’s charities: A
profile. Nonprofit Employment Bulletin.
Smith, S. P. (1976). Government wage differentials by sex. Journal of Human
Resources, 11, 185-199.
Stratton, L. (2002). Examining the wage differential for married and cohabiting men.
Economic Inquiry, 49, 199-212.
Washington Post, Henry J. Kaiser Family Foundation, & Harvard University. (1998).
Survey on American Values. (2,025 telephone interviews conducted July 29 - August
18, 1998).
Weichselbaumer, D. (2003). Sexual orientation discrimination in hiring. Labour Economics, 10, 629-642.
Weisbrod, B. A. (1983). Nonprofit and proprietary sector behavior: Wage differentials
among lawyers. Journal of Labor Economics, 1, 246-263.
Wittmer, D. (1991). Serving the people or serving for pay: Reward preferences among
government, hybrid sector, and business managers. Public Productivity & Management Review, 14, 369-383.
Bio
Gregory B. Lewis is professor of public management and policy in the Andrew
Young School of Policy Studies at Georgia State University and director of the joint
Georgia State-Georgia Tech doctoral program in public policy. He has published
widely on the career patterns and attitudes of public employees, on public support for
lesbian and gay rights, and on morality policy more broadly.
Downloaded from aas.sagepub.com at CAL STATE UNIV SACRAMENTO on September 13, 2010