Abstract
This study investigates Lipset-Rokkan’s “freezing hypothesis” based on the contemporary Indonesia cleavages by analyzing the 2014 election together with the previous elections. This study found the endured cleavages are ethnic and aliran (literally, “stream”), which consist of the competing abangan (nominal Muslim) and santri (devout Muslim) cleavages. Both had been “frozen,” or existed latantly, during the New Order, but they revived afterwards. However, the ethnic cleavage arguably has been “melted” faster than the aliran one as the former seems to be more vulnerable to other political appeals. The aliran, in contrast, is somewhat callous due to its complex basis. This paper also suggests that, because the strength of any individual cleavage may vary from one region to another, capturing the empirically existing cleavages better may require going to regional or district levels instead of being solely reliant on a national-level study. This study underscores that the cleavages in the form of ethnicity and aliran cannot simply be abandoned in the elections in Indonesia. It is even suggested that the aliran may be one of the dimensions of the endogeneity of voters’ religio-political identity.
Introduction
The year 2014 was certainly another essential year for Indonesia’s political development. As the world’s most populous Muslim country and the largest economy in Southeast Asia, Indonesia went to the polls. The election was the fourth parliamentary election since the collapse of the authoritarian Suharto’s New Order in 1998, and the 11th general election in the entire history of Indonesian politics. Overall, Indonesia experienced general elections in three of its four periods of democracy. The first free general election in Indonesia was held in 1955, 10 years after the country proclaimed its independence in the era of so-called liberal or constitutional democracy (1949–1958). Elections were absent in the era of Guided Democracy (1959–1965) when Sukarno, the first president of the Republic of Indonesia, reigned. The second to the seventh elections were held under Suharto, the second president of the Republic of Indonesia, and his regime called New Order (1966–1998), but all the elections were considered unfair. Finally, in the Reformation or the post-Suharto era (May 1998 up to now), Indonesia has conducted democratic general elections four times, in 1999, 2004, 2009, and 2014.
The 2014 general election consisted of 12 national political parties competing in the balloting. 1 Nevertheless, like previous Indonesian elections, the contest was also probably among the most complex of one-day electoral activities around the globe due to the administrative burden that had to be managed. 2 In this contest, Indonesian voters gave the Partai Demokrasi Indonesia Perjuangan (PDIP, Indonesian Democracy Party of Struggle) the victory. The PDIP, which has been twice unsuccessful in the 2004 and the 2009 elections and out of power for 10 years, obtained 18.95% of the vote, followed by the Golkar Party, the former election winner, and the Partai Gerakan Indonesia Raya (Gerindra, Great Indonesia Movement Party), a new rising party, with 14.75% and 11.81% votes, respectively. As the frontrunner in the 2014 election for the parliament, the PDIP’s number of votes was actually smaller than the most votes the party had ever achieved (33.7%) in the first election in 1999, soon after the downfall of Suharto. Votes for the PDIP continued to decline by almost half in the 2004 elections (18.5%) and decreased to the lowest point (14%) in the 2009 election.
However, the trend of party outcomes in this contest seems not only confined to the PDIP. Five of the other eight parties gained votes compared to the 2009 election. Two others received tallies almost identical to the 2009 vote. The newly formed party, the Partai Nasional Demokrat (Nasdem, National Democratic Party), passed the 3.5% threshold of the national vote as the minimum to have seats in the national parliament. Two other small parties, the PBB and the PKPI, failed to reach the threshold for national representation. Finally, the Democratic Party lost significant ground compared to the balloting five years ago. This party, which was formed by and used to be the vehicle for Susilo Bambang Yudhoyono, the former president of Indonesia, was the biggest loser in the 2014 contest.
The “rise and fall” of the parties’ electoral votes might reflect the dynamic of Indonesian cleavage structures. Social cleavages are widely assumed to have a close link with the party system. In other words, voters’ alignment to any party in the contest is often shaped by the existing cleavages’ formations. In a plural society like Indonesia, in which primordial sentiments frequently emerge, the cleavages do not only draw upon different social groups, but also, within various parameters, divide the society into conflicting sections. Because Indonesia is one of the countries that is socially plural, its election and party system have been frequently analyzed through cleavages framework.
In a comparative perspective, election and voter political alignment could not be perceived as the only notion. There are also alternative perspectives on election and voter political alignment due to the varying quality of democracy across countries. Elections in mature democracies, particularly those that generally exist in developed countries, presumably have different features from those held in less developed countries. In contrast, in developing countries, there are also many reasons behind voters’ political alignment during balloting. As shown in the work of Kerkvliet (1996), an election and its dynamics have multiple meanings and significances. Among the important functions of an election are identifying the enduring social cleavages across the nations and distinguishing continuity and change in cleavages reflected in the polls. In addition, as Schiller (2004) has highlighted, voters’ political alignment cannot be seen as merely “independent, rational and enlightened decisions based on their interest” as we frequently perceive in institutionalized liberal democracies. Furthermore, Powel (2000) has stated that elections have many functions and are not exclusively matters of political representation and accountability.
This paper discusses the 2014 Indonesian general election and beyond, which is particularly pertinent to critically appraising the forming social cleavages in contemporary Indonesia. The paper analyzes to what extent the elections have informed us about the contemporary cleavages in the country. As cleavage pattern formation is evolutionary in a society, this paper cannot solely and exclusively employ the data from the 2014 election. Instead, the paper traces back to the previous elections of, to borrow Huntington’s terminology, the second wave of democracy in the 1950s under Suharto’s authoritarian New Order (1971–1997) and the elections of the “third wave of democracy” (1999–2014). Thus, despite its importance, the 2014 election does not hold the exclusive spotlight in the discussion.
The paper utilizes information from a variety of sources. Regarding the election results in the era of Sukarno and Suharto, this paper uses nonsurvey data that had been widely published. For the elections in the Reformation era (1999, 2004, and 2009), the survey data are extensively utilized. For instance, this paper employs a series of studies from three national public opinion surveys 3 conducted by a leading pollster in Indonesia, Lembaga Survey Indonesia (LSI, the Indonesian survey institute), as comprehensively analyzed by Mujani, Liddle and Ambardi (2012). As for the 2014 general election, as a substitute the paper relies on data derived from the outcome of the 2014 exit polls conducted by Saiful Mujani Research and Consulting (SMRC). 4
The paper consists of five sections. The first briefly discusses the theoretical terrain of Lipset and Rokkan that spotlights social cleavages, party systems, and voters’ alignment. This section will critically review the cleavages concept, particularly its relevance to analyzing Indonesian elections. The second identifies the existing cleavages in Indonesia in the 1950s democracy. It explains the configuration of the cleavages and their formation. The third elaborates on some critical junctures under the authoritarian New Order regime and identifies the cleavages at the time. The fourth examines the revived cleavages in the post-authoritarian period (1999–2014). It particularly underscores the continuity and change of the cleavages towards the 2014 election. It will critically evaluate what has changed in dealing with the “frozen” Indonesian cleavage structures. It also highlights changing trends in recent elections. The final part draws some preliminary conclusions.
Social cleavages, party systems, and voters’ alignment
Party systems are rooted in the existing cleavages structure. In this sense, the latter to a great extent characterizes the former in many aspects. Any discussion about party systems up to contemporary analysis on the relationship between the party system and the cleavages structures has almost always referred to Seymor M Lipset and Stein Rokkan’s seminal book (Lipset and Rokkan, 1967). This most cited reference was based on their study of the dynamics of the Western European party system. They convincingly argued that the significant influence of social cleavages on Western European party systems has been patterned or even “frozen” for decades. The extract of their central argument is (1967: 50): “The party systems of the 1960s reflect, with few but significant exceptions, the cleavage structure of the 1920s …. the party alternatives, and in remarkably many cases the party organizations, are older than the majorities of the national electorates.”
The study focused on the established social cleavages as the consequences of large-scale social transformations propelled by two prominent historical events: nation-building and industrialization. In Lipset and Rokkan’s words (1967: 34), these two concepts “in basic characteristics of party system that emerged in the Western European politics during the early phase of competition and mobilization can be interpreted as products of sequential interactions between these two fundamental processes of change” (italics in original). Their analyzes of cleavage structures were taken along territorial and functional axes. They then identified four dimensions of oppositions, as will be elaborated later, in Western European democracies in which party systems took root.
Of the four opposition dimensions, two were the products of the national revolution and the others were shaped by the Industrial Revolution. The national revolution had facilitated the conflict between central nation-builders and the peripheral subject culture and then were known as “center versus periphery” hostility. The other tension was between the church and the centralized state, which was then recognized as the “church versus secular state” conflict. Dealing with this latter conflict, Lipset and Rokkan (1967: 34) state that many countries of Western Europe were all split to the core in the wake of the secularizing French Revolution and without exception developed strong parties for the defense of the Church, either explicitly as in Germany, the Low countries, Switzerland, Austria, Italy, and Spain or implicitly as in the case of the Right in France. The conflict between landed and urban interests was centered in the commodity market. The peasants wanted to sell their wares at the best possible prices and to buy what they needed from the industrial and urban producers at low cost. Such conflicts did not invariably prove party-forming … [d]istinctly agrarian parties have only emerged where strong cultural oppositions have deepened and embittered the strictly economic conflicts.
As noted before, Lipset and Rokkan’s concept of cleavages has been very influential in analyzing the composition of parties. The concept is frequently concerned with the number of parties, interaction among parties, and connections between the parties and the electorates. Regarding the number of parties, cleavages structure is widely seen as a determinant in shaping parties’ configurations. Cleavages might manifest in several forms along socio-economic, religious, ethnicity, and urban-rural parameters. The number of the emerging parties would depend on how many parties were organizationally capable of building their political bases on the existing cleavages. The parties’ interactions would be characterized by either sharp or blunt differences among cleavages structure. It seems that the sharper the cleavages distinction, the more polarized the party system is. Moreover, voters’ alignment during the election would depend on how successful any of the parties are in institutionalizing and politicizing the cleavages to be politically appealing as part of their efforts to mobilize political support. Despite its academic influence, Lipset and Rokkan’s concept has been challenged on several grounds, as can be seen in the works of Inglehart (1984) and Mair (1997), but the challenges were beyond the scope of the present study. Nevertheless, critical review of the concept is suggested in the Indonesian context. As will be elaborated, although the concept is helpful to explain diverse voters’ alignment in Indonesian elections, the concept might not succeed due to its limitations.
Indonesian cleavages in the 1950s
Lipset and Rokkan’s concept of cleavages was useful in describing the configuration of the Western European party system, including its potential conflicts. The model to some extent also explains the characteristics of the Indonesian party system. Of the four suggested cleavage structures, only two derived from the territorial axes are relatively prominent, particularly in the post-independent democratic period (1950–1959), while the other two models are hardly found in the context of Indonesia. One main reason was that during the early period after Indonesian proclaimed its independence in 1945, there had not been a significant economic change in the country due to the absence of industrialization. As a result, no significant economic-based cleavage emerged nationwide.
Looking more deeply at the Indonesian cleavage structures, the 1955 general election provided a reliable and valuable picture. As mentioned, the 1955 balloting was the first general election in the country. It was held on 29 September 1955, and 30 political parties participated in it. This first polling is often seen as one of the most ideal and most open democratic contests in Indonesian history. Besides being inclusive by accommodating all existing political parties in the election commission body, the event was also characterized by democratic principles such as the neutrality of the bureaucracy and the military, the absence of election-related violence in the community, high respect for freedom, and the plurality of participant elections. Accordingly, the 1955 election result genuinely reflected the configuration cleavage structures.
Lipset-Rokkan’s center-periphery cleavage model, which was particularly related to the cultural-ethnic dimension, to a greater extent fits the 1950s party divisions. This can clearly be seen in the 1955 electoral results, as has been described by Feith (1957: 32 and 33), particularly the votes gained by the four major parties: the Majelis Syuro Muslimin Indonesia (Masyumi, Consultative Council of Indonesian Muslim Party); Partai Nasional Indonesia (PNI, Indonesian National Party), representing the “modernist” Muslim; Nahdlatul Ulama (NU, Renaissance of Scholars Party), which maintained the “traditional Muslim perspectives”; and Partai Komunis Indonesia (PKI, Indonesian Communist Party). Table 1 shows that the four major parties gained four-fifths of the total votes. The Masyumi, the largest party, drew its political support mainly from outside Java island. According to the same work of Feith (1957), this party was thus perceived to represent the peripheral Outer Java regions. Meanwhile, the other three major parties—the PNI, the NU, and the PKI—obtained more electoral support across Java island, where the central government resided. In Java region, where 66.2% of the total population lived, the PNI got 85.9%, the NU received 85.6%, and the PKI took 88.6%. More specifically, the strength of these three major parties was concentrated in the Central and East Java areas, where 45.6% of Indonesia’s total population resided. In these areas, of the total votes, the PKI obtained 74.9%, followed by the NU (73.9%) and the PNI (65.5%). Based on the outcome, these three parties were recognized as the “Java-based” parties.
The 1955 election results (nationally).
Notes: PNI (PartaiNasional Indonesia, Indonesian Nationalist Party); Masyumi (MajelisSyuroMuslimin Indonesia, Consultative Council of Indonesian Muslims); NU (NahdatulUlama, Renaissance of Islamic Scholars); PKI (PartaiKomunis Indonesia, Indonesian Communist Party); PSII (PartaiSarekat Islam Indonesia, Islamic Association Party Indonesia); Parkindo (Partai Kristen Indonesia, Indonesian Christian Party); PartaiKatolik (Catholic Party); PSI (PartaiSosialis Indonesia, Indonesian Socialist Party).
Source: Ufen (2006: 8).
The second model of “church-state” cleavage to a certain degree was also relevant. However, to explain Indonesian political parties’ configurations, the application of the model to the Indonesian context needs some modifications. There are several reasons for this. First, as the largest Muslim country, Indonesia has no strong church institution with its rigidly hierarchical structure like those in Western European countries. Yet, an identical cleavage has functionally persisted along religious issues. The main political rivalry did not express itself through different religions, such as Christian versus Islam and the like. Instead the competition was more concentrated in the Islamic community itself.
Dealing with cleavages division, the 1950s Indonesian voters’ choice had mainly been attributed to “aliran” (literally, “stream”) loyalties. It should be noted here, nonetheless, that the aliran is neither simply religious nor ideological. In his study of Modjokuto (East Java) in the 1950s, Clifford Geertz revealed the political weight of a so-called aliran when he defined the aliran as “more than a mere ideology; as it has been described by Geertz (1965: 128) the aliran was a set of interconnected social forms which acted to large group masses of people in a generalized category.” The aliran can thus be seen as a kind of “mode of integration” that may include religious values, social and/or religious practices, beliefs, loyalties, and identity. Bearing all these in mind, it could, therefore, be argued that the application of “secular-religious” division here might need a slight modification since the aliran is not exclusively reliant on the religious element. In terms of social and political groupings, Geertz (1965: 128) has noticeably shown that the aliran-based cleavage comprised “a political party surrounded by a set of sodalities—that is voluntary organization—normally or informally linked to it” such as youth groups, women’s clubs, labor and farmers’ unions, and the like. More importantly, the memberships of these subordinate organizations were almost exclusively limited to aliran.
In Java, aliran had such characteristics. Another study by Geertz (1959) shows that the competing aliran in Java were obvious and strong. Borrowing Geertz’s words (1959: 37), this situation was primarily so because of the so-called “social contrivances of Indonesian nationalism” shaped during the revolution. The constant spread of national political influence, powered by the improvement of communication, had facilitated the competing aliran to strengthen village social divisions. In this sense, the aliran allowed socio-religious divisions within a Muslim community between santri (devout Muslims) and abangan (nominal Muslims). 5 Geertz (1959: 37) underscores that cleavage division was politically significant in shaping or limiting partisan voting behavior, particularly for the four competing major parties. In a similar vein, Wertheim (1987: 116) has shown that pretty firm swapping of political affiliations between the two was most unlikely. In addition, the santri community would quite possibly be leaning politically towards the two major competing Islamic parties, the Masyumi and the NU. In contrast, the abangan community would instead commonly be aligned with one of the two major parties, the PNI and the PKI. Yet, the aliran contestation, and thereby the santri and abangan distinction, prominently existed within Javanese and had limited value outside Java. The 1955 election gains of the four major parties also indicated that the competition of the aliran notably occurred in Java, predominantly between the NU versus the PNI and the PKI. A recent comparative study by Patriadi (2007: 93–98 and 108–110) in two different districts – one in Java and one outside Java – also confirmed that in the 1955 election, the conflicting aliran cleavage was intense in Java and almost absent outside Java.
However, the democratic political system that had fostered the open political contest was short-lived. In 1959, President Sukarno issued a decree to abandon the democratic parliamentary system and adopted an authoritarian regime, known as Demokasi Terpimpin or Guided Democracy. The works of Feith (1962: 515–520) and Lev (1966: 46–59) have argued that this political decree was part of disillusionment with the parliamentary system, which was widely seen as having politically marginalized Sukarno from his power. Furthermore, in 1960, the Masyumi party disappeared after being discredited for being involved in the regional rebellion against the central government in the late 1950s. Following the 1965 alleged PKI Coup, the PKI was also formally eliminated from the political landscape, leaving the two other major parties, the PNI and NU, to enter the New Order period.
Elections under authoritarian new order
With the 1965 abortive military coup led by General Suharto, the authoritarian New Order regime came to power. To strengthen his political power, besides thoroughly cultivating Indonesian politics using various methods of coercion to gain tight control over society, Suharto also backed Golkar to be his political machine instead of the existing parties. Additionally, unlike other authoritarian regimes that generally depend on coercive power, Suharto, as Liddle (1985: 71) has observed, had also created a supportive constituency through various forms of state-patronage practices, including many state-funded development projects. Increased state revenues from the oil boom that began in the early 1970s enabled Suharto to undertake all those programs.
As part of his efforts to gain political legitimacy, Suharto let a general election take place. The general election was held on 5 July 1971, after a 15-year vacuum. Ten political parties participated in the 1971 New Order’s first general election. 6 Since 1971, Suharto had consistently held general elections every five years, and he claimed that they represented a constitutional “democratic” political process. As Table 1 shows, Golkar witnessed a great success in the 1971 election by getting 62.8% of the total vote nationally. Golkar’s big victory seemed to make Suharto confident about constantly manipulating Indonesian politics. In preserving the New Order regime, ascertaining political party leadership, and ensuring the continuous winning of Golkar, in 1973 Suharto also set up a “forced fusion” of political parties to limit the number of parties participating in general elections. This policy forced all the previous existing political parties except Golkar to be part of either the Partai Persatuan Pembangunan (PPP, United Development Party) 7 or the Partai Demokrasi Indonesia (PDI, Democratic Party of Indonesia). 8 Thus, starting with the 1977 election, the electoral system allowed only three parties to participate in elections—Golkar, PPP, and PDI.
As predicted, many observers considered all elections during the Suharto period a facade due to the involvement of military operations in every election, the overly dominant role of the government as election organizers, and the lack of involvement of the community at almost all levels of institutional and electoral process. Some prominent Indonesianists such as Crouch (2010) and Liddle (1996) have argued that those New Order elections were then far from free and fair and that political maneuvering had made Golkar the winner in landslide victories. Table 2 shows that Golkar won by roughly 65% in the elections from 1977 to 1997. In the end, following each election, Golkar dominated the Majelis Permusyawaratan Rakyat (MPR, People’s Consultative Assembly), 9 repeatedly electing Suharto as the President for more than 32 years until 1998, when people power ousted him.
The 1977–1997 Indonesian election results (%).
Source: Suryadinata (2002: 32).
Notes: *) Ten parties participated in the 1971 election. The PPP figure was the combined electoral result of the merged Islamic parties (NU, Parmusi, Perti (Persatuan Tarbiyah Islamiyah, Islamic Education Association) and PSII), while the PDI’s votes were a compilation of five other non-Islamic parties’ votes (PNI, Parkindo, Catholic Party, Murba Party (Indonesian: Murba stands for Musjawarah Ra’jat Banjak), and IPKI).
Because the New Order’s electoral gains were predominantly the result of the government’s political engineerings, it was hard to see them as a genuine reflection of the dynamics of cleavages structures. Although the New Order regime had encouraged the “functional groups” unified into Golkar to eradicate the influence of aliran-based parties, the effort seems to have failed. In other words, New Order repression did not necessarily mean that the aliran was eliminated. Instead, it only made it go underground. Suharto’s resignation on 21 May 1998 brought to an end the longest era of one of the world’s most durable authoritarian regimes, which had lasted for more than three decades. The installed democratic regimes allowed the aliran to revive although they experienced some changes.
Towards the 2014 election: Cleavages in the post-authoritarian period
Following the collapse of Suharto’s New Order in 1998, to gain the recognition and trust of the public domestically and internationally, the new appointed President Habibie immediately made several popular political decisions as part of his efforts to install a democratic political system. Besides releasing political detainees, respecting human rights, and encouraging the establishment of parties, he also accelerated the implementation of an open and democratic election. The polls were held on 7 June 1999, with 48 political parties participating, the largest number ever.
The election was the first openly contested since 1955 and led to the forming of the new DPR and the MPR. 10 The election resulted in the five largest parties emerging. Table 3 shows that from the total votes, the Indonesian Democratic Party-Struggle (PDI-P, the PNI inheritance) won 33.7%, while Golkar gained 22.4%, followed by the National Awakening Party (PKB, the revived NU) which got 12.6%, and the United Development Party (PPP) which obtained 10.7%. Since 1999, there have been three consecutive democratic elections up to the recent 2014 ballot. Since all were held under democratic settings, the post-authoritarian elections are relatively comparable to the 1955 election. Hence, the voting trend of the consecutive elections result can reasonably be the foundation for examining the characteristics of changes in the cleavage structure. However, the analysis cannot exclusively rely on the 2014 election for at least two reasons. Firstly, referring to Lipset-Rokkan’s “freezing hypothesis,” any change of cleavage structures theoretically would be evolutionary. Secondly, as will be elaborated upon below, general trends in the results of the four consecutive elections were relatively stable. It could thus be argued that it is better to place the 2014 election in a wider context and analyze it combined with the previous elections (1999, 2004, and 2009) to identify any changed pattern.
The 1999–2014 election results (% nationally).
Notes: PDIP—Partai Demokrasi Indonesia Perjuangan (Indonesian Democracy Party—Struggle); Golkar—Golongan Karya (Golongan Karya Party); PKB—Partai Kebangkitan Bangsa (National Awakening Party); PPP—Partai Persatuan Pembangunan (United Development Party); PAN—Partai Amanat Nasional (National Mandate Party); PD—Partai Demokrat (Democrat Party); PKS—Partai Keadilan Sejahtera (Justice and Prosperous Party); Gerindra—Partai Gerakan Indonesia Raya (Great Indonesia Movement Party); Hanura—Partai Hati Nurani Rakyat (People’s Conscience Party); Nasdem—Partai Nasional Demokrat (National Democrat Party).
Sources: The outcomes of the elections of 1999, 2004, and 2009 are from Mujani and Liddle (2010: 38); for 2014 they are from Vermonte (2015: 306).
One prominent change in the contemporary cleavage structures from those in the 1950s was related to the division of parties’ electoral support. It seems that in the 1999 election, the votes split pattern was relatively identical to that in the 1950s. Table 3 shows that the four major parties (the PDIP, Golkar, PKB, and PPP) nationally maintained around four-fifths of the total votes, which was very similar to the pattern of the 1955 election. However, interestingly, from 2004 up to 2014, the pattern of electoral votes had been steadily scattered into more other parties. Table 3 shows that the four major parties, with a different configuration, overall gained almost two-thirds of the total national votes, which was approximately a 20% decline from that in the 1999 election. The election outcomes indicated that voters’ political alignments to the parties had been becoming more fluid, reflecting a lessening change in the existing cleavage structures. Yet, as will be elaborated later, some cleavage-based politics to some extent still persist.
A clearer picture of the softening change of the cleavage structure can also be seen in voters’ choices for any single party based on ethnicity. Table 4 shows that the configuration of the ethnic-based cleavage has slightly changed from those in the 1950s. From the major parties, only the PKB, being seen as the revived NU, continued to be a Java-based party, at least until the 2009 election. The outcome in Table 4 indicates that overall it gained votes across the elections, except for the 2014 election 11 which reached roughly 70% in Java compared to 30% in the Outer-Java region. On the contrary, Table 4 also shows that the majority of the parties except the PDIP have been representing the Outer Java voters more by gathering overall electoral votes roughly around 40% in Java and 60% in Outer Java. Compared to the 1950s, these parties seemed to have taken over the previous Masyumi’s political base. In this respect, Golkar to a great extent could be seen as the “new” form of the extinct Masyumi party.
Votes for individual parties (based on ethnic cleavage), 1999–2014 (%).
Sources: Mujani and Liddle (2010: 210). *) SMRC (2014).
The 1999 election result even reflected, with a slight difference, the revived old pattern of voter alignment division between Java and Outer Java. From the total of 27 provinces, the PDIP—as the PNI’s successor—won 10 provinces, while Golkar was victorious in another 15. Interestingly, of the 15 provinces in which Golkar won, all were Outer Islands provinces. Conversely, Golkar suffered electoral losses across all Java provinces, where the PDIP maintained its superiority (Patriadi, 2007: 14). The 1999 voting trends of the Outer Java regions also clearly support the pattern. Although Golkar suffered a 56.3% drop down to 35%, it was still moderately higher than the PDIP’s 31.2%. In contrast, in Java, Golkar faced a 50.9% decline, down to only 16.8%, less than half of the PDIP’s 37.5% (Schiller, 1999; Suryadinata, 2002: 218–223). Considering the rationale behind this voting pattern, it might be reasonable to say that the PPP, PAN, and PKS inherited the defunct Masyumi’s cleavage because they were Islamic-based parties. However, a similar explanation would be problematic to apply the Golkar, the PD, the Gerindra, the Hanura, and the Nasdem because these five parties were not Islamic-based.
Nevertheless, there are some possible alternative explanations for all five parties (the Golkar, the PD, the Gerindra, the Hanura, and the Nasdem). As has been elaborated, as the state’s party, Golkar has received some political benefits. During the durable authoritarian New Order, for more than three decades, the government used various methods to enable the party to take root in society (Patriadi, 2007: 58 and 140–157). Besides using coercive power such as political repression, the authoritarian New Order regime had more sophisticated ways to get political support for Golkar. For example, the government skillfully created a supportive constituency through state-funded and managed development programs such as financing public facilities like infrastructure projects and health, and, most importantly, providing agricultural subsidies for farmers. The other method was to recruit many local traditional leaders as Golkar’s functionaries. These popular policies in turn allowed Golkar to attract a great amount of political support in remote areas, particularly in the Outer Java region. Since the four new parties—the PD, the Gerindra, the Hanura, and the Nasdem—basically originated from the splintered Golkar factions, it is safe to argue that, to some extent, those parties have also benefited from Golkar’s privileges. However, another possible explanation should not be overlooked. The fact that the leaders of the five parties were the prominent national figures widely known to the public might also have contributed to the parties’ moderate electoral support. 12 Additionally, it is widely assumed that the popularity of the four parties has been more reliant on the reputation of their national parties’ leaders.
The other more interesting issue was the phenomenon of the PDIP, which has moved from its previous political basis. Referring to the electoral figures across the four consecutive elections shown in Table 3, after receiving 33.7% of the 1999 total votes nationally, the PDIP suffered a 15.2% decline in the 2004 election. The PDIP then experienced a relatively constant gain in votes up to the recent 2014 election. However, if the PDIP’s stable gain of votes is broken down based on ethnicity, the PDIP had gathered relatively balanced electoral votes both in Java and Outer Java. Table 4 shows that across the four consecutive elections from 2009 to 2014, the party obtained roughly 50%, even in the two different regions. These figures could indicate that the PDIP, seen as the PNI’s heir, has not only been successfully widening its supporting social cleavage, but also has been a step ahead of the PKB—the revived NU—in leaving its 1950s label to become the Java-based party.
In addition, another change of cleavage structure seemed to occur in the aliran division. Although there is, unfortunately, no data about aliran partition in the 2014 election’s exit polls, the available data of the previous three elections (1999, 2004, and 2009) provide a reliable number for analysis. Table 5 shows that nationally, the general trend was that all parties, surprisingly, were inclined to have more santri voters than the abangan one, with a ratio of roughly 75% to 25% of the total votes nationally. Interestingly, there have been contradictory tendencies dealing with the parties’ electoral support in the santri-abangan division. As Table 5 shows, concerning the received electoral support from the santri voters, all Islamic-based parties (PPP, PKB, and PAN) were likely to continually decline, so as a consequence they had a steady increase of abangan voters. While all the largest non-Islamic based parties (PDIP and PD) except Golkar were able to attract more santri voters, those parties consequently have fewer abangan voters. All trends considered, this paper therefore suggests that the aliran-based cleavage division has been becoming insignificant in determining voters’ choices.
Votes for individual parties (based on aliran cleavage), 1999–2014 (%).
Sources: Mujani and Liddle (2010: 193). *) SMRC (2014).
In line with this trend, some researchers claim that the religious factor had been losing its political significance. One argues that the waning political importance of religion might be related to the fact that voters were becoming “rational.” The Indonesia Survey Institute (LSI) survey found that non-Islamic parties were more politically capable, especially in offering appealing programmes in response to the public (Lembaga Survey Indonesia, 2008). In other words, the nationalist parties, particularly the PDIP and PD, were successful in gaining support from santri voters mainly because they could benefit from the increasing numbers of “rational” voters.
However, not all agreed with this argument. Sunny Tanuwidjaja, for example, quarrelled with the claim that the increasing number of “rational” voters has been replacing the political significance of religion, stating that such a conclusion is premature. He vividly argues that the success of the nationalist parties to attract the enlarged santri voters indicated that political Islam is unquestionably still significant, pointing to its influence on the nationalist parties becoming more accomodative towards religious matters (Tanuwidjaja, 2010: 31). This argument seems plausible, particularly if we recall that since the 1980s, the santri voters had steadily increased due to, as Ufen said, the process of santrinisasi in which many abangan (secular) had turned into santri (pious) (Ufen, 2008). As such, it is fairly reasonable to say that the phenomenon might have urged the nationalist parties to be responsive to them to attract political support. In criticising the LSI survey’s “rational” voters argument, Tanuwidjaja raised two crucial points. First, it is important to differentiate “political Islam” from “Islamic parties.” Second, the survey was doubtful as it was more reliant on the “exogenous” judgment than the “endogenous” one 13 of voters’ religious characteristics (Tanuwidjaja, 2010: 35 and 44). These two propositions will be further critically examined later in discussing the findings.
It would be misleading, however, to see the national general voting trend of parties’ electoral support, shown in Table 5, as the single general pattern in every individual region nationwide. As noted before, the religio-cultural division, commonly known as the santri-abangan dichotomy or aliran, had value on Java and not on the Outer Islands. Bearing this in mind, it is not exaggerated to say that there might be a different pattern of the aliran’s political significance from one region to another. To be sure, some studies have supported this assumption. Through a multivariate regression analysis, both nationally and in the specific region of Central and East Java, Dwight Y King found “broad continuities” in the 1955 and 1999 elections (King, 2003: 121–134). He convincingly argued that in the 1999 election, basic cleavages of the santri and abangan division re-emerged, as reflected in the continuing areas of support for the nationalist and Islamic parties. In line with this, the division within Islamic communities, known as the “modernist” and “traditional” dichotomy, also revived. In the areas where the PDIP gained, the higher votes were the political basis of two former biggest nationalist parties: the PNI and PKI. In santri communities, most PKB votes came from NU-dominated areas. PPP votes were higher in any area where Islamic parties (Masyumi, Perti, and NU) received strong political support, while the PAN and PBB received votes in previous Masyumi strongholds.
King’s findings in the 1999 election are also supported by other studies. Through the analyses of the 1999 and 2004 election results across Java regions, particularly in Central and East Java, a similar voting pattern was also found. Although there had been a perception that the aliran had been softening, it was arguably still powerful in shaping voters’ political alignment, particularly with the division in political support for nationalist and Islamic parties (Patriadi, 2007: 371–373). Of them, the most interesting phenomenon was Golkar, which to a great extent has persistently been perceived to represent the abangan community more. This was in line with the fact that since the 1999 election, Golkar had never been victorious in the Java regions. Surely this regional trend was completely different from the national trend. As Table 5 shows, Golkar nationally did indeed maintain “Islamic” characteristics, as the Golkar had a huge number of santri voters. In this respect, King’s multivariate regression analyses captured the Golkar trend (King, 2003: 133) but unfortunately did not explain it and how to relate it to the recent transformation of the abangan voters or the santri-nization phenomenon. Consequently, one question can be raised here: How did Golkar come to have “two different features”?
There is at least one possible explanation for the Golkar phenomenon. The Golkar nationally was perceived as “Islamic” because it gained relatively significant support from the dramatic santri voters, while in Java regions, especially in Central and East Java, the party continued to be seen as the “nationalist” or “secular” party (hereafter, we call this “the Golkar phenomenon”). The party nationally has been successful in attracting more santri voters, as seen in Table 5, because the surveys captured Golkar’s nationally gained votes that predominantly are represented by its santri cleavage in Outer Java regions. As noted, this phenomenon has situated Golkar as the main inheritor of the defunct Masyumi. As the previous “state” party under the authoritarian New Order, unlike in Java regions, Golkar had faced strong opposition, particularly from NU supporters (Liddle, 1978; Ward, 1974); the party in contrast has taken root within society across outside Java regions without meeting serious resistance. There are at least two related reasons behind this situation. First, in the Outer Islands regions, civil society was generally weaker, making the New Order state patronage effective in institutionalizing Golkar. Second, facing a strong state combined with a limited path for political mobility since the disappearance of the Masyumi had encouraged many Muslim local elites in the regions to join Golkar (Patriadi, 2007: 50–51) . As such, in terms of Islamic image, people outside Java commonly saw Golkar as no different from the previous Islamic Masyumi. This once again underscores that the dichotomy of aliran cleavage did matter in Java instead of in the Outer Islands regions.
The relatively strong persistence of the aliran in the Java region, particularly in Central and East Java, and Golkar’s different voting trend nationally and regionally, clouds analysis of recent voting behavior. The claim that the nationally combined decline of the Islamic parties across the post-New Order elections was due to the weakening significance of political Islam was questioned. As mentioned, the LSI survey which concluded that the declining Islamic parties’ votes reflects the more “rational” voters was deemed inaccurate because the survey based its judgment of voters more on exogeneity rather than an endogeneous process of their religious identity. Additionally, it was suggested that political Islam continued to be significant because the voting trend even showed that the nationalist parties have been replacing the Islamic parties to be more accommodating toward Islamic community aspirations (Tanuwidjaja, 2010).
This paper agrees with Tanuwidjaja’s argument, including his crucial proposition that it is important to differentiate between political Islam and Islamic parties. However, knowing that the dichotomy of abangan and santri, known as aliran, to a large extent still persists, particularly in Java regions, it can be argued that the strengthening political Islam evidenced in the national voting trend for the most part was very likely to happen in the Outer Islands regions rather than in the Java regions. In this respect, Golkar’s “double features” may even strengthen the argument. As elaborated before, in non-Java regions, the party had been seen as the heir of the former largest Islamic party, the defuct Masyumi, and nationally it could attract relatively big support from Islamic voters; in Java regions, the Golkar party in contrast continued to be perceived as the “nationalist” or “secular” party and persistently failed to get support from areas with Islamic communities (King, 2003: 133).
Furthermore, Java, specifically Central and East Java regions, were found to have this perception in the 1999 and 2004 elections, and this trend seemed to be persistent in the following elections. Golkar could not even take advantage by dealing with the increasing santri (devout or orthodox Islamic) voters. In other words, in Java regions, the process of “santri-nization” that had changed social composition arguably had no positive correlation with voters’ political affiliations and thus had no significant impact on their voting behavior (Patriadi, 2007: 190–194). Embarking from the argument that the aliran for the most part continues to be significant in shaping voting behavior, the aliran may be considered in Java regions to be one dimension of what Tanuwidjaja called the endogeneity of voters’ religious identity related to their political affiliations.
Specifically related to the 2014 election, another interesting trend should be noted. The 2014 election tells us about an indication that the new parties, such as Gerindra, Hanura, and even the “newly” formed Nasdem party, could obtain relatively significant electoral votes. The Gerindra is exceptional, having experienced a considerable increase from 4.5% in the 2009 general election to 11.8% in the 2014 polls. One interpretation is that voting trends might provide a further signal of the increasing fluidity of voters’ alignment in the Indonesian political landscape for at least two reasons. First, as newly formed parties, there would be little possibility for the parties to get so many votes because of their limited organizational capacity. To institutionalize any party would undoubtedly take a long time. Second, it seems likely that the significant votes gained by the relatively new parties (the Democrat, Gerindra, Hanura, and Nasdem) were instead mainly gained because of the magnetism of their party leaders, all of whom had become well-known national figures.
Nevertheless, another interesting point is related to the recent voting trends up to the 2014 election. It seems that social dynamics, particularly urbanization, to a certain extent politically also mattered, at least for some existing parties. Based on his multivaritate regression analyzes, King argued that “urbanization had a positive influence on voters for PDIP and PAN, but a negative impact on support for PKB and PPP” (King, 2003: 225). However, King’s argument was reliant on data only up to the 1999 election; it should therefore be revised because national voting trends from the 2004 to 2014 elections showed little difference, except for the PAN. As Table 6 shows, after experiencing a relatively big rise in rural support from 54% in 1999 to 66% in 2004, the PDIP went into a steady decline in rural areas, from winning 66% in 2004 to 62% and 51% in 2009 and 2014. As a consequence, it had steadily increasing support in urban areas from 34% in 2004, followed by 38% and 48% correspondingly in 2009 and 2014.
Votes for parties based on regions (% based on rural-urban division).
Sources: Mujani et al. (2012: 227). *) SMRC (2014).
Meanwhile, the PKB and the PPP also retained an identical trend. The PKB had a constant decline in votes in rural areas nationwide since the 1999 election, from 77% in 1999 down to 67%, 58%, and 56% in 2004, 2009, and 2014, respectively, and had therefore an increase in support from 23% in 1999 up to 33%, 42%, and 44% in the 2004, 2009, and 2014 elections. In the meantime, in rural areas, PPP nationally got 74% in 2004 and 70% and 54% in 2009 and 2014, respectively; it thus experienced a steady vote upsurge in urban areas by obtaining 26% in 2004, which rose to 30% and 46% in 2009 and 2014, respectively. By acquiring relatively constant increased votes in the three consecutive recent elections, it can be argued that urbanization had made the PDIP, PKB, and PPP, which had previously been seen as more rural-based parties, more acceptable to urban voters. In other words, urbanization had a steadily increasing political significance, particularly on the small shift of social bases, at least for the three large parties.
Considering the recent trend in the 2014 balloting, it can be argued that in line with the increase of swing voters, the “frozen” cleavages are likely to liquefy in the future. Although how quickly this will occur remains to be seen, such a process would be progressively more common in the future. In this respect, considering the 2014 election among the two prominent cleavages, the ethnicity and aliran-based, the former seems to be declining faster compared to the latter because the ethnicity-based cleavage seems to be more vulnerable to other political appeals. There are at least two supporting empirical facts for this. First, the Golkar phenomenon of being the “new” Masyumi occurred mainly because of state patronage. Second, related to the recent tendency, the four newly formed parties—Democrat, Gerindra, Hanura, and Nasdem—gained relatively significant numbers of votes largely due to the attractiveness of their national party leaders. However, such changes would be rather hard for the aliran, which is somewhat impervious to any other political pleas because, as Geertz had insisted, it is composed of complex “interconnected social forms” including values, beliefs, and ideology.
Conclusion
The paper has applied Lipset-Rokkan’s concept of cleavage to the Indonesian context. The concept to a certain extent was helpful for analysing the existing cleavage structures. However, it should be noted that the concept was formulated from a Western European model and consequently mostly does not work for the case of Indonesia. From the suggested two axes of cleavage—territorial and functional—only the former is mostly workable while the latter was almost entirely absent because Indonesia has not experienced a thorough industrialization.
Additionally, in terms of territorial axes, at least two prominent dimensions of cleavage were relevant, “centre-periphery” and “church-state” tension. However, only the first model fits Indonesia. To apply the second one requires some modifications because the religious tension has been much more complex than that in non-Western Europe. In Indonesia, the tension indeed was not only between religions but also within a single religion, as is evident in the enduring conflicting aliran cleavages. Dealing with Lipset-Rokkan’s “freezing hypothesis” to a certain extent does seem to fully work. Both ethnic and aliran-based had been “frozen” for some time, particularly during the New Order period, and revived afterwards.
Yet, the duration of being “frozen” is not necessarily assumed to be the same from one aspect to other. As has been elaborated, social and economic dynamics that have operated for decades to a certain extent had political significance, as can be observed by the impact of urbanization on the transformation of certain parties’ electoral votes and also the changed ethnic-based cleavage. However, the change of ethnic-based cleavage would arguably be faster than the aliran because the former empirically has been susceptible to “transactional” alignment. Meanwhile, although the aliran experienced a softening change, this “melting” process would be slow because as a solid complex social form, it would be impermeable to any alternative political plea. As has been elaborated, the political significance of the aliran had been proven to be relatively persistent, especially in Java regions. While the New Order regime through various methods had failed to eradicate it, the increased number of pious voters due to the process of santri-nization, at least in Java regions, had also no significant impact on aliran-based voting behavior. It could therefore be argued here that having such strong perseverance, the aliran may need to be seriously taken into account in identifying the endogeneity of voters’ religious identity, particularly in conjuction with the inclination of their political choices.
Unlike the Western experience in which voters’ alignment was “naturally” conditioned by social and economic transformation, the phenomenon of Golkar suggests that in Indonesia, the previous creative state’s political maneuvering, particularly under the authoritarian New Order regime, has in turn made the party take root in society. Additionally, the phenomenon of Golkar provides a valuable theoretical contribution suggesting that identifying the enduring cleavages cannot totally be reliant on a national-level study or survey. The fact is that the strength of cleavages may vary from one region to another. Thus, it is important for us to consider the lower levels, such as regions, districts or even villages. By narrowing the field of study, we can capture better the empirically existing cleavage structures.
Footnotes
Acknowledgements
We would like to express our sincere gratitude to Saiful Mujani, PhD, and Deni Irvani from the Saiful Mujani Research and Consulting (SMRC), Jakarta, Indonesia, for the data and valuable suggestions. We thank R-GIRO Ritsumeikan University and The University of Jember for kind supports. We also thank anonymous reviewers. We are solely responsible for the contents of the paper.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.