ABSTRACT
In Western democracies, decentralization is typically associated with pluralism and demands for minority rights. In other contexts, however, decentralized governance may be instrumental to conservative and exclusionary ideologies. We illustrate this point with an analysis of Indonesia, a diverse and decentralized country. By leveraging an original survey, we find that Islamist individuals are significantly more likely to support decentralization than pluralists. This can be attributed to this country’s legacy of political development. As pluralist elites have long dominated national politics, political Islam has sought to empower regional governments, where Islamist agendas can more easily be implemented.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 To be sure, Indonesian history also presents instances of separatist movements and calls for greater autonomy emerging in Christian-majority areas, such as Papua, Timor Leste and North Sulawesi.
2 Examples of these local regulations include provisions requiring certain dress codes for women, Koran-reading exams for aspiring public servants, mandatory alms for Islamic charities and bans of prostitution, gambling, or the sale and consumption of alcoholic beverages.
3 For an exception, see [redacted for anonymity].
4 The World Bank (International Telecommunication Union World Telecommunication/ICT Indicators Database, accessed April 21, 2021) estimates that 48% of the Indonesian population had access to the internet in 2019.
5 Despite this simplified version, 17.9% of respondents declined to answer it, which indicates that the question may still be relatively complex for some respondents. In the appendix (), we investigate if nonresponse to this question is related to our dependent variable. The estimations show that political ideology (preferences over political Islam) is unrelated to nonresponse, which is only significantly associated with two other variables (satisfaction with democracy and rural residence).
6 These questions were originally asked with a 4-point scale, but given the very low numbers of respondents who choose the ‘not close at all’ options, these responses were recoded into the same category as ‘not very close’.
7 While these measures have been originally developed in the context of Western societies and mostly applied to the United States and Western Europe, 4-point Likert scales such as those used here have also been used to study social identities in low- and middle-income countries and regions, including Eastern Europe (Cebotari Citation2016), China (Sinkkonen Citation2013) and Papua New Guinea (Feeny, Leach, and Scambary Citation2012). The latest iteration of the ISSP National Identity Survey in 2013, which includes these questions, also covers several non-Western countries, such as Mexico, South Korea, India, South Africa and the Philippines. We are therefore confident that this methodological approach is suitable to a variety of empirical contexts, including the Indonesian one.
8 Responses to these questions are positively correlated, which is consistent with our characterization of the Indonesian case as a country in which regional and ethnic identities have not generally developed in opposition to national identity. The appendix reports a full correlation matrix for all variables used in this study ().
9 Both satisfaction with the incumbent president and with democracy were originally measured with a question asking if respondents were not at all satisfied, not very satisfied, somewhat satisfied, or very satisfied. Data was then aggregated to produce a binary indicator.
10 The relationship between identity and support for decentralization, however, may be contingent on the specific region being studied, given that some provinces (such as Aceh and Papua) may have stronger regional identities than others.