Brokers and the Roots of Polarization in Indonesia

• In Event: Comparative Approaches to Party Building (Virtual Session)

Abstract

How do parties with weak brands, limited machines, and no discernable ideology operating in a candidate-centric electoral system generate stable support from voters? I theorize that parties mobilize voters through distinct non-party brokerage networks. These networks differ from one another in their accountability to parties and capacity to deliver votes. Long-term trends in party success and voters’ political identity formation are the result of differences in the types of broker networks present in distinct locales.

To test this new theory of partisanship under brokered mobilization, I implement a parallel survey of candidates and brokers in Indonesia that uses a novel respondent-driven sampling design to measure the variety of brokers in an area, implement multiple measures of their quality, and test whether candidates identify brokers of high quality.

I find that vote brokers operate within separate mobilization networks: one religious, one secular. But where secular brokers interact with both religious and secular voters, religious brokers interact with only religious voters. The separate networks follow religious identity—not partisanship. Candidates’ access to voters is mediated by the type of broker network they can access. Religious brokers have smaller followings but are more likely to deliver. Secular brokers vary widely in both the size of their followings and their trustworthiness.

These results have two key implications for long-term political development in Indonesia. First, they explain how, despite two decades of chaotic and ideologically incoherent party coalitions, voters deliver about the same share of votes to religious and secular parties each year, even as the fortunes of specific parties vary. Second, they explain why pure patronage parties have fallen to the middle tier—their reliance on secular brokers means working with high-variance brokers.

More broadly, these results show that broker quality and broker type—variables that have not been part of the large comparative politics literature on clientelism—are important determinants of candidate success, and shape the long-term development of political identities by mediating voters’ exposure to political parties.

Author

• Seth Nathan Soderborg, Harvard University