Women in Presidential Cabinets 3

Political background.
Writing about the US cabinet, Wyszomirski

(1989: 59) argues that political skills “come from serving in the legislature
Copyright © 2016. Oxford University Press. All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under
or as an executive” (e.g., governor, mayor, head of a government agency),

and through informal “experience as a lobbyist for an interest group.”
(We explore ministers’ group connections in the next chapter.) “An alter-
native source of political skills has historically resided in political parties.
Indeed, party functionaries, even if they have never held elective office,
are likely to be adept political operants” (Wyszomirski 1989: 61). We code
six types of political background capturing this range of types of politi-
cal experience: (1) Political insider—ministers who had previously held a
cabinet post, been a vice minister, or built a career in national govern-
ment (237 ministers, 53.5%). We use political insider as our proxy for
the PCR of political skills because it requires experience at top levels of
government; however, it does not require the minister to have experi-
ence in an elected post, which was infrequent in our data set, though
insider status can come from elected office experience. Ministers who do
not meet the requirements to be coded as political insiders are outsiders.
An outsider may have worked in government, even for many years, but
not in a high-level post or not in the capital (Borrelli 2002).1 (2) Elected
1. Borrelli coded a cabinet secretary as an insider only if they changed directly from
one cabinet post to another, entered directly from the subcabinet, or had an estab-
lished relationship with the Washington, DC community from building a primary
career in national government. We made modifications to accommodate differences
in political career paths and the history of democracy in our Latin American cases. We
relaxed the requirement that a person move directly from a subcabinet post into the
cabinet or directly from one cabinet post to another, but we did not count breaks that
were so long that it was not reasonable to expect the same people would still be play-
ers in national politics, such as having served as a minister prior to the intervention
of a military regime, and returning to the cabinet ten or more years after the restora-
tion of democracy. A career in national government is often shorter in Latin American
countries than in the United States due to lower reelection rates to congress, particu-
larly in Argentina and Colombia. (Chapter 2 offers information about reelection rates.)
Immediate reelection to Costa Rica’s Legislative Assembly is prohibited; consequently
many deputies want to continue their political career through an executive branch ap-
pointment (Taylor 1992; Carey 1996). In Costa Rica, a career in national government
typically means a person worked in the private sector in between top-level posts in
government, which might be considered a revolving door career in the United States.
In Argentina and Chile we code a “career in national government” as having held a seat
U.S. or applicable copyright law.
in congress or a top executive branch post for most of the time since the return to de-
mocracy. We, and also Borrelli, are using the terms “insider” and “outsider” somewhat
differently than de Winter (1991: 46) used them for his study of parliamentary cabi-
nets of Western Europe. For de Winter “parliamentarians of long-standing and [who]
belong to the leadership circles of their party are insiders” while outsiders “have effec-
tively no prior contact with political life before joining the cabinet.” Camerlo and Pérez-
Liñán (2015a) use the term “outsiders” for a minister who does not have party ties or
independent policy expertise and who is prone to behave as a loyalist to the president.
EBSCO[ 106 ] Women
Publishing in Presidential
: eBook Collection Cabinets
(EBSCOhost) - printed on 11/29/2017 2:21 PM via ECONOMIC COMMISSION
FOR LATIN AMERICA AND THE CARRIBEAN (ECLAC LIBRARY)
AN: 1260012 ; Escobar-Lemmon, Maria C., Taylor-Robinson, Michelle M..; Women in Presidential Cabinets :
Power Players or Abundant Tokens?
Account: ns191738
office—ministers who have held elected office at the local, regional, state
or national level (155 ministers, 35.0%).2 (3) Appointed post—ministers
who previously held a high level post (agency head, ambassador, cen-
tral bank chief, minister, vice minister) (236 ministers, 53.5%). (4) Prior
minister/vice minister—individuals who had previously been appointed
minister or vice minister in the same or different portfolio regardless
of whether there was a gap between the appointments (179 ministers,
40.7%). (5) Organizational partisan—people who held a post in their party
(national, local, party office in congress), managed a campaign, or served
as a presidential campaign advisor (141 ministers, 31.8%).3 (6) Campaign
involvement—ministers who played a leadership role in a campaign other
than their own, though that involvement need not immediately precede
their appointment to the cabinet, nor did it have to be the campaign of
the president who appointed them (86 ministers, 19.5%).4 Many minis-
ters have more than one type of political experience so categories are not
mutually exclusive and some are nested (e.g., prior minister/vice minister
within appointed post).
Political connections. We also coded for three types of political connec-
tions: (1) Connected to the president—individuals who had ties to the presi-
dent that could be a formal connection (e.g., campaign manager), personal
friendship or family ties, or who were described in the press as a close con-
fidant (134 ministers, 30.3%). (2) Friend of the president—ministers known
to be a friend of the president or described in the press as a close confidant
(56 ministers, 12.6%). Professional association was not enough to meet
the threshold of “friend.” Thus “friends” are “connected,” but not every-
one who is “connected” is a “friend.” (3) Political family—ministers from
a leading family in the country’s politics (e.g., the Kennedy family in the
United States, the Barco family in Colombia) (32 ministers, 7.2%). These
are families for whom politics is the family business, including traditions
of dominating subnational politics, or prominent party factions associated
with a particular family.
2. We do not disaggregate number or types of elected offices, or level of government.

We also do not examine time in elected office. Few people in our data set held prior
elected office only at the local level (4% of ministers), or only at the local or state/
regional level (6% of ministers), which precludes fine-grained analysis comparing men
and women, particularly across categories of portfolios.
3. Organizational partisan is part of Cohen’s (1988b) typology of party activity in
the United States. Studies of minister background in parliamentary cabinets also code
party leadership as occurring at the national, regional or local level of the party (see de
Winter 1991; Davis 1997).
4. Involvement could be at the national or lower level and includes roles ranging
from campaign chair to policy advisor.
EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017 P2:21

ol i t
PMi cvia
a l ECONOMIC [ 107 ]
S k i l l s COMMISSION
Account: ns191738
DO MEN AND WOMEN DIFFER IN POLITICAL
BACKGROUND OR CONNECTIONS?
The data set overall
Political background. When we group together all 447 ministers in our

database we observe some differences in the political background of the
women compared to the men. First, regarding the PCR of political skills—
measured as qualifying as a political insider—figure 5.1 shows that level of
political experience is more common for men than women (56% of men,
46% of women, p = .084).5
Second, a greater percentage of men than women held elected office
before entering the cabinet (38% of men, 27% of women, p = .035). Since
the executive branch must work with the legislature to pass laws and the
budget,6 lack of elected office experience may be a disadvantage for women.
However, experience with elected office does not appear to be a norm for
cabinet ministers, as 62% of men and 73% of women ministers did not have
prior experience in elected office. As we will see below, the percentage of
ministers with elected office experience varies across countries, which is
likely due at least in part to the number of elected offices that are available
across these countries.
Prior experience in an appointed post is slightly less common for women
(48%) than men (55%), but the difference is not statistically significant (p
= .162), and almost identical percentages of men and women have prior
minister or vice minister experience (39% of women, 41% of men, p = .763).
Executive branch experience was more common for ministers—both men
and women—than elected experience, but almost half of ministers (46%)
had not held an appointed post and thus lacked prior executive branch expe-
rience. It appears that prior experience working for government is not viewed
as essential for all kinds of ministers, at least not in the aggregate. This find-
ing is in line with the conclusion of scholars studying the US cabinet. For ex-
ample, Wyszomirski (1989: 54–56) describes three different types of staffing
strategies used by US presidents: recruitment of political allies, recruitment
of policy expertise, and recruitment of managers. We explore below whether
this finding holds when we look at different types of portfolios.
5. Detailed tables for all political trait analyses presented in figure 5.1 and subse-
quent figures in the chapter, including the number of ministers coded on each trait,
the percentage of women and men with each trait, and the p-values for the difference
of means tests are available from the authors on request.
6. In Argentina, Chile, and Colombia the president has varying capacity to make
policy by decree instead of working through the congress to pass a law (see chapter 2
for more detailed information).
EBSCO[ 108 ] Women
Account: ns191738
Source of Political Skills
Political background
political insider (see note) †

elected office *
appointed post
prior minister/vice-minister
organizational partisan
campaign involvement
Political connections
connected to president
friend of president
political family **
0 20 40 60 80 100
Women
Men All ministers (n = 447)
Figure 5.1 Minister political backgrounds—data set overall

Statistical significance levels: † p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001
Note: The variable political insider is the proxy for political skills in this study.
The percentages of women (30%) and men (32%) with background as

organizational partisans are almost identical (p = .649), as are the percent-
ages of women (22%) and men (19%) who had prior campaign involvement
(p = .454). Some party or campaign leaders are appointed to these cabinets,
which indicates that party or campaign experience can be a valued part of
a political résumé for a cabinet minister. Yet such types of political back-
ground are clearly not a norm in the data set overall.
Political connections. Ministers can have political connections along
with, or possibly instead of political experience. In our data set 30% of
ministers are coded as connected to the president—36% of women, 29%
of men, though this difference is not statistically significant (p = .171).
More men than women are coded as being a friend of the president—14%
of men, 10% of women, though again this is not a significant difference
(p = .342). The small percentage of ministers of both sexes with close
ties to these presidents is noteworthy, though this varies across countries
(see figure 5.3). Finally, while few ministers in our data set come from
political families, this is more common for women (14%) than for men
(5%), p = .003.7
7. MaryAnne Borrelli (personal communication) suggested that coming from a po-
litical family could be a way for women to compensate for their relative lack of formal
political experience, since women are relative newcomers to elected and high-ranking
EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017P2:21

ol i tPMi cvia
Account: ns191738
Clearly ministers put together their careers in different ways. One
precabinet career trajectory is embodied in Mary Peters, US secretary of
transportation (2006–09). Before entering the cabinet she worked in state

and national government, building a career in the Arizona Department
of Transportation, and then as the head of the Federal Highway
Administration, but she was not known for involvement in her party or
campaigns. In contrast, Yolanda Mainieri, minister of women’s affairs in
Costa Rica (1998) entered with party, but not appointed experience. She
had not held a formal government post, instead building a career as a
professor and consultant (e.g., for the UN, the Costa Rican Court, orga-
nizations in Brazil and Norway), she was also an organizational partisan,
having worked in her party through the Social Christian Party’s Women’s
Program. Marta Pinto de Hart, minister of communications in Colombia
(2002–06), lacked experience in government having built a career in the
private sector, but she had worked in her party and on campaigns. Of
course some ministers enter the cabinet with political experience related
to both parties and government. For instance, Martin Zilic, Chile’s min-
ister of education (2006), worked as a professor and hospital director and
then was appointed intendent of Region VIII of the country (an appointive
position somewhat equivalent to governor) before entering the cabinet. He
was also an organizational partisan, having served as party president for
the Christian Democrats in Region VIII. In contrast, Henry Paulson, US
secretary of the treasury (2006–09), typifies the nonpolitical minister. He
worked in government briefly in the early 1970s when he completed his
MBA then spent more than thirty years working in the private sector at
Goldman Sachs, ultimately becoming CEO. Thus, Paulson entered the cabi-
net with policy expertise but possessing none of the political background
traits we analyze in this chapter.
appointed posts. Hinojosa (2009: 388), studying election of women to municipal gov-

ernment offices in Chile, found that more women than men had a family member in
politics. She argues that this is due to local power monopolies having a great deal of
control over nominations, and women are more likely to be able to find a sponsor
from a local gatekeeping group if they are from a political family. Prewitt and Stone
(1973: 133) in their seminal study of elite recruitment in the United States under-
score that throughout history “The political elite literally reproduced themselves” and
coming from such a family increases a person’s opportunity to reach an elite level in
politics. Coming from a political family has historically helped men enter the top ranks
of politics, and our data indicate that it helps women as well. With the increased re-
cruitment of women into politics female family members become a resource for po-
litical families to maintain their presence in power. However, it seems unlikely that
appointment of such women to the cabinet would indicate true elite circulation, where
circulation means “different types of persons” entering government (Prewitt and Stone
1973: 168).
EBSCO[ 110 ] Women
Account: ns191738
In sum, for our data set overall, we observe few significant differences
between women and men in their political background or connections.
However, men are more likely to qualify as political insiders and thus for
us to consider them to be a source of the PCR of political skills. Men are
also more likely to have held elected office. Notably, less than half of the
ministers have several of the forms of political experience we code, and
only about half (54%) are political insiders or held a prior appointed post.
Thus, aggregate data give a strong indication that political skills are not
essential for men or women to be appointed to the cabinet.
Comparison of initial and replacement ministers
For the data set as a whole we noted statistically significant sex effects only
for political insider, elected office, and connection to a political family. All
three differences continue to be significant only in the initial cabinet (see
figure 5.2). Among ministers appointed at the beginning of the adminis-
tration, when diversity in the cabinet is likely to receive media coverage,
more men than women are political insiders (56% of men, 40% of women,
p = .030), but of the replacement ministers almost identical percentages of
women (54%) and men (56%) are insiders (p = .853). In the initial cabinet
42% of the men had previously held elected office compared to just 29% of
political insider (see note) *
elected office †
appointed post †
connected to president *
friend of president
political family **
0 20 40 60 80 100 0 20 40 60 80 100
Women
Men Initial (n = 237) Replacement (n = 209)
Figure 5.2 Minister political backgrounds—initial/replacement ministers


ol i PM
t i cvia
a lECONOMIC [ 111 ]
Account: ns191738
women (p = .066). Among replacement ministers of either sex, prior ex-
perience with elected office is less common. A difference, which was not
apparent in the data set as a whole, is observed in the initial cabinet: more

men held a prior appointed post (52% of men, 39% of women, p = .079),
while for replacement ministers 60% of both men and women held prior
appointed posts (p = .971).
Once again there are no statistically significant differences between men
and women for the organizational partisan or campaign involvement forms of
political background. However, both are less common among replacement
than among initial ministers. This comports with findings in the US lit-
erature about a change in appointment strategy over the course of the ad-
ministration (Mann and Smith 1981; Cohen 1988b: 113–16; Wyszomirski
1989: 47; Nicholls 1991; Borrelli 2002: 42).
Women were about twice as likely as men to come from a political family
(p = .004 initial cabinet, p = .163 replacement ministers). However, we now
observe that in the initial cabinet it is also more common for women (47%)
than men (33%) to be connected to the president (p = .046), while for replace-
ment ministers (both men and women) connections to the president are
less common (average 22%, p = .695).
Some examples illustrate differences between initial and replacement
ministers and how women appointed to initial cabinets appear more dif-
ferent from the men in their political background and connections. In
Costa Rica, Fernando Zumbado was appointed minister of housing at the
start of President Arias’s second term (2006–08). He had political ex-
perience in that he had previously been minister of planning (1977–78)
and minister of housing (1986–90), the latter during Arias’s first term as
president. His female replacement, Clara Zomer, differed mainly in her
sex as she also had political experience, entering the cabinet from a seat
in the legislature (2006–08), and she had also been executive president
of the National Training Institute (1994–98), an independent govern-
ment agency. In contrast, when María Ocaña was appointed minister of
health (2008–09) in President Fernández’s initial cabinet in Argentina,
she lacked much political experience (though she had been elected to
the Chamber of Deputies in 1999 she was not a member of President
Fernández’s party), but she was connected to the president and described
as becoming close to Kirchnerismo in 2003. Her replacement, Juan Luis

Manzur, had more political experience, holding several provincial posts
in the ministries of health, ultimately as the health minister in Tucuman
(2003–07), after which he was elected vice governor of Tucuman, and
unlike Ocaña was not known for being connected to President Fernández
or Kirchnerismo.
EBSCO[ 112 ] Women
Account: ns191738
Overall it is among initial ministers, not replacements, where men and
women differ in political background and connections. In the initial cabi-
net women appear less likely than men to bring the PCR of political skills to
the administration. This appears to be consistent with earlier findings for
the United States that women are appointed who will be dependent on the
president (Wyszomirski 1989: 66–69; Borrelli 2002: 55). This interpreta-
tion may be supported by the greater percentage of women than men in the
initial cabinet coded as connected to the president.8 However, when presi-
dents are replacing ministers or reshuffling their cabinet, expectations of
social control theory and about administration needs for PCRs seem to
apply, as based on all our measures the women are just as likely (or un-
likely) as the men to have some type of political background or connection.
Individual countries
This analysis reveals some ways that countries differ in norms for cabinet
ministers, and figure 5.3 also shows some differences with respect to incor-
poration of women into the cabinet. It is notable that we do not observe
significant sex differences within any country in the percentage of men and
women who are political insiders.9 Within countries women are as likely as
men to bring the PCR of political skills to the administration.
In the data set as a whole more men than women were observed to have
prior elected office experience, but here we see that a significant sex differ-
ence is found only in Chile (25% of men, 6% of women, p = .026). For both
sexes prior elected experience is fairly rare in Chile (18%), Colombia (26%),
and Costa Rica (30%) but quite common in the United States (43%); and
in Argentina more than half of ministers (58%) have prior elected experi-
ence. The prevalence of elected experience in Argentina fits with the mobile
nature of Argentine political careers (Jones et al. 2002; Francheset and
Piscopo 2014), and with the multiple election opportunities across levels of
government in a federal system. In the United States, which is also federal,
prior elected experience is less common, which may be attributable to the
8. For both men and women a higher percentage of initial than replacement minis-
ters were connected to the president: 47% of women appointed in the initial cabinet
were connected compared with 21% of replacements (p = .005); for men 33% of initial
ministers, 24% of replacements were connected (p = .070).
9. This is consistent with Borrelli’s (2002) findings, even with her more strict defini-
tion of political insider. While Dolan (2013: 85) finds that in President Obama’s first
administration more women than men are insiders, we do not analyze the Obama
administration.

ol i t
PMi cvia
Account: ns191738
Copyright © 2016. Oxford University Press. All rights reserved. May not be reproduced in any form without permission from the publisher, except fair
political insider (see note)
elected office *
appointed post *
campaign involvement †
connected to president ***
friend of president
political family * † *
0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100
uses permitted under U.S. or applicable copyright law.
Women Argentina (n = 94) Chile (n = 94) Colombia (n = 83) Costa Rica (n = 97) United States (n = 77)
Men
Figure 5.3 Minister political backgrounds—by country

EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017 2:21 PM via ECONOMIC COMMISSION FOR LATIN AMERICA AND THE CARRIBEAN (ECLAC LIBRARY)
AN: 1260012 ; Escobar-Lemmon, Maria C., Taylor-Robinson, Michelle M..; Women in Presidential Cabinets : Power Players or Abundant Tokens?
Account: ns191738
greater desirability of building seniority in Congress, raising the cost to a
politician of resigning a seat in Congress to join the cabinet.
In every country, statistically equivalent percentages of men and women

had previously served as minister or vice minister, though the percentages
of ministers of both sexes with that type of political background ranges
across our five countries from 31% (in Colombia) to 53% (Costa Rica). Prior
experience in appointed posts ranges from 46% in Chile to 62% in Costa
Rica. The high level of appointed posts in Costa Rica is consistent with the
country’s legal constraints on building a career in elected office, with its
ban on immediate reelection to the legislature and limited opportunities
for election below the national level. Only in Colombia do we find a signifi-
cantly lower percentage of women (32%) than men (63%) held an appointed
post before entering the cabinet (p = .017).
In none of our countries do a majority of ministers have background as
organizational partisans. High-ranking party background is most common
in Chile (43%), which makes sense given the intra-coalition bargaining and
balancing that has helped maintain the Concertación alliance (Carey and
Siavelis 2006). Yet, country-specific analysis reveals no significant sex dif-
ferences for the organizational partisan trait. The same is also true for cam-
paign involvement, except in Colombia where more women (37%) than men
(18%) have campaign experience (p = .096).
Women, more frequently than men, come from political families in
Argentina (p = .010), Colombia (p = .088), and the United States (p = .026). As
in the data set overall, there are no country-specific sex differences in the per-
centage of ministers who are friends of the president who appointed them. We
do, however, observe a large and significant sex difference in the percentage
of US secretaries who are connected to the president: 79% of women but only
22% of men (p = .000).
Country-specific analysis reveals some differences across countries in
the types of political experience that appear to be the norm for cabinet
service. However, we do not observe within any single country’s data set
multiple measures indicating that women appointees bring less political
experience to the cabinet than do the men. In Argentina, Colombia, and
the United States women are more likely than men to have a political con-
nection, which may be an important attribute for a woman being consid-
ered for appointment to one of these scarce posts. Yet the women who are
appointed are typically as likely as the men to have the type(s) of political
credentials that are the norm in their country, and no country exhibited
significant sex differences on our proxy for the PCR of political skills. Thus,
in each country the women appear to conform to the male norm. Analysis
in the conclusion chapter will allow us to assess if this is paired with the

ol i t
PMi cvia
Account: ns191738
women being as effective in their post as the men, which is the prediction
in the upper-left cell of fi
gure 1.1.
DO POLITICAL QUALIFICATIONS VARY ACROSS CATEGORIES

OF POSTS, AND DO WOMEN AND MEN BRING THE SAME
POLITICAL SKILLS TO EACH TYPE OF POST?
So far our analysis indicates that women as a group may have somewhat
less political experience but are more likely than the men to have politi-
cal connections. Those sex differences, however, are only apparent in the
initial cabinet, and in no country do we observe multiple dimensions on
which women are less likely than men to have political experience. Yet
cabinet portfolios have different policy purviews; some portfolios are
high profile and must tackle contentious issues while others have a lower
profile; some portfolios are inherently political in the tasks they perform
while for other portfolios the president may desire (or the norm may be)
political neutrality and an emphasis on technocratic policy expertise. In
sum, the need for political skills may vary across different types of posts
(Wyszomirski 1989: 46–47). We now turn to an examination of political
background and connections and the capacity of ministers to bring the PCR
of political skills to the cabinet within the several broad categories of port-
folios described in c hapter 3.
Economics/S ocial Welfare/C entral categories of portfolios
Dividing portfolios into these three broad categories yields several instruc-
tive findings compared to the data set overall, as well as when compared
to the analysis in c hapter 4 in which we saw that women appointed to
Economics category posts, less frequently than men, brought the PCR of
policy expertise to the administration. What is notable in figure 5.4 is the
near absence of statistically detectible sex-based differences in political
background in any portfolio categories. The only form of political experi-
ence where we observe a significant sex difference is in the Economics cat-
egory of portfolios, where more men than women have prior elected office
experience (28% of men, 13% of women, p = .071). It is also noteworthy
that unlike in the data set overall, and in the subset of initial appointees
where fewer women than men were political insiders, within each of these
three broad categories of portfolios, the percentages of men and women
who were political insiders did not differ statistically. We do not find sex
EBSCO[ 116 ] Women
Account: ns191738
elected office †
appointed post
connected to president **
friend of president * ** *
political family **
0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100

Women
Economics (n = 172) Social Welfare (n = 124) Central (n = 138)
Men
Figure 5.4 Minister political backgrounds—Economics/Social Welfare/Central categories of portfolios

Account: ns191738
differences with regard to our proxy for the political skills PCR. However
the incidence of political insiders varies across categories: 43% of minis-
ters in the Economics category, 51% for Social Welfare, 71% for Central
category posts.10
With respect to political connections, on the other hand, we observe sev-
eral significant sex differences. In the Social Welfare category more women
than men are connected to the president (p = .001), a friend of the president
(p = .003), or a member of a political family (p = .006). This difference is worth
mentioning and something worth exploring in future research because the
Social Welfare category contains many of the posts where women have
traditionally been appointed. By contrast, a significantly greater percent-
age of men than women are a friend of the president in both the Economics
(p = .045) and Central categories (p = .013), and no women appointed to
Economics or Central category posts have this type of political connection.
Overall, it appears that men and women in each of these broad catego-
ries of posts are equally likely to bring political skills to the cabinet. Yet the
percentage of ministers (male and female) with political experience, and
the types of experience they possess, vary across categories. For example,
45% of ministers in the Central category are organizational partisans, but
only 22% of Economics category ministers have that type of political expe-
rience; 50% of ministers in the Central category held elected office, but only
25% of Economics category ministers had done so; and 72% of ministers in
the Central category had held a previous appointed post, compared to 50%
of Social Welfare ministers and 43% of Economics category ministers. This
supports the expectation outlined above that norms about credentials that
enable a minister to do their job vary across types of posts, but that the
same traits signal competence in male and female ministers. Ministers in
the Central category—both men and women—look like “politicos,” while
men and women appointed to Economics category posts are more likely to
be “técnicos” (Camerlo and Pérez-Liñán 2013), which makes it interesting
that the Economics category is where we observed that significantly more
men than women had experience in elected office. In chapter 4 we saw that
Social Welfare category ministers—both men and women—typically have
10. The percentage of insiders in Central category portfolios is significantly differ-

ent from Social Welfare (p = .002) and Economics categories (p = .000) based on both
the Scheffé and Bonferroni multiple comparison test. That fewer ministers (men and
women) in Economics category posts would be political insiders comports with find-
ings for parliamentary cabinets in Western Europe since the Second World War (where
virtually all of these posts were held by men). De Winter (1991: 63) reported that “the
large economic portfolios (treasury, the economy, industry, foreign trade, energy) are
less likely to be held by parliamentarians; these posts tend to be filled by economic
experts recruited from outside the world of politics.”
EBSCO[ 118 ] Women
Account: ns191738
policy expertise. Here we see that political connections are more common
for women than for men in Social Welfare category posts (though still only
a minority of the women in Social Welfare posts have political connections).

As we can see, political connections manifest themselves in different
ways. Michelle Bachelet and Pedro Garcia Aspillaga, both ministers of health
in Chile (2000–02, 2003–06 respectively), have similar political experience
(she had worked as an advisor in the ministries of health and defense, he
worked in various government jobs related to health such as Director of
Metropolitan Health Services), but she was known for her connections to
President Lagos, and had helped run Lagos’s campaign, while he was not
connected to Lagos and came from a different party within the Concertación
coalition. In the United States Roderick Paige and Margaret Spellings both
served as secretary of education for President Bush (2001–05, 2005–09
respectively). Both came from Texas and worked in education in the state
(Paige as superintendent of Houston schools, Spellings on a state education
reform commission), but Paige did not have known formal connections to
President Bush, while Spellings had been the political director of Bush’s first
gubernatorial campaign and then worked as a senior advisor when Bush
was governor. In both examples, both male and female ministers had policy
expertise, but only the women had political connections. Given that Social
Welfare category posts reflect role congruity and are where women are most
commonly found and have been present in cabinets for the longest amount
of time, it is intriguing that this is the category of portfolios where women
are over-represented for having political connections.
Stereotypically masculine/f eminine policy domain portfolios
In contrast to chapter 4, where our comparison of masculine/feminine

policy domains underscored how these categories are different from
Economics/ Social Welfare/Central categories of posts, the findings in
figure 5.5 are very similar to the findings in the previous section. There are
only small changes compared to the previous section in the percentages
of men and women in each broad category who have each type of political
background or connection.
From this analysis we can conclude that women who are appointed both
to posts that fit stereotypes about appropriate policy areas for women
(i.e., policy areas where women are typically evaluated to be competent;
see Alexander and Anderson 1993; Huddy and Terkildsen 1993a, 1993b;
Matland 1994; Koch 2000; Lawless 2004b), and women appointed outside
of stereotypically feminine policy domains are as likely as men to bring the

ol i t
PMi cvia
Account: ns191738

elected office *
appointed post
friend of president ** **
political family † *
0 20 40 60 80 100 0 20 40 60 80 100
Women
Men Masculine (n = 337) Feminine (n = 110)
Figure 5.5 Minister political backgrounds— stereotypically masculine/ feminine policy

domain portfolios
PCR of political skills to the cabinet. In addition, women holding posts with
a stereotypically masculine policy domain resemble the men with regard to
their political background, with the exception of experience in elected office
(25% of women, 40% of men, p = .030).
Ministers in High Visibility category posts
Figure 5.6 shows that there are no measures of political background

where significantly different percentages of women and men who hold
High Visibility category posts have (or lack) a type of political experience
(p-values range from .509 to .964). It is also noteworthy that, with the ex-
ception of insiders (which are more common in High Visibility category
posts),11 political background traits appear to be no more common for min-
isters holding High Visibility category posts than for ministers whose posts
are not in the High Visibility category. Given that many ministers in our
data set lack political experience and because High Visibility category posts
are where extensive political experience would seem most necessary to do
11. Among the group of ministers whose posts were known to be “high visibility”
before they were appointed 59.6% are coded as insiders, while 50.8% of ministers in
posts that were not “high visibility” are insiders (p = .089).
EBSCO[ 120 ] Women
Account: ns191738

elected office **
appointed post
friend of president * *
political family * *
0 20 40 60 80 100 0 20 40 60 80 100
Women
Men High Visibility (n = 138) Not High Visibility (n = 309)
Figure 5.6 Minister political backgrounds—High Visibility/Not High Visibility categories

of posts
the job because the public is expressing concern and thus, the post is likely
to receive much attention in the media and congress, the lack of differences
between High and Not High Visibility category posts is intriguing.
With regard to political connections, for ministers in High Visibility
category posts a significantly greater percentage of women than men are
connected to the president (p = .005), friends of the president (p = .022), or
come from political families (p = .010).12 This could be interpreted as indicat-
ing that women are more likely to receive High Visibility category posts if
they are women who the president knows and who should thus be expected
to remain loyal to the president under pressure. It may also be that the
only female names that make it onto the list in times of crisis are those
the president already trusts. For men, in contrast, a political connection is
less common and appears to be less critical for their appointment to High
Visibility category posts.
Some examples illustrate this sex difference in connections. Janina del
Vecchio served as minister of public security in Costa Rica when it was a
High Visibility category post (2008–10). She had political experience in
12. For posts that are Not High Visibility women are more likely than men to come
from a political family (12% of women, 5% of men, p = .047). Men are more commonly
friends of the president than women (15% of men, 5% of women, p = .013). There is no
significant sex difference for connected to the president.

ol i t
PMi cvia
Account: ns191738
the legislature and as an ambassador. She had also served as cabinet chief
during Oscar Arias’s first term as president and thus was coded as con-
nected to the president when he appointed her as a replacement minister.

In comparison, her predecessor, Fernando Berrocal Soto, had previously
been a minister and an ambassador, though neither appointment was
during Arias’s first term as president. Unlike del Vecchio, he was not coded
as connected to the president, friend of the president, or from a political
family. Another example of women having political connections comes
from Colombia’s Ministry of Labor. Gina Riano, appointed to the Labor
post by President Pastrana (1999–2000), came from a political family
(her father was a Conservative Party leader in their province) and her ap-
pointment was described as helping Pastrana build links to the province
of Boyoca. Angelino Garzón, who subsequently served as minister of labor
for Pastrana (2000–01) was described as an independent and he came to
the post from the Communist union movement, an ex-guerrilla, and from
the left-wing M-19 party.13 Thus, although Riano had political connections,
Garzón appeared to bring a very different set of connections.
Overall in this analysis, as in the other two analyses of post categories,
there is evidence women are well integrated into these cabinets. A greater
percentage of the women appointed to High Visibility category posts have
some sort of political connection, but the women are as likely as the men to
be a source of the PCR of political skills.
DO MINISTERS WITH POLITICAL CONNECTIONS

HAVE POLICY EXPERTISE?
As an additional exploration of whether women and men are held to differ-

ent standards for appointment to a full cabinet-rank post, we look at whether
ministers who have political connections also have policy expertise. Only a
minority of the ministers in our data set have political connections (30%
are connected to the president, 13% are coded as friends of the president, 7%
come from political families). However, since political connections are more
common for women than men,14 we are curious whether the women with
13. Latin America Weekly Report, July 18, 2000.

14. A greater percentage of women ministers in our data set are coded as connected
to the president (35.5% of women, 28.5% of men, p = .171) or as coming from a political
family (13.6% of women, 5.1% of men, p = .003). The percentage of men in our data set
who are coded as friend of the president is slightly larger than the percentage of women:
13.5% of men, 10% of women (p = .342). Carroll (1984: 103), in her study of appoint-
ments to the cabinets of US state governors using data from 1981, measured loyalty in
EBSCO[ 122 ] Women
Account: ns191738
these connections are also qualified for their post with regard to policy ex-
pertise, or if these women are being appointed to posts unrelated to their
education and work background (which would make them more likely to be
tokens in our classification scheme in figure 1.1). This is a topic of concern
because studies of women in the US cabinet in earlier administrations found
the women often lacked “the political resources to act as independent power
brokers” (Borrelli 2002: 55), or “possessed few administrative, expertise, or
political assets that might have enhanced their utility to or influence with the
president” (Wyszomirski 1989: 67). Borrelli (2010: 737) explains that women
are marginalized in a gender-desegregated cabinet, because “[h]‌aving so few
political resources leaves the women secretaries relatively more dependent
on the White House, limiting their power and sending a reassuring message
to those who resist women’s leadership in the public sphere.” This is par-
ticularly likely when “[f]emininity is reinscribed through the appointment
of women to departments with jurisdiction over ‘women’s issues,’ such as
health or education, even when the woman appointees lack related policy or
political expertise.” In essence, they may have been appointed for their likely
loyalty to the president rather than to bring PCRs to the administration.
In the literature on the US cabinet, loyalty is generally thought to come
from working on the president’s election campaign. Loyalty also encom-
passes commitment to the president’s ideology or policy program (Carroll
1984: 101; Weko 1995; Edwards 2001: 83; Auer 2008; Borrelli 2010: 735).
Because our study is cross-national, we need a more general definition of
loyalty to the president that is applicable across different kinds of electoral
and party systems. Thus, we utilize political connections as proxies for loy-
alty especially as we lack the detailed data from each minister to assess
whether they were motivated to join the cabinet because of an ideological
commitment and partisanship is not an appropriate proxy.
While in many partitions of our data set a greater percentage of women
than men have political connections; we see in Table 5.1 that ministers who
have political connections are no less likely than ministers without them to
have extensive experience related to their post. This is true for both women
and men for each type of political connection, and it indicates that women
who have political connections are no less likely than men who have the
same type of political connection to bring the PCR of policy expertise to
the administration.
several ways that are distinct from the measures we use here, including whether the
cabinet member had the same party affiliation as the governor. She found that “women
appointees more often than all appointees shared the party affiliation of the governor,”
which she interpreted to suggest “that governors more often looked for a basic guaran-
tee of loyalty when selecting women for their administration.”

ol i tPMi cvia
Account: ns191738
Table 5.1 DO MINISTERS WITH POLITICAL CONNECTIONS ALSO HAVE
POLICY EXPERTISE?
% of ministers with extensive experience related to

the policy area of their portfolio
(# of ministers in category)
Women Men p-value for row
Type of political connection

Friend of the president
Yes 73% (11) 59% (44) 0.415
No 60% (99) 65% (289) 0.328
p-value for column 0.402 0.444
Connected to the president
Yes 67% (39) 59% (94) 0.381
No 58% (71) 66% (238) 0.180
Political family
Yes 53% (15) 53% (17) 0.983
No 62% (95) 65% (316) 0.619
Some examples illustrate the cases presented in table 5.1. Graciela

Camaño (2002–03) and Alfredo Atanasof (2002), ministers of labor for
President Duhalde in Argentina, both had connections and policy expertise.
Both were known as “Duhaldistas,” and both also had decades-long con-
nections to organized labor. Colombia provides an example where only the
woman had political connections. Consuelo Araújo Noguera was appointed
minister of culture by President Pastrana (2000–01). She had well-known
qualifications for the post due to her publications on Vallenata culture and
her work toward creating a Vallenata music festival. She also came from
a political family, with her father known as a Liberal Party leader in their
province, and she is the aunt of a former senator and of another minister.
Juan Luis Mejía Arango, who served as minister of culture (1999–2000)
also had extensive post-relevant experience (e.g., published works on his-
tory and culture, prior work as director general of the Colombian Institute
of Culture), but he was not known for his connections to President Pastrana
or for his family connections to politics.
In sum, most women with political connections also have policy exper-
tise related to their post. Their political connections may have been what
prompted the president to consider them for an appointment, but they
EBSCO[ 124 ] Women
Account: ns191738
are not likely to appear unqualified for their post if the public or the media
assess qualifications based on the minister’s education and work expe-
rience related to the portfolio. This is a very positive finding in light of

unflattering speculation surrounding the appointment of women with as-
sumptions that they are the wives, girlfriends, or mistresses of presidents
and that they do not “deserve” the post.
CONCLUSION
The findings in this chapter comparing the political experience and connec-
tions of men and women in these presidential cabinets indicate that inte-
gration of women is occurring. The term “gender integration” (as opposed
to “gender desegregation”) comes from the work of MaryAnne Borrelli
(2010) from her study of women in the US cabinet. Borrelli (2010: 737–38)
argues that integration is occurring if women bring the same types of po-
litical experiences and group connections to the cabinet and are appointed
to posts for which they have appropriate background. Integration is ex-
pected to be important for representation of women because when women
in the cabinet have extensive political resources they have their own power
base and a greater capacity to get things done (Borrelli 2002; Annesley and
Gains 2010). The analysis here and in c hapter 4 indicates that women are
as likely as men to have extensive post-relevant experience. The women
are also as likely as the men to have political experience. This means that
the women can bring the PCRs of policy expertise and/or political skills to
the cabinet.15 In addition, while it is more common for women than men
to have political connections, the women with political connections typi-
cally also have policy expertise. This is congruent with Franceschet and
Piscopo’s (2012: 44) findings from their comparison of the background of
men and women in the Argentine Congress. They found that legislators of
both sexes have similarly high levels of education and are political insiders
based on their extensive political experience, and conclude that this evi-
dence “undermines claims that quota women are tokens.”
15. Our finding that women are as likely as men to have post-relevant policy ex-
pertise and political skills may appear to clash with the typology of ministers that is
common in the parliamentary cabinets literature that refers to “generalist” and “spe-
cialist” ministers (see for example Blondel 1991; Davis 1997). However, the parliamen-
tary literature acknowledges that ministers with extensive political skills may also be
policy experts, as for example in the Austrian cabinet where Gerlich and Müller (1988)
categorize ministers as “politicians,” “politically based experts” and “independent
experts.”

ol i tPMi cvia
Account: ns191738
Overall with regard to political background the women in these presi-
dential cabinets conform to the male norm. They are not appreciably more
likely or less likely than the men to have any particular type of political
experience, with the exception of prior experience in elected office. This
lower incidence of women ministers with experience in elected office may
be important for their political power and for the administration. Borrelli
(2010: 744) points to women secretaries becoming like the men in their
elected office experience during the William J. Clinton and George W. Bush
administrations as a sign of gender integration. The large number of women
appointed to the initial Obama cabinet who were incumbents in elected
office was significant because it was the first time women secretaries were
incumbents in elected office (“meeting this male norm”), and because she
views the electoral alliances of cabinet ministers as an important resource
for the administration (p. 716).
The women in these presidential cabinets mainly differ from the men,
and then only in some subsets of our data, in that they are more likely to
have political connections. Regarding the women appointed to the initial
cabinet in the Obama administration Borrelli (2010: 746) writes, “That the
credentials of the women and the men secretaries were becoming more
similar was not merely evidence of women’s ability to match the accom-
plishments of their men colleagues, or of their willingness to conform to
masculine standards. The similarity was also indicative of the president’s
endorsement of highly credentialed women as cabinet secretaries.”
With regard to our theoretical predications in figure 1.1, these findings
indicate that women have backgrounds that are similar to the men. We
expect embedded political elites, usually men, to want to appoint people
like themselves (even if their appointees are women) as a way to protect
against loss of benefits for their political interests and because history in-
dicates the administration will need those types of PCRs. Such minimal
change is more likely to mean elite turnover than true elite circulation
(Prewitt and Stone 1973: 168) and situates on the top row in figure 1.1
(where backgrounds are the same). We predict that these women will be
treated like the men—be equally effective (upper-left cell of figure 1.1) if
integration of women is really happening in these cabinets. Whether treat-
ment is equal or not remains to be tested in Part III of the book.
EBSCO[ 126 ] Women
Account: ns191738
CHAPTER 6
Support resources
When and where presidents appoint ministers
with links to clients of their department
A long with policy expertise and political skills, support resources also
help the administration implement the president’s policy agenda and
confront political problems and crises. Organized interests and sectors of
society are often important constituencies for an administration, and links
to groups help the president, or the president’s party to strategize legisla-
tive initiatives and reelection (Loomis 2009) though different constituen-
cies are salient for different presidents. US cabinet studies often highlight
how Secretary X’s links to a particular group helped the administration
build bridges to key segments of the population. Appointments signal
which groups and organizations will have the president’s ear, especially in
portfolios where different interests may be in conflict (e.g., mining compa-
nies and environmental activists in the mining, energy, and environment
portfolio). For example, Hilda Solis, appointed by President Obama as sec-
retary of labor at the beginning of his first term, had well established links
to organized labor and experience winning elections from California to the
US House of Representatives, both of which helped her build bridges to
labor and to the West Coast region of the country. Yet such connections are
quite new for women, while they have been the norm for men in the cabi-
net in the United States (Borrelli 2010: 740, 746). US cabinet studies that
analyze sex differences in the credentials of secretaries have also found
that women generally have lacked linkages to groups that are clients of
their agency, while it has been more common for men to be group liaisons.
EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017 2:21 PM via ECONOMIC COMMISSION
Account: ns191738
The lack of links (at least until recently) between women cabinet secretar-
ies and clients of their agency has been interpreted as a sign that women
are not being integrated in the US cabinet (Borrelli 2010).

Group links constitute the third component of our analysis of the back-
grounds and connections of cabinet ministers, to determine if the women
that are now appointed to presidential cabinets in greater numbers and
to more diverse portfolios have credentials similar to the men’s. Through
partitions of the data (following the same format as in the preceding two
chapters) we examine whether patterns hold in ministries with links to
occupations where women have been involved longer or are related to
areas where we might expect women to have a special interest. With this
analysis we also complete our assessment of whether women and men
are equally likely to bring political capital resources (PCRs) to the cabinet.
Group links are a source of the PCR of support resources for the admin-
istration, and ministers can have links to various sectors of society and
organized interests, including possibly groups that are viewed as clients
of their ministry.1
We start by comparing men and women in the data set overall to deter-
mine whether women in general are equally likely to have group links. We
then split the data set into initial cabinet and replacement ministers, because
appointment strategies may change over time, which may affect both sexes
equally or have a differential impact on women compared to men. Then we
look at individual countries to see if women in some countries have links
to ministry clients or certain types of group links while in other countries
such linkages are less common for women than for men, or if in some coun-
tries a particular type of linkage is uncommon for both women and men.
Finally, we divide cabinet portfolios into several broad categories to see if
women and men who hold similar types of portfolios have similar types of
group links and are equally likely to bring the PCR of support resources to
the administration.
We might expect group linkages to be particularly important for port-
folios that have clear constituents (what in US politics are often called
“constituency-oriented agencies”), while ministries that handle national
needs (e.g., justice, foreign affairs) may not have organized groups to which
they regularly need to build bridges (see Wyszomirski 1989: 50, 54). The need
for group links may vary across categories of portfolios. We expect women to
be equally as likely as the men within the same category of posts to have
group links and that the women will closely resemble the men in the types of
1. We use the term “clients” broadly, including groups who are dependent upon a
ministry for resources, interest groups, and businesses regulated by the ministry.
EBSCO[ 128 ] Women
Account: ns191738
group link credentials they bring to their posts. If group links are necessary
for the administration to implement its policy agenda and handle crises, the
administration will need its female appointees to be as able as the men to

draw on those resources or the legacy of the administration may be harmed.
An alternative hypothesis is that women ministers appointed to head minis-
tries in policy areas where women have traditionally been employed or where
women have traditionally been seen as having an interest (home and hearth)
will be more likely to bring support resources than women in other types of
posts, which may affect their capacity to be successful.
We conclude this chapter by bringing together chapters 4–6 to explore
how common it is for ministers to bring more than one type of PCR to the
administration. In particular we want to know whether women are as likely
as men to have more than one type of PCR. This will allow us to assess
whether presidents are appointing highly credentialed women to their cab-
inets who have the potential to be powerful actors in the highest levels of
the executive branch.
To preview our findings, with regard to group links we observe a high level
of similarity between the men and women, in our data set overall, and within
broad categories of portfolios. Women are as likely as men to have links to
clients of their ministry, with a few exceptions. Women are, however, notably
less likely than men to have links to business, which is particularly a concern
for women appointed to Economics category posts. We also observe that
very few women appointed to these presidential cabinets have links to wom-
en’s organizations. Finally, our summary analysis shows that it is rare for
ministers—men or women—to have all three types of PCRs; however, men
and women are equally likely to bring multiple PCRs to the administration.
CODING SUPPORT RESOURCES AND GROUP LINKS
We constructed minister bios based on publically available information.

A minister may have a link to an interest group that is not publicly known
(for example, the minister could be a dues-paying member of the Sierra
Club and thus have some quiet environmental interests). However, if cabi-
net appointments are a way for the president to signal, both in their own
country and internationally, which sectors of society or interest groups are

valued constituents of the administration, those signals have to be public
signals. Thus, when it comes to a minister bringing support resources to
the administration, publicly known group links are an appropriate measure
of a minister’s political capital. For this analysis we coded six measures of
S u pp2:21
EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017 or tPMRvia
e s oECONOMIC [ 129 ]
u r c e s COMMISSION
Account: ns191738
ministers’ group links and we use links to clients of the ministry as a proxy
for the PCR support resources.
Group links. Ministers can obtain group links in various ways. A minister
can have a formal association with a group (e.g., as a member or president
of the National Chamber of Agricultural Producers, labor union leader, af-
filiation with an environmental or women’s organization). Work may have
brought a person into regular and close contact with international aid or-
ganizations, lending agencies, or with business by serving on the boards of
various companies.2 A minister can be a member of professional associations
(e.g., the Colegio de Médicos, association of civil engineers). In addition, if
the press touted a new minister as “well known within the business commu-
nity” or “applauded by environmentalists” we coded them as having a link
to that group. We are interested in identifying how these links build connec-
tions with groups in society, so we do not treat political parties as a group
link. Linkages to parties are important, but we capture them in the political
experience traits covered in chapter 5. Similarly, we do not treat previous
appointments in the executive branch, state government, or prior elected ex-
perience in the legislature as building links, since those traits are captured in
the political experience chapter. Additionally, while we recognize that many
appointees are practicing members of religious organizations, we do not code
this membership in any systematic way. Religious affiliation may make a cab-
inet member sympathetic to certain policy positions or issues, but appoint-
ing someone who is simply religiously observant (even if they might be seen
at religious services), is not expected to formally build ties with that group.
A similar example would be the appointment of someone who consistently
donates money to a particular cause, but if their interest in the cause is not
publicly known it does not build formal links to that organization.
In the analysis presented here we focus on six types of group links: (1) links
to any group (283 ministers, 64.4%). We track two types of linkages that are
relatively common: (2) links to business (115 ministers, 26.2%), and (3) in-
ternational links—meaning links to international or multinational organi-
zations (159 ministers, 36.2%). We also track links to (4) women’s groups
and (5) organized labor, but only sixteen ministers have links to women’s
groups and twelve have links to labor. Finally, we pay special attention to
(6) links to ministry clients—capturing if any of the minister’s known group
associations overlapped with the policy purview of their portfolio (185

ministers, 42.1%). If a portfolio encompasses several sectors (e.g., energy,
mining, and environment) a minister only needed links to one sector to
2. Borrelli (2002: 70, 89n5) also considers service on corporate boards as evidence a
cabinet secretary has business issue networks.
EBSCO[ 130 ] Women
Account: ns191738
qualify as having client links.3 Ministerial links are not mutually exclusive
and ties to one organization could count in more than one category: for in-
stance, serving on corporate boards would count as links to business. If the

minister had been appointed to finance/treasury that service would also
provide him or her with links to ministry clients. We use links to ministry
clients as our proxy for the PCR of support resources. Although a minister
may have links to groups that are not clients of their ministry, and thus
could still bring support resources to the administration, if the minister’s
group links cannot be served from their post this is likely to create con-
flicting interests and could diminish both their capability in their post and
their ability to mobilize group support for the administration.4
DO MEN AND WOMEN DIFFER IN THEIR GROUP LINKS?

The data set overall
In the data set overall the women look very much like the men with
regard to their group links, though there are some noteworthy exceptions
(see figure 6.1).5 First, regarding the PCR of support resources, it is im-
portant for equal treatment of women in these cabinets that the percent-
ages of women (45%) and men (41%) coded as having links to ministry
clients did not differ statistically (p = .494). These percentages may un-
derestimate the percentage of ministers with links due to the impossibil-
ity of uncovering all links, though determining all client links should be
equally difficult for women and men; thus, what errors exist are likely to
be random rather than systematic by sex. Moreover, the similarity of per-
centages observed indicates that women in general are no more likely than
men to be appointed to buffer the administration from a group that is not
viewed as a constituency of the administration (Borrelli 2002: 9; 2010).
3. We were unable to obtain enough biographical information for eight ministers to
code group links. One is a woman, and she is from Colombia. The remaining seven are
from Argentina (1), Colombia (5) and Costa Rica (1).
4. Borrelli (2010: 735) in research about the US cabinet argues that a danger for
the president is that a minister will be captured by constituents of their agency or by
the bureaucracy of their agency. To counter this problem US presidents have central-
ized decision making in the White House and screened appointees for loyalty to and
overlapping ideology with the president (also see Weko 1995; Andeweg 2000: 390–91;
Edwards 2001; Auer 2008; Lewis 2008; Bäck et al. 2012: 4).
5. Detailed tables for all group links analyses presented in figure 6.1 and subsequent
figures in the chapter, including the number of ministers coded on each trait, the
percentage of women and men with each trait, and the p-values for the difference of
means tests are available from the authors on request.
S u pp
2:21 e s oECONOMIC [ 131 ]
Account: ns191738
Source of Support Resources
Minister Group Links
ministry clients (see note)
business ***
international
women’s groups ***

organized labor
any type of group link

0 20 40 60 80 100
Women
All ministers (n = 447)
Men
Figure 6.1 Minister group links—data set overall

Note: The variable ministry clients is the proxy for support resources in this study.
It also means that these women are as likely as the men to be able to bring
the PCR of support resources to the administration.
Second, statistically indistinguishable percentages of women and men
had international links (p = .735), and links to any type of group (p = .690).
However, significantly fewer women than men have links to business (12%
of women, 31% of men, p = .000). That could impact the credibility of the
women in certain types of posts. While business links may not be crucial
in some posts (e.g., health or education where other types of groups are
the ministry’s clients), business links could be critical for job performance
in Economics category portfolios (we explore this later in the chapter). It
may also be important for the empowerment of women as ministers that
women bring business links to the administration. If the administration
needs backing, or acquiescence, from business to make and implement its
policies, women may be less important players in the administration if they
are not sources of support resources from business. In this scenario women
would be dependent on the president and have greater potential to be con-
trolled by the president (Borrelli 2010; Wyszomirski 1989: 69).
Third, all the ministers with links to women’s groups were women.
Although that may not be surprising, the fact that only 15% of women
ministers have links to women’s groups (sixteen women) is striking.6
Women who are not known to be affiliated with women’s organizations
6. Examples of the types of women’s organizations to which these women ministers
are known to be affiliated include a leader in a provincial council on women (Argentina),
a researcher at the Casa de la Mujer de Valparaiso (Chile), an activist for women’s em-
powerment (Colombia), a United Nations gender working group (Costa Rica), and the
National Women’s Law Center (United States).
EBSCO[ 132 ] Women
Account: ns191738
could certainly represent women’s interests; however, the ability of the
largely elite and white women7 that are appointed to the cabinet to rep-
resent the diverse types of women and intersectional identities found

in society is a topic of debate (see, for example, Mansbridge 1999: 641;
Young 2000: 136–37; Weldon 2002; Vincent 2004: 74 and 79–81; Childs
2006: 9–11; Htun 2014).8 Still, the small percentage of women in these
cabinets with known links to women’s organizations indicates that these
presidents were not particularly interested in building bridges to organized
women for political support or to signal that women are an important
constituency for their administration (Borrelli 2002: 70). Wyszomirski
(1989: 66) commented on the lack of women in US cabinets who “could
be considered to be representatives of women in any politically organized
sense. Rather they were symbolic representatives of the gender constit-
uency.”9 This would be consistent with social control theory and might
suggest that the women viewed as desirable to appoint will be those who
may not be strong advocates for women’s interests and instead preserve
the status quo while appearing to signal change. Although these cabinets
have appreciable numbers of women appointees (compared to the ex-
treme under-representation of women in the past), most of the women
apparently have not built publicly known links to women’s groups. Given
the lack of women with connections to women’s organizations in our data
set overall, it is perhaps unsurprising that Franceschet (2005: 86) makes
note of the “absence of stronger links between women in formal politics
[in Chile] and those active in the informal political areas (for example,
women’s movements)…”
7. A few men and women in our data set are racial minorities. However, the numbers
of racial minority ministers are small enough that a separate consideration of them is
not possible and a qualitative evaluation is beyond the scope of the study.
8. Franceschet and Piscopo (2012: 44) describe the “double bind” that women politi-
cians face (though they are referring specifically to women elected to legislatures via
gender quotas). They explain that if the women “do not possess similar levels of edu-
cation and political experience to those of men, then they are deemed less deserving
of their positions and treated as tokens. On the other hand, if female legislators are
drawn from the political class, they are criticized for being too elite to represent female
voters adequately.”
9. Davis (1997: 21) also commented on women being appointed as “symbols” rather
than as true representatives of women. She gave the example of the first woman
appointed by President Giscard d’Estaing to lead the women’s ministry in France.
Francoise Giroud was a well-known journalist, but conservative and with limited con-
nections to women’s organizations. She explained that “French feminists argued that
the creation of the post and her appointment were largely a ‘public relations exercise’
rather than a sincere effort to address the concerns of women.” Vincent (2004: 88) and
Franceschet (2005: 104) explain how women in the South African and Chilean cabinets
feel pressure to not advocate for women or identify with the women’s movement.
S u pp2:21
Account: ns191738
Fourth, very few ministers—men or women—have known links to or-
ganized labor: 2.8% of women, 2.7% of men (p = .989), which translates to
three women and nine men. Although we do not observe a sex difference
with respect to labor links, the near absence of ministers with connections
to labor does appear to indicate a lack of interest on the part of these presi-
dents in obtaining support resources relating to organized labor. However,
nine of the twelve are appointed to ministries in the labor and social secu-
rity category, so around 35% of ministers of labor and social security do
have links to labor unions.
An example of a woman entering the cabinet with client links would be
Luisa Fernanda Lafaurie, minister of mines and energy in Colombia (2001–
2002). She worked in the energy sector, which gave her links to ministry
clients via corporate boards of directors, and it also meant that she had
business links, which make her unusual among female ministers. In con-
trast, Jenny Philips Aguilar, minister of finance in Costa Rica (2009–2010),
had extensive experience related to her post, including serving as vice min-
ister; however, like many women ministers, she lacked known connections
to business groups and thus lacked links to clients of her ministry. A typical
man could be Álvaro Rojas Marin, minister of agriculture in Chile (2006–
2008). He was educated in veterinary medicine and agricultural sciences
and was a professor of agricultural economics. He also worked as a consul-
tant for the FAO and other international organizations, giving him links
to ministry clients and to international groups. Similarly, Esteban Brenes
Castro, minister of agriculture in Costa Rica (1998–2000), had a doctor-
ate in agricultural economics and worked in agricultural business and con-
sulting. His links to ministry clients, and to business, came from his work
on the boards of directors of agricultural companies and as director of the
National Chamber of Agriculture and Agro-Industry.
Comparison of initial and replacement ministers
When we split the data set into the initial cabinet and replacement minis-
ters, we observe the same patterns as in the data set overall (see figure 6.2).
Women were as likely as men in both the initial and replacement subsets of
the data set to have links to clients of their ministry (p = .571 initial, p = .726
replacement). There is also sex equality in the likelihood of having inter-
national links (p = .725 initial cabinet, p = .854 replacement), and connec-
tions to any type of group (p = .221 initial, p = .418 replacement). In both the
initial and replacement groups it is less common for women than men to
have links to business, so it appears that most women in these cabinets lack
EBSCO[ 134 ] Women
Account: ns191738

business ** **
international
women’s groups *** ***
organized labor
0 20 40 60 80 100 0 20 40 60 80 100
Women Initial (n = 237) Replacement (n = 209)
Men
Figure 6.2 Minister group links—initial/replacement ministers

credentials to connect the administration to business, regardless of when

they are appointed during the term. In the initial cabinet 32% of men, and
13% of women had business links (p = .004). For replacement ministers 30%
of men, and 11% of women had links to business (p = .008). The percentage
of women ministers with connections to women’s groups decreases slightly
from 16% in the initial cabinet to 13% among replacement ministers, but the
difference between men and women remains statistically significant in both
sub-samples.
The following ministers illustrate this pattern of initial and replace-
ment ministers having similar types of group links: in Colombia, Cecilia
Rodríguez Gonzalez was the initial minister of environment, housing and
territorial development in the first Uribe administration, serving from
2002–2003. She had links to business through her membership in the
Baranquilla Chamber of Commerce and to environmental groups through
BioParque, the latter giving her links to ministry clients. Ramiro Valencia
Cossio was a replacement minister of mines and energy in Colombia, serv-
ing in 2001. He had links to business from serving on the boards of directors
of companies and links to ministry clients from his work with Ecopetrol.
In Argentina, Romina Picolotti was President Fernández’s initial minis-
ter of environment and sustainable development (originally appointed by
President Kirchner, though before the post was full cabinet rank), serving
from 2007–08. She had links to international organizations working on

both the environment and human rights, and these environmental group
connections (e.g., leader of the Center for Human Rights and Environment,
the Committee for Environmental Law in Bonn, Germany) gave her links to
ministry clients. Picolotti’s replacement was Homero Bibiloni (2008–10), a
lawyer with a specialty in environmental law. Bibiloni was described in the
S u pp2:21
Account: ns191738
press as a “militant for the environment,”10 so we code him as having links
to environmental groups.
In sum, in both their initial and replacement appointments, presidents

are choosing women who are as likely as the men to enhance the adminis-
tration’s PCRs in the form of support resources regarding ministry clients,
and while the exact group to whom they were linked might differ, their
connections still provided support resources for the administration. The
exception, however, is that it is uncommon for women to be a source of
support resources related to the business community regardless of when
they were appointed.
Individual countries
Examination of individual countries reveals some ways norms for ministers

differ across our countries, or differ for men and women within countries
(see figure 6.3). The incidence of links to business varies across countries,
ranging from 4% of ministers in Argentina and Chile to 46% in Colombia.
In Argentina the low incidence could be explained by Peronists holding the
presidency for three of the four administrations in our data set, and particu-
larly Kirchner and Fernández, who returned their party to its more populist
roots. The Chilean presidents in our data set are from the Socialist Party
within the Concertación coalition, which also could decrease the prominence
of business links. In Colombia, in contrast, the greater frequency of minis-
ters with business links could be due to President Uribe’s lack of a clearly
defined party machine and his neoliberal economic policies. However,
under-representation of women with regard to links to business is only sta-
tistically significant in Costa Rica (13% of women, 47% of men, p = .002),
though the difference is considerable in Chile (6% of women, 18% of men,
p = .111). International links are far less common for US secretaries (13%)
than for the ministers in Latin American cabinets (range 37% to 46%).11
However, differences in international links between women and men are
only significant in Colombia (16% of women, 53% of men, p = .004).
Colombian women ministers were also less likely than men to have any
type of group links (53% of women, 83% of men, p = .008), or links to cli-
ents of their ministry (26% of women, 48% of men, p = .096). In the United
10. “Quién es Homero Máximo Bibiloni, el sucesor de Picolotti”. Perfil by Javier

Blanco. December 3, 2008 (www.perfil.com).
11. The proportion of ministers with links to international groups is significantly
different between the United States and each of the Latin American cases (p < .032 or
less) based on both the Bonferroni and the Scheffé multiple comparison tests.
EBSCO[ 136 ] Women
Account: ns191738

ministry clients (see note) † †
business **
international **
women’s groups ** ** * *** ***
organized labor *
any type of group link ** †
0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100
Women
Men Argentina (n = 94) Chile (n = 94) Colombia (n = 83) Costa Rica (n = 97) United States (n = 77)
Figure 6.3 Minister group links—by country

Account: ns191738
States women are more likely than men to have any type of group links (86%
of women, 60% of men, p = .073), which was also true for links to clients
(71% of women, 43% of men, p = .054). In all countries, more women than
men had links to women’s groups (p < .012 or less).
It appears that presidents in Argentina and Chile utilize both men and
women to bring the PCR of support resources to their administration since
the only difference we detect is with regard to links to women’s groups. In
the United States, it appears that a greater percentage of women than men
have the PCR of support resources, though we have insufficient informa-
tion to know if the group linkages of women secretaries are equally valued
by the president, or equally valuable to the administration. In Costa Rica,
women are much less likely than men to have links to business, but there
is no sex difference in links to ministry clients. Colombia is where gaps are
most stark, with the women less likely to be a source of client links, links to
the international community, or to any type of group. In sum, with regard
to their capacity to bring the PCR of support resources to the administra-
tion, women appear better integrated into the cabinets of Argentina, Chile,
and the United States than in Colombia or Costa Rica.
DO GROUP LINKS VARY ACROSS CATEGORIES OF POSTS,

AND DO WOMEN AND MEN BRING THE SAME SUPPORT
RESOURCES TO EACH TYPE OF POST?
Our analysis so far indicates support for our predictions that women’s
qualifications will resemble men’s. With the potentially important excep-
tion of links to business, the women are as likely as the men to have group
links, and (except in Colombia) to be sources of the PCR of support re-
sources for the administration. But the group links useful to specific port-
folios vary, and women might only have valuable group links when they
are appointed to traditional posts for women; thus, we need to explore
whether these findings hold when we partition the data set into different
categories of posts.
Economics/S ocial Welfare/C entral categories of portfolios
Several potentially important sex differences appear when we partition

the cabinet into Economics/Social Welfare/Central categories of portfo-
lios (see figure 6.4). In the Economics category in particular, women are
consistently less likely than men to have group links that could add to
EBSCO[ 138 ] Women
Account: ns191738

ministry clients (see note) * **
business * **
international †
organized labor
any type of group link † *
0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100
Women
Men Economics (n = 172) Social Welfare (n = 124) Central (n = 138)
Figure 6.4 Minister group links—Economics/Social Welfare/Central categories of portfolios

Account: ns191738
their credibility in their post or enable them to bring the PCR of sup-
port resources to the administration. A significantly smaller percentage
of women than men in the Economics category have links to ministry cli-
ents (31% of women, 55% of men, p = .017). Women in Economics cat-
egory posts are also less likely than men to have any type of group link
(59% of women, 75% of men, p = .090), links to business (25% of women,
45% of men, p = .036), or international links (22% of women, 40% of
men, p = .054). For instance, José Luis Machinea, minister of finance in
Argentina (1999–2001), had known links to international organizations
through work with the World Bank and the Inter-American Development
Bank, and to business through the Industrial Development Institute of
the Argentine Industrial Union and as a consultant for companies; so he
also had links to ministry clients. In contrast, Felisa Miceli, also minister
of finance in Argentina (2005–07), had extensive portfolio-relevant ex-
perience, but no known links to business or international groups, so she
is coded as not having links to ministry clients. Yet it is notable that even
in Economics category posts women are more likely than men to have
links to women’s groups (p = .000).
In the Social Welfare category of posts, women again were less likely
than men to have links to business (2% of women, 19% of men, p = .007).
However, women in the Social Welfare category were as likely as the men to
have links to ministry clients, thus bringing the PCR of support resources to
the administration. For example, María Soledad Barría, minister of health
in Chile (2006–08), had known links to doctors and to the PanAmerican
Health Organization, providing her with links to ministry clients. Álvaro
Erazo, also minister of health in Chile (2008–10), had known links to
doctors and international organizations (CEPAL as an advisor on health
issues), and thus links to ministry clients. In fact, all Chilean ministers
of health in our data set (three men, two women) had links to ministry
clients.
For the Central category of portfolios women were more likely than
men to have links to ministry clients (48% of women, 21% of men, p = .008)
or links to any type of group (78% of women, 51% of men, p = .018). For
example, Madeleine Albright, US secretary of state (1997–2001), had nu-
merous links to clients of her department, including directing the Center
for Women in Foreign Service at Georgetown University, and having been

a fellow at the Woodrow Wilson Center and at the Center for Strategic
and International Studies. Her successor, Colin Powell (2001–05), came
to the State Department from a military career, and thus while he had
extensive links to the military community, as well as to private organiza-
tions (e.g., America’s Promise Alliance for Youth), he did not have clear
EBSCO[ 140 ] Women
Account: ns191738
links to the foreign policy community, so we do not code him as having
links to clients.
Overall in the Social Welfare and Central categories of posts we see evi-
dence for integration of women with respect to their capacity to bring the
PCR of support resources to the administration. Additionally, in Economics,
not just in Social Welfare category posts, we see a statistically significant
difference in the proportion of women with links to women’s groups. Using
Borrelli’s (2010: 746) language, it appears that presidents are willing to ap-
point “highly credentialed” women to Social Welfare and Central catego-
ries of posts. But this cannot as easily be said for women appointed to the
Economics category of posts. Some of the women appointed to Economics
category portfolios have relevant group links, but most do not. Whether
presidents prefer women to be weak in these posts because they lack ties
to constituents, or if the difficulty of breaking into the upper echelons of
business means that fewer women have these kinds of linkages in the first
place, is a subject for future research. At the same time, statistically indis-
tinguishable percentages of women have links to clients across the three
types of posts suggesting that women’s ability to bring the PCR of support
resources is not dependent upon their appointment to posts where women
have long been appointed.12
Stereotypically masculine/f eminine policy domain portfolios
The patterns observed when the cabinet is partitioned into stereotypically

masculine and feminine policy domain posts resemble the findings for the
data set overall (see figure 6.5). The only significant sex difference is in links
to business, with more men than women having business links in both mas-
culine (p = .005) and feminine policy domain portfolios (p = .086). Overall
though in feminine policy domain posts fewer ministers have links to busi-
ness (13%), than in masculine policy domain posts (31%) (p = .000). The
difference between men and women in links to women’s groups persists
even in masculine policy domain posts.
Regarding the expectation that women would be more likely to have the
PCR of support resources when they were appointed to posts in a stereotyp-
ically feminine policy domain, we do not find this to be the case. While 52%
12. Over half the women (52%) appointed to head Social Welfare category minis-
tries had links to clients, compared to 31% of those heading Economics ministries and
48% heading Central category ministries. However, the difference in the proportion
of women heading these ministries with links to clients is not significantly different
under either the Bonferroni or Scheffé multiple comparison tests.
S u pp
EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017 or t
2:21 PM Rvia
Account: ns191738

business ** †
international
organized labor
0 20 40 60 80 100 0 20 40 60 80 100
Women
Men Masculine (n = 337) Feminine (n = 110)
Figure 6.5 Minister group links—stereotypically masculine/feminine policy domain

portfolios
of women in feminine domain posts and 40% in masculine domain posts

had links to clients this difference is not statistically significant (p = .199).
Unsurprisingly a higher percentage of female ministers appointed in femi-
nine domain posts (22%) have links to women’s groups compared to 10%
in the masculine domain, (p = .076), though what is probably most no-
table here is the small percentage of women ministers with known links to
women’s groups.
In sum, with the exception of business links, in both masculine and
feminine policy domains there is evidence of integration of women. For
example, Consuelo Araújo Noguera, minister of culture in Colombia
(2000–2001), had links to cultural groups, particularly from found-
ing and organizing the Vallenato Legend festival (highlighting religion,
music, dancing, arts and crafts), so we code her as having links to minis-
try clients. Similarly, Juan Luis Mejía Arango, minister of culture (1999–
2000), had links to the Colombian Institute of Culture, the Hispanic
Culture Institute, and he participated in international groups working
on cultural topics, so that he also had links to ministry clients. We see
a similar pattern with Chilean foreign affairs ministers Soledad Alvear
and Alejandro Foxley (2000–04 and 2006–09 respectively). Alvear had
known links to the Chilean Council on Foreign Relations and the Justice
Center of the Americas, while Foxley had known links to the World Bank,
Inter-American Development Bank, the Council on Foreign Relations,
and InterAmerican Dialog, so both were coded as having links to ministry
clients (and international links). These examples illustrate that in both
categories of posts, women and men are sources of the PCR of support
resources for the administration.
EBSCO[ 142 ] Women
Account: ns191738
High Visibility category posts
Women and men appointed to High Visibility category posts are similarly
likely to have links to clients of their ministry (54% of women, 49% of men,
p = .635), and thus to bring the PCR of support resources to the admin-
istration (see figure 6.6). Men are more likely than women to have links
to business (34% of men, 4% of women, p = .002). Especially notable here
is that almost no women appear to have business links, which is compa-
rable to what we found in the Social Welfare category of ministries, or the
feminine policy domain posts, where 2% and 7%, respectively, of women
had business links. This may result from the types of High Visibility cat-
egory portfolios to which women have more commonly been appointed
(e.g., three in education, six in health and social welfare). However, seven
women were in the labor portfolio, two in security, two in presidency, two
in finance, and one in defense, all posts in the stereotypically masculine
policy domain. We also see that more men than women (40% versus 23%)
have international links; however, the difference is not statistically signifi-
cant (p = .102). Yet women are as likely as men to have links to any type of
group (62% of women, 64% of men, p = .800) so it is not as though these
women are not connected. They bring support resources to the administra-
tion, but different types of connections than the men. Finally, we note that
five of the women appointed to High Visibility category posts (19%) have
links to women’s groups.
To illustrate, Sonia Tschorne was appointed minister of housing in Chile
(2004–06) during a high visibility time. She had known links to ministry
clients through the Pan American Federation of Architecture Associations

business ** **
international
organized labor
0 20 40 60 80 100 0 20 40 60 80 100
Women High Visibility (n = 138) Not High Visibility (n = 309)
Men
Figure 6.6 Minister group links—High Visibility/Not High Visibility categories of posts

S u pp2:21
Account: ns191738
and from serving as chair of the committee in charge of incorporating
art into public infrastructure. Fernando Zumbado, minister of housing
in Costa Rica (2006–08), had links to ministry clients from being direc-
tor of the International Center for Human Development, regional director
for the UNDP (also giving him international links), and founder and presi-
dent of the Housing Mortgage Bank. Alexis Herman and Donna Shalala,
US secretaries of labor (1997–2001) and of health and human services
(1993–2000), respectively, during high visibility times, are examples of
ministers with links to women’s organizations. Herman had worked with
agencies advocating for minority women’s employment and worked for the
Women’s Bureau in the Department of Labor. Shalala was recruited from
a women’s network (Borrelli 2002: 126), and her work with the Children’s
Defense Fund and the National Women’s Law Center gave her links to
women’s groups and to clients of her department.
Integration of women is evident in High Visibility category posts if
integration means the minister has links to ministry clients. However,
if integration requires a broader definition of the support resources that
ministers can bring to the administration, such as a network within the
country’s or the international financial sector, it is less clear that integra-
tion is occurring, since these women are far less likely than men to have
business or international links.
DO MINISTERS BRING MULTIPLE TYPES OF PCRS

TO THE ADMINISTRATION?
In this chapter we explored whether ministers have group connections

that enable them to provide the administration with support resources.
The answers we find are generally in the affirmative for our comparison of
women and men, though women are less likely to have links to business.
For the Economics category of portfolios this may mean that the women
lack credibility in their posts. It also means that presidents rely on women
less than men to bring their administration support resources from busi-
ness. However, it is possible that women lacking links to business may be
appointed to Economics category posts precisely because they lack those
links, although in-depth studies would be needed to test this. If women are
seen as new faces and less corrupt, then the absence of links may be delib-
erate to signal there will not be privileged access for the economic elite or
corrupt treatment of cronies. Yet for support resources from other types
of groups it appears that these women have the same capacity as men to
enhance the political capital of the administration.
EBSCO[ 144 ] Women
Account: ns191738
Regarding our theoretical predictions in fi gure 1.1, with the exception
of business links, many women’s backgrounds are similar to men’s, placing
them in the top row of the figure. What remains to be explored is whether
these women are treated equally and are as effective as men once they are in
the cabinet. A few women in our data set stand out from the other women
for their known links to women’s groups. Is that background trait enough
to place them in the bottom row of fi gure 1.1? If it is, will they be treated
as equals with their male colleagues and be incorporated, or will they be
treated unequally, as a token minister to disingenuously appear to link the
administration to women while not really giving them respect and power?
These are questions we address in Parts III and IV of the book.
However, before assessing treatment/effectiveness of women in these
cabinets, we want to explore if women are as likely as men to bring mul-
tiple types of PCRs to the administration, as it is efficient for the admin-
istration if ministers are sources of multiple PCRs (Wyszomirski 1989: 46;
Blondel 1991a, 1991b). Borrelli (2010: 737 and 745–46) views it as evidence
of gender integration, instead of desegregation, when women appointees
have extensive political credentials (e.g., they are Washington insiders, in-
cumbents in elected office, and are connected in ways that can enhance the
electoral alliances of the president).13 Assessing the ability of women in US
cabinets in earlier administrations to bring PCRs to the administration,
Wyszomirski (1989: 67) wrote that the women “possessed few administra-
tive, expertise, or political assets that might have enhanced their utility to
or influence with the president.” Is this still true?
As explained in chapters 4 through 6, for our measure of policy exper-
tise we use the extensive experience variable. As a single but broad measure
of political skills we use the political insider variable. As a single measure of
support resources we use the links to ministry clients variable. Though these
are crude measures of the political capital tool kit of ministers, they permit
us to examine whether PCRs are additive or if ministers, and particularly
women, typically bring only one type of PCR to the administration. We can
also assess which types of ministers most often have none of these PCRs.
Table 6.1 shows how many men compared to women have each possible
combination of these three traits. We look at the data set as a whole and
test for sex differences in PCRs within different broad categories of posts.
It is clearly not the norm for ministers to bring all three types of PCRs
to the administration—only 16.5% have all three, at least when measured
13. Borrelli (2010: 737) also lists women having “strong ties to women’s organiza-
tions and networks” as a sign of gender integration, which does not occur among most
women in our data set.
S u pp2:21
Account: ns191738
Table 6.1 HOW MANY AND WHAT TYPES OF PCRS DO MINISTERS BRING
TO THE ADMINISTRATION?
% of Ministers:
p-value
Types of PCRs Overall Women Men for row
FULL DATA SET (n = 443) (n = 110) (n = 333)

policy expertise + insider + client links 16.5 17.3 16.3 0.806
policy expertise + insider 35.2 32.7 36.0 0.529
policy expertise + client links 37.4 36.4 37.8 0.793
insider + client links 18.6 21.8 17.5 0.309
none 15.2 20.9 13.4 0.055
ECONOMICS CATEGORY POSTS (n = 170) (n = 32) (n = 138)

none 15.7 31.3 12.1 0.007
SOCIAL WELFARE CATEGORY POSTS (n = 122) (n = 47) (n = 75)

none 14.5 12.8 15.6 0.666
CENTRAL CATEGORY POSTS (n = 137) (n = 23) (n = 114)

none 15.2 21.7 13.9 0.340
MASCULINE POLICY DOMAIN POSTS (n = 333) (n = 63) (n = 270)

none 15.4 25.4 13.4 0.015
FEMININE POLICY DOMAIN POSTS (n = 109) (n = 47) (n = 62)

none 14.6 14.9 14.3 0.929
Account: ns191738
Table 6.1 CONTINUED
% of Ministers:
p-value
Types of PCRs Overall Women Men for row
HIGH VISIBILITY POSTS (n = 136) (n = 26) (n = 110)

none 12.3 11.5 12.5 0.893
NOT HIGH VISIBILITY POSTS (n = 306) (n = 84) (n = 222)

none 16.5 23.8 13.8 0.035
Note: Values for women, men, and p-value appear in bold when p < .10.
as we have done so here. Yet, overall, women and men are equally likely to
have all three types of PCRs (p = .806), and this is also true in all the broad
categories of posts that we have used for our analysis, with the exception
of Central category posts where a higher percentage of women (26%) than
men (11%) have all three types of PCRs (p = .044). Building on Borrelli’s
(2010) assessment of gender integration in US cabinets, we view this find-
ing that equal percentages of women and men bring all three types of PCRs
to the administration (even though it is a small percentage of ministers) as
an indication that women are not only present in cabinets in growing num-
bers but also that these presidents appointed women with broad PCRs—
which is evidence of integration of women. The relatively small number of
men and women with all three PCRs suggests that both men and women
are appointed for different mixes of skills they bring to the job.
Several ministers are good examples of appointees who bring all three
types of PCRs to their administrations. Nilda Garré (minister of defense in
Argentina, 2005–10) had policy expertise related to her portfolio coming

from work with NGOs and in the Chamber of Deputies on the Defense
Committee. Her experience serving multiple terms in the Congress and as
vice minister of interior gave her the PCR of political skills. Her work with
human rights groups gave her links to ministry clients because teaching
the military about human rights is a component of the policy purview of
S u pp
2:21 e s oECONOMIC [ 147 ]
Account: ns191738
the Defense Ministry in Argentina and thus made her a source of support
resources (though she was an appointment that surprised the military es-
tablishment). Monica Nagel (minister of justice in Costa Rica, 1998–2002)

had policy expertise related to her portfolio from education and work (law
degree, work as a lawyer and consultant). Her experience as vice minister
of justice, gave her the PCR of political skills, and contributed to her policy
expertise for the post. She had links to ministry clients via international
organizations (e.g., anti-drug organizations, UNICEF), which made her
a source of support resources. For Donald Rumsfeld (US secretary of de-
fense, 2001–06), his extended career in defense, including service in the
US Navy and as secretary of defense during the 1970s (and later as a presi-
dential advisor on arms control), gave him policy expertise. Experience as
a prior cabinet secretary also meant that he had the PCR of political skills.
Links to the Council on Foreign Relations, Center for Security Policy, and
the Hoover Institution gave him links to agency clients and made him a
source of support resources. For Isidro Solis (minister of justice in Chile,
2006–07), his education and work in law and public security for the gov-
ernment gave him policy expertise. Experience as vice minister of mining
and then vice minister of aviation gave him the PCR of political skills. He
held leadership posts in the College of Lawyers, which gave him links to
ministry clients and made him a source of the PCR of support resources.
Table 6.1 also shows that in the data set overall, similar percentages of
women and men had any pair of two types of PCRs, and men and women
did not differ statistically. The combination of political insider + client links
is the least common (only 18.6% of ministers), while slightly more than a
third of ministers have policy expertise + political insider (35.2%) or policy
expertise + client links (37.4%). There is some variance in the frequency of
pairs of PCRs across categories of posts. Ministers who hold High Visibility
category posts more frequently have all combinations of PCRs, so it ap-
pears that presidents are selecting men and women with extensive political
capital.14 This analysis adds additional evidence prompting us to conclude
that women appointed to Economics category posts often lack credibility
for their job, as significantly more men than women in the Economics cate-
gory of posts have both policy expertise + client links (51.1% of men, 28.1%
of women, p = .019). But this analysis also indicates that the women who
14. A binomial test of proportions indicates that it is significantly more common for
ministers in High Visibility category posts, than for ministers in the data set overall
(or for any other category of posts) to have all three PCRs (two-tailed t-test p = .004).
In the High Visibility category thirty-six ministers have all three PCRs. For the data
set overall 16.5% of ministers have all three PCRs, so the expectation would be that
only twenty-two ministers in High Visibility category posts would have all three PCRs.
EBSCO[ 148 ] Women
Account: ns191738
are appointed to Central category posts—though still few in number—
are amply qualified for their job, as significantly more women than men
in Central posts have all three types of PCRs and also have the two-PCR
combination of insider + client links. We also find that significantly more
women than men have the insider + client links combination of PCRs in the
masculine policy domain category of posts.
So far this analysis indicates that integration of women is occurring in
these cabinets. However, the picture is not uniformly rosy as we also find
that more women than men bring no detectable PCRs to the administra-
tion based on the measures employed here (20.9% of women in the data set
overall, 13.4% of men, p = .055). In the broad Economics category of posts
more women than men have no detectable PCRs (31.3% of women, 12.1%
of men, p = .007), which could mean that in Economics posts women may
be showcased for their sex (Borrelli 2002: 54) but not set up for success in
their job.15 We also find that more women lack any observed PCRs in mas-
culine policy domain posts (25.4% of women, 13.4% of men, p = .015), and
in posts that are NotHigh Visibility before they are appointed (23.8% of
women, 13.8% of men, p = .035).
Karen Poniachik (minister of mining and energy in Chile, 2006–08) and
Juan Francisco Lozano Ramirez (minister of environment, housing and
territorial development in Colombia, 2006–09) are examples of ministers
without any of the types of PCRs we study here. Poniachik was educated in
journalism and international relations and worked in journalism, foreign
relations, and foreign investment before becoming minister of mining.
Her known group links were to the Chilean–North American Chamber of
Commerce. While her résumé is impressive, it did not provide her with
policy expertise relevant to her post, or make her a political insider, or give
her publicly known links to ministry clients. Lozano was educated in law
and worked as a lawyer and in various, primarily local/regional government
posts and also was a presidential advisor: so he lacked policy expertise for
his post or political insider status (although he did have connections to the
president). He had connections to Transparency International and to the
newspaper El Tiempo, but neither provided links to ministry clients.
The analysis of group links in the first half of this chapter showed that
the qualifications of women ministers in this study look very similar to
those of the men. Analysis of whether ministers bring multiple types of

PCRs to the administration again shows, in most instances, that the women
resemble the men. Most women will be placed on the top row of figure 1.1.
15. This significant sex difference is still found when we exclude the finance portfolio
(34.5% of women, 16.5% of men have no PCRs, p = .034).
S u pp2:21
Account: ns191738
Based on our more optimistic theoretical predictions, this should make the
women as likely as the men to be “political players” in government. Yet a
notable minority of women, and significantly more women than men, have
none of the PCRs we measure. These women will be placed on the bottom
row of figure 1.1. Will they also be able to be incorporated and be political
players, or are they tokens?
If we find that women with similar backgrounds to the men are able
to be equally effective, that will indicate integration of women with male
norms is occurring in these cabinets (upper-left cell in fi
gure 1.1). We pre-
dicted that women would resemble the men—that is, conform to male
norms other than in their sex—because that would make representation
of this new group less threatening to those who have traditionally held
power. Appointment of women who resemble the men would also mean
the women are likely to bring PCRs to the administration that presidents
expect to need to pursue their policy agenda and solve problems; scarce
appointment opportunities would not be squandered. If instead we find
that these women are less likely than the men to be political players, then
that will indicate that even with increasing numbers of women, cabinets
are gendered institutions that are very limiting to women (upper-right cell
figure 1.1). Regarding women whose backgrounds are different from men,
if they are treated equally and are as effective as men, this will indicate
incorporation of women as women not as surrogate men (lower-left cell
figure 1.1), that gender differences are rewarded and the cabinet as an insti-
tution is transformed (Lovenduski 1998: 340). We take up the question of
minister treatment/effectiveness in the next part of the book.
EBSCO[ 150 ] Women
Account: ns191738
PART III
Four Benchmarks for Comparing

Treatment and Effectiveness of the Men
and Women Appointed to Presidential
Cabinets
Account: ns191738
Account: ns191738
CHAPTER 7
Getting seats at the table, and not just

in the “women’s seat”
I n Part III we turn to how women are treated in their cabinet posts, or how
successful they are in their jobs. Annesley et al. (2014: 3) call for research
on this topic, writing, “[W]‌e want to know when, why, and how women are
recruited as ministers or chief executives. We also want to know what factors
shape their capacity to act or ability to succeed once in office.” We compare
the records of men and women, with each chapter using a different bench-
mark for assessing whether or not we observe equal treatment or success in
these presidential cabinets. Our benchmarks are the following: the types of
posts women receive (chapter 7), how ministers exit their post (chapter 8),
how long they remain in a post (chapter 9), and legislative productivity
(chapter 10). We use our findings on these benchmarks, singly and ultimately
collectively in chapter 11, to draw conclusions about whether integration of
women is happening in these presidential cabinets even without parity.
The analysis in Part II showed that women ministers are typically simi-
lar to the men in their qualifications for their post and the political capital
resources (PCRs) they bring to the administration,1 though there are some
exceptions, especially in the broad Economics category of posts. This leads
us to hypothesize that women will be as successful as the men in their ability
to do their job. Certainly not all cabinet ministers are viewed as successes.
There are high-profile cases of failures—examples of ministers raked over
1. PCRs in our analysis, which are drawn from Wyszomirski’s (1989) study of the
U.S. cabinet, are policy expertise (operationalized as extensive portfolio-related ex-
perience), political skills (being an insider), and support resources (links to ministry
clients).
Account: ns191738
the coals by congress, pilloried in the press, brought down by a scandal.
But, as explained in the literature review in c hapter 1, selection criteria
have evolved over time, such that certain types of people (i.e., elites with
certain background and experience characteristics) are the type commonly
seen in the cabinet. Appointment of people that fit that type does not raise
questions, since such people match expectations concerning the creden-
tials a minister will have. A woman (or a minister from some other his-
torically under-represented group) will often attract additional attention
from the media because her sex (race, ethnicity, etc.) is seen as an unusual
characteristic for a minister, or for a particular portfolio. However, if the
appointee otherwise fits the established (male) norms of “qualifications”
and doing so means a minister has the skills needed to do the job, then
both women and men should be equally likely to be successful as minis-
ters. Referring to our theoretical predictions in figure 1.1, this would mean
that we predict women whose backgrounds are similar to the men’s will be
equally effective as the men (upper-left cell) and we expect to find evidence
of women being integrated into cabinets with male norms.
HOW CAN WE ASSESS IF THERE IS INTEGRATION

OF WOMEN IN PORTFOLIO OPPORTUNITIES?
In this chapter we examine the types of posts women receive, which is our
first benchmark for assessing if there is evidence of equal treatment. If
we find that, despite increasing numbers and women having similar back-
grounds as men, women are still overwhelmingly appointed to posts that
fit stereotyped gender roles—this is a sign that cabinets are gendered insti-
tutions. If women are (relatively) equally represented in all types of posts,
and background traits other than sex predict portfolio assignments, then
on this benchmark we would conclude that there is equal treatment even
while women are a numerical minority in the cabinet.
Why is equal access to all types of posts a benchmark of equal treatment?
Globally, women typically enter the cabinet through social welfare portfolios,
and many studies have found that women are over-represented in those types
of posts (see below). We do not mean to imply that social welfare posts are un-
important. As Tremblay and Bauer (2011: 179) argue, social welfare posts are

important for many reasons, ranging from the size of their budget to the scope
of the population affected by these services—and thus the need to defend
their ministry from budget cuts during times of budget austerity—to the role
these ministries play in defending social rights. Rather, our concern is whether
women also have access to other types of posts.
EBSCO[ 154 ] Women
Account: ns191738
Receiving non-traditional portfolios matters for several reasons. Women
may have policy interests outside social welfare (Drew 2000; Furlong and
Riggs 1996; Schwindt-Bayer 2006, 2010a). Analyzing survey data from

the mid-1990s from high-ranking political elites in twenty-seven industri-
alized countries, Drew (2000: 61) found that both men and women were
equally likely to serve on economic affairs, fiscal policy, and international
relations committees.2 Ministers who hold key posts, such as finance, may
have control over the capacity of other ministers to make and implement
policy (Blondell 1988: 10, 12). Another reason that broader appointments
of women are necessary is explained by Schwindt-Bayer in her analysis of
women’s committee assignments in legislatures in Latin America: “If female
legislators are the predominant political actors working on women’s issues,
this reinforces long-standing stereotypes that these are issues for women
only rather than gender issues or human rights issues that should be priori-
ties for both men and women. It reduces women to an inferior status in the
legislature (Vincent 2004; Franceschet 2005; Macaulay 2006)” (Schwindt-
Bayer 2010a: 16; also see Diaz 2005: 129; Yule 2000: 36). Appointment of
women (almost) exclusively to social welfare portfolios reinforces possible
societal stereotypes that there is an appropriate “hearth and home” policy
domain that encompasses women’s interests and that women are credible
to handle (Bratton 2005; Dolan 2004; Huddy and Terkildsen 1993; Matland
1994; Sanbonmatsu 2002). Sykes (2009: 40) argued that education, health,
and welfare portfolios “threaten to become regendered as the ‘women’s
posts.’ ” Irish Minister of Education Mary Hanafin complained, “I can go to a
European Council meeting now, and all the Education Ministers are women.
It’s kind of branding almost” (interview 24 June 2005 in Sykes 2009: 40).
Conversely, Martin (1997: 52) argued that “a president may achieve a greater
degree of symbolic representation … where few women may be appointed
but those who are receive the highest administration posts.”
First we look at the percentage of women and men appointed to differ-
ent broad categories of posts. Not surprisingly we find significant sex dif-
ferences and under-or over-representation of women in some categories
of posts. More surprising though is where we do not find sex differences
in treatment. Next we employ multivariate analysis to explore if sex differ-
ences are still apparent when we control for PCRs and other background
2. Women were more likely than men to want to serve on committees handling labor
or health topics, while men were more likely to want to serve on internal affairs com-
mittees (Drew 2000: 61). In contrast, a study of Danish municipal councils found that
women were more likely to want to serve on the Children’s Committee, while men
wanted to serve on the Finance Committee, or the Technical Committee whose juris-
diction includes roads and sewage services (Baekgaard and Kjaer 2012).
EBSCO Publishing : eBook Collection (EBSCOhost) - printedGon

e tt i n g S e at
11/29/2017 s at
2:21 PM tvia
h eECONOMIC [ 155 ]
Ta b l e COMMISSION
Account: ns191738
traits. We examine whether characteristics that are the male norm are also
associated with a woman getting appointed to a particular type of post. To
preview our findings, while there are posts in which women are rarely ap-
pointed, when we control for minister background sex is no longer a signifi-
cant predictor of the type of post received. However, some PCRs or other
background traits affect women and men in different ways, which reveals
some ways that treatment of women and men differs.
WHERE ARE THE WOMEN? DO THE POSTS WOMEN

RECEIVE INDICATE EQUAL TREATMENT?
Women typically were first appointed to cabinets in a social welfare post—

the education portfolio was particularly common (three of the six women
appointed to presidential cabinets in Latin American countries during the
1950s were education ministers)—or as minister of women’s affairs (Luna
et al. 2008). There were exceptions, such as the first appointment of a
woman to the US cabinet: Frances Perkins as secretary of labor during the
Franklin Roosevelt administration (1933–45). The first woman appointed
to the cabinet in Panama, Maria Santodomingo de Miranda (in 1950) was
also in the labor portfolio (Luna et al. 2008). Generally, women were ini-
tially appointed in posts that fit gender-role stereotypes, conforming with
the maternal image and legitimacy society grants women in the sphere of
home and family (Chaney 1979; Huddy and Terkelson 1993a, 1993b; Dolan
2004; Sykes 2009). Numerous studies of representation of women in both
presidential and parliamentary democracies have found that women are
disproportionally appointed to social welfare posts or “feminine” posts
(Lovenduski 1986; Thiebault 1991; Moon and Fountain 1997; Reynolds
1999; Studlar and Moncrief 1999; Borrelli 2002, 2010; Escobar-Lemmon
and Taylor-Robinson 2005; Sykes 2009: 40; Annesley and Gains 2010;
Buckley and Galligan 2011: 142–43; Krook and O’Brien 2012; Tremblay and
Stockemer 2013). Davis (1997: 28) concluded in her path-breaking study of
women’s representation in cabinets in fifteen parliamentary countries in
Western Europe (using data from 1968–1992) that, “Under Dogan’s (1989)
conception of government as a system of concentric circles, women are
more likely to appear in the outer ring, less likely to appear in the second
tier, and least likely to be present at the core.”
This repeated finding that women are primarily appointed to feminine
policy domain or social welfare posts, and the frequent interpretation that
women face a glass ceiling for the highest prestige posts in the cabinet’s
inner circle, has led to debate about the importance of social welfare posts,
EBSCO[ 156 ] Women
Account: ns191738
and the “element of circularity in labeling all ‘women’s portfolios’ as low in
prestige” (Studlar and Moncrief 1999: 380). It has also prompted questions
of whether women appointed to stereotypically feminine policy domain

posts are receiving the posts they actually want, or that women are taking
the only posts that they can get. One way to address this question is by
looking at the match between the minister’s pre-cabinet career field and
their post. If occupation can be used as a proxy for interest in a topic, then
78% of women (and 77% of men) in our data set are likely to have posts
they desired or were pleased to accept, rather than being forced to take
an ill-fitting opportunity to move into the cabinet.3 Thus, we need to em-
phasize that by looking at post assignments we can only determine if all
posts are equally open for appointment of women, or if there appear to be
glass ceilings. In the absence of interview data, we cannot assess whether
women (or men) received the post that they most wanted.
We begin our analysis by exploring whether patterns found in other re-
gions or earlier decades appear in the countries/administrations in our study.
To test for whether women are significantly over-or under-represented in
some types of portfolios we conduct binomial tests of proportions using
Economics/Social Welfare/Central categories of posts; stereotypically mas-
culine/feminine policy domain, and High Visibility category posts (see
chapter 3 for explanation of portfolios included in each broad category). This
test allows us to determine if women hold posts in a category proportional
to their representation in the data set. For example, 24.6% of the ministers
in our data set are women, so if women are proportionally represented in
posts with a stereotypically masculine policy domain, 24.6% of those posts
should be held by women. We look at representation of women in each of
these categories of posts for all ministers in our data set, initial ministers,
replacement ministers, and in each country (see table 7.1).
We observe that women are over-represented in the broad Social Welfare
category of posts for the data set overall and among both initial and re-
placement ministers. However, women are not over-represented in Social
Welfare posts in Colombia, Costa Rica, or the United States. Women are
only significantly under-represented in the broad Economics category of
posts in Costa Rica. In the Central category women are under-represented
for the entire data set, but under-representation is only observed in
Argentina and Chile when we examine individual countries.
3. This assessment is based on the ministers we coded as having some work experi-
ence, or education and some work experience related to their portfolio, though the
extent of their experience may not meet our measure for “extensive experience” (i.e.,
having the PCR of policy expertise). We thank Jureé Capers for bringing this point to
our attention.

e tt i n g S e at
11/29/2017 s at
2:21 PM t h eECONOMIC
via [ 157 ]
Ta b l e COMMISSION
Account: ns191738
Table 7.1 ARE WOMEN OVER- OR UNDER-R EPRESENTED IN SOME TYPES OF PORTFOLIOS?
Economics category Social Welfare category Central category

Women Women Women

% women p-value p-value p-value
In data set Observed # (%) Expected # (2-tailed) Observed # (%) Expected # (2-tailed) Observed # (%) Expected # (2-tailed)
All ministers 23.5* 32 (19%) 40 0.150 47 (38%) 29 0.000 23 (17%) 32 0.070

Initial ministers 24.8 17 (19%) 22 0.267 26 (38%) 17 0.017 14 (19%) 18 0.342
Replacements 22.2 15 (18%) 19 0.430 21 (39%) 12 0.005 9 (14%) 14 0.134
Argentina 19.8 4 (19%) 4 1.000 12 (40%) 6 0.010 3 (7%) 8 0.050
Chile 34.8 12 (34%) 12 1.000 14 (58%) 8 0.019 6 (18%) 12 0.046
Colombia 19.7 8 (21%) 7 0.838 4 (25%) 3 0.537 3 (14%) 4 0.600
Costa Rica 23.7 4 (9%) 10 0.030 11 (37%) 7 0.129 7 (35%) 5 0.289
United States 18.2 4 (11%) 6 0.384 6 (25%) 4 0.424 4 (22%) 3 0.555
Masculine Policy Domain Feminine Policy Domain High Visibility category
Women Women Women

% women p-value p-value p-value
In data set Observed # (%) Expected # (2-tailed) Observed # (%) Expected # (2-tailed) Observed # (%) Expected # (2-tailed)
All ministers 24.6 63 (19%) 83 0.011 47 (43%) 27 0.000 26 (19%) 34 0.138

Initial ministers 26.2 34 (19%) 46 0.048 28 (45%) 16 0.001 14 (19%) 19 0.186
Replacements 23.0 29 (18%) 37 0.159 19 (40%) 11 0.010 12 (19%) 15 0.550
Argentina 19.8 9 (12%) 15 0.140 10 (44%) 5 0.008 7 (20%) 7 1.000
Chile 35.1 20 (27%) 26 0.179 13 (62%) 7 0.019 8 (38%) 7 0.820
Colombia 24.1 12 (19%) 16 0.383 8 (44%) 4 0.053 2 (7%) 7 0.043
Costa Rica 24.7 11 (16%) 17 0.095 13 (46%) 7 0.014 4 (13%) 7 0.203
United States 18.2 11 (19%) 10 0.863 3 (15%) 4 1.000 5 (20%) 5 0.796
A binomial test of proportions is performed to determine if there are significantly more or fewer women in a category than would be expected based on the percentage of women in the data
set or relevant subset of the entire data set. To illustrate, women are 24.6% of our data set, so if women are proportionally represented across categories of portfolios women will be 24.6% of
the ministers in every category.
* The % women is 23.5% for the data set when examining Economics/Social Welfare/Central categories because the culture portfolio is not included in this analysis, as it does not fit well in any
of the three general categories. The slightly smaller data set without the culture portfolio has fewer women than the complete data set.
Account: ns191738
Looking at stereotypically masculine/feminine policy domain posts we
see that women are over-represented in feminine policy domain posts for
the data set overall and in all subsets of the data set, except the United
States. However, for masculine policy domain posts, while women are
under-represented for the data set overall, closer inspection indicates this
under-representation is primarily found among initial appointments, and
the only country where the sex difference is even marginally statistically
significant is Costa Rica.
What we observe regarding women’s representation in High Visibility
category posts could indicate integration of women in these cabinets, even
though women still hold a minority of cabinet seats.4 Women are not under-
represented in High Hisibility posts for the data set overall, for initial or
replacement ministers, or in any country except Colombia. This finding is
surprising because ministers in charge of High Visibility category posts
could be anticipated by the president to encounter conflict and public criti-
cism and possibly to have to manage protest against government policy or
management. Women are typically viewed as compassionate and sensitive,
while men are viewed as aggressive and able to take charge (Alexander and
Anderson 1993; Huddy and Terkildsen 1993a, 1993b; Heilman 2001; Eagly
and Karau 2002; Lawless 2004b; Banwart 2010; Dittmar 2012: chapter
2). Based on gender role stereotypes, we expected that women would be
unlikely appointees to these posts. Although women are not equally rep-
resented in these presidential cabinets in the sense of having half of the
seats, these presidents do not appear to be keeping women out of High
Visibility category posts.
Table 7.1 confirms earlier findings in the literature that women are
over-represented in Social Welfare posts but also adds new insights that
we interpret as indicating that women are being integrated into these
cabinets. Women are still over-represented in posts that conform with
gender stereotypes about areas where women have expertise: home,
family, and social welfare more broadly. However, women are not as con-
sistently under-represented outside the stereotypically feminine domain,
and women are not consistently kept out of Economics and Central posts
or High Visibility category posts. This initial analysis indicates more evi-
dence than the literature led us to anticipate that there is equal treatment
on this benchmark.
4. Posts are coded as High Visibility category because their policy purview includes
topics flagged in public opinion surveys as “the most important problem facing the
country.” See table 3.2 for portfolios coded as High Visibility and the years in which
that coding occurred.

e tt i n g S e at
11/29/2017 s at
2:21 PM tvia
h eECONOMIC [ 159 ]
Ta b l e COMMISSION
Account: ns191738
DO PCRS OR OTHER BACKGROUND CHARACTERISTICS PREDICT
POSTS, AND DO THEY HAVE THE SAME IMPACT ON WOMEN’S
APPOINTMENTS AS THEY HAVE ON MEN’S APPOINTMENTS?
Women are over-represented in Social Welfare category posts and stereo-

typically feminine policy domain posts, but men are appointed to those
posts, too. For example, twenty-two of thirty-two education ministers in
our data set are men. Women in these five countries are being appointed
to Economics and Central category posts: for example, five of the twenty-
two foreign relations ministers are women, and five of the twenty-seven
defense ministers are women. Women also are appointed to posts that are
known to be High Visibility when the appointment is made. To better un-
derstand how background is related to different types of appointments,
and in particular to explore whether the effect of a background trait is the
same for women as for men (which we would consider to be additional evi-
dence of equal treatment) we conduct multivariate analyses to predict the
type of post a minister receives.
Continuing the practice from Part II of the book we measure the PCR of
policy expertise with our “extensive experience” variable, the political skills
PCR with our “political insider” variable, and the support resources PCR
with our “links to ministry clients” variable. We also include in our analysis
several background traits for which, under some circumstances, we found
significant sex differences in Part II of the book: government (vs. private)
sector career, connected to president, organizational partisan, and prior
elected office.5
Appointment to Economics/S ocial Welfare/C entral

categories of portfolios
This analysis divides posts into three broad categories, so we use multino-
mial logistic regression, including interactions of each PCR and background
trait with minister sex to allow us to directly compare how background
affects likelihood of appointment for men and women (see table 7.1A,
column 1 in the chapter appendix for the full model results). We illustrate
the importance of sex after taking into account a minister’s background

and credentials, upon the kind of post a minister receives graphically in
figure 7.1. We calculate the change in probability for appointment to each
5. We do not include all background traits measured in Part II of the book in the
analysis presented here or in subsequent chapters both because of a desire for parsi-
mony and because these variables capture a broad range of experiences.
EBSCO[ 160 ] Women
Account: ns191738
0.6
0.4
Change in Probability
0.2
–0.2
–0.4
–0.6
Economics Social Welfare Central
Figure 7.1 Effect of being a woman on appointment to Economics/Social Welfare/Central

category posts
Note: The bars in this figure reflect the range covered by the 95% confidence intervals associated with taking
the first difference and simulating the probability of receiving an Economic/Social Welfare/Central category
post for the change from male to female minister based on model 1, table 7.1A. Other variables are set to
baseline values (see footnotes 6 and 7 for details).
kind of post by assigning modal values for all our variables6 and detect
significant differences across the sexes only with regard to appointment
to Central posts. This graph shows how if nothing else about the base-
line minister changes except that they are a woman instead of a man the
probability of receiving each kind of post changes. There is not a differ-
ence distinct from zero in the probability that ministers of either sex
with the modal background characteristics receive either an Economic or
a Social Welfare category post. However, for Central category posts, given
the modal combination of background characteristics, the difference in
probability between men and women is different from zero with women
having lower probabilities.7
6. Our baseline minister reflects the mode in our entire data set: a man, appointed
to the initial cabinet, with the PCRs of policy expertise and political skills, from a gov-
ernment sector career because a majority of ministers in our data set have these traits.
The baseline minister lacks the PCR of support resources, is not connected to the presi-
dent, is not an organizational partisan, and has not held elected office. The baseline
minister is from Costa Rica because Costa Rica contributes the most ministers to our
data set, though the choice of country does not drive results.
7. All graphs in this chapter use simulated probabilities retrieved from Clarify ver-
sion 2.1 by Tomz, Wittenberg, and King (2003). We are reporting the change in prob-
ability for that outcome given a change in one independent variable and holding every-
thing else constant at its modal value.

e tt i n g S e at
11/29/2017 s at
2:21 PM t h eECONOMIC
via [ 161 ]
Ta b l e COMMISSION
Account: ns191738
Some PCRs and background traits appear to operate differently for men
and women. Figure 7.2 graphs the impact of each of the three PCRs on
the probability that a man and a woman receive each category of posts.8
For both men and women having versus not having the policy expertise
PCR (graph 7.2a) does not significantly change the probability a minister
is likely to be appointed to an Economics, Social Welfare, or Central cat-
egory post. And because the bars overlap each other, there is no difference
between men and women. This is not surprising because most ministers
in our study—men and women—have extensive post-related experience.
In graph 7.2b, in contrast, we see that moving from being an insider (i.e.,
having the political skills PCR) to an outsider, decreases the probability for
both women and men of receiving a Central category post, but this change
has no effect upon Economic and Social Welfare category posts. Insider
status influences the type of post a person receives, but this PCR appears
to operate the same way for both men and women. In addition, for women
the bars for Economic, Social Welfare, and Central category posts overlap,
indicating that having the PCR of political skills does not affect the prob-
ability of receiving a post in one category versus another. For men there
is no overlap in the bars for Economic and Central category posts, which
indicates that lacking the political skills PCR has a significantly different
(and negative) effect upon the chances of a man receiving a Central rather
than an Economics category post. In graph 7.2c we see that moving from
not having to having client links (the support resources PCR) makes a
woman less likely to receive an Economics category post, while for men it
decreases their probability of getting a Central category post. In addition,
the three bars for women overlap, indicating that moving from not having
to having client links does not affect the probability a woman receives a
post in one category versus another. For men there is no overlap in the
bars for Economics and Central posts or Social Welfare and Central posts.
This indicates that for men the effect of moving from not having to having
client links is different in these two instances but that the effect is the same
for Economic as it is for Social Welfare category posts.
Figure 7.3 shows sex differences in the influence of some of the other
four background traits but not all of them. Coming from a private sector
instead of a government career increases the probability for both men and
women that a minister receives an Economic category post (graph 7.3a).
8. In figure 7.2 and the rest of the figures in this chapter, if a bar crosses the zero
line it indicates that moving from not having to having a characteristic (for instance
changing from not having to having the PCR of support resources) does not produce a
change in the probably of receiving a particular type of post that is statistically distinct
from zero.
EBSCO[ 162 ] Women
Account: ns191738
(a) Change from having PCR#1 (policy expertise)
to not having PCR#1
Central
(Women)
Central
(Men)
Social Welfare
(Women)
Social Welfare
(Men)
Economics
(Women)
Economics
(Men)
–0.5 –0.4 –0.3 –0.2 –0.1 0 0.1 0.2 0.3 0.4
(b) Change from having PCR#2 (political skills)

to not having PCR#2
Central
(Women)
Central
(Men)
Social Welfare
(Women)
Social Welfare
(Men)
Economics
(Women)
Economics
(Men)
–0.5 –0.4 –0.3 –0.2 –0.1 0 0.1 0.2 0.3 0.4
Figure 7.2 Continued
For men only, coming from a private sector career decreases the chances
of getting a Central category post. For women only, coming from a private
sector career decreases the chances of getting a Social Welfare category
post. Thus, we detect sex-based differences in the influence of government
vs. private sector careers with regard to Social Welfare and Central category

e tt i n g S e at
11/29/2017 s at
2:21 PM tvia
h eECONOMIC [ 163 ]
Ta b l e COMMISSION
Account: ns191738
(c) Change from not having PCR#3 (support resources)
to having PCR#3
Central
(Women)
Central
(Men)
Social Welfare
(Women)
Social Welfare
(Men)
Economics
(Women)
Economics
(Men)
–0.5 –0.4 –0.3 –0.2 –0.1 0 0.1 0.2 0.3 0.4
Figure 7.2 Effect of whether a person has a PCR on appointment to Economics/Social

Welfare/Central category posts
Note: 95% confidence intervals associated with the change in probability based on changes in each PCR based
on model 1, table 7.1A. Other variables are set to baseline values (see footnotes 6 and 7 for details).
posts, but not Economic. We speculate that these sex differences in the
effect of private sector careers for Central and Social Welfare category posts
may be due to women’s greater or lesser prevalence in some occupations
that are natural pipelines for posts in that area.
Being connected to the president appears to have an impact on the
type of post men receive but not women (graph 7.3b), suggesting some
sex-based differences in the effect of this variable. For women, the effect
of having versus not having connections to the president is not statis-
tically distinguishable from zero for any post type. For men, however,
changing from not being connected to the president to being connected
significantly increases the chances of receiving a Central category post
but decreases the chances of receiving a Social Welfare category post. It
appears presidents put men with whom they have connections in Central
posts but not in Social Welfare category posts. Possibly this is because
Social Welfare category posts are viewed as the least prestigious posts
for men. In some sub-samples of our data set more women than men
have connections (see chapter 5), but this analysis indicates that is not
a requirement—or helpful—for a woman in getting any broad category
of posts.
EBSCO[ 164 ] Women
Account: ns191738
The effect of changing from government to
(a)
private sector career
Central
(Women)
Central
(Men)
Social Welfare
(Women)
Social Welfare
(Men)
Economics
(Women)
Economics
(Men)
–0.6 –0.5 –0.4 –0.3 –0.2 –0.1 0 0.1 0.2 0.3 0.4 0.5 0.6
(b) The effect of changing from not being connected to the

president to being connected to the president
Central
(Women)
Central
(Men)
Social Welfare
(Women)
Social Welfare
(Men)
Economics
(Women)
Economics
(Men)
–0.6 –0.5 –0.4 –0.3 –0.2 –0.1 0 0.1 0.2 0.3 0.4 0.5 0.6
Figure 7.3 Continued
Being an organizational partisan (having held a party leadership post)

only matters in one case: moving from not holding a party leadership post
to being an organizational partisan decreases a man’s, but not a woman’s,
chances of receiving an Economics category post (graph 7.3c). This is con-
sistent with presidents often wanting to depoliticize Economics posts

e tt i n g S e at
11/29/2017 s at
2:21 PM tvia
h eECONOMIC [ 165 ]
Ta b l e COMMISSION
Account: ns191738

Women in Presidential Cabinets 3

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Women in Presidential Cabinets 3

Uploaded by

Copyright:

Available Formats

Political background.

Writing about the US cabinet, Wyszomirski

or as an executive” (e.g., governor, mayor, head of a government agency),

2. We do not disaggregate number or types of elected offices, or level of government.

EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017 P2:21

The data set overall

Political background. When we group together all 447 ministers in our

political insider (see note) †

Figure 5.1 Minister political backgrounds—​data set overall

The percentages of women (30%) and men (32%) with background as

EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017P2:21

transportation (2006–​09). Before entering the cabinet she worked in state

appointed posts. Hinojosa (2009: 388), studying election of women to municipal gov-

Comparison of initial and replacement ministers

Source of Political Skills

Figure 5.2 Minister political backgrounds—​initial/​replacement ministers

EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017 P2:21

apparent in the data set as a whole, is observed in the initial cabinet: more

as becoming close to Kirchnerismo in 2003. Her replacement, Juan Luis

EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017 P2:21

Source of Political Skills

Figure 5.3 Minister political backgrounds—​by country

In every country, statistically equivalent percentages of men and women

EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017 P2:21

DO POLITICAL QUALIFICATIONS VARY ACROSS CATEGORIES

Economics/​S ocial Welfare/​C entral categories of portfolios

Source of Political Skills

0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100

Figure 5.4 Minister political backgrounds—​Economics/​Social Welfare/​Central categories of portfolios

10. The percentage of insiders in Central category portfolios is significantly differ-

a minority of the women in Social Welfare posts have political connections).

Stereotypically masculine/​f eminine policy domain portfolios

In contrast to c­hapter 4, where our comparison of masculine/​feminine

EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017 P2:21

political insider (see note)

Figure 5.5 Minister political backgrounds—​ stereotypically masculine/​ feminine policy

Ministers in High Visibility category posts

Figure 5.6 shows that there are no measures of political background

political insider (see note)

Figure 5.6 Minister political backgrounds—​High Visibility/​Not High Visibility categories

High Visibility category post (2008–​10). She had political experience in

EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017 P2:21

nected to the president when he appointed her as a replacement minister.

DO MINISTERS WITH POLITICAL CONNECTIONS

As an additional exploration of whether women and men are held to differ-

13. Latin America Weekly Report, July 18, 2000.

EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017P2:21

% of ministers with extensive experience related to

Women Men p-​value for row

Type of political connection

Some examples illustrate the cases presented in table 5.1. Graciela

rience related to the portfolio. This is a very positive finding in light of

EBSCO Publishing : eBook Collection (EBSCOhost) - printed on 11/29/2017P2:21

are not being integrated in the US cabinet (Borrelli 2010).

administration will need its female appointees to be as able as the men to

CODING SUPPORT RESOURCES AND GROUP LINKS

We constructed minister bios based on publically available information.

country and internationally, which sectors of society or interest groups are

associations overlapped with the policy purview of their portfolio (185

stance, serving on corporate boards would count as links to business. If the

DO MEN AND WOMEN DIFFER IN THEIR GROUP LINKS?

ministry clients (see note)

women’s groups ***

any type of group link

Figure 5.1 Minister political backgrounds—data set overall

transportation (2006–09). Before entering the cabinet she worked in state

Figure 5.2 Minister political backgrounds—initial/replacement ministers

Figure 5.3 Minister political backgrounds—by country

Economics/S ocial Welfare/C entral categories of portfolios

Figure 5.4 Minister political backgrounds—Economics/Social Welfare/Central categories of portfolios

Stereotypically masculine/f eminine policy domain portfolios

In contrast to chapter 4, where our comparison of masculine/feminine

Figure 5.5 Minister political backgrounds— stereotypically masculine/ feminine policy

Figure 5.6 Minister political backgrounds—High Visibility/Not High Visibility categories

High Visibility category post (2008–10). She had political experience in

Women Men p-value for row

Figure 6.1 Minister group links—data set overall

Figure 6.2 Minister group links—initial/replacement ministers

from 2007–08. She had links to international organizations working on

Figure 6.3 Minister group links—by country

Economics/S ocial Welfare/C entral categories of portfolios

Figure 6.4 Minister group links—Economics/Social Welfare/Central categories of portfolios

Stereotypically masculine/f eminine policy domain portfolios

Figure 6.5 Minister group links—stereotypically masculine/feminine policy domain

Figure 6.6 Minister group links—High Visibility/Not High Visibility categories of posts

Argentina, 2005–10) had policy expertise related to her portfolio coming

tablishment). Monica Nagel (minister of justice in Costa Rica, 1998–2002)

Riggs 1996; Schwindt-Bayer 2006, 2010a). Analyzing survey data from

Women typically were first appointed to cabinets in a social welfare post—

Women are over-represented in Social Welfare category posts and stereo-

Appointment to Economics/S ocial Welfare/C entral

Figure 7.1 Effect of being a woman on appointment to Economics/Social Welfare/Central