With the rise of democratic regimes in many parts of the world over recent decades, scholarly attention has increasingly turned to assessing the quality of democracy (Kramon, 2013: 252). One of the important measures of the quality of democracy is free and competitive elections. A widely held view among scholars is that electoral clientelism has become a major impediment to such democratic elections. Indonesia is a perfect example. It has made important advances towards democratic consolidation with four consecutive national elections since 1999 and thousands of local elections since 2005. In reality, however, the development of democracy in Indonesia has been burdened by pervasive forms of patronage distribution, especially vote buying. This is especially the case since the introduction of the open-list proportional voting system that has opened a Pandora box of massive money politics within the democratic system. Consequently, the ill-suited design of electoral system has aggravated the accountability of democratic institutions and policy representation in Indonesia.

As a former student activist who took to the streets along with thousands of protesters demanding democratic reforms in 1998—which ended in the birth of a newly democratic regime—in making this study I have been in part motivated by normative concerns about the impact of vote buying on the quality of democracy in my country. More importantly, this study has also been driven by the strong impression I received when I was conducting 13 months’ fieldwork in Indonesia. I increasingly felt that the dominant literature on vote buying and turnout buying, which were developed in other settings, was simply insufficient to explain the ubiquity of vote buying in Indonesia. Much of the scholarship on vote buying is based on the Latin American experience. As a consequence, theories of vote buying assume that parties determine how money is distributed. In Indonesia, by contrast, individual candidates do the vote buying. In Indonesia’s open-list ballot system, intraparty competition is fierce, and candidates see party peers as their most proximate threat. Under these conditions, candidates are compelled to rely heavily on personal networks rather than party machines. Intense competition between a large number of candidates also means they only need a small share of the popular vote in order to win office. This differs from the situation in Latin America, where candidates and parties tend to pursue majorities or large constituencies. For all of these reasons, the established theories can offer only limited insight into the machinations of vote buying in Indonesia.

Under such different conditions, and in a context where candidates in Indonesia are largely insecure about their prospects of victory, I have argued in this book that candidates have clear incentives to use vote buying as a means of chasing a small margin of victory. In settings where elections are shaped by intraparty competition as a result of the open-list system, candidates busy themselves fighting on the ground against their internal party rivals in the hunt for personal votes. The desire to defeat their co-partisans makes them risk averse when selecting targets for their material rewards. As a result, most politicians and brokers say that their preferred targets are partisan, loyalist voters, which is strongly evocative of the core-voter strategy that has been devised in the Latin American context.

Yet, as I explained in Chap. 4, my voter survey showed that while such partisan voters are more likely to be targeted than non-partisan voters, in reality the vast majority of vote buying—in absolute terms—happens among undecided voters. This is particularly the case because the aggregate level of mass partisanship in Indonesia is relatively small. Only 15% of Indonesians feel close to a party. This limited number of partisans is also highly contested among internal rivals. Accordingly, my findings show that although politicians tending to target constituents who they think are truly loyal, most brokers end up distributing to voters who receive benefits but don’t reciprocate with votes. For various reasons, therefore, we cannot conclude that Indonesian electoral clientelism is based on the pursuit of party loyalist strategies.

In order to explain this combination of features, in Chap. 5, I offered an additional account to the scholarly literature on vote buying by combining the party loyalist model with a role for personal networks. I argued that in Indonesia, candidates and brokers actually intend to target partisan voters, but in reality they mostly distribute patronage to people who are connected to personal networks. Though they think of these people as ‘loyalists,’ in fact, they might lack any sense of loyalty to the candidate. I call this mixture a ‘personal loyalist’ approach. Though it acknowledges that candidates largely depend on personal networks to identify voters to target with vote buying, this approach doesn’t rule out the importance of party loyalists, seeing also a significant role for personalised partisan voters (i.e. those who possess a sense of loyalty to both the party and individual candidate within the party).

The personal loyalist approach is best explained in the context of open-list systems, as indicated earlier. Under such circumstances, the pressure to collect personal votes is intense. Candidates seek to personalise party constituents to defeat their internal party rivals. However, given only a tiny proportion of Indonesians are aligned with parties, such intraparty competition pushes candidates not to depend solely on party loyalists to win elections. To be sure, despite their limited numbers, such voters are also fought over among multiple candidates from the same party, but their numbers are limited. Accordingly, having personalised their party constituents, politicians seek to expand their electoral base and extend their vote-buying reach through personal connections mediated by non-party brokers. In the process, candidates frequently confuse personal loyalists as partisans, and misconstrue people with personal connections to their brokers as loyalists, too. At the same time, expanding electoral bases in this way can make vote buying susceptible to broker predation. Agency loss produces massive rent-seeking behaviours by brokers, making the problems of vote-buying distribution severe (see Chap. 6). It is often the case that brokers exaggerate the number of loyalists, even deceive their candidates on this issue, so that they can engage in predation. As a result, many of the people who are identified through personal networks mediated by brokers are in fact not even loyal to the candidate. These two factors in combination—confusion of personal connections with loyalty and agency loss—contribute to the large amount of targeting of uncommitted voters revealed by my study.

Interestingly, notwithstanding these factors, candidates still insist on spending a large amount of money on vote buying to pursue such spuriously loyalist voters. Recall that as many as a third of voters nationwide are exposed to the practice, making the aggregate level of vote buying in Indonesia the third highest in the world (Chap. 2), as indicated by voter surveys taken over the last decade. Yet these problems, plus other factors, clearly undermine the effect of vote buying on electoral behaviour. Offers of money ‘only’ influenced the vote choice of roughly 10% of the total electorate (see Chap. 7). Here lies the key to why vote buying remains so important: while this effect may appear small, in Indonesia’s highly competitive electoral landscape, that 10% matters immensely. Across the country, the average margin of victory by which winning candidates defeated their co-partisans was only 1.65% (with the winning margin here defined as the percentage of votes cast for a party in a constituency which separated the lowest-placed winner from the highest-placed loser on a party list). The 10% swayed by cash are more than enough to make a difference to electoral outcomes, both in the aggregate and in the case of many individual races. Most politicians, therefore, feel vote buying can play a decisive role in determining the electoral outcome that counts: whether or not they win a seat in the legislature. As a result, many pursued this strategy with enthusiasm.

Chasing a narrow margin victory, however, isn’t the only possible answer for why candidates still engage in vote buying, despite high levels of leakage. The prisoner’s dilemma offers a potential answer for the widespread practice of vote buying in Indonesia, even if such strategy is proven to produce low returns, and even if they aren’t in a close battle. Fearing that their opponents will distribute handouts, candidates might still find it in their interest to engage in vote buying (see Chap. 7). Every candidate might be better if no one bought votes, but if just one person defects and buys votes, then everyone else loses. When the fate of candidates relies on the others’ actions, vote buying is often seen as their best chance of stopping other candidates from winning votes, despite its inefficiency of delivery and wastage. This corresponds with the narrative among candidates and brokers that despite vote buying often appears to be objectively inefficient, such tactics still be relatively more efficient than other electoral strategies (Chap. 7). These three explanations—vote buying as a function of narrow victory margins, politicians end up facing a prisoner’s dilemma, and the perceived significance of vote buying in driving votes, relative to other electoral strategies—in combination contribute to explain how and why vote buying is so prevalent in Indonesia.

8.1 Theoretical Implications

I have argued that the personal loyalist strategy helps to explain patterns of vote buying in Indonesia. What are the primary theoretical implications of this personal loyalist approach and how do they contrast with predictions of existing theories? Table 8.1 provides a stylised summary of five major models of vote buying: my own personal loyalist model, plus the following four models: swing-voter model, core-voter model, informational model, and norms of reciprocity model.Footnote 1

Table 8.1 Different models of vote buying: Contrasting theoretical predictions
Full size table

The swing-voter logic predicts that uncommitted voters or weakly opposed voters would be the preferred target of campaign largesse in order to persuade them to vote for the benefactor party or candidate and change the game (Lindbeck and Weibull, 1987; Dixit and Londregan, 1996; Stokes, 2005). This is based on the underlying assumption that a loyal voter is already captive. In contrast, the core-voter hypothesis argues that parties and candidates tend to target their own party supporters because such voters are the most predictable source of votes. The rationales behind the ‘core-voter’ model vary, ranging from risk aversion on the part of candidates (Cox and McCubbins, 1986), mobilising lukewarm supporters for turnout (Nichter, 2008), the endogeneity of partisan loyalties to electoral handouts (Diaz-Cayeros et al., 2012) to broker predation (Stokes et al., 2013). Unlike the swing- and core-voter models whose key parameter in distributing benefits largely depends on voters’ partisan proximity to the machine or to its opponents, the norms of reciprocity model doesn’t require voters to have strong ideological attachments. Instead, according to this model, clientelistic practices produce a sense of moral obligation or indebtedness on the part of beneficiaries to vote for the distributing candidate in exchange for the reward (Finan and Schechter, 2012). Finally, drawing from much of the African elections, Kramon (2013) developed the informational model of vote buying, arguing that vote buying is a mechanism for politicians to establish credibility with voters regarding the distribution of patronage and private goods in the future. It serves primarily to convey information to voters that candidates are credible and able to provide future rewards.Footnote 2

Each line of Table 8.1 shows what theoretical preconditions or outcomes are to be expected by each of these models. The first group of ‘Contextual Factors’ sets out the features of electoral competition under which the model is assumed to be applicable. The second, ‘Nature of Vote Buying’ summarises what form vote buying takes in each model. Finally, the table summarises effects of vote buying in terms of both vote choice and turnout.

For simplicity’s sake, let’s focus here only the aspects of my personal loyalist approach that have the most notable theoretical implications. The first distinctive feature of my model concerns the political context, specifically the role of political parties. I have demonstrated from the outset that the existing theories of vote buying have emerged in contexts different from that in Indonesia. Table 8.1 makes these differences visible in a number of fields, such as party organisation, electoral system, party base, partisanship, and voting system. In particular, much of the extant literature on the swing- and core-voter models relies on the underlying assumption that party machines have the capacity to enforce vote-buying agreements, and that they engage in ongoing constituency service (Stokes, 2005; Kramon, 2013). In contrast, the other three models (i.e. the norms of reciprocity and informational models as well as my personal loyalist model) don’t require well-organised party organisations. In Kenya, as in many African countries where the informational theory was produced, the main proponent of the approach Eric Kramon (2013) shows that socially embedded and well-organised parties don’t exist. Similarly, the reciprocating model is built on the assumption that political parties are weak. Drawing from an excellent study conducted in Paraguay, the key theorists of the model Frederico Finan and Laura Schechter (2012) demonstrate that political parties in the country are weakly organised, not strongly ideologically oriented, and less embedded in society. Accordingly, Paraguayan politics tend to be extremely personalised (Rizova, 2007). My personal loyalist model arises in a similar setting. Vote buying in Indonesia, therefore, doesn’t rely upon strong party organisations but can also be mediated through a variety of informal networks. This study has provided evidence that institutional factors matter. The non-party organisation of vote buying in Indonesia is largely a product of the open-list system, which incentivises intraparty competition and prompts candidates to invest in building campaign teams that rely on personalised networks.

Still with regard to the role of parties, much of the literature on swing- and core-voter models broadly assumes that only one machine has the ability to engage in clientelist exchange (Nichter, 2010). This assumption may be justified in the context of contemporary Latin America, where mostly only one party can take advantage of the state resources and social networks required for clientelism (Stokes, 2009: 12; Nichter, 2010: 97). This, however, doesn’t really fit in Indonesia where there is cartelised party system, so that no party is locked out of state resources (Slater, 2004; Ambardi, 2009). Note that all parliamentarians independent of party affiliation have opportunities to access resources allocated for their constituencies (Farhan, 2016). More importantly, the expectation of the personal loyalist strategy is that multiple candidates from the same party or other parties compete against one another even in the same neighbourhoods or households. This runs contrary to the existing literature on vote buying, which assumes that each machine tends to cultivate separate networks with distinct constituencies (Stokes, 2005: 324; Gans-Morse et al., 2014: 17). In Indonesia, by contrast, as Aspinall and his collaborators (2017: 2) put it, “competing network machines often overlap within the same geographical locations and social milieus.” Many areas, especially in Java, had become free-for-all battlegrounds in which multiple candidates fight for personal votes in every village, neighbourhood, and laneway (Aspinall et al., 2017: 12). Multiple candidates from multiple parties also compete against each other to recruit brokers and determine their base areas. In sum, borrowing Stokes’ (2005: 324) words, these ‘duelling machines’ not only compete to purchase votes, but they also avidly bid up the price of votes to outbid rivals (see Chap. 6; also, Aspinall et al., 2017).

A second distinctive feature of the personal loyalist model concerns the purpose of the vote-buying transaction. The swing-voter model assumes that vote buying is an exchange of a reward for a vote choice: a voter receives money and in return votes for the giver. Conversely, according to the core-voter school, the payment isn’t to ‘buy’ a vote, but rather to mobilise supporters to turn out. In the norms of reciprocity model, the expectation is that such exchange serves either in a model of vote buying (persuading swing voters) or turnout buying (mobilising core voters). The informational theory views the transaction as a mechanism to convey a signal of candidate credibility with respect to future performance (Kramon, 2013). The personal loyalist strategy offers a slightly different story. In Indonesia’s extremely competitive election settings, where only the winner takes home the prize of office, vote buying serves as a means of providing a small margin of victory. Minor shifts in support whether as a result of buying lukewarm supporters for turnout (core-voter model), or purchasing the support of uncommitted voters (swing-voter model) can make a difference in the electoral outcome. Most politicians in Indonesia realise that vote buying doesn’t ensure victory in an election, but they do believe that it increases their chance of winning in a closely contested election. The dominant narrative among politicians is that if they don’t engage in vote buying and others do, they will certainly lose. Hence, although vote buying doesn’t always produce the vote that was hoped for, candidates still have incentives to pursue vote buying because in the context of fierce campaigns like Indonesia the value of each vote collected through such exchange can potentially make the difference between winning and losing.

A third point where my thesis about the personal loyalist strategy differs from other approaches is with regard to targeting strategies. Given that the primary purpose of vote buying under the swing-voter model is more to sway voters’ decisions than increase their turnout, the theory therefore predicts that politicians tend to target voters who are ideologically unattached or weakly opposed supporters. In contrast, the core-voter theory observes that politicians will try to target lukewarm supporters to persuade them to turn out on voting day. While the swing- and core-voter models employ an ideological test to explain who is or isn’t targeted by clientelist parties, the reciprocity model predicts that political machines target voters who are intrinsically reciprocal. According to this line of reasoning, machines will not distribute their largesse randomly across the electorate but will look at whether voters feel indebted to reward those who have helped them in the past (Greene and Lawson, 2012). By contrast, the informational model argues that the targeting of handouts will be relatively diffuse. Given that under such a model vote buying is a mechanism to signal credibility with respect to future rewards, cash distribution is therefore less targeted at specific types of voters (Kramon, 2013). In Kenya, where the literature on the informational model is based, distribution of cash and gifts can take place in public and one voter can receive multiple payments from multiple candidates, a strong indication that vote buying is distributed in a diffuse manner. An expectation of my own approach is that politicians and brokers express strong intentions to target constituents they think are truly loyal, but in reality, they mostly target uncommitted voters who will not always reciprocate with support. This is largely because the concept of loyalty in Indonesia is ambiguous and has multiple dimensions ranging from kinship, ethnic and religious ties, receipt of patronage to connection via brokerage networks. When candidates and brokers claim to be targeting partisan, loyalist voters, they don’t only rely exclusively on partisan loyalties but also judge the target in terms of personal networks. As a result, they misidentify non-partisans as partisans because they assume personal connections are partisan leanings. But the key is that they target persons who are connected by personal networks to their brokers, even if some of these persons, in fact, lack even a sense of personal loyalty to the candidate.

As with the informational model, my approach contrasts with the assumptions that underlie the two dominant theories of vote buying, which suggest that politicians distribute particularistic rewards in a highly targeted way to specific types of voters guided by the partisan preferences of the recipients to the machine or to its opponents (Diaz-Cayeros et al., 2012). Furthermore, both the swing- and core-voter models require ideological parties and ideological voters (Amick, 2016). In contrast, it is extremely difficult in Indonesia to target individuals based on their political preferences, not only because the population is becoming less attached to parties, with only one in every ten people in the country who feel close to a political party (see Chap. 4) but also because parties are becoming less ideological and are weakly rooted in society (see Chap. 1). Tomsa (2010) claims that most Indonesian parties are now presidentialist in essence and no longer represent sharply defined ideological constituencies. The current models of vote buying were developed in a setting where some parties possess very strong ‘base areas’—for example, working class areas in Argentina that have been voting for Peronists for decades, or the long-time rural voter base of the Party of the Institutionalised Revolution (PRI) in Mexico (see Chap. 5). A few parties in Indonesia, like PDI-P and PKB, have quite a strong tradition of having partisan voters; most don’t. All of this distinguishes Indonesia from the Latin American cases on which the literature is based and therefore explains why the targeting of vote buying is distinctive in Indonesia.

Finally, existing theories also differ with regard to the impact of vote buying on voting behaviour. Most existing models predict that the exchange will influence vote choice; only the core-voter approach differs in this regard, instead focusing on mobilising the turnout of passive loyalists. Although the dominant theories of vote buying admit the effectiveness of such a practice on vote choice, they have no consensus in the answer to why vote buying sways individual’s vote decision (Kramon, 2013: 68). The expectation of the swing-voter model is that despite targeting those who are ideologically unattached, machines are able to detect who has kept with the vote-buying bargain and enforce the deal (Stokes, 2005: 322). As noted above, in this view, parties have the capacity to monitor the recipients and ensure they reciprocate with votes because they are bottom-heavy and socially embedded in local communities (Stokes, 2005). In the reciprocity model, vote buying is effective because it is a moment of retrospective evaluations of candidates, when voters feel a moral obligation to vote only for those who have provided them with rewards (Kramon, 2013; Greene and Lawson, 2012) and because of their hopes of maintaining close social relations and receiving future rewards (Finan and Schechter, 2012). Meanwhile, the informational model has little to do with retrospective evaluations (Kramon, 2013). Instead, the framework largely depends on prospective expectations of politicians, where voters expect to receive patronage goods in the future.

Overall, the dominant models assume that cash handouts will affect vote choice and motivate the recipients to turn out on voting day. Though my study doesn’t directly address this debate, it does take up one issue which has been largely neglected by the scholars who have engaged in it. Thus, scholars proposing these approaches have largely neglected the critical issue of the magnitude of the effect on vote choice and turnout: how big is the effect of cash gifts on both vote choice and turnout and how should we interpret it? The conventional literature on vote buying generally fails to quantify the effect, assuming that vote buying will automatically result in higher turnout or vote share. The expectation of the personal loyalist strategy is that the impact of vote-buying exchanges may appear insignificant in numerical terms but could still be influential in determining electoral outcomes. Recall that the meaning of the effect of vote buying, whether small, medium, or large, varies by context. A 10% effect on voting decisions might be small in some contexts, but this number will be sufficiently high to clinch a victory by a vote buyer in many highly competitive election settings, as in Indonesia. This is similar to how advertisements work. They might be unable to make all viewers buy the product being advertised, but at least some people get interested, translating into increased sales. Likewise, vote buying may be ‘ineffective’ in yielding significant votes to the extent anticipated by the buyer, but the votes that are flipped may be more than enough to make him or her win.

8.2 Policy Implications

Having discussed some theoretical dimensions of this study, we are now in a position to review the policy implications. This book clearly has implications for democratic accountability, in general, and reform of electoral institutions, in particular. In Chap. 1, we began the research by delving into an explanation of institutional arrangements in the post-Suharto era, seeing these as shaping the supply side of vote buying. Changing electoral rules clearly shapes the propensity of candidates to engage in vote buying. Therefore, it is appropriate for this study also to offer ‘supply-side’ remedies to discourage parties or candidates from adopting vote-buying strategies.

This study found that patronage distribution has been central to election campaigns in Indonesia. The results demonstrate that most candidates pursue vote buying because they see this strategy as affecting the outcomes of many competitive elections, where small changes in the voting calculus can alter the final results. An obvious conclusion is that the open-list system, in which a small number of personal votes are expensive yet critical for politicians, has been responsible for the growing prominence of cash handouts during political campaigns in Indonesia. As discussed in Chap. 1, vote buying was unheard of in Indonesia’s first free election after the fall of Suharto in 1999 when Indonesia adopted a fully closed-list system (where voters can only vote for parties and don’t have power over the order in which party candidates are elected). The practice of vote buying began to flourish in the 2009 election, after the introduction of a fully open-list system in that election. The fact that each candidate not only competes with candidates from other parties but also with other candidates from his or her own party has exacerbated the practice.

Although this list isn’t exhaustive, we can speculate that the open-list system has three far-reaching implications for Indonesia’s political system as a whole. First, the open-list system obviously makes elections more candidate-centric because they create incentives for the cultivation of a personal vote. The results in Chap. 1 show that official statistics of parliamentary elections over the period of 2004–2014 exhibit a clear linear trend towards an increase in the share of personal votes cast and a decline of party votes. The finding from a series of voter surveys conducted by my polling institute, the Indikator, in 56 electoral districts in February 2014 with a total of 43,510 respondents also shows that candidates have become more important. Likewise, a voter survey conducted by LSI and La Trobe University in North Sulawesi and Maluku in October 2012 suggested that voters were more likely to vote based on specific candidate attributes than on the party affiliation of the candidates.Footnote 3 This corresponds with the results as presented in Chap. 6 that the candidate’s qualities and personal reputation are far more important than the programme of the party for which the candidate runs. The increasing role of candidates should be read in the context of the open-list PR system with its emphasis on intraparty competition (see Chap. 7). Under such circumstances, candidates are forced to compete against each other and differentiate themselves from their internal party rivals, including by vote buying. Accordingly, they need to run well-structured but expensive success teams and expend money on cash payments, club goods, and other handouts.

Second, it follows that given that the open-list elections create incentives for candidates to pursue clientelist strategies, money has become the most important foundation of political success. Indeed, more money doesn’t guarantee victory, but it does increase the chance of it (Aspinall et al., 2015). This was very evident during the 2014 campaign, which most candidates dubbed as the most ‘brutal’ election in Indonesian history. Zuhairi Misrawi, a DPR candidate from PDI-P with NU background jokingly put it, “there is a new Islamic jurisprudential maxim in politics: ‘Al-fulus tuhyin nufus, ma fi fulus manfus’ (money will extend your life. If you don‘t have money you will die politically)” (Informal Communication, 2 July 2016). As Indonesia has moved away from party-centred to candidate-centred campaigns, candidates themselves have had to engage in costly mobilisation efforts, including running advertisements, commissioning surveys, building campaigns, mobilising constituencies, and buying votes. The popular view is that given such massive costs, only better-resourced candidates can do well in elections. These expensive electoral processes, coupled with the fact that most parties are less ideological, mean that parties are open to nominating candidates from outside the ranks of their own cadres. The Forum of Citizens Concerned about the Indonesian Legislature (FORMAPPI) revealed that in the 2014 election, only 33% of the candidates could be classified as party cadres. Almost half of the total candidates (3241 out of 6607) running for national parliamentary seats had business backgrounds, with many joining the parties just a few months before the election (FORMAPPI, 2013).Footnote 4 The entry of such business people has come at the expense of candidates with backgrounds in political and social activism, who are considered to be unable to fund their own campaigns (Budiman Sudjatmiko, Interview, 29 April 2014). Such findings correspond with those of a study from Centre of Political Studies, University of Indonesia (PUSKAPOL UI) that documented two important aspects of the backgrounds of those elected in 2014. About 58.86% of MPs had backgrounds in business or the private sector, or they were entrepreneurs or professionals,Footnote 5 suggesting that money is important for winning elections. PUSKAPOL also revealed that 77 out of 560 elected candidates came from wealthy political dynasties (Republika, 9 October 2014). Around seven out of them were among the top ten candidates receiving the biggest share of votes in 2014.

Third, by contributing to more candidate-centred elections, the open-list system has also jeopardised the relationship between voters and parties, making party cues less important and directing voters towards the short-term appeal of candidates. This electoral system has also contributed to the rapid decline of party loyalty (Chap. 5) and diminution of the image of parties in the public eye (Chap. 6). Survey data show that the number of Indonesians who feel close to a political party has declined significantly from 86% in 1999 to a mere 10.1% in February 2019. This number is comparatively low, making Indonesia a country with one of the lowest partisanship levels in the world. Interestingly, the level of party identification started collapsing in 2004 (see Chap. 4), when the country introduced the semi-open proportional system, and for the first time, ballot papers featured candidate names as well as party logos, allowing voters to vote for a particular candidate rather than just for a party. This parallel development of declining party loyalty and the adoption of candidate-centred elections has made it difficult for parties to mobilise voters on the basis of programmatic campaigns and policy positions.

Accordingly, this study recommends changes in institutional design, law enforcement, and voters’ education in order to reduce the practice of vote buying:

  1. 1.

    Changes in the electoral system from open-list proportional system to closed-list proportional system. In order to reduce vote buying and move towards more programmatic politics, Indonesia needs to also move towards a more party-based system of electoral competition. This can be made possible by revising the current electoral system and returning to closed-list proportional representation. Shifting the arena of competition from intraparty competition between individual candidates, to interparty contests, should reduce the need to generate personal votes through vote buying. More importantly, in closed-list multimember districts, voters choose among parties, and the rank order of the candidates in the party list is determined by the party. Given that citizens aren’t allowed to express a preference for any particular candidate within each list, this type of electoral system tends to produce party-centric elections. In such a system, voters should increasingly turn their attention to party policies rather than personalities, enhancing party cohesion, reducing internal disputes, and centralising party leadership (Suwarso, 2016). As Norris (2006: 105) argues, closed-list elections “encourage politicians to offer programmatic benefits, focused on the collective record and program of their party, and to strengthen cohesive and disciplined parliamentary parties.” In an environment where voters have less choice to determine the fate of individual candidates, and where campaigns are more focused on party platforms than on personal reputations and connections, we can expect a reduction in the importance of money in determining electoral outcomes.

    Timor-Leste is an interesting example of how the closed proportional system reduces the incentives of parties or candidates to use bribery tactics in elections. Although Timor-Leste’s voter-side demand for vote buying is high, where 32.7% of citizens consider such strategies are acceptable in the 2017 legislative elections, the incidence of vote buying in that country was considered as very rare occurrences. Only 4% of voters claimed to be targeted by money politics.Footnote 6 This happened because the closed proportional system used in Timor-Leste has made political parties the main players during the election. As a result, the supply side of vote buying is limited by design. Of course, the closed proportional system isn’t without its flaws. In the 1999 elections in Indonesia, this system was proven to encourage elite oligarchs. Additionally, there were numerous reports of wealthy candidates purchasing winnable slots on party lists by bribing party leaders. But even accounting for this weakness, the diminished incentive for widespread vote buying in a closed-list system makes it preferable to the easily exploitable status quo. The closed proportional system is, at least, able to isolate and confine the incidence of vote buying at the party elite level instead of at the grassroots voter level, increasing the role of the party in electoral battles which ultimately leads to increase in the institutionalisation of political parties. In order to reduce elite oligarchs, a closed proportional system can be accompanied by primary elections or conventions within the party to screen the respective legislative candidates.

  2. 2.

    The single-member district electoral system can also be considered as an option. Studies have shown that these single-member districts would encourage political accountability to larger constituencies and are proven to reduce the practices of vote buying and pork-barrel politics (e.g. Maisrikrod, 2002: 196). This happens because only one representative is elected in one electoral district so that incentives to do vote buying are greatly diminished. The district system would also involve the winner takes all electoral dynamic where the candidate who gets the most votes represents the electoral district without taking into account the difference in vote tallies. Because in the district system only one representative is elected in one electoral district, the incentive for candidates to use money politics, which commonly materialised within multimember constituency system, would be highly reduced.

  3. 3.

    If an open proportional system is maintained, then it is necessary to have an electoral redesign that allows this system to become a disincentive for candidates to buy votes. One of them is by reducing the district magnitude. In an electoral district where district magnitude is high, candidates need only few votes to be entitled to get a seat, and they can therefore engage in vote buying and win (Carey and Shugart, 1995; Chang, 2005). In contrast, the fewer the seats that are contested, the less incentive there is for candidates to buy votes. The reason is because there will be more votes needed to secure a seat and less number of political actors who would battle for the remaining seats in higher magnitude electoral districts due to the shrinking district magnitude.

  4. 4.

    Whatever reform is undertaken, vote-buying strategies will run rampant if there is no law enforcement. Based on the dozens of DPR (national legislature)-level candidates I interviewed, most of them stated that this practice was rife because they believed there were no strict sanctions for the perpetrators. Thus, strict and severe sanctions are required as a deterrent effect to the perpetrators and also to give a signal to the candidates that they will face serious repercussion when conducting money politics.

  5. 5.

    Given the high level of voter-side demand for vote buying, it is necessary to systematically educate voters so that citizens are much less tolerant of this practice. KPU (Election Commission Agency), BAWASLU (Election Supervisory Body), mass media, and NGOs must work together to carry out voter education about the dangers of vote buying, especially the potential for perverse accountability (Stokes, 2005). In democracy, people have the right to hold politicians accountable. But if politicians have purchased their votes, then the voters are actually held accountable because they have exchanged their democratic mandates at cheap value.

Overall, this book has argued persuasively that electoral systems matter in explaining the ubiquity of vote buying. I have shown that Indonesia’s open-list voting system is creating a situation where one-third of the electorate engages in vote buying. Accordingly, in the context of transition stages towards a mature democracy, one has to cautiously attend the party and electoral system. Especially within new democratic countries, the political parties are still in a generally weak and vulnerable state. If there is some incongruous predicament between the party or electoral system and the current condition, it would only result in long-term damage towards the political system, that is, sluggish transition and feeble party institutionalisation, as well as unbridled money politics.

8.3 Concluding Remarks

In the last part of this book, I highlight my contribution to the study of vote buying and how it might make a difference. Overall, this study makes four contributions to the comparative literature on vote buying. First, in terms of methodology, this study has used multiple methodologies by combining quantitative and qualitative approaches to determine the patterns of vote buying in Indonesia. Most previous works on clientelist exchanges are either purely ethnographic or they rely heavily on survey data (Kramon, 2013: 253). Focusing solely on one type of approach is susceptible to methodological problems. The general challenge of the former is to establish the extent of clientelism in a population and to determine causal inferences. The methodological problem of the latter is in understanding a phenomenon or the context in which the data are collected. Mixed methods of the sort I used in this study can offset these weaknesses and have allowed me to develop a broader perspective of vote buying in Indonesia. My surveys and qualitative research allow me to determine the scope of vote buying, trace its causal mechanisms, test novel hypotheses, provide interpretations in a meaningful way, and adjudicate between claims and debates in the existing literature. In particular, by relying on voter surveys, my study has demonstrated that the level of vote buying in Indonesia is high by international standards, and such practice is effective in producing slightly—but electorally consequential—higher rates of turnout and vote share that are enough for most candidates to secure victory. No other methodological approach can so systematically establish the extent and the effectiveness of vote buying. In addition, supported by a list-experiment, my study has also showed that direct survey items about vote buying in Indonesia aren’t subject to response bias. Moreover, unlike much of the literature on vote buying which largely relies on voter surveys, my analysis isn’t based solely on the demand side of vote buying. By utilising broker and politician surveys, my study also puts emphasis on the supply side in order to understand vote buying from the perspective of the actors who orchestrate it.

Second, by embedding my quantitative findings in rich empirical findings drawn from my qualitative fieldwork, I have been able to demonstrate that investigating the political context in which vote buying operates is crucial to really understanding how the mechanism work in practice. Much scholarly theorising on vote buying is based on empirical observations drawn from several Latin American countries. In these countries, as already noted above, the organisation of vote buying is party-based, and political parties aren’t only socially embedded, but they are also well organised. The Indonesian case, however, is very different. This study has explained the prevalence of vote buying in the Indonesian setting where—partly as a result of the personalised electoral system and intense intraparty competition that occurs there—candidates rely on personal networks rather than their party, where partisan ties are weak, and where personalised loyalties matter far more. I have showed that the meaning of ‘partisanship’ in the framework of the Indonesian context—where a lot of connections between voters and candidates aren’t mediated by political parties but by informal brokerage networks—is far more nuanced than we might expect. One may relate the findings from Indonesia to neighbouring countries such as the Philippines and Thailand, where much of the literature suggests that political parties are generally weak and where vote-buying practices are prevalent. Julio Teehankee, for instance, shows that after the Second World War and long period of dictatorship under Ferdinand Marcos, political parties in the Philippines “suffered from weak internal organisation, structure, and discipline, which resulted in weak party loyalties and constant party-swiching” (2012: 190). Schaffer (2007: 3) also shows that in the Philippines, clientelist practices have also become a central feature of electoral politics. However, one major point of divergence is that constituency service in the Philippines plays a greater role in the interplay between parties and voters at the grassroots than in Indonesia (Berenschot, 2015: 560). Additionally, Filipino voters were more likely to be attached to parties with 42% feeling close to political parties in 2005, compared to Indonesian with only 15% of the electorate who did so in 2014 (see Chap. 5). Similarly, despite sharing a lot of similarities with Indonesia with regard to the widespread occurrence of vote buying (Hicken, 2002; Callahan, 2005), weak party organisation, and the use of personal campaign strategies (Hicken, 2002), political parties in Thailand are more likely to engage in constituency service than their counterparts from Indonesia. In her excellent ethnographic research in Thailand, Bjarnegård (2009: 123) shows that although they show little interest in promoting party policy, parties in the country actively arrange drainage, roads, electricity, and admission in school for their party supporters, suggesting that parties largely take care of the needs of community which typically happens beyond elections (Berenschot, 2015). In Indonesia, by contrast, individual candidates present themselves as caring for the interests of the community and offer ‘concrete’ benefits to their constituents, especially during election time. This greater role of individual candidates makes vote buying in Indonesia more personal (i.e. non-party-based) and might distinguish it from the existing literature, especially that which stems from Latin American cases.

Third, the fine-grained analysis of this study also contributes a more detailed understanding of how targeting of vote buying in Indonesia works, arguing that empirical evidence reviewed in this project fits differently to the pathway of the lively debate in the literature between proponents of the core- and swing-voter models. My research design, which primarily relies on tracking individual voter surveys, allows us to conclude that although there is a greater likelihood of party loyalists being targeted, the absolute majority of vote buying happens among non-partisans, given the relatively small number of party loyalists in Indonesia. In fact, my rich probability sample of low-level politicians and brokers, and in-depth interviews with national politicians, provide evidence strongly evocative of the core-voter argument. I have highlighted three pieces of empirical evidences explaining the gap between the politicians’ intentions to focus on party loyalists and the fact that in total numbers, swing voters are more targeted than loyal supporters. First, candidates and brokers tend to exaggerate the number of partisan voters. Second, loyalty is an amorphous concept and has multiple dimensions in Indonesia. Third, agency loss results in both unreliable brokers and unreliable voters, confirming the classic problem of vote buying as an uncertain business. These are the points for departure for the personal loyalist approach I put forward. Despite candidates and brokers’ claims that they were targeting loyal voters, it turned out that they were often providing benefits to basically uncommitted voters.

Lastly, my study has endeavoured to advance our understanding of the logic and motivations behind candidates’ insistence on pursuing vote buying, regardless of the fact that the targeting of vote buying is so misdirected, and there are principal-agent problems among politicians and brokers, and between brokers and their voters. I have shown throughout this study that despite such intrinsic problems of clientelist exchange, vote buying remains an attractive investment and has been widely practised in Indonesia. This is largely because candidates believe that the voting decisions of a sufficient proportion of voters can be swayed by rewards. Even if the overall number of such voters is small, it is often more than enough to provide a vote buyer with a narrow winning margin. In addition, applying the prisoner’s dilemma to interactions between voters and party machines, the presence of other candidates doing vote buying to win elections may explain why they insist to distribute handouts, even if these actually don’t purchase vote. Further, much of the emerging conventional wisdom among candidates is that vote buying is a more reliable strategy in winning votes, relative to other mobilisation tactics. By emphasising the role of vote buying as a mechanism to provide a small margin of victory, and by understanding the electoral motives behind candidates’ distribution of rewards, we can better understand how and why vote buying has become so rampant in Indonesia.