Beyond the Symbolism: Organisational Networks and the Conditional Electoral Power of Nahdlatul Ulama

INTRODUCTION

Nahdlatul Ulama (NU) has shaped Indonesian politics for decades. Its mass base — concentrated in East Java and Central Java but stretching across the archipelago — makes it one of the most coveted constituencies in any national election. Since Reformasi, parties and presidential candidates alike have competed to claim NU’s backing, often treating it as an electoral prize waiting to be unlocked.

Despite its enormous membership, Nahdliyin (members and affiliates of NU) have consistently split their support across multiple parties and across Presidential candidates with very different ideological profiles.-

But why does NU’s numerical strength so rarely produce electoral consolidation?

The 2024 election deepened this puzzle. PKB, the party most directly associated with NU, remained a significant actor in parliament. But in the presidential race, Nahdliyin preferences fragmented dramatically, with support shifting between candidates in ways that had little obvious connection to NU’s formal positions. NU as an organisation does not issue formal presidential endorsements the way a political party does; this is consistent with its khittah — the 1926 doctrine, reaffirmed at the 1984 Situbondo congress, under which NU formally withdrew from practical politics. Anies Baswedan–Muhaimin Iskandar (PKB’s chairman) failed to win over NU voters decisively, while Prabowo Subianto–Gibran Rakabuming, who had no ideological connection to NU, ended up winning broad support across the NU electorate, largely through the mobilising power of then-President Joko Widodo.

Making sense of these patterns requires moving beyond the idea — implicit in the aliran-based accounts that will be discussed later — of NU as a monolithic bloc and distinguishing between two different forms of NU affiliation: cultural proximity (exercised by those who identify as NU or see themselves as traditionalist santri) and active organisational membership (exercised by those who participate in NU institutions, its autonomous bodies and its networks).

Following convention in the literature on religious voting, I use affiliation as the broader, umbrella term for any form of attachment to NU — encompassing both cultural proximity and active membership – while membership refers specifically to formal or active participation in NU’s organisational structures. All members are therefore affiliates, but not all affiliates are members; the two terms are not interchangeable.

Cultural proximity may predict electoral behaviour, but active membership is generally the stronger and more consistent predictor. NU’s electoral influence runs primarily through its organisational machinery rather than through symbolic identity alone. ‘Elites’ are defined as the top tier of NU and NU-affiliated political leadership — principally the Rais Aam and Chairman of NU’s Central Executive Board (PBNU), prominent national and regional kiai, and PKB’s party leadership — whose public endorsements function as cues for the broader network.

NU’s electoral effects are strongest where organisational ties are dense and elite cues are coherent, and weakest where affiliation is purely symbolic.

SITUATING THE DEBATE

Geertz’s tripartite division of Javanese society into santri (exemplified Islamic piety), abangan (foregrounded animist elements in the Javanese syncretic mix), and priyayi (aristocrats) segments is mapped onto distinct political orientations, with NU representing the traditionalist santri stream. After Reformasi, NU’s political expression found a natural home in PKB, which won significant NU-affiliated support in the 1999 and 2004 elections. Two broad approaches have shaped how scholars understand NU’s electoral role. The first — rooted in Clifford Geertz’s aliran framework — holds that socio-religious cleavages in Indonesia are relatively stable and continue to structure political behaviour.

The second approach challenges how much of this survives today. Liddle and Mujani, among others, documented a shift in Indonesian voting behaviour toward candidate-centred, performance-based voting — a dealignment (dealiranisasi) from traditional social identities. Presidential elections since 2004 have accelerated this trend, putting individual candidates front and centre and weakening parties as primary political anchors. For NU voters, this means the connection between religious identity and the ballot has grown looser and more contingent; split-ticket voting has become common.

Both perspectives miss a crucial point: “NU” is not homogeneous. There is a difference between those who attend tahlilan gatherings and privately consider themselves NU, and someone who sits on a local NU committee, participates in Muslimat NU events, and receives organisational communications during campaign season. Both forms of attachment carry some network character, but they differ sharply in density and formality: the first is rooted in loose, informal communal practice with little organisational infrastructure, while the second is embedded in a dense, formal network built to carry information, signals, and social pressure. These two types of NU attachment produce different electoral effects and conflating them has resulted in much of the confusion in the literature. These two forms of attachment are not necessarily discrete: active members are, almost by definition, also culturally proximate, so the categories are better understood as nested rather than mutually exclusive. Which form is more electorally decisive likely depends on context — network density, the coherence of elite cues, and the competitiveness of the race itself.

DATA AND METHOD

My analysis draws on two main data sources. The first is a large-scale provincial survey conducted across East Java (N = 10,800) and Central Java (N = 10,100) in December 2023, using multistage random sampling and face-to-face interviews. The margin of error for each provincial survey is approximately ±1.1 per cent at the 95 per cent confidence level. The second is a national survey covering all provinces with oversampling in 18 provinces (N = 5,500), conducted in late January to early February 2024, about two weeks before election day on 14 February 2024.

NU affiliation was measured in several ways. In the provincial surveys, I asked respondents about their Islamic mass-organisation affiliation, their self-identification as traditionalist santri or modernist santri (or priyayi/abangan), and the frequency with which they practise rituals associated with NU qunut (standing in worship) in Shubuh prayers, tahlilan/yasinan, grave visitation (ziarah kubur), and attending haul commemorations of prominent kiai (clerics). These three dimensions — organisational affiliation, cultural self-identification, and ritual practice — give a rich picture of what being “NU” means on the ground.

In the national survey, I focus on two key variables: cultural proximity to NU (whether respondents feel close to NU) and active NU membership (whether they participate in NU or its autonomous bodies). Both measures are self-reported: cultural proximity captures respondents’ own subjective sense of closeness to NU, while active membership captures self-reported participation rather than independently verified organisational records. Active membership, being anchored in concrete, recallable behaviours (attending meetings, holding a position), is likely less prone to self-report bias than cultural proximity, which is inherently a matter of subjective self-identification.

The provincial survey data, collected two months before the vote, are used for descriptive profiling rather than inferential modelling. For each electoral outcome, I run a multivariate logistic regression model controlling for gender, age, ethnicity, education, urban/rural location, and approval of President Joko Widodo.

KEY FINDINGS

NU in Its Heartland: East Java and Central Java

NU dominates its two core provinces: 78 per cent of Muslim respondents in East and Central Java identified with NU, compared with 54.6 per cent nationally and just 4.4 per cent with Muhammadiyah. This figure comes from a direct self-identification question, rather than a composite measure; its relationship to ritual practice is examined below. Overall, 65.7 per cent identified as santri, with traditionalist santri — the group closest to NU’s self-understanding — the largest subgroup in the two provinces.

Traditionalist santri have a distinctive socio-demographic profile. They tend to be older, less educated, have lower incomes, and are concentrated in rural areas — classic wong cilik (common people or grassroots). Modernist santri, by contrast, skew towards younger, more urban, more educated, and more likely to hold white-collar or professional jobs. These are not just religious distinctions; they reflect different social strata with different political implications.

NU’s hold also shows up in ritual practice. When I constructed an index from the four ritual measures mentioned above — qunut (standing in worship) during Shubuh prayers, tahlilan/yasinan, grave visitation (ziarah kubur), and attending haulcommemorations of prominent kiai (clerics) — around 55.5 per cent of Muslim respondents in the two provinces scored high enough to be classified as “NU” on practice-based grounds, more than half of the local Muslim population not only claims NU identity but actively lives it.

Perhaps the most practical implication is that parties and candidates who want NU votes cannot rely on claiming NU identity from the outside. What matters is whether they can engage NU’s actual organisational infrastructure, which figures they can bring on board, which networks they can activate, and whether they can sustain consistent messaging through those channels all the way to election day.

National Patterns: Cultural Proximity, Membership, and Fragmentation

Nationally, 54.6 per cent of Muslim respondents claimed some form of closeness to NU. But cultural proximity does not mean organisational engagement. Only about 16 per cent of Muslim respondents nationally reported being active members of NU or its affiliated bodies. Another 17 per cent said they were members but were not active. Nearly two-thirds had no formal NU membership at all.

This gap between cultural proximity and organisational engagement matters enormously for electoral politics. The 16 per cent who are active members can potentially be mobilised through NU’s networks; the much larger pool of cultural identifiers is far more loosely tethered and is responsive to other influences.

Tracking vote intentions among NU-affiliated respondents over the campaign period reveals a striking shift. Earlier in the campaign, Ganjar–Mahfud led among NU respondents, reflecting Ganjar’s positioning as Jokowi’s political heir. But as election day neared, support shifted sharply and consistently toward Prabowo–Gibran, a pattern consistent with the “Jokowi effect”, namely the president’s active endorsement of Prabowo, which carried particular weight among voters who had previously followed Jokowi’s lead. Note that Jokowi was popular among NU voters. Anies–Muhaimin never gained traction despite Muhaimin’s PKB leadership, and Mahfud MD’s NU credentials as Ganjar’s running mate were not enough to stem the migration.

What Cultural Proximity Actually Predicts

The regression results for NU cultural proximity show real but uneven effects across parties (Figure 1). The two clearest positive signals are PKB (β ≈ 0.8) and PPP (β ≈ 0.94), both with confidence intervals sitting entirely to the right of zero. PDIP and Golkar show positive but modest coefficients (β ≈ 0.2–0.3), with confidence intervals that are barely significant. On the negative side, PKS is the sharpest case (β ≈ −0.6, CI entirely left of zero), reflecting the well-documented ideological distance between NU’s traditionalism and PKS’s scripturalism. NasDem, PAN, and Demokrat cluster similarly (β ≈ −0.4 to −0.5), while Gerindra sits just left of zero with a CI straddling null — too uncertain to draw firm conclusions.

Figure 1: Effect of Feeling Part of NU on Party Choice

For presidential choice (Figure 2), Ganjar–Mahfud shows the only clearly significant positive effect (β ≈ 0.41, CI fully right of zero), reflecting Ganjar’s regional base in Central Java and East Java — a historic NU stronghold with longstanding ties to PDIP — rather than any NU-specific credential carried by Mahfud himself. Anies–Muhaimin is negative (β ≈ −0.25, CI mostly left of zero), suggesting NU-identifying voters were actually less inclined toward this ticket despite Muhaimin’s credentials, likely reflecting high NU satisfaction with Jokowi and their reluctance to accept Anies’s change-oriented campaign narrative. Prabowo–Gibran sits near zero with a CI straddling null, confirming that Prabowo’s NU support was not rooted in NU identity but was driven by Jokowi’s endorsement.

Figure 2: Effect of Feeling Part of NU on Presidential Ticket Choice

Why Active Membership Often Has Stronger Effects

Active NU membership tells a sharper and more polarised story throughout (Figures 3 and 4), though not uniformly so: PPP is a notable exception, discussed below. For party choice, PKB again produces the strongest signal (β ≈ 0.75), with tight estimates and entirely positive. Golkar shows a clear positive effect (β ≈ 0.45) that was barely visible in the cultural proximity model. I interpret this as being consistent with a pragmatic, network-brokerage connection — plausible given Golkar’s patronage-based organisational character and the likelihood that active members occupy brokerage roles within NU’s own networks — rather than an identity-based one, though this mechanism is not tested directly here and remains a question for future research.

PPP remains positive (β ≈ 0.35), while PDIP’s coefficient is too imprecisely estimated to confidently claim it differs from zero. This imprecision is itself informative: unlike PPP or PKS, PDIP’s pull on NU voters does not appear to run through individual-level cultural or organisational attachment at all — consistent with my earlier finding that its NU-linked support (via Ganjar–Mahfud) is better explained by regional and historical alignment in Central Java than by individual attachment to NU.

PPP’s weaker showing here compared to the cultural-proximity model (β ≈ 0.94 to β ≈ 0.35) is interesting. It suggests that cultural identification with NU — untethered from organisational embeddedness — may carry a more conservative, traditionalist-Islamic valence that aligns with PPP’s appeal, whereas active membership reflects a more pragmatic, network-driven alignment that favours Golkar instead. PKS stands out dramatically on the negative side (β ≈ −1.1), by far the largest effect in the plot, reflecting a deep organisational and ideological divide. Gerindra (β ≈ −0.34) and Demokrat (β ≈ −0.55) are also clearly negative. NasDem and PAN produce negative point estimates but with wide CIs –statistically inconclusive.

Figure 3: Effect of Active NU Membership on Party Choice

For presidential preference, active membership sorts the three tickets more decisively than cultural proximity did (Figure 4): one clear positive effect, one clear negative effect, and one precise null — compared to the more ambiguous, borderline pattern cultural proximity produced, where Anies–Muhaimin sat in a marginal middle ground and Prabowo–Gibran was statistically indistinguishable from zero rather than clearly negative. Ganjar–Mahfud is the only clearly significant result (β ≈ 0.35, CI fully right of zero) — consistent with Ganjar’s regional and historical alignment with NU-PDIP networks in Central Java and East Java. Prabowo–Gibran is negative (β ≈ −0.30, CI left of zero); active NU members were less likely to back the pair despite PBNU chairman Staquf’s personal endorsement, plausibly because the organisational channel that mobilises active members, PKB, was simultaneously backing Anies–Muhaimin. This suggests active members responded to a competing organisational signal closer to their own political machinery than to Staquf’s individual cue. Anies–Muhaimin sits near zero with a wide CI: despite Muhaimin’s PKB chairmanship, active NU membership had no discernible effect on support for his ticket in either direction.

Figure 4: Effect of Active NU Membership on Presidential Ticket Choice

Overall, both variables point in the same direction, though with different degrees of force. Active membership generally produces stronger effects, both in drawing voters toward PKB and Golkar, and in pushing them away from PKS, Gerindra, and Demokrat. Cultural proximity works along the same lines but with weaker explanatory power. This reinforces the argument: NU’s electoral relevance runs through both symbolic identity and organised socio-political networks, with effects generally sharper among those more organisationally embedded.

CONCLUSION

The 2024 election confirms that NU voters cannot be won simply by invoking the right symbols or picking the right running mate. NU’s electoral significance is conditional on whether organisational networks are activated, whether elite signals are consistent, and whether a candidate or party can appeal to both the activist core and the much larger cultural periphery.

The evidence here points toward a network-activation model: NU’s electoral effects are strongest where organisational ties are dense and elite cues are coherent, and weakest where affiliation is purely symbolic. The Jokowi effect was significant in 2019, as it was in 2024; what distinguished 2019 was that a coherent elite signal channelled and amplified that appeal specifically through NU’s organisational networks. Structural and cultural affiliates moved together as a result, and Jokowi won decisively in NU’s heartland, consistent with district-level evidence that Ma’ruf’s nomination eliminated the relative disadvantage Jokowi had faced in NU-dense, Javanese-majority districts in 2014. In 2024, fragmentation was the dominant story: under those conditions, the network could not deliver a consistent signal, and Jokowi’s personal influence filled the vacuum.

The survey evidence does not support the argument that NU’s leadership successfully mobilised the organisation’s resources behind Prabowo–Gibran, and that this mobilisation was a decisive factor in the ticket’s victory. If Staquf’s endorsement had translated into effective organisational mobilisation, cultural proximity to NU should predict support for Prabowo–Gibran in the same way it predicts support for PKB or PPP elsewhere in the data. It does not: the regression estimate for Prabowo–Gibran is statistically not significant (Figure 2), meaning NU-identifying voters were no more likely to back the ticket than otherwise similar non-NU voters. What moved NU voters toward Prabowo–Gibran was not NU’s endorsement machinery but Jokowi’s effect.

For future research, the most important questions concern the micro-mechanisms through which NU networks actually transmit political signals: through kiai, through pesantren networks, through women’s organisations like Muslimat and Fatayat NU, or through the looser, more informal communal ties described above as the less-dense end of the same network spectrum. A related question for future work is whether these factors interact in a systematic or hierarchical fashion rather than simply operating in parallel. Perhaps the most practical implication is that parties and candidates who want NU votes cannot rely on claiming NU identity from the outside. What matters is whether they can engage NU’s actual organisational infrastructure, which figures they can bring on board, which networks they can activate, and whether they can sustain consistent messaging through those channels all the way to election day. Getting that right is harder than it looks, and harder than a simple identity claim would suggest.

---

This is an adapted version of ISEAS Perspective 2026/60 published on 20 August 2026. The paper and its references can be accessed at this link.