When the Indonesian Middle Class Takes to the Streets

The World Bank’s latest Indonesia Economic Prospects (June 2026) reached a compelling and unsettling conclusion: between 2018 and 2025, fewer workers earned middle-class livelihoods while more educated workers had gone into informal employment. International scholars continue to debate the definition and measurement of the middle class, but the World Bank’s findings are consistent with the first two authors’ research, based on a broad range of consumption and job quality indicators, which found similar evidence of Indonesia’s shrinking middle class. Both the World Bank and the authors use Indonesia’s official definition of the middle class: households with per capita expenditure between 3.5 and 17 times the national poverty line. This should provide some basis for some serious reflection on the part of the authorities. In addition to a shrinking middle class, Indonesia is also witnessing an increase in protests and riots on the streets, with the latest mass student protests taking place in mid-June. The Armed Conflict Location and Event Data Project (ACLED) recorded a sharp increase in the number of demonstrations at the district level in Indonesia in 2018, followed by continuous increases. This was followed by a spike in 2022 and another in 2024 (Figure 1). ‘Demonstrations’ in our analysis refer only to protests and riots, although the ACLED’s dataset includes other types of conflicts, such as military battles, explosions/remote violence, and violence against civilians.

Nexus Between the Middle Class and Protests

Figure 1: The number of demonstrations has increased sharply in Indonesia in 2018 and again in 2022

The data patterns suggest that rising demonstrations might derive from the shrinking middle class. The authors conducted a study to examine this hypothesis.

The middle class have been likened to “professional complainers”: they can be vociferous in their criticism about the government, whether on the street or in the digital realm. Using a nationally representative survey, Wihardja and Wibisana show that compared to other income-class groups, the middle class is more proactive in voicing their aspirations and less fearful of criticising the government. They also have greater political interest, engagement and expectation of the government compared to other income-class groups, according to the same survey. Using this argument, it is reasonable to assume a positive relationship between middle class shrinkage and the increased number of demonstrations taking place on the Indonesian streets.

However, empirical evidence is lacking to show this relationship. In his recent talk titled “Why Development Becomes Harder: The Political Economy of the Possible”, Basri shows a positive relationship between middle class shrinkage and social unrest intensity. He used ACLED data across emerging markets from 2019 to 2024. Countries that experience a larger decline in the share of the middle class tend to register higher levels of social unrest. This is one of the first studies conducted to bridge the gap in the literature. The authors’ study extends this hypothesis by using provincial-level data in Indonesia between 2015 and 2024.

At the provincial level, six provinces stand out in terms of increased numbers of demonstrations between 2015 and 2024: Jakarta, West Java, Central Java and East Java, North Sumatra and South Sulawesi (Figure 2). Across provinces and years, the average number of demonstrations in Indonesia was 52 demonstrations per year, with a maximum of 492 demonstrations (for Jakarta in 2024). Around 53 per cent of all provincial-level observations between 2015 and 2024 experienced a decline in the middle-class population, while around 58 per cent experienced a decline in the share of the middle class (out of the total population).

This lesson should resonate with other middle-income countries in region, especially Vietnam and the Philippines, which recently attained upper middle-income status.

The authors looked at the relationship between middle class shrinkage (both in terms of population and share declines) and the increase in the number of demonstrations in the provinces. The “unconditional” relationship considers the correlation between the middle class share and public demonstrations. However, various other factors, such as each province’s education levels, internet usage, urbanisation, and unique features of each year and province, can also affect public demonstrations. To isolate the effect of the middle class shrinkage on the number of demonstrations from these other factors, we compute the “conditional” relationship between the middle class share and public demonstrations controlling for these other factors. The results are consistent with the authors’ hypothesis that the middle class shrinkage increases the number of demonstrations in Indonesia.

Protest Prone Provinces

Figure 2: Demonstrations intensified the most in Jakarta, West Java, Central Java and East Java, North Sumatra and South Sulawesi

Figure 3 plots the change in middle class shares against the change in the number of demonstrations, with the fitted line representing the unconditional relationship. The link is sharp and consistent, moving from right to left: that is, as a province’s middle class thins, the number of demonstrations climb. Put differently, there are three more demonstrations for every percentage point of middle class share lost. This relationship holds up even after accounting for education, income, urbanisation, and internet use as well as unique features of each province and year: provinces whose middle class share shrank saw roughly 13 more demonstrations compared to those where the share held steady.

One twist stands out: provinces with wider internet access saw fewer demonstrations, not more. This echoes Basri’s cross-country finding across 29 Asian countries: when restrictive government censorship chokes off grumbling online, a shrinking middle class pushes discontent onto the streets instead. This finding may sound counterintuitive, as social media is often seen as an amplifier and transmitter of middle-class grievances. However, the fact that people are able to express their grievances online, such as by posting laughable memes or making excoriating comments, may actually prevent people from taking their grievances to the streets. Moreover, by channelling their grievances online, they can do it anonymously without the immediate risks of being beaten up or arrested on the streets.

Sobering Implications

Figure 3: A shrinking middle class goes with more protests across Indonesia (2015-2024)

Policy Implications

The authors’ study shows that in Indonesia, middle class shrinkage has serious political implications. Indonesia has had a stop-start relationship when it comes to attaining upper middle-income status: it entered the ranks in 2019, lost it momentarily in 2020 due to the Covid-19 pandemic but regained it in 2022.

Indonesia’s case is pertinent. In addition, having achieved upper middle-income status is as much about resilience as it is about social upward mobility. It is not only about elevating lower income groups into the ranks of the middle class, but also preventing the middle class from regressing. As Wihardja and Kong show, this resilience depends on three interlocking factors: job quality, access to high-quality public services and institutional trust. In Indonesia, a mismatch between rising expectations and limited fiscal capacity to improve public services is exacerbated by perceived inefficiency and corruption in public services (which compel middle-class households to turn to private services). This lesson should resonate with other middle-income countries in region, especially Vietnam and the Philippines, which recently attained upper middle-income status. Put differently, government policies should be able to sustain middle class expansion and their economic security.

2026/205