The effects of primaries on electoral performance: France and Italy in comparative perspective

Abstract

For some years now the clamour for the democratisation of internal party politics has been spreading. This issue has been addressed in Italy and France. In the run-up to the 2013 parliamentary election in Italy and the 2012 presidential election in France, leftist coalitions held open primary elections to select their chief executive candidates, with Pier Luigi Bersani and François Hollande finally emerging as nominees. In this article, we examine whether the levels of turnout and competition in the party primaries are associated with the results of the general election. Using data at the level of the Italian province and the French département, we show that the results of the general election were affected by the primaries. We show that turnout enhanced electoral performance, while competition depressed it. The positive effects of turnout were greater than the negative effects of competition.

http://link.springer.com/article/10.1057/s41253-016-0007-4

Full text

Original Article
The effects of primaries on electoral
performance: France and Italy
in comparative perspective
Marino De Luca
a,
*and Fulvio Venturino
b
a
Department of Social and Political Science, University of Calabria, Via P. Bucci, Cubo 0,
87046 Arcavacata di Rende, CS, Italy.
E-mail: delucam@unical.it
b
Department of History, University of Cagliari, Via San Giorgio 12, 09124 Cagliari, Italy.
E-mail: fventurino@unica.it
*Corresponding author.
Abstract For some years now the clamour for the democratisation of internal party
politics has been spreading. This issue has been addressed in Italy and France. In the
run-up to the 2013 parliamentary election in Italy and the 2012 presidential election in
France, leftist coalitions held open primary elections to select their chief executive
candidates, with Pier Luigi Bersani and Franc¸ois Hollande finally emerging as nomi-
nees. In this article, we examine whether the levels of turnout and competition in the
party primaries are associated with the results of the general election. Using data at the
level of the Italian province and the French de´partement, we show that the results of the
general election were affected by the primaries. We show that turnout enhanced elec-
toral performance, while competition depressed it. The positive effects of turnout were
greater than the negative effects of competition.
French Politics (2016). doi:10.1057/s41253-016-0007-4
Keywords: primary election; elections; Italy; France; participation; competition
Primary Elections and Their Critics
For more than a couple of decades, the literature on political parties has suggested
that parties have been suffering from an internal crisis (Dalton and Wattenberg,
2002; Dalton et al,1984; Ignazi, 2014; Webb et al,2002). This literature suggests
that representation in democratic, Western-style countries is being challenged by
new kinds of political actors, such as social movements, and internet-led
democracy. As a consequence of these developments, researchers have noticed a
Ó2016 Macmillan Publishers Ltd. 1476-3419 French Politics
www.palgrave.com/journals

steady decline in both the membership of parties and voter turnout over the past
half century (Franklin et al,1992; Franklin, 2004).
Although their support is declining and the character of their political activism is
undergoing rapid changes (Scarrow, 2014; van Haute and Gauja, 2015), these
‘parties without partisans’ are not inert when faced with a loss of legitimacy.
Rather, they react to these challenges by adopting new forms of organisation and
new patterns of internal democracy. These reforms have sometimes been criticised
as a manipulation of party members by party leaders (Katz, 2001). Whether or not
that is the case, the role of the ordinary people has generally increased to affect all
the functions performed by political parties, including the selection of candidates
and party leaders.
The introduction of primary elections is the most noticeable reform adopted by
parties to expand the inclusiveness of their decision-making processes. In a
nutshell, a party adopting a primary system is accepting that its candidates and/or
leaders are chosen by party members or by all the citizens. In the former, the
selection is defined as a ‘closed’ primary election, also known as a one-member-
one-vote (OMOV) process; the latter is called an ‘open’ primary.
A party is composed of several related branches, and if political recruitment
changes from an oligarchic to a democratic form, then this change is likely to affect
the whole organisation. The consequences brought about by the use of primary
elections can be very complex, and unsurprisingly research was initially conducted
with reference to party politics in the USA. However, the early results were
disappointing for the supporters of the intra-party democracy: parties adopting
primaries tasted defeated in the ensuing general elections, especially when
primaries are competitive and/or negative (Hacker, 1965).
Today, primaries are not a special characteristic of American politics. This tool
of democracy is now also used extensively in Latin America (Carey and Polga-
Hecimovic, 2006), Western Europe (Barbera
`et al,2015; De Luca and Venturino,
2015), Asia (Narita et al,2015), and Africa (Ichino and Nathan, 2013). The
expansion of the set of cases has also made the assessment of the political
consequences produced by the primaries on political parties, party systems, and the
political systems at large more problematic. In contrast with most American
studies, primaries are now said to lead to parties avoiding major internal conflicts,
to improving the candidates’ image, and their fund-raising capabilities. Thus,
primaries are nowadays often considered to be an asset for electoral success, rather
than a liability as previously believed.
Primary Elections and Their Consequences: An Overview
Due to the mounting importance of this topic, several studies dealing with the
methods for selecting leaders and candidates have been published in recent years.
De Luca and Venturino
Ó2016 Macmillan Publishers Ltd. 1476-3419 French Politics

One of the most important is the comparative research of Cross and Blais (2012)on
the leadership of 25 parties in English-speaking democracies from 1965 onwards. A
second piece of research, coordinated by Pilet and Cross (2014), analysed the
selection of party leaders in 12 European countries, plus Australia, Canada, and
Israel. Another research project, directed by Sandri et al (2015), analysed the
impact of primaries on the functioning of political parties in some European and
Asian countries.
In comparison with this empirically oriented research, Hazan and Rahat (2010)
aimed to provide a stronger theoretical underpinning to the study of primary
elections. According to them, the candidate selection methods differ in their levels
of inclusiveness and generate noticeable consequences for four dependent
variables: participation, representation, competition, and responsiveness. Their
conclusions are quite pessimistic. At the very least, no method of selection can
simultaneously maximise party democracy across the four dimensions. At best, it
may be possible to obviate the most damaging consequences by adopting a mixed
system, but the growth of inclusiveness is likely to generate major pathologies, and
therefore party democracy cannot work in principle.
Hazan and Rahat (2010) have been criticised for their thin and sometimes
anecdotal evidence. Whatever the case may be, all of the scholars operating in this
new field of research have adopted a theoretical perspective strongly based on their
work. For instance, Indriðason and Kristinsson (2015) have empirically examined
the four dimensions proposed by Hazan and Rahat in reference to the Icelandic
case. Although little known and rarely quoted, Iceland is a European country where
the primary elections have been in vogue for decades, and the tradition is therefore
deep-rooted. Thus, Indriðason and Kristinsson (2015) make use of an impressive
data set covering the four major Icelandic parties at all parliamentary elections
from 1971 to 2009. In contrast to Hazan and Rahat (2010), they reach a more
positive conclusion about primaries and their consequences. For instance, primaries
in Iceland did not restrain female representation, nor they did damage the
cohesiveness of parliamentary parties. Instead, they promoted the rejuvenation of
the parliamentary political class and boosted party membership. In the same vein,
the study by Ramiro (2016) on Spain finds that primaries provide an electoral
bonus. The main goal of Ramiro’s research is to examine whether the OMOV
primaries held by the Socialist Party affected the party’s electoral result. His
research shows that parties promoting primaries have obtained better results than
those that did not.
Similar inconclusive results have been reached by scholars studying primary
elections outside Europe. On the one hand, several researchers have shown that
primaries’ competitiveness alone does not impact on parties’ electoral perfor-
mances (Carty and Eagles, 2003; Galderisi et al,2001). Nevertheless, referring to
US primaries, Carey and Enten (2011) and Adams and Merrill (2008) endorse the
mainstream idea that candidates are unappealing to the general electorate because
The effects of primaries on electoral performance
Ó2016 Macmillan Publishers Ltd. 1476-3419 French Politics

they are chosen by ideologically extreme selectorates. Furthermore, parties
adopting primaries are damaged, especially when primary campaigns are negative,
causing factional conflicts and deteriorating candidates’ image. On the other hand,
several scholars have argued that primaries may have a positive impact on parties’
electoral performance and public image. This happens because an inclusive
candidate selection process will probably enhance sympathisers’ and voters’
mobilisation in the following general election (Hazan and Rahat, 2010).
In this article we address the relationship between the use of primaries and
parties’ electoral performance in the following general election. To do so, we deal
with the open primaries of the two main left-wing parties in Italy and France,
namely the Democratic Party (PD) and the Socialist Party (PS). The article
proceeds as follows. In the next section we detail the rules and the political context
of the primaries promoted by the PD in Italy and the PS in France. Then, we
describe our approach and sketch the data we make use of. Finally, we present our
results and their theoretical implications.
Leftists Go to Primaries: The Democratic Party, the Socialist Party,
and the Selection of the Chief Executive Candidate
Italy: Towards the 2013 Parliamentary Election
After the demise of the party system in the 1990s, Italian parties adopted different
strategies. On the one hand, major rightist parties, Forza Italia – also named Popolo
delle Liberta
`– and Lega Nord, were created to promote the political interests of
their self-enthroned leaders. On the other, the leftist parties reformed the
procedures for selecting candidates and leaders by injecting a greater inclusiveness
among the electorate. As a consequence, the leftist Italian parties have become the
most enthusiastic promoters of open primaries in Western Europe.
At the national level, the first national primary election was launched in 2005 by
the centre-left coalition for its prime ministerial candidate. In this case, 4.3 million
selectors turned out, with Romano Prodi winning both the primary and the
following 2006 parliamentary election. After that, the Democratic Party since its
inception in 2007 relied on primaries to select its first leader, Walter Veltroni
(Pasquino, 2009; De Luca and Venturino, 2010), his successor, Pierluigi Bersani, in
2009 (Pasquino and Venturino, 2010), and, finally, in June 2012 Bersani himself
announced that a primary election would be held in the autumn to choose the chief
executive candidate for the 2013 general elections.
Thus, in November and December 2012, a leftist coalition, named Italia. Bene
Comune, was created through the agreement of four parties: in addition to the
leading Democratic Party, they were Left Ecology Freedom (SEL), the Democratic
Centre (CD), and the Italian Socialist Party (PSI). A board representing all of these
De Luca and Venturino
Ó2016 Macmillan Publishers Ltd. 1476-3419 French Politics

parties – and including their leaders – screened the candidates, allowing five of
them to contest the election. Although its statute clearly stated that the party leader
was the only admissible nominee, the PD fielded three candidates: the incumbent
leader Pier Luigi Bersani; the mayor of Florence, Matteo Renzi; and the regional
councillor, Laura Puppato. The other candidates were the SEL leader, Nichi
Vendola, and the CD deputy Bruno Tabacci. All citizens could participate after a
pre-registration, an endorsement to support the coalition manifesto, and a
contribution of two Euros.
As shown in Table 1, no candidate secured a majority of the three millions of
voters attending the first round. The following week, the top two candidates,
Bersani and Renzi, competed in a second round, with the latter finally gaining
60.9% of the vote.
On December 30, after the primaries for the premiership, semi-open primaries
were promoted only by PD and SEL to select their candidates for parliament. This
prolonged cycle gave the leftist coalition great momentum; nevertheless, the results
of the February parliamentary elections were disappointing (Seddone and
Venturino, 2015). In fact, the coalition led by Bersani was unable to reach a
majority in both Houses and in the following months it broke up, with Bersani
himself being persuaded to step down from the party leadership. The PD entered a
coalition government, led by the Democrat MP Enrico Letta, with rightist and
centre parties, led respectively by Silvio Berlusconi and the former premier Mario
Monti.
France: Towards the 2012 Presidential Election
The sequence of Italian-style primaries seems to have produced appealing results
for other European parties. Inclusive selections of candidates are extraneous to
French parties, but in 2006 the PS organised a closed presidential primary election
for the first time – which was won by Se
´gole
`ne Royal – in which only enrolled
members had the right to vote (Dolez and Laurent, 2007; Le Gall, 2007).
Table 1: Results of the centre-left coalition primary to select the prime minister candidate in Italy in
2012
Candidates First round Second round
N Valid votes % N Valid votes %
Pier Luigi Bersani 1395,096 44.9 1,706,457 60.9
Matteo Renzi 1,104,958 35.5 1,095,925 39.1
Nichi Vendola 485,689 15.6
Laura Puppato 80,628 2.6
Bruno Tabacci 43,840 1.4
Total valid votes 3,110,211 100 2,802,382 100
The effects of primaries on electoral performance
Ó2016 Macmillan Publishers Ltd. 1476-3419 French Politics

Afterwards, writing in the French press intellectuels, practitioners and political
theorists urged the Socialist Party to adopt the Italian primary system (De Luca,
2015). Finally, in 2009, a detailed report entitled Pour des primaires ouvertes et
populaires was presented by the party’s renewal secretary, Arnaud Montebourg.
Later, the party officially adopted a primary framework similar to the Italian
Democratic Party. Approaching the 2012 pre´sidentielle, it ran an open primary
explicitly inspired by the Italian experience.
1
The so-called primaires citoyennes were held in October 2011 (Lefebvre, 2011;
De Luca, 2014) and were organised by the Socialist Party and the Radical Party of
the Left (PRG) using a two-round electoral system. For prospective voters, it was
necessary to pre-register on an electoral list, to make a contribution of one Euro,
and to sign a charter of party values. As listed in Table 2, six candidates ran in the
first round. As no candidate obtained 50% of the votes, the top two candidates,
Franc¸ois Hollande and Martine Aubry, contested the second round, with Hollande
securing 56.6% of the vote. Hollande then defeated the Gaullist incumbent, Nicolas
Sarkozy, in the run-off for the presidential election, and a few weeks later the leftist
parties secured a large majority in the Assemble´e Nationale in support of the
government led by socialist prime minister Jean-Marc Ayrault.
Table 3identifies the main similarities between the regulations of the two
primaries
2
: (a) in both cases, a run-off is necessary if no candidate secures more
than 50% of the votes in the first round; (b) both primaries are open, as both party
members and non-formally enrolled citizens are allowed to vote; (c) there is a fee in
both primaries, two Euros in Italy and one in France; (d) a signature is mandatory in
both primaries. There are also some differences: (a) a pre-registration is required to
vote only in the Italian case; (b) people less than 18 years old cannot vote in Italy,
while they are allowed in France; (c) in France, migrants can vote if they are party
members, while Italy has more restrictions; (d) candidates need to be endorsed by
national or local representatives in France, by simple voters in Italy.
Table 2: Results of the centre-left coalition primary to select the presidential candidate in France in
2011
Candidates First round Second round
N Valid votes % N Valid votes %
Franc¸ois Hollande 1,036,767 39.2 1,607,268 56.6
Martine Aubry 805,936 30.4 1,233,899 43.4
Arnaud Montebourg 455,536 17.2
Se
´gole
`ne Royal 183,343 6.9
Manuel Valls 149,077 5.6
Jean-Michel Baylet 17,030 0.6
Total valid votes 2,658,667 100 2,841,167 100
De Luca and Venturino
Ó2016 Macmillan Publishers Ltd. 1476-3419 French Politics

Hypotheses and Measurements
We wish to shed light on the relationship between primary elections and the general
election, but to do so a number of ancillary points need to be developed. First, as
the main independent variables we focus on two characteristics of the primary
elections: turnout and competition. In our view, these variables are related to the
electoral performance of the promoting parties in an intuitive way. Parties will
achieve the best results in the general elections when there is a high turnout in the
primaries and low competition. In both cases, the rationale is easy to understand.
High turnout at the primaries shows that candidates are popular. Furthermore, a
large number of activists are important to fight a post-modern centralised campaign
(Norris, 2000) and to build a positive image in the public (Noelle-Neumann, 1993),
increasing their chances of electoral success. In a nutshell, crowded primaries draw
the interest of a great number of citizens who are then likely to participate in the
ensuing general election.
Table 3: Rules governing primaries in Italy (2012) and France (2011)
Italy 2012 France 2011
Election Two rounds, with a run-off if candidates
secure less than 50% of the vote
Two rounds, with a run-off if candidates
secure less than 50% of the vote
Participation All Italian citizens All French citizens
Registration Be enrolled in the Italia. Bene Comune
register between 4 and 25 November
2012
Be registered in the electoral lists before 31
December 2010
Under 18 Not admitted Admitted if attains 18 years at the time of
the 2012 presidential election;
alternatively, admitted if member of a
party
Foreigners EU citizens resident in Italy and citizens of
other countries in possession of a valid
residence permit
Admitted if member of a party
Contribution 2 Euros 1 Euro
Signature Support for the call of the coalition Italia.
Bene Comune and for the Charter of
Intents
Support for the values of freedom, equality,
fraternity, laity, justice, solidarity, and
progress
Candidature Before 25 October 2012, 20,000 signatures
of voters of the centre-left (no more than
2000 in each region)
Either, endorsement from 5% of the
Socialist MPs (17 endorsements), 5% of
members of the National Council (16
endorsements), 5% of regional
councillors (100 endorsements) from at
least 10 de
´partement and four regions, or
5% of socialist mayors of cities with
more than 10,000 inhabitants (16
endorsements) from at least four regions
The effects of primaries on electoral performance
Ó2016 Macmillan Publishers Ltd. 1476-3419 French Politics

While high levels of turnout are likely to boost parties, primaries featuring high
levels of competition may damage them. This hypothesis is well studied in
individual-level theory (Anderson et al,2005). Very competitive primaries entail
the selection of a nominee who wins only a small margin of votes more than his
challengers. In turn, a great number of voters have seen their favourite candidate
eliminated, leaving them with two alternatives in the general election. Disgruntled
but loyal voters will support the nominee from their party as the second best
candidate, while those completely disappointed will desert the polling station, and
sometimes will vote against their own party candidate. Thus, the greater the
competition at the primary elections and the greater the number of potential
deserters, the lower the number of votes gained by the nominee in the general
election.
So far we have discussed turnout and competition at primaries as our main
independent variables; however, both can be measured according to different
procedures. Our choices are as follows. For turnout in the primary elections, we
calculate the percentage as a ratio between the number of selectors and the number
of people enfranchised in the general election.
3
For the level of competition, several different indexes are also available (for a
review cfr. Kenig, 2008). To assess the level of competition in the Italian and
French primaries we prefer the classic effective number of candidates measure
proposed by Laakso and Taagepera (1979). This seems to be more practical
measure compared with other indexes that use only the result of the front runner, or
the difference between the two best competitors, i.e. the closeness of the contest.
4
In addition to the two main predictor variables, we make use of a couple of
control variables to detect the possible existence of spurious relationships. As the
result of a party in a given election is necessarily affected by the result of that party
in the previous election, we consider the percentage of the vote gained by the main
left-wing party in 2007 and 2008 for the Italian and French cases. In practice, we
collect the percentage vote for the Democratic Party and the Socialist Party in the
general election preceding those under examination at the level of the Italian
province and French de´partements.
5
In so doing, we aim to control for the
autoregressive effects created by the continuity of the vote for a party in a given
territory. Finally, we consider rates of party membership as a proxy for the strength
of the parties’ organisation. This control variable is calculated as the ratio between
formally enrolled members and inhabitants at the level of province and
de´partements.
6
Some additional information about our research design is still needed. To begin
with, in both Italy and in France the primaries to select the candidates for the role of
chief executive have been held under a two-round system. In both cases, our
analyses make use of data pertaining to the first round. There, as we have already
said, our territorial unity of analysis is province for Italy and the de´partement for
France. We have chosen these levels for several reasons: first, Italian and French
De Luca and Venturino
Ó2016 Macmillan Publishers Ltd. 1476-3419 French Politics

promoters used them to spread over the results of their primaries; second, they
make it easy to match the results of the primary and general elections; third, parties
define their local branches according to them; fourth, they allow us to analyse about
100 cases both in France and in Italy, thus escaping the small-n problem.
Finally, we describe our dependent variables. We analyse two multiparty systems
where several left-wing parties coexist that, due to the constraints of the electoral
laws, come together in pre-electoral coalitions. This is because, over and above the
Democratic Party in Italy and the Socialist Party in France, primaries have involved
other coalition partners.
7
In any case, when examining the electoral results of the
parties after their primaries, we have preferred to disregard small parties and to
focus on the pivotal Democratic and Socialist parties. In both cases, as regards the
general election, we considered the 2013 parliamentary elections in Italy and the
2012 presidential election in France. All data have been collected at provincial and
departmental level.
Results
Turnout, Competition, and General Elections: A (More) General Picture
To start with, we regress the vote for the leading parties in Italy and France at the
general election against primary turnout and the two afore-mentioned control
variables. The results reported in Table 4show that in both cases a control variable,
namely the vote gained in the previous general election, explains a substantial
amount of the parties’ support, while the membership rate reaches no statistical
significance. In both cases, primary turnout is a key predictor – higher in Italy than
in France – while the high r
2
demonstrate a good fit of the models overall.
Similarly, in Table 5the parties’ electoral performances are regressed against
competition in the primaries and the same set of control variables. As expected, for
Table 4: Primary turnout and other predictors of electoral performance in Italy and France
Italy France
B SE Sig. B SE Sig.
Constant 6.543 1.531 0.000 -2.537 1.646 0.127
Primary turnout 0.947 0.134 0.000 0.776 0.142 0.000
Vote to PD 2008/PS 2007 0.337 0.054 0.000 0.571 0.043 0.000
Membership rate -0.151 0.371 0.686 -2.642 1.359 0.055
Adjusted R
2
0.824 0.841
SE 2.622 1.713
F171.123 168.898
N110 96
The effects of primaries on electoral performance
Ó2016 Macmillan Publishers Ltd. 1476-3419 French Politics

the main independent variable negative correlations prevail in both models. But
while the competition is not statistically significant for Italy, it substantially
contributes to the goodness of fit of the model for France. Regarding the control
variables, previous votes prevail, while the membership rate continues to be
irrelevant in France but it becomes quite important in Italy.
Turnout and Competition: What Matters More?
So far we have discovered that primary turnout has improved the fortune of both
parties under examination, while primary competition has played a damaging role
in the French case, but not in the 2013 Italian parliamentary election. We have also
noted that the positive effects of the turnout seem to be greater than the negative
effects of the competition, but this point relies on only cursory and therefore
unsatisfactory evidence. We now address this crucial point with greater accuracy.
Table 6details the results of two regressions based once more on the Italian and
French data. The variables are the same as the previous analyses, the major
difference being the concurrent use of primary turnout and competition as
predictors. To assess the relative weight of each independent variable, the
table shows also standardised regression coefficients (beta). In both cases, primary
turnout has the expected sign, and it continues to be a good predictor of the parties’
electoral performance. This appears to be a trustworthy piece of evidence showing
the positive role of open primaries when robustly attended by citizens. However,
the case for competition is less clear. It shows a negative sign in both cases, but the
score for Italy is not significant, while in France it actually depresses the vote for
the Socialist Party, although its negative consequences are less relevant than the
positive effects of turnout. All things considered, parties should not be too anxious
about the possible effects of competition when there are several viable candidates.
Turning to the control variables, votes gained in the previous elections are very
important predictors in both cases, being the second best after primary turnout in
Table 5: Primary competition and other predictors of electoral performance in Italy and France
Italy France
B SE Sig. B SE Sig.
Constant 5.629 3.949 0.157 4.373 2.785 0.120
Primary competition -5.666 6.198 0.363 -12.344 3.156 0.000
Vote to PD 2008/PS 2007 0.556 0.053 0.000 0.656 0.041 0.000
Membership rate 1.148 0.398 0.005 -0.345 1.347 0.798
Adjusted R
2
0.743 0.841
SE 3.165 1.713
F106.295 168.898
N110 96
De Luca and Venturino
Ó2016 Macmillan Publishers Ltd. 1476-3419 French Politics

Italy and the most important in France. At first glance, what is striking is the major
role of path dependency in Italy, the astonishing results of the critical 2013
parliamentary election notwithstanding, when the populist newcomer, the Movi-
mento 5 Stelle, won 25% of the votes. As regards the role of party members, in the
Italian case it is not significant, while in the France it has a little negative
importance. Thus, left-wing parties seem to gain more votes where the Socialist
Party has less members, stressing the main role of external selectors.
Discussion
Primaries have only recently been used outside the USA, and their occurrence in
European countries is so far restricted to a small number of cases. Furthermore,
there is little work on party performance after primary elections. In this article, we
have compared two major left-wing parties operating in large countries, Italy and
France, where primaries have a high political value. Our approach entailed the use
of data collected at the level of the province and the de´partement. We have reached
some interesting results. Indeed, both cases demonstrated that the impact of the
primaries is relevant for the main parties. Furthermore, by analysing two of the four
dimensions identified by Hazan and Rahat, participation and competition, our data
show conflicting conclusions. First of all, citizen participation in the primaries
seems a great starting point to a successful general election. According to previous
studies, although no method of selection can maximise party internal democracy on
all four dimensions at the same time, open primaries with a high level of
participation appear to enhance electoral performance. As regards competition, the
theoretical insight questioning the commitment of followers of the losing primary
candidates shows different results. In the Italian case, competition is not significant,
while in France it has a negative impact on electoral performance, although with
Table 6: Turnout versus competition as predictors of electoral performance in Italy and France
Italy France
B SE Beta Sig. B SE Beta Sig.
Constant 9.140 3.312 – 0.007 7.143 2.334 – 0.003
Primary turnout 0.943 0.135 0.556 0.000 0.832 0.125 0.343 0.000
Primary competition -4.548 5.141 -0.040 0.378 -13.810 2.613 -0.200 0.000
Vote to PD 2008/PS 2007 0.335 0.054 0.420 0.000 0.529 0.039 0.684 0.000
Membership rate -0.237 0.384 -0.041 0.538 -3.448 1.205 -0.134 0.005
Adjusted R
2
0.824 0.877
SE 2.624 1.506
F128.274 170.756
N110 96
The effects of primaries on electoral performance
Ó2016 Macmillan Publishers Ltd. 1476-3419 French Politics

moderate values. This shows that in France losers may be discouraged to support
the nominee, while in Italy losers are in principle more loyal, such that their
disgruntlement is not a determining factor for the result of the general election.
8
Nevertheless, both variables in the same model indicate – at least in the French case
– a greater prevalence of participation over competition. This shows that popular
involvement matters more than the underdogs’ disappointment, and that there is a
greater likelihood that parties which embrace internal democracy will score better.
Often party officials are afraid because they believe primaries will damage their
parties. Thus, if they cannot avoid them, they prefer a small number of candidates
and restricted competition. If our conclusions are confirmed by future studies,
rather than be afraid for their own careers, parties should reverse their conservative
strategies: more primary candidates, more competition, more participation, and
finally an enhanced electoral performance.
Notes
1 In the French case the Socialist Party has drawn up an internal social network for improving
mobilisation (De Luca and Theviot, 2014).
2 The reader should be aware that so far Italy and France did not adopt any public regulation, and
therefore in both countries the primaries are a voluntary and self-regulated activity.
3 Turnout in the primary elections is often calculated as a ratio between the number of selectors casting
a vote and the number of votes gained by the nominee in the ensuing general election. As our
following analyses also make use of the nominee’s percentage votes to tap the dependent variable,
we disregard this solution to prevent a possible problem of reciprocal causality.
4 Kenig (2008) has proposed a new index where the effective number of candidates is weighted by the
sheer number of competitors. Although promising, this index would be redundant to be used here.
This is because we compare several districts in a single primary election where the sheer number of
candidates is a constant rather than a variable.
5 Own elaborations based on official data supplied by the Italian and French Ministry of the Interior.
6 Own elaboration based on official data supplied by the Italian Democratic Party and French Socialist
Party.
7 As we have already said, they are Left Ecology and Freedom and the Democratic Centre in Italy; the
Radical Party of the Left in France.
8 The same result is found by Seddone and Venturino (2013) by examining two leader selections
organised by the Democratic Party in 2007 and 2009. Both selections were multistage processes with
an open primary being the key phase.
References
Adams, J. and Merrill, S. (2008) Candidate and Party Strategies in Two-Stage Elections Beginning with
a Primary. American Journal of Political Science 52(2): 344–359.
Anderson, C. J., Blais, A., Bowler, S., Donovan, T. and Listhaug, O. (2005) Losers’ Consent. Elections
and Democratic Legitimacy. Oxford: Oxford University Press.
De Luca and Venturino
Ó2016 Macmillan Publishers Ltd. 1476-3419 French Politics

Barbera
`, O., Lisi, M. and Teruel, J. R. (2015) Democratising Party Leadership Selection in Spain and
Portugal. In: G. Sandri, A. Seddone and F. Venturino (eds.) Party Primaries in Comparative
Perspective. Farnham: Ashgate, pp. 59–84.
Carey, J. and Polga-Hecimovic, J. (2006) Primary Elections and Candidate Strength in Latin America.
The Journal of Politics 68(3): 530–543.
Carey, J. and Enten, H. (2011) Primary Elections. In: J. Colomer (ed.) Personal Representation.
Colchester: ECPR Press, pp. 81–98.
Carty, R. K. and Eagles, M. (2003) Local Conflict within National Parties: The Case of Canada. Paper
presented at the Annual Meeting of the American Political Science Association, Philadelphia.
Cross, W. and Blais, A. (2012) Politics at the Centre. The Selection and Removal of Party Leaders in
Anglo Parliamentary Democracies. Oxford: Oxford University Press.
Dalton, R. J. and Wattenberg, M. P (eds.) (2002) Parties without Partisans. Political Change in
Advanced Industrial Democracies. Oxford: Oxford University Press.
Dalton, R. J., Beck, P. A. and Flanagan, S. C. (eds.) (1984) Electoral Change in Advanced Industrial
Democracies. Realignment or Dealignment? Princeton: Princeton University Press.
De Luca, M. (2014) Les e
´lecteurs socialistes dans les primaires pre
´sidentielles a
`Paris. French Politics,
Culture & Society, 32(2): 123–142.
De Luca, M. (2015) The French Socialist Party: Membership between Closed and Open Primaries.
Partecipazione e Conflitto, 8(1): 264–283.
De Luca, M. and Theviot, A. (2014) French Primary Elections and the Internet, the Social Network of
the Socialist Party, the Coopol. International Journal of E-Politics 5(3): 46–65.
De Luca, M. and Venturino, F. (2015) Democratising Candidate Selection in Italy and France. In: G.
Sandri, A. Seddone and F. Venturino (eds.) Party Primaries in Comparative Perspective. Farnham:
Ashgate, pp. 129–144.
De Luca, R. and Venturino, F. (eds.) (2010) Mobilitazione e partecipazione elettorale. Un’analisi delle
‘‘primarie’’ per l’elezione del primo segretario del Pd. Rome: Aracne.
Dolez, B. and Laurent, A. (2007) Une primaire a
`la franc¸ aise: La de
´signation de Se
´gole
`ne Royal par le
parti socialiste. Revue franc¸ aise de science politique 57(2): 133–161.
Franklin, M. N. (2004) Voter Turnout and the Dynamics of Electoral Competition in Established
Democracies since 1945. Cambridge: Cambridge University Press.
Franklin, M. N., Mackie, T. and Valen, H. (eds.) (1992) Electoral Change. Responses to Evolving Social
and Attitudinal Structure in Western Countries. Cambridge: Cambridge University Press.
Galderisi, P. F., Ezra, M. and Lyons, M. (eds.) (2001) Congressional Primaries and the Politics of
Representation. Lanham, MD: Rowman and Littlefield.
Hacker, A. (1965) Does a ‘‘Divisive’’ Primary Harm a Candidate’s Election Chances? American
Political Science Review 59(1): 105–110.
Hazan, R. Y. and Rahat, G. (2010) Democracy Within Parties. Candidate Selection Methods and their
Political Consequences. Oxford: Oxford University Press.
Ichino, N. and Nathan, N. (2013) Do Primaries Improve Electoral Performance? Clientelism and Intra-
Party Conflict in Ghana. American Journal of Political Science 57(2): 428–441.
Ignazi, P. (2014) Power and the (ll)legitimacy of Political Parties. An Unavoidable Paradox of
Contemporary Democracy?. Party Politics 20(2): 160–169.
Indriðason, I. H. and Kristinsson, G. H. (2015) Primary Consequences: The Effects of Candidate
Selection through Party Primaries in Iceland. Party Politics 21(4): 565–576.
Katz, R. S. (2001) The Problem of Candidate Selection and Models of Party Democracy. Party Politics
7(3): 277–296.
Kenig, O. (2008) Democratization of Party Leadership Selection: Do Wide Selectorates Produce More
Competitive Contests? Electoral Studies 28(2): 240–247.
Laakso, M. and Taagepera, R. (1979) ‘‘Effective’’ Number of Parties. A Measure with Application to
Western Europe. Comparative Political Studies 12(1): 3–27.
Lefebvre, R. (2011) Les primaires socialistes. La fin du parti militant. Paris: Raison d’Agir.
The effects of primaries on electoral performance
Ó2016 Macmillan Publishers Ltd. 1476-3419 French Politics

Lefebvre, R. (2013) Le militantisme socialiste n’est plus ce qu’il n’a jamais e
´te
´.Politix (2): 7–33.
Lefebvre, R. (2015) La proximite´ en politique: Usages, rhe´toriques, pratiques. Rennes: Presses
Universitaires de Rennes.
Le Gall, G. (2007) Le primaire au PS: L’irre
´sistible sscension pre
´sidentielle de Se
´gole
`ne Royal. Etat de
l’opinion, September 2005-December 2006: 47–74.
Narita, Y., Nakai, R. and Kubo, K. (2015) Democratising Party Leadership Selection in Japan and
Taiwan. In: G. Sandri, A. Seddone and F. Venturino (eds.) Party Primaries in Comparative
Perspective. Farnham: Ashgate, pp. 129–144.
Norris, P. (2000) A Virtuous Circle. Political Communications in Postindustrial Societies. Cambridge:
Cambridge University Press.
Noelle-Neumann, E. (1993) The Spiral of Silence: Public Opinion –Our Social Skin. Chicago:
University of Chicago Press.
Pasquino, G. (eds.) (2009) Il partito democratico. Elezione del segretario, organizzazione e potere.
Bologna: Bononia University Press.
Pasquino, G. and Venturino, F. (eds.) (2010) Il partito democratico di Bersani. Persone, profilo e
prospettive. Bologna: Bononia University Press.
Pilet, J. B. and Cross, W. P. (eds.) (2014) The Selection of Political Party Leaders in Contemporary
Parliamentary Democracies. London: Routledge.
Ramiro, L. (2016) Effects of Party Primaries on Electoral Performance. The Spanish Socialist Primaries
in Local Elections. Party Politics 22(1): 125–136.
Sandri, G., Seddone, A. and Venturino, F. (eds.) (2015) Party Primaries in Comparative Perspective.
Farnham: Ashgate.
Scarrow, S. (2014) Beyond Party Members: Changing Approaches to Partisan Mobilization. Oxford:
Oxford University Press.
Seddone, A. and Venturino, F. (2013) Bringing Voters Back in Leader Selection. The Open Primaries of
the Italian Democratic Party. Modern Italy 18(3): 303–318.
Seddone, A. and Venturino, F. (2015) The Partito Democratico after the 2013 elections: All change?.
Journal of Modern Italian Studies 20(4): 474–490.
van Haute, E. and Gauja, A. (eds.) (2015) Party Members and Activists. London: Routledge.
Webb, P., Farrell, D. M. and Holiday, I. (eds.) (2002) Political Parties in Advanced Industrial
Democracies. Oxford: Oxford University Press.
De Luca and Venturino
Ó2016 Macmillan Publishers Ltd. 1476-3419 French Politics