Which matters more in the electoral success of Islamist (successor) parties – religion or performance? The Turkish case

Abstract

Does the electoral success of Islamist parties depend on the support of religious voters or does it owe as much or more to their performance in dealing with key political and economic issues? The repeated electoral success of an Islamist-rooted party, the Justice and Development Party (AKP) of Turkey, provides an important opportunity to answer this question. Using a nation-wide survey conducted in 2011 in Turkey, our findings suggest that in addition to religiosity the party’s performance with respect to social services, the economy and democracy were determining factors in the AKP’s success. We also found that the popularity of political leaders has an independent effect on party preference. We discuss similar tendencies in the aftermath of the Arab Spring elections where the Islamist parties emerged as the major winners.

Keywords

democratization islamist parties moderation religion Turkey

Introduction

The Arab Spring and subsequent elections have revitalized discussions about democratization and how new governments can overcome the previous regime’s authoritarian legacy, democratize their political structures and, equally importantly, fix their crippled economies. The first democratic elections of the Arab Spring in Tunisia and Egypt resulted in the success of the Islamist parties, leading many to ask whether these parties can adhere to democratic principles and resolve the economic problems confronting their countries (The Economist, 2012; Saif and Abu Rumman, 2012). Although the military coup in Egypt and the mass protest against the al-Nahda government in Tunisia in summer 2013 may divert academic and popular attention from the electoral success of the Islamist parties, it is still important to know why Islamist parties have been able to garner a significant share of the popular vote in free and fair elections.

Studying the Turkish case can provide insight into how an Islamist (successor) party¹ can turn into a successful electoral party and manage to maintain its electoral success. In June 2011, Turkey’s pro-Islamist party, the Justice and Development Party (AKP), became the first party in over half a century to win three successive elections. Since its stunning breakthrough in the November 2002 election, the party has seen its share of the popular vote increase with each election. In 2002, it was only thanks to the vagaries of the Turkish electoral system that the party was able to form a single-party government with a mere 34 percent of the popular vote. The subsequent victories have rested on much more substantial foundations: in 2011, 49.8 percent of Turks voted for the AKP, up from 46.7 percent in 2007. As other Muslim-majority countries in the Middle East grapple with the challenges of democratic transition, the electoral success of Turkey’s pro-Islamist party bears closer examination.

Studies of Islamist parties in the Arab Middle East and Turkey have tended to attribute their rise to religion, rapid urbanization, economic crisis, and the role of international factors (e.g. the Iranian revolution of 1979, the West, etc). With some notable exceptions, little scholarly attention has been paid to the sources of Islamist parties’ electoral support (Brumberg, 2002; Clark, 2004; Gülalp, 2001; Gümüşçü, 2010, 2013; Önis, 1997; Sokhey and Yıldırım, 2013; Tessler, 1997; Volpi, 2009). Is it really religion that drives their success? Once in power, do they behave like any other incumbent party that seeks to maximize its support on the basis of its performance in office? Or is there more to it than this? We argue that it is essential for scholars to analyze Islamist parties’ appeal to voters and identify the factors that drive their electoral success. Given that other Islamist movements in the Middle East look to the Turkish case as a source of lessons, if not models, understanding why the Turkish party has been so successful at the polls can provide important new insights (Auf, 2013; Peterson, 2013). How can we explain the success of an Islamist party in an environment where – as in the newly democratizing Arab countries – suspicion about its intentions abound and its capability to stay in power is questioned?

In this article, we examine the factors that influenced support for the AKP in 2011. Analyzing vote intentions at the end of the party’s second term in office not only provides insights into the sources of the AKP’s appeal to a once skeptical Turkish public; more importantly, it can help us understand how, once in power, a party with Islamist roots can succeed in increasing its vote share and winning re-election. We pay particular attention to the role of religiosity and performance evaluations. Does the party’s electoral success depend on the support of religious voters or does it owe as much or more to its performance in dealing with key political and economic issues? We argue that religion is only part of the explanation for the AKP’s success. Using data from a nation-wide survey conducted in 2011,² we show that the party’s performance with respect to social services, the economy and democracy were also determining factors.

This study makes several contributions to the democratization and political parties literatures. First, it indicates that religion is only one factor in the electoral appeal of pro-Islamist parties. Second, it contributes to the study of the inclusion-moderation hypothesis (Brocker and Künkler, 2013) by showing that once pro-Islamist parties come to power they can maintain their hold on office by successfully addressing voters’ ‘bottom-line’ concerns (Grzymala-Busse, 2002; Schwedler, 2006; Sokhey and Yıldırım, 2013; Tezcür, 2010).³ In this respect, the AKP case suggests parallels with Italy’s Christian Democratic Party and communist successor parties such as Poland’s Democratic Left Alliance (SdRP) and Hungary’s Magyar Socialist Party (MSzP). Once these parties moderate their ideologies and are included in democratic political life they come to behave pragmatically much like catchall parties (Gryzmala Busse, 2002; Sokhey and Yıldırım, 2013; Warner, 2013). Third, unlike most studies on the inclusion-moderation hypothesis, we use survey evidence to examine how voters react to the performance of an Islamist successor party.

This study also contributes to the literature on the AKP. First, in looking beyond the role of religiosity and economic evaluations to focus more broadly on the AKP’s performance in office, it adds an important new dimension to our understanding of the bases of the party’s electoral support. Second, this study shows that contrary to the claims of some scholars the AKP appealed to voters by democratizing the political system, at least in voters’ eyes. Finally, it highlights the importance of the party leaders’ popularity in explaining the party’s electoral success.

Islamist parties in the Middle East

A number of factors have been adduced to explain the success of Islamist parties in the Middle East (Brumberg, 2002; Clark, 2004; Volpi, 2009; Wickham, 2002; Wiktorowicz, 2004). Rapid urbanization, corruption, the patronage of major powers over these regimes, the impact of the global economic crisis and growing unemployment, especially among the young, have all been credited with creating fertile ground for the emergence of strong Islamist movements. Given the economically volatile context in the region, studies have emphasized the role of social welfare programs run by Islamist groups, which provide services to the needy in the periphery and to the new middle class in the semi-periphery (Wickham, 2002; Wiktorowicz, 2004). These studies argue that Islamist opposition groups have benefited from economic liberalization and the decline of repression to gain an advantage over their secular rivals by providing alternative political venues for disgruntled citizens (Clark, 2004; Wickham, 2002). Their discourse of social justice and freedom from oppression, as well as their success in delivering social services, has appealed to many in the Muslim-majority countries (Brumberg, 2002; Clark, 2004).

In those Middle Eastern countries where elections have been more or less free and open, such as the 1989 Jordanian election and the 2005 Egyptian election, Islamist parties have emerged as the most powerful opposition groups even though they did not contest every electoral district (Hamid, 2011). Widespread corruption and growing poverty and inequality had discredited the traditional parties and, apart from weak secular left parties, the Islamist movements/parties were the only ones that had not been tested in government. Their organizational strength and their ability to provide social assistance to the needy gave the populace a signal that they could rule the country more efficiently than corrupt governments (Wickham, 2002).

Finally, the Arab Spring elections brought victory to Islamist parties in Tunisia, Egypt and Morocco for the first time in their history. Many observers emphasized that the economic problems are so entrenched in these countries that the governing Islamist parties had to tackle these first if they wanted to be considered credible political actors. These parties were faced with the challenge of enacting radical political and economic policies in order to eradicate corruption, provide economic security for their citizens and promote political liberalization (Fuller, 2012). Facing such hard choices, the parties turned to Turkey as a model for designing the policies that would ensure their credibility in the eyes of both their citizens and the international community (Maddy-Weitzman, 2011; Moussaoui, 2011; Ülgen, 2011). What the AKP has done to increase its vote share after coming to power has been of considerable interest to Islamist parties in the Middle East. The 2013 military coup against Mursi’s presidency in Egypt and the mass protests against al-Nahda rule in Tunisia make it all the more important to uncover the factors that have enabled an Islamist party to repeat its success and stay in power for over a decade.

Islamist parties in Turkey and the rise of the AKP

It is not surprising that Islamist parties in the Middle East would look to Turkey: the AKP is the only party with roots in Islamist politics that has not only gained power but remained in power for three successive elections. Scholars have attributed the electoral success of Turkey’s Islamist parties to a variety of factors: rapid urbanization starting in the 1950s, the embrace of religion as a result of the weakening of traditional identities, the rise of a new devout provincial business class and, beginning in the 1980s, the role of Islamist organizations in providing social services, especially to the poor and marginalized (Demiralp, 2009; Eligür, 2010; Gülalp, 2001; Gümüşçü, 2010; Önis, 1997; Sokhey and Yıldırım, 2013; Sunar and Toprak, 1983; Tuğal, 2009; Yavuz, 1997).

Pro-Islamist parties are not new to the Turkish electoral scene. The first explicitly Islamist political party, the National Order Party (MNP), was founded in 1970. It attracted the support of small shopkeepers and various economically disadvantaged conservative groups (Sunar and Toprak, 1983). Banned by the Constitutional Court, the party was succeeded by the National Salvation Party (MSP). In 1974, the MSP gained 48 seats in Parliament with 11.8 percent of the vote and established a short-lived coalition government with the left-of-center Republican People’s Party (CHP). Despite attracting only 8.6 percent of the vote and getting only 24 seats in the subsequent election, the party once again joined a short-term coalition government in 1977. The military coup of 1980 closed all of the parties including the MSP and banned its leader from politics. The party leadership went on to found the Welfare Party (RP), but it remained a marginal electoral force until 1991 when it won 62 seats in Parliament with 16.9 percent of the vote. The party managed to win a plurality of the popular vote in 1996 and formed a coalition government. However, growing alarm on the part of the armed forces, the judiciary and the media culminated in the banning of the RP by the Constitutional Court in 1998 on the grounds that the party had violated the constitutional obligation to respect secularism. The RP was succeeded by the Virtue Party (FP) until it, too, was banned by the Constitutional Court for promoting Islamic fundamentalism.

Two pro-Islamist parties emerged to take its place, but the AKP under the leadership of Recep Tayyip Erdoğan quickly asserted its supremacy. Thanks in part to an electoral system that favors the winning party and a 10 percent electoral threshold, the AKP obtained nearly two-thirds of the seats in Parliament in the 2002 election with only 34 percent of the vote and formed the first single-party government in over a decade.

The AKP represented a much more moderate brand of political Islamism than its predecessors. Indeed, some observers have likened it in important respects to Western Europe’s Christian Democrat parties, its anti-establishment orientation notwithstanding (Hale, 2005). Erdoğan himself has rejected that parallel, preferring to describe the AKP as a conservative democratic party rather than a pro-Islamist one.

What induced this moderation has been the subject of debate. Tezcür (2010) argues that strategic interests, the threat of state repression and ideological change on the part of Islamist politicians and intellectuals combined to encourage ‘the evolution of Islamists into Muslim reformers’ (p. 69). Gümüşçü (2010) rejects the notion that state repression encouraged moderation. She points out that the closing of four parties by the Constitutional Court in recent decades had failed to have a moderating effect. She also cites the Egyptian experience where state repression did not induce Islamist parties to moderate their rhetoric. Like Demiralp (2009), she attributes the transformation of the Islamist party into a conservative democratic party to economic liberalization, which has fostered the emergence of a devout provincial bourgeoisie ‘with vested interests in liberalism and democracy’ (p. 836). Sokhey and Yıldırım (2013) make a similar argument. Drawing parallels between the AKP and Hungary’s MSzP, they argue that economic liberalization fosters party moderation, but only when liberalization is associated with an open and competitive economy that favors the development of small- and medium-sized enterprises.

However we choose to characterize the party, the AKP’s roots are clearly Islamist. Religious parties have been central to the development of the inclusion-moderation hypothesis. However, as Brocker and Künkler (2013) emphasize, ‘a moderate party is not necessarily a less religious one’ (p. 181). This raises the question of just how powerful a factor religion is in the party’s repeated electoral success. What keeps voters voting for a party with roots in Islamist politics? Is it their religious devotion, the party’s record in office or a popular leader?

Religion

Even in the absence of an electorally viable Islamist party, religion has played a significant role in Turkish voters’ choice of party. In the 1960s and 1970s, religious voters typically supported the Justice Party (AP), while their secular counterparts favored the left-leaning CHP (Ergüder, 1980–1981). Indeed, ‘[t]he debate over secularism versus religion…represents major and deep-seated divisions in Turkish society: modern versus traditional, progressive versus conservative, and rationalist versus religious’ (Yavuz, 1997: 64). Yavuz (1997) argues that Islam became politicized because, rather than separating religion and politics, secularism ‘subordinated religion to the political realm’ (p. 65).

The divide between religious and secular voters has been central to the center-periphery approach to explaining party preferences and voting patterns in Turkey (Kalaycıoğlu, 1994). Pioneered by Mardin (1973), this framework depicts Turkish politics in terms of a cleavage that pits a nationalist, centralist and secular Kemalist state elite against a variety of ethnic, religious and regional groups that identify with traditional conservative religious values. Self-identified religiosity has repeatedly emerged as a significant predictor of party preference with the center traditionally voting for liberal or left-wing parties, while the periphery has preferred conservative or right-wing parties (Esmer, 1995; Kalaycıoğlu, 1994).

Religiosity has also proved to be associated with a preference for the AKP. Başlevent and his colleagues (2005) report a strong positive relationship between religiosity and AKP vote intentions in the 2002 election. Similarly, Çarkoğlu (2010) found reported frequency of religious practice to be a significant factor in the choice between the AKP and every other party in the run-up to the 2007 election. This was especially true of the choice between the AKP and its main rival, the center-left CHP. There was also a strong relationship between AKP vote intentions and support for a Shari’a-based religious state in Turkey. In 2011, by contrast, one study reports that religious practice failed to have a significant effect on AKP vote intentions (Çarkoğlu, 2012). It is possible, though, that this reflects the way that the vote intention model was specified: the inclusion of reported vote for the AKP in the previous election effectively creates a lagged dependent variable, which means that the independent variables are capturing the factors that are related to change in vote preference.⁴

Performance evaluations

The economy

Recent studies of voting behavior in Turkey have focused a good deal of attention on the role of economic evaluations in explaining support for the AKP. Given the context of the party’s electoral breakthrough in 2002, it is hardly surprising that economic voting models came into vogue in the wake of the AKP’s ascent to power. That election was held just as the country was emerging from a deep economic crisis. Under the governing center-right coalition, the annual inflation rate had hit 68.5 percent the year before the election.⁵ The incumbents were so discredited that none of the three coalition partners succeeded in winning a single seat. Their fate would certainly seem to exemplify the reward-and-punish calculus whereby voters punish incumbents for bad economic times just as they reward them for good times.

Surprisingly, Çarkoğlu (2012) concludes that the outcome of the 2002 election was not determined by negative evaluations of the incumbents’ economic performance; ideology played a more decisive role. This is not to say that economic performance evaluations were unimportant. An additive scale comprising retrospective and prospective egocentric and sociotropic evaluations plus evaluations of current economic conditions had a significant negative relationship with intended vote for the incumbents. Focusing instead on the choice between the AKP and each of the other parties, another study of vote intentions in that election confirmed the importance of economic evaluations, though egocentric evaluations proved to be better predictors than sociotropic ones (Başlevent et al., 2005). The authors conclude that ‘economic voting for the AKP is basically a reaction to personal suffering caused by the recent economic crisis’ (Başlevent et al., 2005: 558).

Subsequent studies of vote intentions have shown that voters are also willing to reward the incumbent when the economy is doing well. By 2003, economic conditions had clearly improved and voters who perceived that Turkey’s economy was in better shape were ready to reward the AKP (Başlevent et al., 2009). The economy was doing even better in 2007 and this boosted support for the AKP, with sociotropic retrospective evaluations proving as they did in 2003 to be more important than prospective and/or egocentric ones (Çarkoğlu, 2010).

Economic conditions were also favorable in 2011. The global economic crisis had had only a limited impact on the Turkish economy. Unemployment had peaked at 16.1 percent during the slump of 2009, but the rate was down to 9.4 percent by the time of the election; the annual inflation rate was just over 10 percent; and, among the big economies, only China and India were experiencing faster economic growth. The AKP campaigned on its economic record and it appears to have been rewarded. A study that looked at vote intentions for the AKP versus all other parties combined found that positive economic evaluations did indeed help the AKP, though the impact was weaker than it had been in 2007 (Çarkoğlu, 2012).

Political performance

But how does the impact of economic evaluations compare with that of other dimensions of performance? The reward-and-punish calculus is not limited to economic performance; according to retrospective voting models, how well the incumbent has dealt with other ‘bottom-line’ concerns can shape voters’ choice of party as well (Fiorina, 1981; Healy and Malhotra, 2013). Prime among these in the Turkish context is likely to be how well the AKP has done in guaranteeing personal safety and security. Along with the state of the economy, the conflict in the Kurdish-majority southeast and the constitution both figured prominently in the 2011 campaign.

Kurdish ethno-nationalism has been described by some observers as one of the biggest challenges faced by the Turkish state in the last two or three decades (Belge, 2011; Sarıgil, 2010). Under the AKP, Kurds have enjoyed more autonomy and freedom than they had under previous governments. This culminated in the ‘Kurdish opening’ in the summer of 2009, though the AKP government stopped short of recognizing Kurdish as an official language or allowing education in Kurdish (Karakoç, 2013). In 2007, voters who identified the ‘Kurdish problem’ as one of the most important problems facing the country were much less likely to intend to vote AKP, opting instead for the ultra-nationalist National Action Party (MHP) (Çarkoğlu, 2010). Fearing the MHP would surpass the 10 percent threshold required for representation in parliament, the AKP adopted a harsher stance toward the Kurds in 2011. Accordingly, we need to ask whether voters approved of the party’s performance in dealing with the Kurds and how this played into their choice of party.

Another potential dimension of evaluation is the AKP’s performance with respect to democratization (Gümüşçü, 2013; Yavuz, 2009). The AKP campaigned on the slogan, ‘Vote for the AKP. Write Your Own Constitution.’ Although some constitutional amendments had been approved by a 58 percent majority in September 2010, Turkey’s constitution remained little changed in its fundamentals from the text drafted by the army following the 1980 military coup. Still, the constitutional amendments did include ending immunity for leaders of military coups, which was just one of the steps that Prime Minister Erdoğan had taken since 2002 to reduce the influence of the military over Turkish politics. Despite the popular approval of these amendments and the AKP’s pledge to introduce more democracy, there was concern in some quarters about the plan to rewrite the constitution to create a French-style presidential system. Critics also raised concerns about corruption and press freedom. Given the prominence of the constitutional issue, we could expect assessments of the AKP’s performance vis-à-vis democratization to have affected party preferences, especially when it came to the choice between the AKP and the CHP.

One of the forces that favored more democracy had been the desire to join the European Union (EU). The RP was strongly opposed to accession, but the AKP saw the EU’s insistence on democratic pluralism and norms of human rights as conditions of membership as a resource for exerting pressure on the state elite to achieve religious freedoms (Tanıyıcı, 2003). Moreover, accession to the EU was in the interests of the new provincial business class, which constituted an influential AKP constituency (Gümüşçü, 2010). A variety of political reforms during the party’s first term in office led to the opening of negotiations with the EU at the end of 2004. In 2002, voters who favored Turkey’s accession to the European Union were less likely to prefer the AKP (Başlevent et al., 2005), but by 2003 there were signs that pro-EU voters were switching toward the party (Başlevent et al., 2009) and the AKP’s victory in 2007 has been interpreted as an endorsement of the negotiations (Çarkoğlu, 2010). However, by 2011 the negotiations had stalled. Views about membership did not appear to influence vote switching between the 2007 and 2011 elections (Çarkoğlu, 2012), but there is still a question as to whether voters approved of the party’s handling of the issue and whether this influenced their party preference in 2011.

Social policy performance

Voters may well have had other aspects of government performance on their minds. Since coming to power, the AKP has substantially reduced the cost of healthcare for low-income families. The AKP government has also instituted monthly cash payments for low-income families to enable them to afford to send their children to school. Other reforms include payments to enable pregnant women to receive regular health checkups. These initiatives enabled the AKP to portray itself as the party of the people, providing services that other parties had failed to provide when they were in power. This proved to be a powerful discourse (Gümüşçü, 2013). While these initiatives have proved popular, the AKP was facing a challenge from the left in 2011. Its main rival, the CHP, was championing social democratic values and the need for a strong social safety net. Moreover, there have been criticisms of some of the social policy reform initiatives on gendered grounds. For example, Kılıç (2008) argues that while some of the reforms represent an improvement in the perception of women’s status, others may result in a worsening of women’s situation. Arat (2010) goes further. She is particularly critical of the 2008 Social Security and General Health Insurance Amendment on the grounds that some of the changes may ‘undermine women’s economic independence and opportunities for work, make them more dependent on their husbands and encourage them to stay at home’ (Arat, 2010: 873–874). Despite the possible long-term negative consequences for gender equality and female labor force participation in Turkey, in the short-term the reforms have increased poor families’ access to hospitals and increased cash assistance to the needy and this may well have increased the party’s appeal among voters.

Leadership

A final dimension of performance relates to the prime minister. Voters may well decide whether or not to support a party based on how well its leader has performed in office (Fearon, 1999; Healy and Malhotra, 2013). The AKP is very much a leader-driven party (Tezcür, 2010) and its electoral success has been attributed to Erdoğan’s personal popularity (Başlevent et al., 2005). Oligarchic leadership has been a characteristic of political parties in Turkey (Turan, 1988). Throughout the country’s political history there has been a strong association between party leaders and their party and its ideology. However, the impact of leader evaluations has not figured in analyses of party preferences in the past decade. It is particularly interesting to examine the role of these evaluations in 2011. While the prime minister was being criticized in some quarters for an increasingly autocratic style, the CHP’s new leader, Kemal Kılıçdaroğlu, was reforming and revitalizing his party. By contrast, the MHP leader, Devlet Bahçeli, was a familiar figure on the electoral scene, having taken over the party leadership in 1997 following the death of the MHP’s founding father, Alpaslan Türkeş. In short, there are good reasons to expect the popularity of the party leaders to have a strong impact on party preference.

Data and methods

In order to assess the relative importance of religion and the various performance evaluations, we draw on data from a survey conducted by GENAR polling agency in Istanbul between January 5 and January 15, 2011, just six months before the June 2011 election. The survey used a random probability sampling design and was conducted face-to-face with a representative sample of 2168 eligible voters in 17 provinces and 47 towns in these provinces. The margin of error for the survey was plus or minus two percentage points. Since the survey was conducted before the 2011 election, the question asked about vote intentions if an election were held ‘this Sunday.’ Note, though, that the percentage that would vote for the AKP (50.4 percent) is very close to the party’s actual vote share in the 2011 election (49.8 percent).⁶

The analyses of vote intentions are based on a multi-stage bloc-recursive model (Miller and Shanks, 1996; Thomassen, 2005). The estimation strategy involves entering blocs of variables sequentially into a regression analysis, starting with the most causally distant variables. The basic idea behind this model is that some of the factors that affect vote intentions, such as how voters evaluate the party leaders, are closer in time to the formation of a vote intention, while other factors, such as voters’ social background characteristics, are more distant. These longer-term factors can have a direct effect on vote intention: a highly religious voter, for example, may choose the AKP simply because of its Islamist roots. However, social background characteristics can also affect party preference indirectly by influencing more proximate factors. A highly religious person, for example, is more likely to approve of the AKP’s performance with respect to democratization because of the expansion of religious rights. If all of the potential explanatory variables are entered into a single equation, the impact of causally prior variables will be underestimated since a single equation can provide only the direct, or unmediated, effects of the causally prior variables. By estimating the model in stages, it is possible to estimate the total impact of each explanatory factor, as opposed to only that portion that is not mediated via more proximate factors.⁷

The first bloc of variables⁸ includes age, income, education, state employment and sex. To evaluate the extent to which religious voters supported the AKP, we use a measure of self-reported religiosity that captures both belief and practice. In the case of sex, we make a distinction between women who are confined to the domestic sphere and those who are not, the assumption being that the former will be much more accepting of patriarchal religious values that promote traditional sex roles. Men serve as the reference category. A dummy variable is also included for members of the minority Alevi religious community. Traditionally left-leaning, Alevis have long supported the secular state (Başlevent et al., 2005; Çarkoğlu, 2010). In 2011, they had even more reason to favor the centre-left CHP because the party’s new leader was himself an Alevi.⁹

The second bloc of variables consists of performance evaluations. We examine the effects of evaluations of the AKP with respect to the economy, terror, the Kurdish problem, democratization, the EU and social policy.¹⁰ The final variables to be added to the model are leader evaluations. Given the comparative nature of leader evaluations, these take the form of relative ratings that indicate how the leader compares with the most liked of the remaining leaders (including minor party leaders).

The vote intention models are estimated using multinomial logistic regression. The choice set is restricted to the three parties that won representation in Parliament: the AKP, the CHP and the MHP. Including vote intentions for other parties (e.g. the pro-Kurdish BDP) would create problems with small cell sizes or even empty cells. This would lead to estimation problems, resulting in inflated and unstable estimated effects for those parties. The coefficients estimated by multinomial logistic regression lack a straightforward, intuitively obvious interpretation. In order to facilitate interpretation, we use the models to estimate marginal effects.¹¹

Results

Social background and party preference

The AKP’s Islamist roots are clearly an important component of its appeal (see Table 1). The estimated probability of expressing a vote intention for the party is 30 points higher for voters who are very religious, compared with voters who perform religious rituals infrequently. This gap grows to 43 points when the very religious are compared with those who do not believe in religion. With almost half the sample (48 percent) describing themselves as very religious, the importance of this factor in the AKP’s electoral success is obvious. The appeal of the CHP to secular voters is equally clear: their estimated probability of expressing a vote intention for the party is 39 points higher than it is for voters who are very religious. Secular voters, though, are very much in the minority and so, of course, are members of the Alevi religious community who are also much more likely to say they would vote CHP than AKP or MHP.

Table 1.

Social background characteristics and vote intentions.

	AKP vs CHP	AKP vs MHP	CHP vs MHP
Age	–0.03 (0.01)***	0.00 (0.01)	0.04 (0.01)***
Education	–0.48 (0.07)***	–0.30 (0.10)**	0.18 (0.10)^a
Income	–0.11 (0.05)*	0.00 (0.08)	0.11 (0.08)
Female	–1.19 (0.18)***	0.48 (0.34)	1.67 (0.33)***
Housewife	0.92 (0.20)***	0.28 (0.39)	–0.64 (0.38)^a
State official	0.58 (0.29)*	1.03 (0.47)*	0.45 (0.49)
Alevi	–3.78 (0.45)***	–1.77 (0.63)**	2.01 (0.52)***
Religiosity	0.79 (0.08)***	0.53 (0.10)***	–0.26 (0.10)*
Constant	1.77 (0.35)***	0.70 (0.48)	–1.07 (0.48)*

Notes: Nagelkerke’s adjusted R² = 0.33. Log-likelihood = –1126.42. Number of cases = 1469. The column entries are multinomial logistic regression coefficients with robust standard errors shown in parentheses.

***p < 0.001; **p < 0.01; *p < 0.05; ^ap < 0.10.

The pattern of effects for religiosity suggests that the center-periphery cleavage is still a defining characteristic of party support in Turkey. This is confirmed by the continued importance of education. This is the other key variable in the center-periphery approach (Kalaycıoğlu, 1994). Educated voters are more likely to identify with the secular values of the Kemalist state and education has traditionally been correlated with a leftist orientation (Esmer, 2002). This pattern was very much in evidence in 2011. The more educated voters are, the more likely they are to favor the CHP; the lower their level of education, the more likely they are to choose the AKP (cf. Başlevent et al., 2009; Çarkoğlu, 2010, 2012). Even allowing for the effects of other social background characteristics, the estimated probability of an intended AKP vote is eight points higher for voters who only have an elementary school education and 15 points higher for voters who are illiterate or have not completed elementary school, compared with middle school graduates. Given the low levels of educational attainment on the part of Turkish voters (44 percent of the sample had an elementary school education or less), the uneducated are clearly an important source of AKP votes. The CHP, meanwhile, fares best among university graduates: the probability of expressing an intention to vote CHP is 16 points higher and the probability of an intended AKP vote is 18 points lower for those with a university education, compared with those whose education ended with middle school.

Income also plays into the choice between the AKP and the CHP, but its effect is much more modest. Given that the AKP government has introduced a number of programs since 2002 that are targeted at low-income families, the weak income effect might seem surprising. But, like the RP (Yavuz, 1997), the AKP promotes Islamic solidarity as opposed to ethnic or class solidarities. Religion cuts across class lines. Indeed, the new provincial entrepreneurial class has proved to be a highly influential constituency of the AKP and the party has tailored some of its policies accordingly.

Age is a more important correlate of vote intention. The younger voters are, the more likely they are to opt for the AKP (cf. Başlevent et al., 2005, 2009; Çarkoğlu, 2010, 2012). All else being equal, the estimated probability of an AKP vote intention drops an estimated 18 points between the mid-20s and the mid-60s while the probability of expressing an intention to vote for the CHP increases by 23 points. There is also a significant age effect for the MHP: voters in their mid-20s are almost twice as likely as voters in their mid-60s to choose the MHP.

As predicted, women who are confined to the domestic sphere are significantly more likely to opt for the AKP. What is striking, though, is that these women are more likely than men to prefer the AKP to the CHP. Even allowing for the effects of other social background characteristics, their probability of expressing the intention to vote AKP is 14 points higher. The pattern is very different for women who are not confined to the domestic sphere: they are significantly more likely than men to choose the CHP over both the AKP and the MHP. The estimated gender gap in CHP vote intentions is 21 points. The corresponding figures for AKP and MHP vote intentions are 13 points and eight points, respectively. We can only speculate about what is driving this gender gap. It is possible that the AKP is seen as a threat – whether real or perceived – to gender equality as a result of ‘the propagation of patriarchal religious values’ (Arat, 2010: 869) by the party’s conservative religious cadres.

The final noteworthy effect relates to state officials. Government workers are significantly more likely than other voters to choose the AKP over the CHP and especially the MHP. This lends weight to the observation that Islamist values are spreading at the administrative level (Arat, 2010).

Performance evaluations and party preference

Important as religion is to support for the AKP, it would be a mistake to attribute the party’s appeal to its Islamist roots alone. The party’s performance in office has also contributed to its electoral success. Voters are clearly satisfied with the party’s handling of social programs (education, health and social assistance) and there is no evidence of any widespread dissatisfaction when it comes to dealing with the economy, the European Union or democratization (see Table 2). The same is true of dealing with the Kurdish problem in Northern Iraq and with the issue of terrorism more generally.

Table 2.

Performance evaluations (major party voters).

	Mean	Standard deviation
European Union	0.51	0.28
Kurdish problem	0.48	0.30
Social programs	0.58	0.34
Economic performance	0.52	0.37
Democracy	0.48	0.37
Terror	0.47	0.37
Number of cases = 1469

Note: All variables have been re-scaled to run from 0 to 1.

These evaluations matter and they matter even taking account of religiosity and other social background characteristics (see Table 3). They are certainly influenced by religiosity,¹² but they have independent effects on vote intentions. The most important aspect of the party’s performance clearly relates to democracy, an issue that was front and center in the run-up to the 2011 election. It is a critical factor in the choice between the AKP and the CHP. The estimated probability of an AKP vote intention is 24 points higher for a voter who rates the party very highly on this dimension, compared with a voter who is very dissatisfied, while the estimated probability of a CHP vote intention is 24 points lower.

Table 3.

Performance evaluations and vote intentions.

	AKP vs CHP	AKP vs MHP	CHP vs MHP
European Union	1.62 (0.48)***	1.78 (0.53)***	0.15 (0.50)
Kurdish problem	1.50 (0.46)**	2.97 (0.55)***	1.47 (0.53)**
Social programs	1.87 (0.60)***	1.15 (0.71)	–0.71 (0.60)
Economic performance	2.65 (0.50)***	1.76 (0.59)***	–0.90 (0.57)
Democracy	2.87 (0.50)***	1.82 (0.61)***	–1.05 (0.61)^a
Terror	0.81 (0.52)	1.34 (0.62)**	0.54 (0.59)
Constant	–3.76 (0.66)***	–4.66 (0.73)***	–0.90 (0.55)

Notes: Nagelkerke’s adjusted R² = 0.74. Log–likelihood = –651.45. Number of cases = 1469.

The column entries are multinomial logistic regression coefficients with robust standard errors shown in parentheses. Controls are included for social background characteristics.

***p < 0.001; **p < 0.01; *p < 0.05; ^ap < 0.10.

Almost as important are evaluations of the party’s handling of the economy. As the large literature on economic voting suggests, many voters appear to engage in a reward-and-punish calculus, depending on how they evaluate the incumbent’s economic performance. As we expected, voters who were negatively affected by economic conditions – the ‘losers’ – were more likely to punish the AKP by opting for the CHP or MHP. The ‘winners’, on the other hand, were more likely to reward the party. Being very satisfied (as opposed to very dissatisfied) increased the probability of expressing an intention to vote AKP by an estimated 22 points and decreased the probability of an intended CHP vote by almost the same amount. The AKP’s handling of social programs matters as well when it comes to deciding between the two parties, with a high level of satisfaction increasing the probability of favoring the party by 14 points.

All of the dimensions considered so far have a negligible effect on the decision to opt for the ultranationalist MHP. They mainly play into the choice between the AKP and the CHP. The pattern changes when we look at the AKP’s handling of the terrorism issue and Turkey’s external relations. The terrorism issue does not have a significant effect on the choice between the AKP and the main opposition party, but it does influence the choice between the AKP and the MHP. The estimated probability of an MHP vote intention is six points higher for voters who are very dissatisfied with the AKP’s performance, while the probability of expressing an intention to vote AKP is eight points higher for those who are very satisfied. Its handling of the Kurdish problem in northern Iraq is even more of a factor in the choice between the two parties. Voters who are highly dissatisfied with the party’s approach have a 14-point lower probability of saying they would vote for the party and the main beneficiary is the MHP. In hopes of limiting support for the ultranationalist party, the AKP had taken a harder line toward the Kurds, but this was clearly not enough to satisfy the MHP’s supporters.

Finally, the estimated probability of choosing the AKP was 13 points higher for voters who endorsed the party’s approach to Turkey’s relationship with the European Union. Opposition to the AKP’s stance gave a modest boost to the probability of supporting the other two major parties.

Leader evaluations and party preference

Prime Minister Erdoğan’s leadership is also a factor in the party’s electoral success. He is clearly the most highly rated of the major party leaders (see Table 4). Moreover, he is the only leader among the three whose ratings are positive on average. The MHP leader is the least popular of the three: fully 45 percent of voters gave him a rating of zero, compared with 38 percent for the CHP leader and 31 percent for the AKP leader. When it comes to vote intentions, the key consideration is how the leaders are evaluated relative to one another. The relative ratings make it clear that both the CHP and MHP leaders had many more detractors than fans.¹³

Table 4.

Leader evaluation (major party voters).

	Mean	Standard deviation
AKP leader	0.53	0.43
CHP leader	0.39	0.40
MHP leader	0.24	0.31
AKP leader (relative rating)	–0.01	0.74
CHP leader (relative rating)	–0.28	0.69
MHP leader (relative rating)	–0.57	0.47
Number of cases = 1453

Notes: Leader evaluations have been rescaled to run from 0 to 1. The relative leader evaluations are the difference between voters’ rating of the leader and their ratings of the most liked of the remaining leaders (including minor party leaders).

Leader evaluations clearly matter, but they are not necessarily a decisive consideration for voters. The estimated probability of an AKP vote intention is eight points higher for a voter who feels very positively about Erdoğan, compared with a voter who rates him no higher than any of the other leaders. The comparable figures for the CHP and MHP leaders are 11 points and eight points, respectively. Had voters had no preference for one leader over the others, the AKP’s share of vote intentions would have been over six points lower and the CHP and MHP shares would have increased by over three points each.¹⁴

Table 5.

Relative leader evaluations and vote intentions.

	AKP vs CHP	AKP vs MHP	CHP vs MHP
AKP leader	3.56 (0.63)***	3.01 (0.65)***	–0.31 (0.61)
CHP leader	–3.00 (0.67)***	1.56 (0.64)*	4.54 (0.80)***
MHP leader	–1.05 (0.72)	–4.55 (0.79)***	–3.07 (0.85)***
Constant	0.10 (1.24)	–1.43 (1.11)	–1.17 (1.12)

Notes: Nagelkerke’s adjusted R² = 0.94. Log-likelihood = –217.13. Number of cases = 1469. The column entries are multinomial logistic regression coefficients with robust standard errors shown in parentheses. Controls are included for social background characteristics and for performance evaluations.

***p < 0.001; **p < 0.01; *p < 0.05; ^ap < 0.10.

Not surprisingly, evaluations of the AKP leader are influenced by evaluations of his party’s performance. Indeed, when leader evaluations are added to the model, the performance evaluations cease to have a significant effect on the choice between the AKP and its main rival, which indicates that their effects are being channeled via evaluations of the leaders. The more voters approve of the AKP’s performance in government, the more likely they are to give the party’s leader a high rating. There is one notable exception, however: evaluations of the AKP’s performance with respect to democracy continue to have a substantial and statistically robust effect, even controlling for leader evaluations. The same is true of the impact of religiosity.

Concluding discussion

Although the AKP presents itself as a conservative Muslim democratic party it is fair to say that the party’s roots lie in Islamist politics, going back to the RP (Akdoğan, 2004; Yavuz, 2009). The example of the AKP tells us how one moderate Islamist party has been successful in obtaining a plurality of votes in three successive elections in a country that many have labeled an unconsolidated democracy (Özbudun, 1996). Our findings have some important implications for the inclusion-moderation hypothesis and for the democratization, political party and comparative politics literatures more generally.

Using the case of the AKP, we have analyzed how a moderate Islamist party can continue to get reelected once it achieves power through democratic elections. In doing so, our study adds an important dimension to work on the inclusion-moderation hypothesis by focusing on voters’ reactions. Like Italy’s Christian Democratic Party (Warner, 2013) and the communist successor parties in Hungary and Poland, the AKP came to power as a result of moderating its ideology. Our results suggest that the key to consolidating this electoral success is following through by implementing pragmatic policies that appeal to different electorates, as the AKP did by appealing to housewives through its social policies, to economic liberals through its policies toward the EU and so on.

At the same time, our findings confirm that, ‘becoming more moderate does not necessarily mean less religious’ (Brocker and Künkler, 2013: 181). For many religious voters, the AKP’s Islamist roots are reason enough to support the party. Their support is clearly a key ingredient in the AKP’s continuing electoral success.¹⁵ Religiosity has a powerful independent impact on the probability of choosing the party that persists regardless of voters’ evaluations of the party’s leader and its performance in office. Equally clearly, though, this is only part of the explanation. The party’s appeal extends beyond religious voters and the reason has to do with its performance in office, especially with respect to democratization, social policy and the economy.

Among these policy domains, the importance of the party’s performance with respect to social policy is particularly interesting. Like the AKP and its predecessor, the RP, many Islamist parties and groups in the Middle East have been successful in providing social programs to the needy and this has garnered them votes (Clark, 2004; Wickham, 2002; Wiktorowicz, 2004). Morocco’s JDP is a case in point. Designing social programs at the national level and using financial resources, such as cash transfers to the needy and free or affordable health services, helps to attract voters to the party. Given the importance of voters’ evaluations of the party’s performance, it will be important for future studies to investigate the factors that shape these evaluations, especially the role of opinion leaders such as religious authorities, media cartels and business leaders.¹⁶

A second implication of this study is that Islamist parties can appeal beyond their traditional electoral base and attract votes from secular/liberal groups in society. While Turkey’s traditional Islamist parties objected to the EU as a ‘Christian club,’ the AKP has taken a pro-EU stance, just as some communist successor parties in Eastern Europe did in the 1990s (Gryzmala-Busse, 2002). This policy shift on the part of the AKP has increased its popularity among urban educated voters. The party’s democratization initiatives have also served to broaden its appeal. A similar trend can be seen in the Arab Spring countries where the Islamist parties have used pro-democratic/freedom discourse and lessened their anti-west discourse, sending a positive signal to the urban educated classes as well as the international community. Although this might seem like a short-term tactic, the Turkish case shows that this change of discourse can create expectations among voters, which may force Islamist parties to keep their promises. In turn, the parties can benefit electorally.

As a result of these and other policies, the AKP has been able to establish a cross-class coalition, receiving support from a variety of groups, ranging from housewives to public employees to pro-EU, mostly urban, educated voters. These policies have served to attract votes away from other parties. The message is clear. If rival parties, whether liberal or moderate conservative, cannot develop equivalent policies, Islamist parties will fill the gap and become more successful than their rivals.

Finally, this study has some implications for understanding electoral behavior outside the long-established democracies. First, it highlights the importance of taking performance evaluations into account. Low literacy rates in some developing countries may militate against issue voting. Levels of political knowledge are likely to be low and so many voters will not be acquainted with the parties’ policy positions. However, they can tell whether the threat of terror, say, is growing or receding, or whether political freedoms are expanding or contracting, and punish or reward the incumbent accordingly.¹⁷

Most studies in developing countries have also tended to overlook the importance of the leaders’ influence on vote choice despite the fact that parties, especially successful ones, tend to be leader dominated. The relative popularity of the party leader can tell us a good deal about why people vote the way they do. The findings emphasize the importance of the popularity of the party leaders in influencing voting behavior in developing countries and we argue that the leadership factor should be taken into account in studies of vote choice in the forthcoming democratic elections in the Middle East.

To conclude, the AKP’s success provides some important clues as to how an incumbent Islamist party can be electorally successful and how the expansion of social programs, sound economic policies and democratization initiatives can enhance the appeal of Islamist parties beyond their traditional conservative religious base. If these parties can manage to improve the condition of the needy, create the impression that their economic policies are successful and that the country is becoming more democratic, they may well succeed in garnering support from a variety of societal groups and we could see similar Islamist successes in the Arab Spring and other Middle Eastern countries when free and fair elections are held. In the near future, we will be able to tell whether these assertions hold and whether Islamist parties’ strategies converge across countries or not.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author biographies

Elisabeth Gidengil is Hiram Mills Professor in the Department of Political Science at McGill University. She is the founding director of the inter-university Centre for the Study of Democratic Citizenship. Her research interests include voting behavior and political engagement with a particular interest in gender, social class and ethno-racial diversity. She was a member of the Canadian Election Study team from 1992 to 2008. Her latest book is Canadian Democracy from the Ground Up: Perceptions and Performance.

Ekrem Karakoç is an Assistant Professor in the Department of Political Science at Binghamton University, SUNY. His primary field is Comparative Politics with a focus on comparative political economy, political behavior and democratization. His works have appeared in a number of journals including World Politics, Comparative Politics, Comparative Political Studies, European Political Science Review, Electoral Studies, Turkish Studies, Canadian Journal of Political Science.

Appendix: Question wording

500 TL and below	4001 TL – 4500 TL
501 TL – 1000 TL	4501 TL – 5000 TL
1001 TL – 1500 TL	5001 TL – 5500 TL
1501 TL – 2000 TL	5501 TL – 6000 TL
2001 TL – 2500 TL	6001 TL – 6500 TL
2501 TL – 3000 TL	6501 TL – 7000 TL
3001 TL – 3500 TL	7001 TL – 7500 TL
3501 TL – 4000 TL	7500 TL and above

References

Akdoğan

(2004) AK Parti ve Muhafazakar Demokrasi. Istanbul: Alfa Yayinlari.

Arat

(2010) Religion, politics and gender equality in Turkey: Implications of a democratic paradox? Third World Quarterly 31(6): 869–884.

Auf

(2013) Why the Muslim brotherhood fell politically (Part II). Available at: www.acus.org/blogs/egyptsource/why-the-muslim-brotherhood-fell-politically-part-ii accessed on September 3, 2013).

Başlevent

Kırmanoğlu

Senatalar

(2005) Empirical investigation of party preferences and economic voting in Turkey. European Journal of Political Research 44(4): 547–562.

Başlevent

Kırmanoğlu

Senatalar

(2009) Party preferences and economic voting in Turkey (now that the crisis is over). Party Politics 15(3): 377–391.

Belge

(2011) State building and the limits of legibility: Kinship networks and Kurdish resistance in Turkey. International Journal of Middle East Studies 43(1): 95–114.

Brocker

Künkler

(2013) Religious parties: Revisiting the inclusion-moderation hypothesis 1 – Introduction. Party Politics 19(2): 171–186.

Brumberg

(2002) The trap of liberalized autocracy. Journal of Democracy 13(4): 56–68.

Clark

(2004) Islam, Charity, and Activism: Middle-Class Networks and Social Welfare in Egypt, Jordan and Yemen. Bloomington, IN: Indiana University Press.

10.

Çarkoğlu

(2010) Ideology or economic pragmatism? Profiling Turkish voters in 2007. Turkish Studies 9(2): 317–344.

11.

Çarkoğlu

(2012) Economic evaluations vs. ideology: Diagnosing the sources of electoral change in Turkey, 2002–2011. Electoral Studies. DOI: 10.1016/j.electstud.2012.02.005.

12.

Demiralp

(2009) The rise of Islamic capital and the decline of Islamic radicalism in Turkey. Comparative Politics 41(3): 315–336.

13.

Demiralp

(2012) The odd tango of the Islamic right and Kurdish left in Turkey: A peripheral alliance to redesign the centre? Middle Eastern Studies 48(2): 287–302.

14.

Dirilen-Gümüş

(2010) The effect of religiosity on political ideology via value types and personality traits: A comparison between Turkey and USA. Procedia: Social & Behavioral Sciences 5: 12–17.

15.

Eligür

(2010) The Mobilization of Political Islam in Turkey. Cambridge: Cambridge University Press.

16.

Ergüder

(1980–1981) Changing patterns of electoral behavior in Turkey. Boğaziçi University Journal 8–9: 45–81.

17.

Esmer

(1995) Parties and the electorate: A comparative analysis of voter profiles of Turkish political parties. In: Çiğdem

E, Cevat

Gareth

Feroz

(eds) Turkey: Political, Social and Economic Challenges in the 1990s. Leiden: EJ Brill, pp. 74–89.

18.

Esmer

(2002) At the ballot box: Determinants of voting behavior in Turkey. In: Esmer

Sayarı

(eds) Politics, Parties and Elections in Turkey. Boulder, CO: Lynne Rienner, pp. 91–114.

19.

Fearon

(1999) Electoral accountability and the control of politicians: Selecting good types versus sanctioning poor performance. In: Manin

Przeworski

Stokes

(eds) Democracy, Accountability, and Representation. Cambridge: Cambridge University Press.

20.

Fiorina

(1981) Retrospective Voting in American National Elections. New Haven, CT: Yale University Press.

21.

Fuller

(2012) Arab Spring: Now begins the education of Islamist politicians. Christian Science Monitor, May 1. Available at: www.csmonitor.com/Commentary/Global-Viewpoint/2012/0501/Arab-Spring-now-begins-the-education-of-Islamist-politicians, accessed on May 23, 2012.

22.

Gerber

Green

(1999) Misperceptions about perceptual bias. Annual Review of Political Science 2(1): 189–210.

23.

Grzymała-Busse

(2002) Redeeming the Communist Past: The Regeneration of Communist Parties in East Central Europe. Cambridge: Cambridge University Press.

24.

Gümüşçü

(2010) Class, status, and party: The changing face of political Islam in Turkey and Egypt. Comparative Political Studies 43(7): 835–861.

25.

Gümüşçü

(2013) The emerging predominant party system in Turkey. Government and Opposition 48(02): 223–244.

26.

Gülalp

(2001) Globalization and political Islam: The social bases of Turkey’s welfare party. International Journal of Middle East Studies 33(3): 433–448.

27.

Hale

(2005) Christian democracy and the AKP: Parallels and contrasts. Turkish Studies 6(2): 293–310.

28.

Hamid

(2011) Arab Islamist parties: Losing on purpose? Journal of Democracy 22(1): 68–80.

29.

Healy

Malhotra

(2013) Retrospective voting reconsidered. Annual Review of Political Science 16: 285–306.

30.

Kalaycıoğlu

(1994) Elections and party preferences in Turkey: Changes and continuities in the 1990s. Comparative Political Studies 27(3): 402–424.

31.

Karakoç

(2013) Ethnicity and trust in national and international institutions: Kurdish attitudes toward political institutions in Turkey. Turkish Studies 14(1): 92–114.

32.

Kılıç

(2008) The gender dimension of social policy reform in Turkey: Towards equal citizenship? Social Policy & Administration 42(5): 487–503.

33.

Maddy-Weitzman

(2011) Tunisia's morning after: Middle Eastern upheavals. The Middle East Quarterly 18(3): 11–17.

34.

Mardin

(1973) Center periphery relations: A key to Turkish politics? Daedalus 102(1): 169–191.

35.

Miller

Shanks

(1996) The New American Voter. Cambridge, MA: Harvard University Press.

36.

Moussaoui

(2011) Moderate Arab Spring Islamists look to ‘Turkish model’ for new democracies. Available at: http://news.nationalpost.com/2011/12/02/moderate-arab-spring-islamists-look-to-turkish-model-for-new-democracies/ (accessed May 24, 2012).

37.

Önis

(1997) The political economy of Islamic resurgence in Turkey: The rise of the welfare party. Third World Quarterly 18(4): 743–766.

38.

Özbudun

(1996) Turkey: How far from consolidation? Journal of Democracy 7(3): 123–138.

39.

Peterson

(2013) Egypt’s coup shakes Brotherhood’s Islamic partners in Turkey. The Christian Science Monitor, July 10 (accessed September 3, 2013).

40.

Saif

Rumman

(2012) The economic agenda of the Islamist parties. Available at: http://carnegieendowment.org/files/islamist_econ.pdf (accessed June 21, 2012).

41.

Sarıgil

(2010) Curbing Kurdish ethno-nationalism in Turkey: An empirical assessment of pro-Islamic and socio-economic approaches. Ethnic and Racial Studies 33(3): 533–553.J

42.

Schwedler

(2006) Faith in Moderation: Islamist Parties in Jordan and Yemen. Cambridge: Cambridge University Press

43.

Sokhey

Yıldırım

(2013) Economic liberalization & political moderation: The case of Islamist and Communist parties. Party Politics 19(2): 230–255.

44.

Sunar

Binnaz

(1983) Islam in politics: The case of Turkey. Government and Opposition 18(4): 421–441.

45.

Tanıyıcı

(2003) Transformation of political Islam in Turkey: Islamist welfare party’s pro-EU turn. Party Politics 9(4): 463–483.

46.

Tessler

(1997) The origins of popular support for Islamist movements. In: Entelis

(ed.) Islam, Democracy, and the State in North Africa. Bloomington, IN: Indiana University Press, pp. 93–126.

47.

Tezcür

(2010) The moderation theory revisited: The case of Islamic political actors. Party Politics 16(1): 69–88.

48.

The Economist (2012) Unfinished business: Revolutions have hurt the wallets of bosses and workers alike. February 4. Available at: www.economist.com/node/21546018.

49.

Thomassen

(2005) The European Voter: A Comparative Study of Modern Democracies. New York: Oxford University Press.

50.

Tuğal

(2009) Passive Revolution: Absorbing the Islamic Challenge to Capitalism. Stanford, CA: Stanford University Press.

51.

Turan

(1988) Political parties and the party system in post-1983 Turkey. In: Evin

Heper

(eds) State, Democracy and the Military: Turkey in the 1980s. Berlin: Walter de Gruyter, pp. 63–80.

52.

Ülgen

(2011) From inspiration to aspiration: Turkey in the new Middle East. Available at: http://carnegieendowment.org/files/turkey_mid_east.pdf (accessed May 24, 2012).

53.

Volpi

(2009) Political Islam in the Mediterranean: The view from democratization studies. Democratization 16(1): 20–38.

54.

Warner

(2013) Christian democracy in Italy: An alternative path to religious party moderation. Party Politics 19(2): 256–276.

55.

Wickham

(2002) Mobilizing Islam: Religion, Activism and Social Change in Egypt. New York: Columbia University Press.

56.

Wiktorowicz

(2004) Islamic Activism: A Social Movement Theory Approach. Bloomington, IN: Indiana University Press.

57.

Yavuz

(1997) Political Islam and the welfare (refah) party in Turkey. Comparative Politics 30(1): 63–82.

58.

Yavuz

(2009) Secularism and Muslim Democracy in Turkey. Cambridge: Cambridge University Press.