post-arab spring elections: a comparison of vaa results · 2014-05-07 · preferences and party...
TRANSCRIPT
1
Authors:
Ali Çarkoğlu, Koç University, [email protected]
Andre Krouwel, [email protected]
Kerem Yıldırım, Koç University, [email protected]
Post-Arab Spring Elections: A Comparison of Voting Advice
Application Results
Abstract:
The Arab Spring protests shook several Middle Eastern countries and authoritarian
leaders of both Tunisia and Egypt were toppled as a result. Relatively competitive elections
that followed in both countries were both surprising as well as being revealing of the ensuing
new regimes and developing political dynamics in these countries. This paper aims at
enlightening the nature of political cleavages that have influenced the outcome of the post-
Arab Spring elections in five countries. We developed voting advice applications (VAAs) for
parliamentary elections in Egypt, Tunisia, Morocco, Turkey and Israel as well as for the
presidential elections in Egypt to tap into voter opinions. Users of VAA’s help us to profile
voters and assess the nature of political cleavages and voters’ preferences in the critical
post-Arab Spring elections. Obviously, VAAs do not extract representative samples of the
population, but they do provide us with important clues about the level and impact of
political activism within five different Middle Eastern societies, at least among the politically
active, well-educated and younger sections of the society. By aligning voter opinions to
deeper-lying cleavage dimensions – socially conservative versus progressive and economic
left versus right – we try to explain party/candidate choice in post autocratic political
structures. We not only find different cleavage patterns emerge in all five countries, which is
not surprising considering the historical differences in the issue agendas, party systems as
well as socio-economic characteristics, but also diagnose reflections of party system maturity
as opposed to unsettled fluidities. Our research design allows us to compare the Moroccan
elections under a consolidated authoritarian regime, with the cases of Tunisia and Egypt that
represent cases of (partly) successful weakening and replacement of ruling elites as well as
the Turkish and Israeli cases which are the two most mature democratic countries in the
region. In this respect, our findings indicate that these five Middle Eastern countries have
different levels of party responsiveness towards voters’ preferences. Not surprisingly the
Moroccan case reflects the least mature cleavage structure that reflects expected
responsiveness of the voters towards party positions. The Egyptian and Tunisian cases also
do not correspond to a socially embedded, responsive party system compared with the Israeli
and Turkish cases.
Keywords: Voting Advice Applications (VAAs), Electoral Responsiveness, Arab
Spring, Policy Congruence, Democratization, the Middle East
2
Introduction
Tunisian and Egyptian autocratic regimes were toppled in early 2011 as a result of
mass revolts that shaped the so-called Arab Spring. As a consequence of the Arab Spring
movement, founding elections were held in these two countries. In Tunisia, a constituent
assembly was elected whereas in Egypt, a government responsible for writing a new
constitution was formed as a result of the parliamentary elections. All in all, transition to a
more democratic regime type was on its way in these two countries. However, transitioning
regimes are exposed to reversals and serious setbacks in democratization efforts. Given the
lack of consolidated and institutionalized formal opposition parties in these two countries
(Langohr, 2004), the reversals that followed were not surprising. As we will exemplify the
unsettled nature of political landscape and unconsolidated nature of political cleavages were
clearly reflected in our findings from the Voting Advice Applications (VAAs). Reversals and
political crises abound in the two countries. In Egypt, Morsi’s elected regime was toppled by
a military intervention while in Tunisia, unrest and governmental crisis goes on as two high
profile politicians from the leftist secular opposition were assassinated.
Many different causes of such political crises can be diagnosed. Our objective here
will not be to entangle all the socio-political causes that underlie the historical roots of the
resulting anti-democratic tendencies that followed the Arab Spring. In order to provide a
comparative grounding to all efforts to answer this question, this paper will contrast these two
transitioning cases with the more settled democracies of Israel and Turkey as well as with the
vigilant autocracy of Morocco. Our comparative study and findings thereof suggests a simple
explanatory framework for the problems that are prevalent in the transition period or the
“boiling hot summer” that pursued the Arab Spring. A critical explanatory variable is shaped
by political parties’ electoral responsiveness. By comparing two relatively stable and
consolidated electoral democracies of the region, Turkey and Israel with the transitioning
3
cases, we show that the post-Arab Spring elections in Tunisia and Egypt reflect serious
setbacks and limitation in regards to responsiveness and policy congruence in the emerging
party systems. In addition to this comparison, including Morocco into this paper increases the
comparative leverage, as Morocco provides a case of a monarchical and non-democratic
regime with a sui generis party system.
Our comparative strategy also incorporates a methodological innovation by utilizing
and benefiting from VAAs as itsdata source. VAAs may have limitations in inference based
on a representative sample. However, their substantive utilization in different areas such as
spatial voting models and party mapping research may disclose new methodological
opportunities and challenges and thus contribute to the development of research that use
VAAs. Despite its limitations VAAs allow us to gather an admittedly unrepresentative but
quite large samples of respondents in all five countries that clearly reflect the spatial
cleavages in the regional democracies. We thus aim to illustrate the extent to which the
spatial political map for these five countries capture the political cleavages that shape
political competition in these systems. Our analyses suggest that the ability of the political
space that emerges from the VAAs capture the foundations of the party systems as a
mechanism of representation and resolution of underlying cleavages in the respective party
systems.
Cleavages and Electoral Responsiveness in the Middle East
Electoral responsiveness is the political parties’ ability to understand and react to the
voters’ preferences. Parties have varying levels of electoral responsiveness based on the party
elite’s ability to interpret mass policy preferences. Also, the literature based on Miller and
Stokes (1963) suggests that the parties’ capacity to provide voters with desired policy
outcomes for all voters determine their electoral responsiveness. For instance, Ho et. al.
(2013) argue that the salience given by the political parties to valence issues renders the case
4
of Taiwan democratization a success story. According to this argument, Taiwan could not
institutionalize its electoral system in a democratic regime without provision of necessary
incentives to the larger public. However, elite capacity and valence issues are not the only
two explanations for differing levels of electoral responsiveness. Dixit and Londregan (1996)
as well as Cox and McCubbins (1986) have suggested two competing models for explaining
selective distribution of incentives. Both of these models suggest that patronage and capacity
to deliver incentives to a limited number of voters may also play an important role in
determining the level electoral responsiveness. Club goods and patronage ties may decrease
electoral responsiveness as clientelistic parties may find it hard to reach out to a larger
constituency (Piattoni, 2011).
All of these three explanations suggest that parties respond to the voters’ preferences
if the party system and distribution mechanisms are institutionalized. For instance, political
systems which lack established opposition parties may find it difficult to provide different
elite groups with means to evaluate the voters’ preferences and respond accordingly.
Similarly, institutions may influence two things in a related manner. First of all, existing
institutions may influence party systems and thus shape the convergence between voter
preferences and party agendas. Secondly, they can also influence the policymaking processes,
determining how the voters’ preferences are translated into policy outcomes (Dalton, 2011).
The impact of consolidated institutions has been an important topic for research in advanced
democracies. For instance, early findings indicate that the voters in more proportional
electoral systems are more satisfied with democracy (Anderson, 1998; Lijphart, 1999; Norris
1999). Additionally, electoral responsiveness comes in two versions. There can be
majoritarian or proportional mechanisms to establish congruence between parties and voter
preferences. However, both election systems require a level of institutionalization in order to
function properly. Without such functions, democratic systems cannot foster electorally
5
responsive outcomes, which produce governments that reflect the voters’ policy preferences
(Blais & Bodet, 2006; Powell, 2013). Previous research in this area suggests that the
congruence between political responsiveness and the electorate’s reaction to this
responsiveness relates to the larger system efficiency in democratic regimes (Costello,
Thomassen, & Rosema, 2012; Soroka & Wleizen, 2004). However, most of the empirical
research on this topic is based on advanced industrial democracies of Europe and Northern
America.
A central concern of electoral responsiveness relates to issue voting. Downs (1957)
has argued that the voters choose the party that is closest to them in the left-right cleavage.
Unlike some contenders (such as McAllister, 2007; Karvonen, 2010) models based on
Downsian issue-based voting minimize the impact of the political leaders and charismatic
figures. Against these spatial voting models, some have argued that rather than the distance
between voters and parties in a given issue, the direction of (dis)agreement is more critical
(Rabinowitz & Macdonald, 1989). In any case, when trying to account for the party choice,
both theories focus on preferences of the voters. In the Downsian case, self-placement of the
voter himself as well as the alternative parties along a single dimension in electoral surveys
suggests a distance whereas in the contending theoretical account, the distance is not as
important as the direction of agreement/disagreement between the voter and the parties.
Based on these two theoretical accounts which focus on voters’ policy preferences, this paper
also deals with issue voting and considers the role of such models for the Middle Eastern
cases in a comparative manner.
Middle Eastern cases are understudied with regards to the impact of institutionalized
parties, issue based voting, social cleavages and responsiveness. This is not unexpected given
that the region is considered as one of the least democratic regions in the world. The two
exceptional Middle Eastern cases are Israel and Turkey with their relatively higher levels of
6
democratic consolidation. Deegan-Krause (2007) indicates that only Turkey and Israel were
able to produce competitive party systems and therefore, studying cleavage structures in the
Middle East usually focus only on these two cases. According to this line of research, both
countries are moving towards a full convergence of social cleavages with the structural and
institutional elements of vote choice. For instance, Shamir and Arian (1999) argue that
identity issues are becoming more prevalent for defining the Israeli vote. This also causes a
mixture between policy preferences and party support while fortifying the existing cleavages.
A similar phenomenon appears in Turkey as well. Hazama (2003) has argued that the
“social cleavages and the party system in Turkey seem to be heading for convergence”.
Similarly, Çarkoğlu and Hinich’s spatial analysis reveals the importance of two cleavages in
the country. Their findings indicate that two dimensions; one based on secularism vs.
Islamism and a second on the Turkish nationalism vs. the Kurdish identity, capture the
essence of Turkish political space. They conclude that “the Turkish case is yet another
evidence of the low dimensional nature of ideological spaces in modern democracies”
(Çarkoğlu & Hinich, 2006).
The Israeli and Turkish cases suggestthat the most democratic countries in the Middle
East are experiencing a convergence between party platforms and cleavages. In other words,
there is a functioning interaction between parties and the voters. As social cleavages
consolidate, institutionalized and formally recognized parties are able to recognize these
cleavages and act in accordance to these division lines. Although the Turkish case has
experienced several disruptions and reversals in repeated military coups such as the 1980
military intervention, the political system was able to conserve its competitive electoral
structure, albeit with a current predominant party.1 A similar regime, an even more stable and
durable one, exists in Israel. Israel was able to be the most resilient democratic regime in the
1 See Hale (1993) on Turkish military regimes. For discussions of more recent interventions, see
7
region according to the most recent Polity 4 report even in the face of the colossal security
issues which may create serious setbacks and impediments for the Israeli democracy.
Therefore, our expectation is that these two countries are home to the most responsive
political parties in the region.
On the other side of the spectrum, we were able to conduct a VAA in Morocco, which
is a monarchy with serious setbacks and limitations in regards to democratic elections.
According to Tessler and Gao (2005), Moroccan public was one of the most supportive
countries in the Arab World towards democracy. In fact, country has witnessed a controlled
opening and political reform in the late 1990’s and early 2000’s and it has established an
electoral system in which Islamist Justice and Development Party (JDP) was integrated into
the party system. However, by the means of gerrymandering the Berber-majority electoral
districts, encouraging intra-party splits in the opposition and integrating the new generations
of political and economic elites into the monarchical system, the kingdom was able to shield
itself from any electoral, oppositional challenge. According to Sater (2009), King
Mohammad VI was unable to initiate democratization despite the initial expectations.
Although mass protests and upheavals against the regime during the Arab Spring threatened
the status quo, the king was able to control the political crisis by holding early parliamentary
elections and initiating a constitutional reform. But these two actions did not change the fact
that the Moroccan political parties are hindered from responding fully to the voters’
preferences as the Kingdom limits the party platforms and their electoral campaigns. Up until
recently, before Mohammad VI promised these political reforms, the king was able to appoint
and dismiss all of the relevant political figures. Even if the cabinet and the parliament gained
some clout vis-à-vis the king, these reforms did not have any impact on the fact that parties
are incorporated into the monarchy. In other words, political parties in the country are still
relying on the king’s political power and his personal discretion. Even if the politically
8
relevant political parties are relatively older compared with the post-Arab Spring Tunisia and
Egypt, the fact that the party elites are incorporated into the monarchy is an important deficit
for electoral responsiveness. For instance, right after the 2007 legislative elections, the king’s
close friend Fouad Ali El Himma established Authenticity and Modernity Party (AMP),
successfully integrating independent MPs and political notables into the monarchy’s political
agenda (Liddell, 2010). Also, Morocco lacks clear-cut cleavages compared to the Turkish and
Israeli cases, besides the role of Islamism in the country. These two issues, namely the
incorporation of parties to the monarchy and lack of well-established political cleavages,
create serious deficits for the electoral responsiveness in Morocco (Amghar, 2007).
Despite revolutionary changes within the Arab Spring movement the Tunisian and
Egyptian cases have not yet completed their transition periods. Three post-Arab Spring
elections were held in these two countries, namely Tunisian constituent assembly elections
and Egyptian legislative and presidential elections. Most of the political parties were
established or legally recognized after respective revolutions in the two countries. It is
already well-established in the relevant literature that the Muslim Brotherhood as well as
other Islamist movements have an enormous social network for responding to the material
and service needs of their constituents (Brocker & Künkler, 2013; Clark, 2004; Diamond,
2010). Institutionalization of the incentive provision mechanisms through vast social
networks gave the Muslim Brotherhood and its political party, Freedom and Justice Party
(FJP) an advantage in Egypt. FJP was able to benefit from its networks and its experience in
formally/informally contesting the Mubarak regime. Similarly, al-Wafd party (the oldest
political party in Egypt) had an aggressive electoral campaign, reaching out to voters through
television and other media. However, neither the Islamist nor the secular fronts were able to
present a substantiated electoral platform. According to Hassan, party manifestos “addressed
Egypt’s most pressing problems like unemployment, education and healthcare with extremely
9
general phrases and promises of increased spending without specifying how the suggested
measures will be financed” (Hassan, 2013). Indeed, ideological divisions and polarization
between Islamists and secularists obscured substantive policy preferences.
In Tunisia, a similar campaign period is observed. The country also witnessed an
autocratic regime under Ben Ali, similar to Mubarak’s Egypt. After Bourgiba was declared
unfit for presidency, Ben Ali was appointed in his place in 1987. Early on in his term, he
initiated reforms by legalizing previously banned political parties and releasing some Islamist
political prisoners. However, “…by the early 1990s, Ben Ali was practicing the full
Bourguiba, cult and all…” (Henry, 2004). But, Tunisia remained as one of the most likely
candidates to experience a liberalizing transition according to Noland’s spatial and socio-
demographic models (2005). This likelihood was realized in early 2011 and Ben Ali was
forced to leave his office. The first free and fair elections to determine the seat distribution of
the constituent assembly was held in late 2011, approximately nine months after Ben Ali’s
government was ousted. Most of the political parties were founded or recognized after the
revolution, except from secular leftist Democratic Forum for Labour (DFL) and Liberties and
Progressive Democratic Party (LPDP). This did not prevent Ennahda from becoming the first
party in the elections. They were able to win 89 seats out of 217. Yet, the political
developments after the elections were not signs of a successful democratic transition. As
expected in the aftermath of founding elections during periods of transition, Tunisia
experienced serious turmoil and political impasses similar to the situation in Egypt. These
political crises in both countries indicate that they may not be considered as full democracies.
In all of the five cases, one can argue that the parties are not historically bound and
institutionalized relative to the advanced industrial countries. However, as the Turkish and
Israeli democracies as well as several Islamist parties in other three countries suggest, there is
variation across our cases in respect to party institutionalization. However, given that the
10
Turkish and Israeli cases have a history of democratic electoral experiences, our argument is
that the Egyptian and Tunisian cases will resemble Morocco in respect to the electoral
responsiveness and the level of social embeddedness of political parties in the post-Arab
Spring elections.
Method and Data
In order to measure convergence between parties and their likely voters, we have
devised six VAAs in our five cases. In these six applications, a question taps into the
propensity to vote for a given party (PTV). This question asks how likely the respondent is to
vote for each political party. Answer categories range from 0 (not likely at all) to 10 (very
likely to vote for the party). This PTV question is asked for each relevant party. Respondents
who have selected high levels of PTV for a certain party do not necessarily have to vote for
that party. However, PTV is a good indicator of how voters judge different parties. In cases
of advanced and institutionalized democracies, high levels of PTV imply congruence
between the party and the voter (van Dijk, Krouwel & Boiten, 2013). However, this may not
be the case in transitioning countries. We have also measured the Euclidean distance between
the voter and four largest political parties in each country for two main political dimensions.
This distance is based on 30 predetermined policy preference questions that are relevant for
the electoral campaign in a given country. These questions tap into both the economic left-
right and the social change vs. conservatism cleavages. Answers for the parties were
determined from their election manifestos and the measurement was based on the spatial
distance of the respondent from political parties. Thanks to these two operationalizations, we
were able to calculate the average distance between the voter and four largest parties for each
order of the PTV question. Very much in line with the Downsian spatial voting framework
this calculation should suggest a negative correlation between the respondent's likelihood to
vote for a certain party and his/her distance to that party. In other words, the higher the
11
likelihood of a respondent to judge a party positively as a high PTV score, the lower the
respondent’s distance to that party should be in the spatial plane where we measure the policy
preferences in two dimensions.
[INSERT TABLE 1 HERE]
We have selected the four largest parties in our five cases (see Table 1). The total
share of the electorate as represented by these four parties that we included in our analyses
differed across cases. This variance clearly suggests different effective number of parties
(ENP) across these party systems. The highest vote share we had across our cases was the
2011 Turkish general elections in which the parties had a share of 95.39% of the total votes
whereas the lowest was 2013 Israeli elections with 49.06% of the total votes. In fact,
calculations of ENP according to Laakso and Taagepera’s (1979) formula indicate that
Turkey had the lowest value of ENP with 2.96 parties while Israel had the largest ENP with
8.68 parties. See Table 1 for details of the selected parties across cases.
We have used six VAAs across five Middle Eastern countries during the post-Arab
Spring period. As mentioned above, Tunisian and Egyptian cases represent countries where
the Arab Spring protests have erupted and changed the political system considerably. For the
Egyptian case, we were able to analyze both the parliamentary and presidential elections and
their respective VAAs. Turkish and Israeli cases are the two most stable countries in the
region where a relatively institutionalized and functioning political system exist. On the other
hand, Morocco is an Islamic monarchy, ruled by King Mohammed VI. Our ensuing analyses
help us identify the extent to which spatial voting models apply to different types of elections
and in our case illustrate how we can compare different types of elections with each other.
More specifically we identify to what extent similar levels of electoral responsiveness exist in
Egypt, independent of the nature and type of compared elections.
12
To use these six VAAs for relevant statistical analyses efficiently, we have conducted
a logic check and removed those respondents who were not located within the country or
those who did not use the VAA in their native language (Arabic, Hebrew or Turkish for
respective countries). After doing necessary logical checks and removing a portion of the
respondents (25% of respondents on average), we still had a considerable number of
respondents. In order to complete the VAA, respondents had to finish all the necessary 30
questions asked to measure the policy preferences, which were then used to measure the
distance between parties and the respondent himself. However, respondents were free not to
answer the PTV questions about the four largest parties. Therefore, the final data sets were
comprised of those respondents who answered the issue questions together with the PTV
questions in their respective three native languages within the geographical boundaries of
each one of the five countries.
Summary statistics (Table 2) indicate that there was a considerable amount of
variance between countries in respect to the number of respondents. The highest number of
respondents was in Israel with 85,727 respondents who fit the criteria of the logic check.
Israeli VAA is followed by the Egyptian presidential election with 63,820 respondents.
Given the limits of online sampling techniques, one methodological criticism against
conducting research with VAA data can be about sampling and representativeness. Such a
criticism should be well taken, and it is possible that VAA samples may never claim
representativeness. But still, such applications enable researchers to get a grasp of the
situation without representing the whole population. Datasets collected online generally
suffer from problems of under-coverage and self-selection that can potentially bias estimates
(Bethlehem, 2010; Hooghe and Teepe, 2007). Yet, there are also benefits to opt-in web
surveys. First, our VAA for each of these elections allowed us to gather unparalleled amounts
of data from larger number of voters than traditional surveys, combining wide-ranging
13
information about their political behaviour, opinions and background characteristics. In
addition, studies have shown that computerized self-administration reduces measurement
error relative to other modes of data collection, increasing both the level of reporting and the
report accuracy of opinions and attitudes compared with more “conventional” surveys
(Kreuter et al., 2008; Sakshaug et al., 2010). Moreover, online survey questions are answered
more truthfully and carefully compared to interviewer-administered surveys (Olson, 2006).
This combination of self-selection and self-administration leads to a pool of respondents less
likely to misreport their preferences and behaviour, and thus measurement errors should be
smaller (Sakshaug et al., 2010). After recording their initial vote intention, the VAA offered
users an instant ‘reward’ in the form of ‘advice’ of which candidate most closely represents
the users’ own preferences, which intensifies the incentive to be truthful and accurate.
[INSERT TABLE 2 HERE]
More practically, we could not find nationally representative samples that cover our
five cases with comparable policy preference and PTV questions in our case. Given the
difficulties of conducting nationally representative electoral surveys in non-democratic
countries, Middle Eastern cases are especially suitable for utilizing VAAs.2 Additionally,
VAA questionnaires are more suitable for research on electoral responsiveness, issue voting
and spatial models since there are thirty questions in each VAA to measure the distance
between parties and voters unlike traditional election surveys which usually have far less
questions on these issues. As such, VAA data provide researchers with more precise
placement of voters and parties along different political dimensions. Our ensuing analyses
show that studies based on VAA samples also have face validity that can help us explain
important political developments in many different parts of the world.
2 It is important to underline that there are efforts for representative surveys in the region as well. One
prominent example is the Arab Barometer. However, neither Arab Barometer nor other similar
surveys, such as the World Values Survey, have similar questions or items that are comparable across
our cases.
14
Findings
Our findings are based on PTV questions and the distance between parties and voters.
For each case, we measured the average distance between four largest parties and their
relevant PTV orders. We expect that as the distance of individuals from the party positions
increases their likelihood to vote for that party should decline. In Israel, Turkey and Egyptian
parliamentary elections our graphical depiction clearly shows this expected trend. In other
cases such expected tendency is not observed.
[INSERT GRAPH 1 HERE]
In the Turkish case, the four largest parties shared 95.39% of the total votes. This is
the largest sum of vote shares among our five countries. This is not unexpected given that
there is a 10% national threshold and a predominant party system in which voters act
strategically.3 For all of the four relevant parties, distance towards a political party in two
dimensional space decreases as the voter’s tendency to vote for that party increases. This is
expected in countries where the parties are socially embedded and the voter’s propensity to
vote for a party overlaps with their distance to the party in the two dimensional space. The
only slightly exceptional case can be the Nationalist Action Party (NAP). Distance to the
NAP does not decrease by gradations as voters become more inclined to vote for the party.
However, there is still a clear decline even if it is not gradated. This lack of congruence
between voters and the party can be explained by several alternative factors. First of all, NAP
represents the ultranationalist far right ideology of Turkey. It is the third largest party in the
country which had about 13% of total votes in 2011. Although the nationalist issue items are
included in our 30 issue questions, the final political map may not be grasping the salient
issue dimension for the NAP.
3 See Çarkoğlu (2011) Gümüşcü (2013) and Müftüler-Baç and Keyman (2012) on the emerging
predominant party system in Turkey.
15
[INSERT GRAPH 2 HERE]
Israeli general elections of 2013 is similar to the Turkish case in the sense that four of
the most important parties’ voters and their distance vs PTV tendencies fit to our expected
pattern reflecting social embeddedness. However, the Israeli case had the smallest sum of
vote shares for the four parties with 49% of total votes unlike the Turkish case. Potentially
the remaining more than half of the electorate in Israel may be mobilized by factors not
grasped by our two dimensions and hence these parties may not reflect the expected pattern
we observe in our analyses. Even if these two cases represent the two opposite sides of the
spectrum4 for party fractionalization, they share an important empirical finding. Similar to the
Turkish case, Israeli parties are also socially embedded exhibiting the expected pattern of
linkage between distance in the issue space and PTV. Therese is relatively high levels of
responsiveness and issues that are relevant for party supporters find a reflection from the
party platforms. An interesting finding relates to Yesh Atid, the second largest party after
Likud/Israeli Beitenu electoral alliance. Country specialists were not expecting Yesh Atid to
gain so much prominence and seats in the Knesset. Indeed, the party platform also signals to
such a prominence: in every different PTV order, respondents were closer to the party on
average compared to the other three largest parties. This signals to the fact that Yesh Atid’s
electoral campaign was based on change, reformation and efficiency that were “catch all”
phrases during the 2013 Israeli electoral campaign period. The party gained prominence
thanks to its charismatic leader, Yair Lapid who is a well-known figure in the Israeli media.
Compared to relatively more institutionalized and older parties in the Israeli party system,
Yesh Atid was able to translate its original and unusual discourse into unexpected results in
the ballot.
4 As mentioned above, value for the effective number of parties was also the lowest in Turkey (2.96)
indicating that the 10% national threshold and emergence of a predominant party system in Turkey
makes the country less fractionalized compared to Israel where the threshold is 2%.
16
[INSERT GRAPH 3 HERE]
Tunisian constituent assembly elections were held on 23 October 2011, 10 months
after the revolution. All of the four largest parties were established before the revolution and
they were relatively more institutionalized compared to some newcomers in other cases such
as Yesh Atid in Israel and Free Egyptians Party in the Egyptian parliamentary elections.
However, this relative organizational strength does not directly translate to socially
embedded parties or a party system that is open to competition in issues besides from
Islamism vs. secularism. In this respect, Tunisian case is an ideal case, showing problems in
democratic transition processes. Tunisian elections do not show a gradated spatial proximity
between parties and voters. None of the four largest parties could tap into the voters’ issue
preferences in accordance with proximity. The likelihood to vote for the Tunisian parties
appear to remain more or less constant irrespective of the distance in the issue space.
Compared with the Turkish and Israeli cases where voters who are more likely to vote for a
given party are closer to the party in the issue space (except for the Turkish NAP), politically
relevant Tunisian parties hint at a finding that is suggested by earlier theoretical work on
democratization and party institutionalization. Transition periods are not finalized with the
coming of elections. On the contrary, such elections carry the risk of reversal with them.
Although we have not witnessed such a reversal in the Tunisian case right after the elections,
it is clear that party platforms do not respond to voters’ preferences. Limitations in
responsiveness and divergence in the issue space indicates that the post-Arab Spring Tunisia
is still in the process of transition and it requires further deliberative processes for the
establishment of socially embedded, responsive parties.
One interesting example in the Tunisian case is the Democratic Forum for Labor and
Liberties. Before the revolution, this party played a secondary role during Ben Ali’s tenure.
Unlike Ennahda and Congress for the Republic, Democratic Forum was already recognized
by the Tunisian government before the revolution. Also unlike the PDP, it did not boycott the
17
rigged elections. This earlier institutionalization gave Forum an advantage on responsiveness
and it hints at the relatively higher level of its embedded nature. This does not suggest that
Ennahda and other parties were not mobilizing different political groups at all. However,
electoral participation and involvement in formal political processes was an important
experience for the Forum. This does not imply that they are more advantageous in earning
votes. Indeed, Ennahda was able to get 37% of the total votes whereas the Forum got only
7%. In other words, political parties that were banned from inclusion into the formal political
processes still had a considerable support in the post-revolutionary Tunisia. But this support
did not translate into a clear-cut positioning in the issue space. Inclination to vote for a given
party in the Tunisian elections is not associated with closer issue positioning. A similar
limitation on political responsiveness and social embeddedness is also visible in the
Moroccan and Egyptian cases as well.
[INSERT GRAPH 4 HERE]
For this study, Morocco represents a yardstick to gauge the distance between a
monarchical Middle Eastern regime in which the Arab Spring uprisings were softened by
general elections held in 2011 and the King Mohammad VI implemented a constitutional
reform package. Referendum for the constitution was passed by a sizeable majority of the
voters (98.5%) and the king’s decisions saved him from a faith similar to that of Egyptian and
Tunisian presidents. Although turnout in the constitutional referendum was relatively high,
around 75%, turnout for the elections held in late November of 2011 was the lowest in the six
elections across our five cases with a 45%, indicating that the overall level of internal
political efficacy is low in Morocco. This is not unexpected given that the king has preceded
the legislative and executive powers prior to the reform. In addition to the problems of
accountability and internal efficacy, Morocco also faces serious limits in competition and
contestation. The king was able to appoint key political offices up until recently. He exercises
18
ultimate political, civil and religious authority in the country and thus it is not unexpected to
argue that Morocco represents the least democratic, i.e. the most autocratic case for our
comparison.
The Moroccan party system is home to several different party types. First of all,
Istiqlal and National Rally of Independents represent parties that are closer to the monarch.
While Istiqlal directly defends the monarchy and traditional authority of the monarch, NRoI
incorporates candidates and independent political figures that are closer to the palace. On the
other hand, the Justice and Development Party represents the moderate Islamism in Morocco.
Just as its name suggests, it is ideologically influenced by its Turkish counterpart, the Turkish
JDP which has been ruling Turkey for the last eleven years. Moroccan JDP gained
prominence in the late 1990s and became the second largest party in the legislative elections
held in 2007. Besides from JDP, our empirical results indicate that no other relevant party is
socially embedded in the Moroccan party system. In fact, even JDP’s average distances per
order are not gradated as expected. For instance, those respondents who have indicated that
they are less likely to vote for JDP (order of 1) are closer to the party in the issue space than
those who selected that they are more likely to vote for JDP (order of 6). This unusual
gradation can be an artificial effect of the VAA opt-in sampling given that only 3.5% of all
the respondents indicated that they are in the order of 1 for the JDP’s PTV score. However,
the fact that JDP has been the only viable alternative to that of the monarchy’s political clout
makes the party an exception in Morocco. In fact, Pellicer and Wegner (2012) argue that the
JDP is able to monopolize votes of the educated middle class in the country. No other party is
able to do that. After the party won the elections as a result of King Mohammad IV’s political
and constitutional concessions, it was able to hold onto the power unlike what we have been
witnessing in Egypt after the coup against Morsi’s government or in Tunisia where political
19
turmoil has been continuing thanks to assassinations of important political figures from the
opposition.
[INSERT GRAPH 5 HERE]
Egyptian parliamentary election was held in three stages between 28 November 2011
and 11 January 2012. This election was a turning point for the country as it was the first free
and fair election after the 1952 military coup by the Free Officers movement. This was also
the first election for the Muslim Brotherhood, an important Islamist movement not only for
Egypt but for the whole region. The Brotherhood participated with the Freedom and Justice
Party (FJP) and gained 37.5% of the votes followed by the radical Islamist Al-Nour Party
with 27.8% of the votes. In total, these two Islamist parties gained more than 65% of all the
votes in this election. In fact, FJP was the only party among all parties that had more
respondents who had PTV score of 6 or higher than those lower than 5. VAA results support
the argument that Islamism vs. secularism was the main issue for this election. There were
questions in the Egyptian parliamentary election VAA tapping into to the role of Sharia,
place of the Coptic minority as well as the role of women in politics. VAA results indicate
that the standard deviation of respondents’ score along the conservatism vs. change axis was
higher than their score along the economic left-right axis. Results from this VAA indicate
that both FJP as well as Al Nour responded to important issues, as their supporters would
expect. In this respect, both parties show a gradated decline in average distance among
different PTV orders even if Al Nour is categorically more distant to the voters compared to
the FJP. Similarly, Free Egyptians Party (FEP) shows the same pattern of social
embeddedness. FEP represents the socially liberal and secular groups in the country. Its
constituency is made up of the Coptic minority as well as secular civil society and prominent
intellectuals. Given that such parts of the civil society were the driving force during the
formative periods of the Egyptian revolution, the party was able to become the fourth largest
20
one in the elections. In addition, it was able to respond to the concerns and preferences of
liberal and secular voters. New Wafd (NW) is also a secular party but the difference between
it and FEP is the fact that NW is the oldest party in the country. However, the party was in
turmoil before and during the elections, the electoral coalition they struck with the FJP was
cancelled right before the elections and the NW had to participate and compete with the
alternative secular Egyptian bloc, led by Egyptian Social Democratic Party and FEP. In this
respect, the NW’s platform and electoral campaign strategy was comparatively inadequate.
Turmoil and internal conflict within the party translated into lack of responsiveness to the
voters. The fact that the party is the oldest in Egypt may have been an advantage against the
alternative FEP. However, even if the NW gained eighty thousand more votes than the
Egyptian Bloc, their party platform was the least responsive in the Egyptian parliamentary
elections.
[INSERT GRAPH 6 HERE]
Egypt was a semi-presidential country and 2012 presidential election was the first of
its type in which presidential candidates could compete for the office. Four largest candidates
got 86.6% of the votes and in the runoff, Morsi and Shafik competed for the presidency.
However, the VAA was devised before the elections and majority of the respondents used the
VAA before the first round. Questions that tap into the two axes were somewhat similar to
the earlier VAA for the parliamentary elections. However, some questions were revised and
others were added in accordance with the changes in issue saliency and political
developments in the country. Results indicate that none of the candidates were responsive
towards the voters in respect to the 30 questions asked to measure the two axes. For instance,
there was neither substantive nor statistically significant difference between different PTV
orders. For instance, those who indicated that they are not very likely to vote for Morsy
(order of 2) had an average distance of 1.44 whereas respondents who indicated that they are
21
very likely to vote for Morsy (order of 10) had a higher average distance of 1.45. Similar
examples are observed for other candidates as well. Therefore, one question is why such a
lack of responsiveness and social embeddedness prevailed in the presidential election
whereas the Egyptian parliamentary elections is closest to what we would expect in a
functional electoral system, except for Israel and Turkey.
Answer to this question may lie in artifacts of the VAA, differences in presidential vs.
parliamentary elections or the changing political conditions in Egypt. Regarding the
deficiencies in the VAA, it is outright possible to rule out the impact of the number of
respondents who have used the VAA (refer to the Table 2). It may also likely be that the
selected questions to tap into the two axes became irrelevant over the course of the
presidential election campaign. However, longitudinal analysis of responses indicates that the
respondents’ spatial distribution stayed the same over time. Regarding the limits of the VAA,
one may also argue that the selected questions do not measure the expected concept. In other
words, it is possible that there are limits in the construct validity. This criticism requires
further methodological studies on VAAs.
Besides from the problems that can be associated with VAAs, there can also be other
substantive issues. First of all, presidential election campaign can be different from political
parties for several reasons. For instance, an important division was the new political figures
vs. the old elite. This division was not very prevalent in the parliamentary elections as the old
elite was barred from participating in the elections. The second candidate in the first round,
Ahmed Shafik was the final prime minister during Mubarak’s presidency. In fact, the runoff
between Morsy and Shafik witnessed a clear division in the country; Morsy was able to win
the presidency with a 3.4% margin. In this respect, a considerable amount of presidential
election campaign was devoted to the differences between the old elite and the opposition.
This division is not fully delineated in the issue space and presidential candidates were not
22
responding to social change or economic left-right cleavages. This is a problem associated
with the Egyptian case indicating that the presidential election was not based on political
platforms or policy preferences. Rather, the campaign witnessed polarization of the country,
where Islamist pro-revolutionary political figures and the old elite were competing. This
polarized social structure is an alternative explanation to the artifacts of the 2012 Egyptian
VAA. Further research can help us explain whether the VAA design or substantive shifts and
conjunctures in the country explain the lack of presidential candidates’ responsiveness.
Discussion and Conclusions
These five cases and six elections suggest that there are contextual and regional
determinants for the social embeddedness and electoral responsiveness of political parties.
Table 3 is a summary of the country graphs and it refers to the correlation between PTV
orders and the average distance. Expectation from a socially embedded, responsive and
democratic election would be to have high values of negative correlation. In other words,
respondents are expected to be more likely to vote for a party as they get closer to that party
in the issue space. Or as their distance in the issue space increases voters are expected to
become less likely to vote for that party.
It is clear that the Turkish and Israeli cases are the most electorally responsive party
systems in the larger MENA region. This is not unexpected given the consolidated
democratic regimes and institutionalized elections in Israel and Turkey, albeit the Turkish
case is still ridden with several deficiencies stemming from party closures and coups. With
the exception of the NAP, all of the largest political parties in these two countries respond to
the voters in the issue space in the expected way. On the other hand, Moroccan elections of
2011 shows a non-responsive party system that may also help us explain alienated voters
with relatively low levels of internal political efficacy and lowest level of turnout across the
six elections. Monarchical nature of the Moroccan politics and injection of monarchist ideas
23
into the electoral campaign as well as the king’s religious and security-related powers
indicate that the Moroccan parties may not be embedded into the issue space, except for the
JDP. This is also not surprising given the fact that Islamism is a salient dimension in many of
the countries in the region and JDP represents such an ideology in Morocco. JDP is able to
monopolize a considerable portion of the voters thanks to its place along the Islamism vs.
secular cleavage.
Between Morocco, Turkey and Israel, we have two cases which held the first free and
fair elections in the post-Arab Spring period. Tunisia and Egypt are the two countries,
additionally with Libya, which witnessed a significant political fracture because of the Arab
Spring. Three elections held in these two countries help us measure the impact of the Arab
Spring. Empirical results from our findings are mixed. First of all, it is clear that the both
countries are less responsive than the Turkish and Israeli elections. The only election that
resembles these two elections is the Egyptian Parliamentary elections. Further research is
required to explain whether this resemblance was an artifact of the Egyptian VAA or if there
were causes for the responsiveness. However, the Egyptian presidential election as well as
the Tunisian elections were not very responsive. It is not possible to argue that the parties
corresponded with the issues that were relevant for the electorate. An interesting finding is
about Islamist parties in the region. As mentioned above in the findings section for the
Moroccan JDP, Islamist parties are the most responsive. FJP and Nour in Egypt as well as
Ennahda in Tunisia have relatively high values of electoral responsiveness. This may be
related to several issues. First of all, the Islamist movements were able to establish their own
incentive provision networks. Also, these movements established their own political elites in
these three countries. These political elites were able to benefit from the Islamism vs.
secularism cleavage. Even if FJP and Ennahda were not legally recognized before the
24
revolutions in Egypt and Tunisia, they benefitted from vast social networks in order to
respond to their constituents.
[INSERT TABLE 3 HERE]
The Egyptian presidential election is an outlier case in this comparison. Further research is
required to explain why this was the case. Even if the parliamentary elections in Egypt were
relatively responsive, this is not the case in the presidential election. This difference may be
related to the differences between first-order and second-order election types as well as the
nature of presidential elections in Egypt. Contextual and theoretical reasons would assist
further research in this topic.
Table 4 introduces the same relationship in a multivariate context with demographic
controls. After controlling for age, education level and gender we still see that distance in the
issue space is inversely related to PTV. The signs of the distance variable conform to
previous findings with correlation analyses and only the Tunisian CPR appears to have an
insignificant distance coefficient. The magnitudes of the distance coefficients show that the
strongest embeddedness in the two-dimensional space appears in Turkey followed by Israel
and the Egyptian parliamentary elections exhibit the third strongest responsiveness in the
expected direction. A similar observation can also be made with the degree of explanatory
power these models have reflected in the r-square values. Israeli case has the highest average
r-square value with 0.27 followed by Turkey with 0.23. Egyptian case follows closely with
0.2 r-square value. However, in the Turkish case the NAP and in the Egyptian case the Wafd
appears with lowest r-square values suggesting that the two-dimensional space we use in
these models do not adequately capture the space of competition for these two parties as they
do for the others.
In summary, this paper indicates that the Arab Spring was not a success story for
electoral processes. Egyptian and Tunisian cases had serious limitations in respect to their
25
electoral responsiveness as well as cleavage convergence. This is not unexpected given that
the countries had only the founding elections. Further institutionalization of political parties,
a free media in which parties are able to inform the voters about their campaigns as well as
contestation and inclusion into the political elite circles may be crucial for development of a
responsive electoral system in which political parties are agents that represent their
principals.
Additionally, this substantive research utilized VAAs to collect the necessary data for
empirical analyses. VAAs are an especially new phenomenon in countries besides from the
advanced industrial democracies. As far as we know, Egyptian, Moroccan, Turkish and
Tunisian VAAs were the first of their kind in these countries while the VAA in Israel was the
second of its kind. Therefore, VAA research in the Middle East is experiencing its first steps.
As the response rates and attrition because of the logic check indicates in the methods and
data section, further research is required to develop VAAs, especially in non-Western
contexts. However, our results indicate that the VAAs have face as well as predictive validity
and they can be especially useful if utilized for preliminary and exploratory research. In
addition, VAAs can also be used in an experimental survey studies, increasing its leverage
and validity. Our substantive findings, especially the comparison of Egyptian parliamentary
and presidential elections, indicate that further methodological research in assessing the
impact of VAAs on the validity of research is crucial. Opt-in online sampling in VAAs may
cause artefactual inflation or deflation of the findings. Matching VAA respondents with
nationally representative samples can be a solution to this shortcoming. However, this may
not be always possible, especially if the research is on non-democratic or semi-democratic
countries where national electoral survey research is not always possible.
26
References:
Amghar, S. (2007). Political Islam in Morocco. CEPS Working Document, No.297. Centre
for European Policy Studies.
Anderson, C. J. (1998). Party systems and satisfaction with democracy in the new Europe.
Political Studies, 46(3), 572–88.
Blais, A., & Bodet, M. A. (2006). Does proportional representation foster closer congruence
between citizens and policy makers? Comparative Political Studies, 39(10), 1243–
1262.
Bethlehem, Jelke 2010. Selection bias in web surveys. International Statistical Review 78(2),
161-188.
Brocker, M., & Künkler, M. (2013). Religious parties Revisiting the inclusion-moderation
hypothesis1-Introduction. Party Politics, 19(2), 171-186.
Clark, J. (2004). Social movement theory and patron-clientelism. Islamic social institutions
and the middle class in Egypt, Jordan, and Yemen. Comparative Political Studies,
37(8), 941–968.
Costello, R., Thomassen, J., & Rosema, M. (2012). European Parliament elections and
political representation: Policy congruence between voters and parties. West European
Politics, 35(6), 1226-1248.
Cox, G. W., & McCubbins, M. D. (1986). Electoral politics as a redistributive game. Journal
of Politics, 48(2), 370-389.
Çarkoğlu, A. (2011). Turkey’s 2011 General Elections: Toward a Dominant Party System?
Insight Turkey, 13(3), 43-62.
Çarkoğlu, A., & Hinich, M. J. (2006). A spatial analysis of Turkish party preferences.
Electoral Studies, 25(2), 369-392.
Dalton, R. J. (2011). Left–right orientations, context, and voting choices. Citizens, Context
and Choice: How Context Shapes Citizens’ Electoral Choices, eds. R.J. Dalton and C.
J. Anderson. New York: Oxford University Press.
Deegan-Krause, K. (2007). New dimensions of political cleavage. The Oxford Handbook of
Political Science, eds. R. J. Dalton and H. D. Klingemann. New York: Oxford
University Press.
Diamond, L. (2010). Why are there no Arab democracies? Journal of Democracy, 21(1), 93-
112.
van Dijk, C., Krouwel, A., & Boiten, M. (2013). Do parties and voters pursue the same thing?
Paper presented at 2nd European Conference on Comparative Electoral Research.
27
Dixit, A., & Londregan, J. (1996). The determinants of success of special interests in
redistributive politics. Journal of Politics, 58(4), 1132-1155.
Downs, A. (1957) An Economic Theory of Democracy. New York: Harper.
Gümüşçü, Ş. (2013). The Emerging Predominant Party System in Turkey. Government and
Opposition, 48(2), 223-244.
Gürsoy, Y. (2013) The Impact of EU-Driven Reforms on the Political Autonomy of the
Turkish Military, in G. Avcı and A. Çarkoğlu (eds.) Turkey and the EU: Accession and
Reform, Routledge, pp. 85-100.
Hale, W. (1993) Turkish Politics and the Military, Routledge.
Hassan, M. (2013). Elections of the People’s Assembly, Egypt 2011/12. Electoral Studies,
32(2), pp. 370-74.
Hazama, Y. (2003). Social cleavages and electoral support in Turkey: Toward convergence?
The Developing Economies, 41(3), 362-387.
Henry, C. 2004. Tunisia’s Rogue Ben Ali Regime. Paper presented at the conference
Winning the Prize for Repression: The Real Rogue States, Kennedy School of
Government, Harvard University, Cambridge, MA, October 14–16.
Ho, K., Clarke, H. D., Chen, L. K., & Lu-Chung Weng, D. (2013). Valence politics and
electoral choice in a new democracy: the case of Taiwan. Electoral Studies. 32(3), 476-
481.
Hooghe, M., & Teepe, W., 2007. Party profiles on the web: an analysis of the logfiles of non-
partisan interactive political internet sites in the 2003 and 2004 election campaigns in
Belgium. New Media & Society, 9 [6]: 965-985.
Karvonen, L. (2010). The personalization of politics: A study of parliamentary democracies.
Colchester: ECPR Press.
Kreuter, F., Presser, S., & Tourangeau, R. (2008). Social desirability bias in CATI, IVR, and
Web surveys: The effects of mode and question of sensitivity. Public Opinion
Quarterly, 72(5): 847–865.
Laakso, M., & Taagepera, R. (1979). Effective number of parties: A measure with application
to West Europe. Comparative Political Studies, 12(1), 3-27.
Langohr, V. (2004). Too much civil society, too little politics: Egypt and liberalizing Arab
regimes. Comparative Solitics, 36(2), 181-204.
Liddell, J. (2010): Notables, clientelism and the politics of change in Morocco, The Journal
of North African Studies, 15(3), 315-331.
Lijphart, A. (1999). Patterns of democracy: Government forms and performance in thirty-six
countries. New Haven and London: Yale University Press.
28
McAllister, I. (2006). The Personalization of Politics. The Oxford Handbook of Political
Behavior, eds. R. J. Dalton and H. D. Klingemann. New York: Oxford University
Press.
Marshall, M G., & Jaggers, K. (2010) Polity IV Country Reports for Egypt, Israel, Morocco,
Turkey and Tunisia, accessed on August 10, 2013,
http://www.systemicpeace.org/polity/polity06.htm. Polity IV Project.
Miller, W. E., & Stokes, D. E. (1963). Constituency influence in Congress. The American
Political Science Review, 57(1), 45-56.
Müftüler-Baç, M., & Keyman, E. F. (2012). Turkey Under the AKP: The Era of Dominant-
Party Politics. Journal of Democracy, 23(1), 85-99.
Noland, M. (2005). Explaining Middle Eastern Authoritarianism. Working Paper 05-5,
Institute for International Economics.
Norris, P. (1999) Institutional explanations for political support. Critical Citizens: Global
Support for Democratic Governance, ed. P. Norris. New York: Oxford University
Press.
Pellicer, M., & Wegner, E. (2012). Socio-economic voter profile and motives for Islamist
support in Morocco. Party Politics, Published Online, March 15 2012.
Piattoni, S. (2001). Clientelism in historical and comparative perspective. Clientelism,
Interests, and Democratic Representation: The European Experience in Historical and
Comparative Perspective, ed. S. Piattoni. Cambridge University Press.
Powell, G. B. (2013) Electoral responsiveness, party government, and the imperfect
performance of democratic elections. Party Governance and Party Democracy, eds. W.
C. Müller and H. M. Narud. New York: Springer.
Rabinowitz, G., & Macdonald, S. E. (1989). A directional theory of issue voting. The
American Political Science Review, 83(1), 93-121.
Sakshaug, J. W., Yan, T., & Tourangeau R., (2010). Nonrepsonse error, measurement error,
and mode of data collection. Public Opinion Quarterly, 74(5), 907–933.
Sater, J. N. (2009). Parliamentary elections and authoritarian rule in Morocco. The Middle
East Journal, 63(3), 381-400.
Shamir, M., & Arian, A. (1999). Collective identity and electoral competition in Israel. The
American Political Science Review, 93(2), 265-277.
Soroka, S. N., & Wlezien, C. (2010). Degrees of democracy: Politics, public opinion, and
policy. Cambridge University Press.
Tessler, M. A., & Gao, E. (2005). Gauging Arab support for democracy. Journal of
Democracy, 16(3), 83-97.
29
TABLES AND GRAPHS
5 Likud and Israel Beitenu formed an electoral alliance on 25th of October, 2012, approximately 3
months before the election. Of the 31 seats won by the coalition, 20 were given to members of Likud
while 11 were Israel Beitenu seats.
Table 1- 6 Elections, 4 Largest Parties and Their Vote Shares
TURKEY - 2011 ISRAEL - 2013 EGYPT – PARLIAMENT 2011
JDP 49.83 Likud
Israel Beitenu5
23.34 FJP 37.5
RPP 25.98 Al-Nour 27.8
NAP 13.01 Yesh Atid 14.33 Wafd 9.2
PDP 6.57 Labor 11.39 FEP 8.9
Total Vote Share
Effective Number of Parties
95.39
2.96
49.06
8.68
83.4
4.2
TUNISIA - 2011 MOROCCO - 2011 EGYPT – PRESIDENTIAL 2012
Ennahda 37.04 JDP 22.8 Morsi 24.78
CPR 8.71 Istiqlal 11.9 Shafik 23.66
Forum 7.03 NROI 11.3 Sabahi 20.72
PDP 3.94 AMP 11.1 Aboul Fotouh 17.47
Total Vote Share
Effective Number of Parties
56.72
4.78
57.1
7.95
86.63
4.92 (First Round)
30
Table 2- Summary Statistics
Variable
Name n Mean
Std
Dev.
Mi
n Max
Variable
Name n Mean
Std
Dev.
Mi
n Max
EGYPT Parliament MOROCCO
dist_fjp 59605 0.93 0.56 0 4.47 dist_jdp 19688 0.83 0.49 0 4.61
dist_nour 59605 1.67 0.61 0 5.32 dist_istiqlal 19688 0.68 0.42 0 3.13
dist_wafd 59605 1.32 0.54 0 3.66 dist_rally 19688 1 0.46 0 3.34
dist_fep 59605 1.91 0.61 0 4.58 dist_amp 19688 1.14 0.44 0 3.63
ptv_fjp 42181 5.43 4.1 0 10 ptv_jdp 16009 5.98 4.04 0 10
ptv_nour 36456 3.45 3.97 0 10 ptv_istiqlal 14770 1.4 2.62 0 10
ptv_wafd 35175 3.32 3.54 0 10 ptv_rally 14239 1.8 2.89 0 10
ptv_fep 34704 4.04 3.88 0 10 ptv_amp 14270 1.3 2.56 0 10
EGYPT President TUNISIA
dist_morsy 63820 1.49 0.48 0 3.71 dist_nahda 11253 1.26 0.56 0 3.44
dist_shafik 63820 1.91 0.5 0 4.06 dist_congress 11253 1.22 0.55 0 2.92
dist_sabahi 63820 1.19 0.47 0 4.17 dist_forum 11253 1.69 0.59 0 2.55
dist_fotouh 63820 0.98 0.4 0 3.56 dist_pdp 11253 1.48 0.52 0 2.75
ptv_morsy 35959 2 3 0 10 ptv_nahda 9441 2.82 3.97 0 10
ptv_shafik 37112 1.75 2.9 0 10 ptv_congress 8507 5.33 4.21 0 10
ptv_sabahi 36519 4.47 3.56 0 10 ptv_forum 8676 2.4 3.32 0 10
ptv_fotouh 37569 3.91 3.54 0 10 ptv_pdp 8728 1.34 2.85 0 10
ISRAEL TURKEY
dist_likud 85727 1.5 0.8 0 3.96 dist_jdp 60004 0.62 0.34 0 4.01
dist_beit 85727 1.86 0.77 0 4.32 dist_rpp 60004 0.64 0.31 0 4.05
dist_yesh 85727 0.92 0.54 0 3.93 dist_nap 60004 0.8 0.37 0 3.69
dist_labor 85727 1.6 0.8 0 4.45 dist_pdp 60004 0.75 0.36 0 4.21
ptv_likud 62190 3.99 3.6 0 10 ptv_jdp 49485 3.25 3.99 0 10
ptv_beit 60646 2.14 2.99 0 10 ptv_rpp 49737 5.03 3.68 0 10
ptv_yesh 60396 4.24 3.36 0 10 ptv_nap 47317 4.83 3.64 0 10
ptv_labor 61546 4.45 3.41 0 10 ptv_pdp 46277 3.08 3.42 0 10
31
Table 3 - Correlation Between PTV orders and the Voters’ Distance to the Party
(All values are significant, p<0.001)
ISRAEL TURKEY
Likud -0.64 JDP -0.54
Israel Beitenu -0.50 RPP -0.51
Yesh Atid -0.39 NAP -0.26
Labor -0.49 PDP -0.49
EGYPT PARLIAMENT MOROCCO
FJP -0.49 JDP -0.31
Nour -0.47 Istiqlal 0.012
WAFD -0.11 NRoI -0.11
FEP -0.53 AMP -0.16
TUNISIA EGYPT PRESIDENTIAL
Nahda -0.24 Morsy -0.04
CPR 0.03 Shafik -0.14
Forum -0.16 Sabahi -0.01
PDP -0.07 Aboul Fotouh -0.01
32
Table 4 –Explaining the Probability to Vote (PTV) Scores with Distance in the Issue
Space And Demographic Controls
Asterisks indicate level of statistical significance, * p<0.05, ** p<0.001, *** p<0.0001
constant Distance Female Age Education r2 N
Turkey
JDP 30.345*** -3.913*** -0.511*** -0.020*** -0.134*** 0.32 47761
RPP 0.554 -3.704*** 0.481*** 0.031*** 0.050* 0.25 48006
NAP 31.319*** -1.404*** -0.388*** 0.008*** -0.163*** 0.07 45728
PDP 27.401*** -2.780*** -0.262*** -0.005*** -0.133*** 0.27 44725
Israel
Likud 8.559*** -2.879*** -0.416*** -0.013*** 0.090*** 0.43 55900
Israel Beitenu 6.388*** -1.896*** -0.400*** -0.013*** -0.017* 0.27 54543
Yesh Atid 5.874*** -2.432*** 0.061* 0.002* 0.129*** 0.15 54284
Labor 7.325*** -2.066*** 0.001 0.006*** 0.048*** 0.25 55311
Morocco
JDP 7.882*** -2.538*** 0.463*** -0.011* -0.045*** 0.11 11496
Istiqlal 1.897*** -0.084 -0.642*** 0.017*** 0.029 0.02 10589
NRoI 4.074*** -0.715*** -0.920*** -0.010*** 0.078* 0.03 10195
AMP 4.095*** -0.981*** -0.453*** -0.012*** -0.079** 0.03 10261
Tunisia
Ennahda 5.743*** -2.707*** 1.242*** -0.018*** -0.385*** 0.09 7452
CPR 5.263*** 0.275 0.321*** -0.048*** 0.131 0.02 6765
Forum 5.349*** -2.135*** -0.434*** 0.024*** 0.192** 0.06 6893
PDP 3.398*** -0.959*** -0.719*** 0.043*** 0.020 0.04 6932
Egypt President
Morsi 3.426*** -0.264*** -0.627*** -0.002 -0.039 0.01 28355
Shafik 3.176*** -0.773*** 0.138*** 0.019*** -0.135*** 0.03 29067
Sabahi 3.008*** -0.089* -0.228*** -0.006** 0.348*** 0.003 28979
Aboul Fotouh 3.491*** -0.106* -0.198*** -0.013*** 0.199*** 0.003 29822
Egypt Parliament
FJP 9.188*** -3.556*** 0.699*** 0.008*** -0.332*** 0.25 27596
Al Nour 10.021*** -2.959*** 0.939*** 0.005* -0.586*** 0.25 23996
Wafd 5.890*** -0.647*** 0.238*** 0.024*** -0.515*** 0.02 23287
FEP 11.330*** -3.147*** -0.543*** 0.015*** -0.186*** 0.28 22780
33
GRAPH 1:
GRAPH 2:
34
GRAPH 3:
GRAPH 4:
35
GRAPH 5:
GRAPH 6: