Do American presidential nomination candidacies run party campaigns, or
candidatecentered campaigns? Do the candidates who thrive in the current
system run factional campaigns, or do they form broad party-based
coalitions? In this study, we look at the campaign organizations of the
leading contenders for the 2000 presidential nominations, and find that
parties and coalition-building are alive and strong in this area of American
politics.
Twenty years ago, Nelson W. Polsby argued that the presidential nominating
process ushered in by the McGovern-Fraser reforms had produced a system in
which factions, not coalitions, dominated (1983; see also Ranney 1975).
Since then, instead of worrying about who might be taking over the party
(Kirkpatrick 1976), research focused on
Media and Momentum and on money (Orren and
Polsby, ed., 1987; Bartels 1988; Mayer 1996). Parties were generally seen as bodies
that were taken over (Oldfield 1996) as a result
of presidential elections, rather than as participants in the selection
process.
Recent research by Bernstein (2000) and by Marty Cohen, David Karol, Hans Noel, and John Zaller
(2001) has raised questions about whether accepted conclusions
about the nomination process in the 1970s and 1980s still apply to the
process in the 1990s and 2000s (see also Steger
2000; Adkins and Dowdle 2002). Students
of parties have increasingly recognized the importance of an Expanded
Party (Bernstein 1999) composed of
informal networks of party activists, campaign professionals, consultants,
and the staffs of elected offi-cials as well as formal party organizations
(Schwartz 1990; Trish
1994; Kolodny and Logan 1998; Bernstein 1999; Monroe
2001). This Expanded Party allows, for example, for the placement
of party resources into congressional candidacies without the necessity of a
formal endorsement.
The rules of the game remain the same. Formal party organizations are still,
for all practical purposes, barred from playing a direct role in the
nomination process. What has changed is that robust party networks have
found ways to surmount those barriers. For the presidential nomination
process, early reports about the Expanded Party focus on two ways parties
can influence nomination outcomes: through elite endorsements which provide
resources directly and which provide useful signals to other party leaders,
and through direct involvement in candidate campaign organizations (Cohen et al. 2001; Bernstein
2000). It is the latter that we look at in this paper. As Cohen,
Carol, Noel, and Zaller point out, what is important here is whether there
exist a fairly limited group of party activists and political professionals
to which candidates for president must turn. For them, the question is
whether candidates who seek the nomination must compete for scarce
resources; in this case, the resource of interest is people to organize and
operate attempts to win the nomination. We concur and add a second important
reason why it is important whether candidate campaign organizations are
filled with people from the party network. We believe it changes the nature
of the candidacies themselves, and subsequently the nature of the
presidencies that emerge from those candidacies.
Party-Centered Campaigns
We examine the candidate campaign organizations of the four most
successful candidates for the 2000 nominations: Republicans George W.
Bush and John McCain, and Democrats Al Gore and Bill Bradley. For each
campaign, we identified a total of 17 key leadership positions.1
Not every position was filled within each
campaign, and in some campaigns more than one person held the
same position when one person replaced another. In the latter
cases, we include both people in our data.
For each of
the people who occupied these positions, we constructed career histories
based on public records, primarily articles available through the
LexisNexis database, profiles in the
Federal and
Congressional Staff Directories, and Federal Election
Commission records, accessed through PoliticalMoney-Line.com. These
sources were sufficient to give us at least a good sketch of virtually
all of the careers involved. This includes people such as Donna Brazile,
Mark Fabiani, Mark Penn, and Carter Eskew from the Gore campaign; Gina
Glantz, Anita Dunn, and Will Robinson from the Bradley campaign; Joe
Allbaugh, Karen Hughes, Karl Rove, and Fred Steeper from the Bush
campaign; and Rick Davis, John Weaver, Bill McInturff, and Mike Murphy
from the McCain campaign. We recognize that our professional biographies
are apt to be incomplete, but are confident that we have sufficient
information about the most important players because the top people in
the campaigns are also the ones who attract the most press attention.
Beyond that, our coding scheme makes it likely that missing data will
hurt, not help, our hypothesis. Our basic concern is whether, for
example, Bush staff are primarily loyal to Governor Bush or whether they
are Republicans who happen to be working for Bush at the present time.
Most media profiles tend to stress the player's personal connections,
since they are part of a natural story line about how the candidate and
the campaign operative hooked up together. Therefore, any bias
introduced by incomplete biographies will tend to result in false
codings for candidate-centered personnel.
Our interest here is in determining whether those within these four
candidate campaign organizations are candidate loyalists (that is, loyal
to the candidate for whom they were involved in 2000), party loyalists,
both, or neither. To that end, we code each campaign official according
to three levels of candidate loyalty: whether prior to the 2000 election
cycle he or she had any, five years, or ten years of connections to the
candidate.2
We consider those who have
been with George W. Bush since his 1994 gubernatorial campaign
as passing this toughest test because Bush's political career is
so short compared to the other three candidates. We consider
veterans of the Clinton-Gore administration or campaigns to be
personally loyal to Gore only if there was evidence of a
specific connection with the Vice President.
Possible
connections include positions in earlier campaigns, financial
contributions to previous campaigns, or staff positions with that
politician.
We define party loyalty as having some connection with a formal party
organization or being associated with a series of same-party candidates.
We will code campaign personnel as party loyal under this definition if
they have worked for, contributed to, or held a position in a formal
party organization, or if they have given money to more than five
candidates.
Party Loyalty and Candidate Loyalty
Table 1 reports the percentage of each campaign
organization that had prior ties to the candidate. It shows that by the
time candidates are ready to run for president, they have quite a few
people to call on who have supported them in the past. The low number
here is Al Gore, and still almost a quarter of his campaign organization
had previous ties to the vice president. That number may be artificially
low because of the coding decision to omit Clinton-Gore service from the
calculation, because Gore's campaign was made up of Clinton-Gore
veterans, not strictly Gore people. At the same time, the overall totals
indicate a limited role for candidate loyalists. About two out of five
people within all the campaigns had some prior experience with their
candidate. But when we apply the five-or ten-year tests, that number
falls off dramatically, to about one in six (for the five-year test) and
one in seven (for the tenyear test). Longtime successful politicians run
for president in part by calling on previous supporters for their
candidate campaign organization. But personal connections are not
enough, and few of these previous supporters are those who have
dedicated their political careers—or even a large fraction of their
political careers—to that candidate. And, therefore, while candidate
connections are common within these candidacies, real candidate loyalty
is a small fraction of the operation.
While candidate connections are common, party loyalty is also common. To
begin with, only two of the 64 people involved in these campaigns had
any record of partisan violation (such as a Bradley person having
previously worked for or contributed to Republicans). People involved in
the elite level of presidential campaigns have plenty of political
experience, but it is almost always strictly limited to one party or the
other. Table 2 reports on our more active
version of party loyalty, which involves connections with formal party
organizations or multiple party politicians. Bradley's campaign
contained the fewest party-connected personnel, but that still accounted
for over half (54%) of the people within the Bradley campaign. The other
three campaigns contained far more party loyalists. Overall, about seven
in 10 of our 64 cases were party loyalists by this definition. Again, we
consider these very conservative estimates. Our biographies are not
complete, and we think it is far more likely that we would have missed a
job with the Arizona Republican Party (for a McCain supporter) than a
contribution to a Democrat. Indeed, we have no record here of
contributions to state and local candidates; it is quite likely, we
believe, that at least some of our campaign personnel might have
qualified as party-loyal if we had records of their contributions to
state and local candidates.
We can see the limits of candidate loyalty by examining the intersection
between the two groups. Table 3 reports these
data. The longer the relationship with the candidate, the less likely
campaign personnel are to fall into our category of party loyalty. Among
those with no prior connection to the 2000 candidate, about three out of
four are party-loyal, while those with the longest candidate loyalty are
about evenly split between party loyalists and those who do not meet
that definition. Indeed, we suspect that the lack of “party loyalty” in
some of these cases is simply an artifact of short careers.
With 70% of campaign personnel party loyal, it is clear that candidates
do, in fact, turn to the pool of party activists and campaign
professionals when they put together their campaigns. But the skills
needed to conduct a presidential nomination campaign are not found
exclusively with those we identify here as party loyalists. We agree
with Cohen, Karol, Noel, and Zaller that it is wrong to regard
presidential campaigns as simply candidate-centered, because these
campaigns are closely tied to the Expanded Party. But at the same time
each of these campaigns does have an element of candidate loyalty. We
would describe these campaigns as primarily party oriented but
with strong candidate-centered elements.
Ready for 2004? What role will the Expanded Party play
in the presidential hopes of these five Democratic candidates?
From left to right: Sen. John Kerry, Vermont Gov. Howard Dean,
Rep. Richard Gephardt, Sen. Joseph Leiberman, and Sen. John
Edwards. Photo: AP/Evan Vucci.
Factions
When candidates have no need to attract the various groups that make up
each of our large, diverse political parties they will be less likely to
be able to compromise and govern effectively (Polsby
1983). A candidatecentered campaign is a form of a factional
campaign, and a candidate who manages to build a true candidatecentered
campaign organization is likely to run a White House that fails to hear
groups within the party, regardless of the lessons learned or not
learned during the campaign. Because the post-reform system requires
that candidates build large organizations in order to contest the
nomination, we can see just by observing their candidate campaign
organizations many of the party factions (if any) that the candidate
intends to represent—as well as whether the candidate is attempting to
mobilize a faction or form a broader coalition.
Our unit of analysis here is the campaign: we want to determine whether
each of the campaigns appears to be a coalition, a party faction, or
simply the organization of an unattached candidate. In order to do that,
we examined each individual within the campaigns in order to determine
whether he or she appears to be a representative of a particular party
faction. Evidence we considered included the politicians and
organizations with which he or she has had any involvement. We also look
at the group as a whole to see whether any common threads emerge.
The Gore Campaign
We were most interested in two aspects of the Gore campaign: the
Clinton influence and the extent to which the campaign was a
platform for the Democratic Leadership Council.
Almost every one of the 22 people within the Gore campaign had some
connection to President Bill Clinton. Less than a quarter of the
Gore team, however, appeared to be strongly associated with the
outgoing president; only one, for example, had been with Clinton
before the 1992 campaign. Gore, like Clinton, embraced many of the
policies advocated by the Democratic Leadership Council (DLC), but
his campaign was not a DLC operation. There were several people from
the Gore organization who had connections to the DLC, but only
somewhere between one and three of the 22 had a strong DLCrelated
career background. The DLC provides a list of all current elected
officials affiliated with that organization, and a select number of
former elected officials (including Gore and Clinton) who were so
affiliated. We looked at each campaign contribution made by a member
of the Gore campaign in order to assess how closely the campaign was
linked to the New Democrats, and found that 17 of the 49
contributions for which we could determine whether or not the
recipient was DLC-affiliated went to DLC candidates.3
We include each candidate-cycle separately, so
contribution to the same Member of Congress in 1992 and 1994
would count as two contributions. We do not here include
contributions to Gore, Bill Clinton, or Hillary Clinton,
since the goal here is to separate the Clinton-Gore team
from the DLC influence.
Certainly, those within the
Gore campaign contributed to many New Democrats—but there were also
contributions to liberals such as Senator Ted Kennedy and
Congressman Barney Frank.
Overall, then, we find that the Gore campaign is best described as
strongly coalition-style. We see little evidence of a factional
campaign or a candidatecentered campaign. Several factions are
represented within the top levels of the campaign organization,
which is exactly what we would look for in a coalitionstyle
campaign. The campaign had surprisingly few people who were
candidate loyalists, and a healthy three-fourths of organization
personnel had party connections. The Gore campaign did have strong
Clinton administration and DLC influences, but these appear to be
influences within a coalition, not the point of the campaign.
The Bradley Campaign
Former Senator Bill Bradley's campaign was characterized by two
elements: personal loyalty to the candidate and apparent antipathy
to President Bill Clinton. As discussed above, Bradley's campaign
had both the lowest degree of party loyalty and the highest amount
of candidate loyalty, at least by the loosest definition. Few, if
any, specific interests within the party appear to be represented by
campaign staff and officials. What unites the campaign, however, is
an almost complete lack of connection to the outgoing Clinton
administration. Only one person from the Bradley campaign held a job
in the executive branch (not the White House) during the Clinton
years; one other person gave a donation to Clinton in 1992. None
donated to Hillary Clinton's Senate run in 2000.
That may have been a function of the liberal slant of the campaign.
While about a third of the candidates who received contributions
from members of the Gore campaign were affiliated with the DLC,
Bradley campaign personnel contributions to New Democrats were rare.
(Only nine of 50 contributions went to DLC members.) Only one person
from the Bradley campaign had any other professional tie to any DLC
member.
For a Democrat to be liberal is hardly evidence of a factional
campaign. But when combined with the absence of connections to the
outgoing administration, the relatively small number of party
connections, and the high levels of personal loyalty, we consider
the Bradley campaign far closer to the model of a factional
campaign.
The Bush Campaign
The Bush campaign appears to be divided between those who appear to
be national Republicans and a number of individuals who share an
overlapping constellation of interests: Texas, energy industries,
and the Bush family. Of these, energy appears to be limited to two
or three individuals. The Texans appear to be just a subset of the
Bush loyalists. This is probably the consequence of a campaign run
by a governor; one cannot have worked for George W. Bush for any
length of time prior to 2000 without either being or becoming a
Texan. As reported in Table 2, this group
(Bush loyalists, using the loosest definition) comprises close to
half of the organization—but many of those are party-connected
activists and campaign professionals for whom Bush may be simply the
most recent job. We see no evidence of representatives from other
organized interests within the party, but since the Bush campaign
has the highest percentage of those we coded as party loyalists, we
do not believe that the absence of obvious interest group
representatives within the campaign is problematic.
The Republicans do not have an organization comparable to the DLC
that we can use to test whether the Bush campaign represented a
conservative faction within the party. Instead, we looked at
Christian Coalition support scores for members of the 106th Congress
(1999–2000) who received contributions from Bush organization
personnel. Of the contribution recipients, 58 had Christian
Coalition support scores between 90 and 100; five had scores ranging
from 80 to 89; and four had scores below 80. This is not
particularly lopsided compared to the universe of all Republican
members of Congress; for example, within the Texas delegation, only
one member of the House falls between 80 and 89, and one falls below
80. The Bush campaign was a conservative campaign, but we find
little evidence of ideological factionalism there.
We do, however, find evidence of a strong Bush family influence on
the campaign. Nine of the 14 individuals considered here had
connections to former President Bush. We cannot determine to what
extent this is a natural consequence of the in-fluence of the most
recent Republican presidency or if it indicates domination of the
campaign by a family faction. Overall, the Bush campaign has both
factional and coalition elements; the former is characterized by
Bush family loyalty, while the latter is indicated by a large party
presence within the presidential campaign.
Disparate Support. The Bush campaign drew support
from state, family, and energy ties. McCain garnered support
from the Republican Party's moderate faction. Photo:
AP/Elise Amendola-Pool.
The McCain Campaign
The McCain organization also had a healthy supply of connections to
the most recent Republican administration. Half of the people under
consideration here had given donations to former President Bush,
worked in his campaigns, or worked in his administration. Unlike
Bradley's camp, the McCain campaign had a normal concentration of
party-centered and candidate-centered elements; that is, it was not
substantially different from the Bush and Gore campaigns. All of
these are good indications that the McCain campaign was a
coalitionstyle campaign.
However, the evidence when we turn to ideology is striking. McCain's
Christian Coalition score during the 106th Congress was a
conservative 91. McCain's organization, however, appears to have
been far less sympathetic to those ideas. Fewer than half of those
receiving contributions from people within the Mc-Cain organization
received Christian Coalition scores of 90–100; the distribution
breaks down as 19 with very high scores, five with scores from 80 to
89, and 17 with scores below 80. McCain organization contributions
lean heavily to moderate Republicans, to the extent that it becomes
necessary to consider McCain as the candidate of that faction within
the GOP.
Discussion
We argue that when examining the nomination process, what goes on within
candidacies is as important as what happens outside of them. Thus while
Cohen, Karol, Noel and Zaller are certainly right to emphasize the ways
in which party networks are activated in endorsements, we would argue
that this element alone would be insufficient to reach the conclusion
that a partycentered process is “Beating Reform.” We also have to show
that candidates must organize party-centered candidacies, and
furthermore candidates must build party coalitions instead of seeking to
be the winning faction. A party that chooses between factions does not
control nominations. A party that actively chooses between factions is a
step up from a party that turns the choice between factions over to a de
facto lottery, but such a party still fails to provide the real benefits
of party politics. What happens within campaigns matters not just
because it indicates party control of the external process of choosing
between candidates, but because it can be part of a process of building
party candidacies. The transition from factional, personalized
candidates to party-oriented, coalition-seeking candidacies is the
hallmark of the Expanded Party.
The present study is still preliminary, but the results are promising.
Two of our four campaigns—the two that had a fairly easy march to the
nomination—appear to be partycentered, coalition-style organizations.
While candidate loyalty is still quite important within these
organizations, especially in the Bush campaign, the candidacy is open to
party activity. We don't think that this will come as a surprise to
anyone who followed the nomination process in 2000. Gore and Bush both
seized the center in as many ways as possible; McCain and Bradley
occupied the fringes of their parties. But neither of the winning
campaigns were simply the largest or the best placed faction; in both
cases, the winning campaign was the one that most fully made the
transition to coalition politics.
We believe that factions can be found within the organizations of
coalition-builders (for it is difficult indeed to build coalitions
without factions), but characterizing those factions in anything but
very broad terms proved difficult. We suggest that one of the most
important foci for further party research is exploration of the types
and varieties of faction found within our Expanded Parties, and the
development of better tools for identifying these factions and studying
their behavior.
