The Deinstitutionalization of Empowerment Infrastructure in Indonesian Urban Development: A New Face of the Tyranny of Participation

Participation has long been the dominant paradigm in development, especially through community-driven development (CDD) approaches adopted by more than 105 countries as a primary instrument for poverty alleviation (S​a​g​u​i​n​,​ ​2​0​1​8). The World Bank has invested billions of dollars in this approach, and Indonesia has become one of its largest laboratories through the Urban Poverty Alleviation Program (Program Penanggulangan Kemiskinan di Perkotaan, P2KP) and the National Community Empowerment Program in Urban Areas (Program Nasional Pemberdayaan Masyarakat Mandiri Perkotaan, PNPM), which reaches more than 70,000 villages and urban wards across the archipelago. Both programs are based on empowerment philosophy, participatory facilitation methods, and a robust facilitator training system. Since their introduction in the late 1990s, this approach has placed community facilitation at the heart of program success in Indonesia, echoing evidence from comparable CDD settings in the Philippines (L​a​b​o​n​n​e​ ​&​ ​C​h​a​s​e​,​ ​2​0​1​1). However, from the beginning, the idea of participation has not been free from criticism. C​o​o​k​e​ ​&​ ​K​o​t​h​a​r​i​ ​(​2​0​0​1​) introduced the thesis of the tyranny of participation and argued that participation could become tyrannical when participatory methods masked power relations and reduced community involvement to a meaningless procedure.

To orient readers unfamiliar with the Indonesian context, this section briefly introduces the programs examined in the study. Launched in the late 1990s in response to the Asian financial crisis, P2KP was a World Bank-supported urban initiative that channeled community investment grants (Bantuan Dana Investasi; BDI) to Community Self-Reliance Boards (Badan Keswadayaan Masyarakat; BKM) at the urban-ward level and supported revolving microcredit, social grants, and small-scale infrastructure. From 2007, it was scaled up and rebranded as PNPM Urban Area, the urban window of the national PNPM platform coordinated by the then Ministry of Public Works in partnership with local governments, which were obliged to provide a co-financing share (Joint Affairs District Fund/DDUB). Around 2014, the program transitioned through P2KKP and oriented toward slum management, and in 2016, it was reconstituted as the National Slum Upgrading Program (Kotaku) under the Directorate General of Human Settlements (Cipta Karya). The reconstitution of the program as Kotaku reflected an explicit national policy agenda: the government’s “100-0-100” target under the 2015–2019 National Medium-Term Development Plan (RPJMN). This target called for universal access to safe water, zero slum areas, and universal access to adequate sanitation. It was intended to deliver rapid, measurable, and verifiable improvements in urban living conditions, standardize service delivery across cities, and integrate community-based programs with national infrastructure investment. The transition also promised clearer accountability for tangible outputs and full financing from the national state budget (Anggaran Pendapatan dan Belanja Negara, APBN). Whereas P2KP/PNPM emphasized multiyear, cycle-based community empowerment, Kotaku was financed entirely through APBN, operated through annual physical targets, and explicitly prioritized the reduction of slum areas. Consequently, its program logic shifted decisively toward physical infrastructure delivery. Implementation in all eras has relied on a tiered corps of facilitators (technical experts, senior facilitators, and ward facilitators) who bridge national policy and community practice.

The transition from P2KP/PNPM to Kotaku shifted program priorities toward measurable physical outputs and annual implementation targets (Z​u​b​a​i​d​a​h​ ​e​t​ ​a​l​.​,​ ​2​0​2​3). This change raises questions about whether the institutional mechanisms that previously supported community empowerment have been maintained, particularly given evidence that infrastructure-oriented participatory programs do not necessarily benefit the poorest groups (G​r​i​l​l​o​s​,​ ​2​0​1​7).

Research has examined this issue from several perspectives. G​r​i​l​l​o​s​ ​(​2​0​1​7​) and S​a​g​u​i​n​ ​(​2​0​1​8​) demonstrated that participation in development programs was vulnerable to elite capture and failed to reach the poorest. J​a​k​i​m​o​w​ ​(​2​0​1​8​a​) and J​a​k​i​m​o​w​ ​&​ ​H​a​r​a​h​a​p​ ​(​2​0​1​6​) highlight the experiences of volunteers and community workers, revealing the moral dimensions and subjectivities of actors in the field. Z​u​b​a​i​d​a​h​ ​e​t​ ​a​l​.​ ​(​2​0​2​3​) pointed out the weak policy network in the implementation of Kotaku. However, these studies generally focused on beneficiaries, program outcomes, and policy structures, but have not yet thoroughly explored how facilitators experienced shifts in program characteristics across eras, as well as the impact of the loss of empowerment infrastructure on their capacities and community self-reliance. Facilitators are carriers of institutional memory across program eras, and their experiences could reveal hidden shifts in official policy documents.

This study was positioned within this research gap. Positioning these concerns within the international debate sharpens their significance. The Indonesian CDD experience has been a global reference point alongside the Philippines and India, where comparable programs have been scrutinized for elite capture and uneven pro-poor reach (D​a​s​g​u​p​t​a​ ​&​ ​B​e​a​r​d​,​ ​2​0​0​7; F​r​i​t​z​e​n​,​ ​2​0​0​7; S​a​g​u​i​n​,​ ​2​0​1​8). Indonesian scholarship has further documented how Indonesian participatory development planning forums (musrenbang) have oscillated between deliberative promise and patronage (A​s​p​i​n​a​l​l​,​ ​2​0​1​3; W​i​d​i​a​n​i​n​g​s​i​h​ ​&​ ​M​o​r​r​e​l​l​,​ ​2​0​0​7), and how rural and urban governance has converged on a “project system” that privileges disbursement over process (L​i​,​ ​2​0​1​6). Against this backdrop, the present study investigated not only whether participation failed, but also why its institutional preconditions eroded, from a perspective applicable beyond Indonesia to any context in which donor-funded CDD was institutionalized and later re-engineered.

This study examines how urban community empowerment programs changed from P2KP/PNPM to Kotaku, how these changes affected facilitator capacity and community self-reliance, and how facilitators adapted to the new program environment. It contributes to the literature by examining the erosion of facilitator training, facilitation methods, and learning mechanisms as an institutional pathway through which participation may become increasingly procedural.

To support readers unfamiliar with the Indonesian policy context, Table 1 summarizes the key program-specific terms and abbreviations used throughout the manuscript.

Participation in development refers to the involvement of communities in decisions that affect their lives, ranging from token involvement to complete citizen control. A​r​n​s​t​e​i​n​ ​(​1​9​6​9​) mapped this gradient through the ladder of participation, to distinguish among non-participation, tokenism (informing, consultation, and placation), and citizen power (partnership up to citizen control). Participation that halts tokenism is a formality without the genuine transfer of power. This understanding is sharpened by C​o​o​k​e​ ​&​ ​K​o​t​h​a​r​i​ ​(​2​0​0​1​), who warned that participation could become tyrannical through the tyranny of participation. They identified three dimensions of tyranny:

(1) the tyranny of decision making and control, when facilitators or elites dominated;

(2) the tyranny of the group, when group dynamics marginalized the voices of vulnerable individuals;

(3) the tyranny of methods, when standardized participatory techniques displaced other more appropriate approaches.

These three forms of tyranny explain how participatory spaces could lose their substantive meanings. Combined with the A​r​n​s​t​e​i​n​ ​(​1​9​6​9​) ladder of participation, both offer tools to distinguish empowering participation from merely procedural participation. This framework is suited to examine mentoring programs that position facilitators as central actors in producing or reducing the quality of participation. These three dimensions serve as the main lens for analysis and synthesis in this study.

This study defines the central term “deinstitutionalization” precisely. Drawing on institutional theory (O​l​i​v​e​r​,​ ​1​9​9​2), deinstitutionalization denotes the process by which an established institutional practice loses legitimacy and erodes until it is no longer taken for granted.

In this study, the term is operationalized along three concrete and observable components:

(1) the weakening of formal institutions, specifically the facilitator training and certification system;

(2) the loss of the normative framework, the empowerment philosophy, and reflective values that defined the role of the facilitator;

(3) the decline of capacity-development mechanisms, the multiyear learning-by-doing cycle, and intergenerational mentoring. “Empowerment infrastructure” refers to the institutionalized bundle of these three components; its deinstitutionalization is not a single rupture but the cumulative attrition of all three. This specification distinguishes the concept from cases where it is used more loosely​ (e.g., mere budget cuts) and makes it analytically testable against the field data.

This framework was used to examine how changes in facilitator training, normative orientation, and learning mechanisms affected the substantive quality of participation across program eras. It also guided the analysis of whether the erosion of these institutional conditions shifted participation from substantive empowerment toward procedural compliance.

Several studies over the past decade have highlighted the experiences of field practitioners. Jakimow (2018c) examined how community-based volunteers in urban community-driven development (CDD) programs in Medan negotiated the impossibilities of the program, compromised their participatory idealism for the sake of outcomes, and borne the burden of failure that devolved onto them. These findings were related to the responses and adaptations of the facilitators. J​a​k​i​m​o​w​ ​(​2​0​1​8​a​) demonstrated that facilitators often experienced affective injuries in the form of complaints, accusations, and lack of appreciation, which gradually eroded their capacity and personal resilience. This dimension of subjectivity was reinforced by J​a​k​i​m​o​w​ ​&​ ​H​a​r​a​h​a​p​ ​(​2​0​1​6​), who revealed that involvement in facilitation could become a space for self-formation among lower-class volunteers, and by J​a​k​i​m​o​w​ ​(​2​0​1​8​b​), who analyzed the empowerment effects within state-led development. These studies affirmed that the capacity of facilitators and community self-reliance were highly dependent on the experience, motivation, and institutional support of facilitators. The findings indicated that the erosion of institutional support, including the philosophy and training systems for facilitation, could undermine facilitator capacity and the quality of participation.

In contrast, studies focusing on program and participation quality provided important points of comparison. Z​u​b​a​i​d​a​h​ ​e​t​ ​a​l​.​ ​(​2​0​2​3​) found that the implementation of Kotaku in Palembang was hampered by weak collaboration among actors and unclear roles, resulting in unmet physical targets. G​r​i​l​l​o​s​ ​(​2​0​1​7​) demonstrated that in participatory budgeting in Surakarta, allocations of infrastructural spending were biased toward areas with fewer poor residents, indicating that physical orientation did not automatically favor those most in need. S​a​g​u​i​n​ ​(​2​0​1​8​) reinforced this finding by showing that participation in CDD was vulnerable to elite capture and failed to reach poor groups, while A​k​b​a​r​ ​e​t​ ​a​l​.​ ​(​2​0​2​0​) assessed that participatory planning practices were often suboptimal when measured by sustainable development outcomes. Several studies have also noted that the shift from human empowerment to physical interventions changed how programs were implemented in the field, although this change was rarely directly examined from the experience of facilitators across different programs. Despite its richness, the literature generally portrayed the experience of beneficiaries, program outcomes, or policy structures at a single time. No study has specifically examined, from the facilitator’s perspective and diachronically, how program characteristics change across eras and how the deinstitutionalization of empowerment infrastructure impacts facilitator capacity and community self-reliance.

Situating the present study within Indonesian scholarship on public participation further clarifies its contribution. A substantial body of work warned that participatory arenas were prone to elite capture: D​a​s​g​u​p​t​a​ ​&​ ​B​e​a​r​d​ ​(​2​0​0​7​) and F​r​i​t​z​e​n​ ​(​2​0​0​7​) showed that without careful design, collective action in Indonesian CDD could be appropriated by local elites, while studies on development planning meeting deliberation (A​n​i​n​d​i​t​o​ ​e​t​ ​a​l​.​,​ ​2​0​2​2; K​a​r​m​a​n​ ​e​t​ ​a​l​.​,​ ​2​0​2​2) and on urban poverty strategies in Yogyakarta (M​u​r​w​a​n​i​ ​e​t​ ​a​l​.​,​ ​2​0​2​2; R​o​i​t​m​a​n​,​ ​2​0​1​9) uncovered persistent tensions between bottom-up aspiration and patronage. These accounts largely explained participatory failure from the demand side, focusing on who captured the process. The present study was considered complementary but distinct: it traced a supply-side condition, namely the institutional preconditions that, once dismantled, left even well-intentioned participation hollow. This was the specific gap the study occupied.

This study employed a qualitative approach with an interpretive paradigm, viewing social reality as a construction of meaning formed and negotiated through experience, language, and inter-actor interaction (A​y​t​o​n​ ​e​t​ ​a​l​.​,​ ​2​0​2​3). This paradigm was chosen because the research focused on the experiences and interpretations of facilitators regarding changes in urban community empowerment programs, an arena marked by interpretation, role negotiation, and power relations. The research was designed with a single case study that is holistic and context-bound (Y​i​n​,​ ​2​0​1​8; T​o​m​a​s​z​e​w​s​k​i​ ​e​t​ ​a​l​.​,​ ​2​0​2​0). The case was delimited to the facilitation practices of empowerment across different program eras in one city, with clearly defined boundaries of space, time, key actors, and arenas of practice. The analysis was guided by the tyranny of participation framework (C​o​o​k​e​ ​&​ ​K​o​t​h​a​r​i​,​ ​2​0​0​1) and the concept of deinstitutionalization of empowerment infrastructure; thus, the study traced how participation was produced or reduced in practice, rather than quantitatively measuring program outcomes. The “how” and “why” research questions rendered the case study the most appropriate strategy, as the phenomenon under investigation was intricately intertwined with its context and could not be separated from the social situation that shaped it (Y​i​n​,​ ​2​0​1​8).

This study was conducted in the city of Makassar. This location was purposively chosen because from 2004 to 2023, Makassar served as a pilot project and intervention target for various urban community empowerment programs, namely P2KP, PNPM Urban Self-Reliance Program, P2KKP, and Kotaku. This two-decade span highlights the representativeness of Makassar for examining shifts in program characteristics across eras, as the city has undergone facilitation using various approaches. Data collection was carried out from January to December 2025: in-depth interviews and document study from January to October 2025, and field observations from July to December 2025. All coding and analysis were completed between January and April 2026. All findings, tables, and figures reported in this article are therefore based on the complete, fully collected and analyzed dataset.

To contextualize the site, Makassar, located on the southwestern coast of Sulawesi Island, is the capital of South Sulawesi Province and the largest city in eastern Indonesia. With a population of roughly 1.4–1.5 million, it is the principal economic gateway of the eastern archipelago, marked by rapid urbanization, dense coastal and riverside settlements, and persistent pockets of slum housing that made it a priority target across successive urban programs. Makassar is an instructive case for three reasons: first, it has hosted all four program generations (P2KP, PNPM Urban Self-Reliance Program, P2KKP, and Kotaku) continuously for two decades; second, it retains a cohort of facilitators who serve across these eras and thus carry the institutional memory the study seeks; and third, its mix of poverty-alleviation and slum-upgrading mandates allows the paradigm shift from empowerment to physical works, especially visible in practice.

Participants were selected through purposive sampling to recruit information-rich informants regarding facilitation practices (A​m​e​s​ ​e​t​ ​a​l​.​,​ ​2​0​1​9). Brief profiles of informants were shown in Table 2.

The 17 informants were assigned to one of two analytical groups according to their primary program orientation rather than their complete service histories. The poverty-alleviation-oriented group comprised three technical experts and seven senior facilitators who entered during the P2KP/PNPM generation, whereas the slum-management-oriented group comprised seven neighborhood facilitators classified as primarily associated with P2KKP/Kotaku.

Table 2 reports the informants’ actual service histories, which overlap across program eras: the 3 technical experts and 7 senior facilitators served across P2KP, PNPM, P2KKP, and Kotaku; FK-03 served across PNPM, P2KKP, and Kotaku; and the other six neighborhood facilitators served in Kotaku.

Despite this cross-era experience, FK-03 was analytically assigned to the slum-management-oriented group based on the informant’s primary program orientation and was counted only once. The resulting groups therefore comprised 10 and 7 informants, totaling 17 without double-counting.

The informants represented technical expert, senior facilitator, and neighborhood facilitator roles, with some also performing city-coordination functions. This grouping supported contextual comparison across program generations and enabled the study to trace how facilitators interpreted program mandates and negotiated rules and resources under different structural conditions. Informant sufficiency was assessed during data collection based on the depth and recurrence of information rather than participant numbers alone.

This study foregrounds facilitators because they carry institutional memory across eras and are uniquely positioned to reveal the diachronic erosion of empowerment infrastructure that official documents conceal. Because the design is facilitator-centered, the perspectives of end-beneficiaries (urban-ward residents, ward officials, and local community leaders) are not directly represented; the implications of this scope for the community-self-reliance findings are consolidated in the Limitations subsection.

• Data Analysis

The interviews were recorded with the informants’ consent and transcribed verbatim to build a traceable base of evidence. The analysis was conducted using iterative coding that combined deductive and inductive strategies. Deductive codes were derived from the conceptual framework, namely the three dimensions of the tyranny of participation, as well as indicators of the deinstitutionalization of empowerment infrastructure (philosophy, methods, and mentoring-training systems). Inductive codes were developed from the field narratives. A codebook was constructed to define the scope, boundaries, and indicators of each code to ensure consistent and testable application (R​o​b​e​r​t​s​ ​e​t​ ​a​l​.​,​ ​2​0​1​9). Analytical memos and audit trails recorded the evolution of categories and coding decisions (B​i​n​g​h​a​m​,​ ​2​0​2​3). The coded data was subjected to within-case analysis using pattern matching, explanation building, and rival explanation testing techniques to strengthen the quality of the inferences (Y​i​n​,​ ​2​0​1​8). The linkage between the data and concepts was carried out in stages so that every finding proposition could be traced from quotation to category coding. The analysis was finalized only after all interviews had been transcribed and coded, so that the patterns, tables, and figures presented in the following section rest on the complete dataset rather than on preliminary or partial results.

Data were collected using three complementary methods: in-depth interviews, document analysis, and field observations. In-depth interviews served as the primary source, positioned as guided conversations that maintained an analytical focus, while allowing informants ample space to freely tell their narratives and meanings (Y​i​n​,​ ​2​0​1​8). The interviews were guided by a semi-structured protocol derived from the research questions and were conducted in either Indonesian or local languages according to the informant’s preference.

All interviews were audio-recorded and transcribed verbatim in the original language; the excerpts reported here were subsequently translated into English, and each quotation was cross-checked against its source-language transcript by the research team to preserve fidelity of meaning. In the translated quotations, square brackets mark clarifying insertions and ellipses indicate omitted material. The interviews explored the facilitators’ experiences of program changes, explanations of actions, and adaptation strategies.

Archival documents were used to corroborate the interview findings and were treated as evidence requiring verification rather than being accepted at face value. The materials comprised the Mobilization and Personnel Placement Decrees (SK Mobilisasi dan Penempatan Personel); recipient decrees for BDI and government assistance for communities (Bantuan Pemerintah untuk Masyarakat; BPM) issued by the Directorate of Settlement Area Development (PKP); and three slum-designation decrees: Makassar Mayor Decree No. 150.05/1341/Kep/IX/2014 of 22 September 2014, Makassar Mayor Decree No. 826/653.2/2018 of 5 February 2018, and South Sulawesi Governor Decree No. 956/III/2020 of 31 March 2020. Together, these documents provided evidence on personnel deployment, the extent of officially designated slum areas, and BDI and BPM allocations during 2018–2021.

Observations were conducted in several areas of Makassar City, including slum neighborhoods, to understand the context of ongoing empowerment practices in the field. These field observations were conducted from July to December 2025 across representative slum settlements in Makassar, including the coastal and Tallo riverbank neighborhoods (e.g., urban wards of Tallo, Buloa, and Kaluku Budoa) as well as city-center, industrial-area, and peri-urban slum pockets, spanning the four slum typologies recognized in the Mayor's decrees.

The observations focused on the physical condition of settlements, infrastructure works, including drainage, sanitation facilities, and upgrades to uninhabitable housing (Rumah Tidak Layak Huni, RTLH), as well as ongoing facilitation and empowerment practices. Because direct observation could capture only the contemporary Kotaku and post-Kotaku context, evidence on the earlier P2KP/PNPM era was necessarily derived from retrospective interviews and archival documents; this evidential asymmetry across program eras is acknowledged among the study's limitations. Photographs were taken as part of the field documentation, as shown in Figure 1 and Figure 2.

The interviews were recorded with the informants’ consent and transcribed verbatim to build a traceable base of evidence. The analysis was conducted using iterative coding that combined deductive and inductive strategies. Deductive codes were derived from the conceptual framework, namely the three dimensions of the tyranny of participation, as well as indicators of the deinstitutionalization of empowerment infrastructure (philosophy, methods, and mentoring-training systems). Inductive codes were developed from the field narratives. A codebook was constructed to define the scope, boundaries, and indicators of each code to ensure consistent and testable application (R​o​b​e​r​t​s​ ​e​t​ ​a​l​.​,​ ​2​0​1​9). Analytical memos and audit trails recorded the evolution of categories and coding decisions (B​i​n​g​h​a​m​,​ ​2​0​2​3). The coded data was subjected to within-case analysis using pattern matching, explanation building, and rival explanation testing techniques to strengthen the quality of the inferences (Y​i​n​,​ ​2​0​1​8). The linkage between the data and concepts was carried out in stages so that every finding proposition could be traced from quotation to category coding. The analysis was finalized only after all interviews had been transcribed and coded, so that the patterns, tables, and figures presented in the following section rest on the complete dataset rather than on preliminary or partial results.

Data validity was maintained through the criteria of trustworthiness: credibility, transferability, dependability, and confirmability (N​y​i​r​e​n​d​a​ ​e​t​ ​a​l​.​,​ ​2​0​2​0; Y​o​o​n​ ​&​ ​U​l​i​a​s​s​i​,​ ​2​0​2​2). Credibility was strengthened by source triangulation. The formal interview sample comprised only the 17 facilitators listed in Table 2; 5 fellow facilitator team members, one BKM program-recipient coordinator, and 2 officials from the municipal government implementation unit (satuan kerja, Satker) were consulted as supplementary triangulation and verification sources, not as separately coded interview informants; these supplementary consultations were audio-recorded and transcribed. Triangulation was also possible within the 17-informant sample itself: because several of these facilitators had served together in the same program-implementation teams, their accounts of shared events and decisions could be cross-checked against one another, further strengthening credibility. This enabled the cross-verification of facilitator narratives at the levels of micro-practices, community institutions, and program governance. Method triangulation was conducted using interviews and document analysis. Discrepancies between sources were not weaknesses automatically considered but rather data read on contextual dynamics and power relations. Member checking and peer debriefing reduced interpretive bias, while a case study database and chain of evidence were maintained to allow readers to trace the links among the quotations, categories, and propositions of the findings (Y​i​n​,​ ​2​0​1​8).

This study upheld ethical principles by obtaining participants’ consent, including permission for recording prior to the interviews. Participation was voluntary, and informants had the right to decline to answer or withdraw at any time without consequences. The identities of informants were protected with codes (e.g., TA-01, SF-01, and FK-01) and data were securely stored for academic purposes only. The presentation of quotations preserved the confidentiality and dignity of the participants, particularly since some narratives touched on work relationships and sensitive policies. The study also observed the principle of non-maleficence by ensuring that the involvement of participants did not pose risks to their professional relationships.

In qualitative research, the researcher is the primary instrument as reflexivity is essential for maintaining the trustworthiness of interpretations (N​y​i​r​e​n​d​a​ ​e​t​ ​a​l​.​,​ ​2​0​2​0). The researcher’s proximity to the field of empowerment facilitation provided not only access and contextual understanding, but also the potential to introduce bias. Therefore, the researcher articulated their position, restrained preconceptions through reflective journaling, and subjected their interpretations to audit trails, peer debriefing, and cross-source evidence. These reflective notes became part of the audit trail that enabled the tracing of the researcher’s interpretative decisions. Thus, the researcher’s closeness to the field was positioned as a source of analytical sensitivity that is managed transparently rather than as an uncontrolled influence.

The identities of the informants were anonymized using alphanumeric codes in accordance with the principles of research confidentiality. The roles of informants were grouped into three categories, namely technical expert, senior facilitator, and neighborhood facilitator with various specializations. Their brief profiles and the individual codes for identification are listed in Table 2, related to the quotations in the Results section.

No formal qualification requirements distinguished the P2KP/PNPM and Kotaku eras. Facilitators were required to hold at least a bachelor's degree and to have a minimum of two years' experience in community empowerment, together with role-specific competencies or certification; technical facilitators, for example, were expected to master the preparation of the Budget Cost Plan (Rencana Anggaran Biaya, RAB). In practice, however, informants reported that facilitator recruitment increasingly emphasized formal diplomas over demonstrated competence.

The transition from P2KP/PNPM to Kotaku was structural rather than linear, spanning six interrelated dimensions: program paradigm, funding, implementation cycles, recruitment, team composition, and disbursement and reporting. Facilitators’ accounts show that these changes collectively shifted the program’s focus from community capacity building to physical target achievement.

The program paradigm was the most prominent dimension of this change. Facilitators described a shift from qualitative empowerment to number-based projects. SF-01 asserted, “Community empowerment was strong in PNPM, but Kotaku is more of a project; it is about targets and disbursement. In PNPM, the community was empowered”. FK-01 also stated, “It used to be about empowerment, now it’s a project. Projects take time. With empowerment, unless the community is truly empowered, capable, and ready, it’s not finished”. SF-02 reinforced this by mentioning a change in approach, “We used to be qualitative, now it’s shifted to quantitative.” This shift alters the focus of objectives, as explained by SF-07, who said the program has “[The focus] has already moved to slum-area management”. Success indicators were now measured by the calculated reduction in slum areas, not community self-reliance.

The funding scheme has undergone fundamental changes in terms of budget sources and governance. SF-02 stated, “PNPM and Kotaku differ in terms of grant funds and budget disbursement. One is a loan, one is pure APBN”. The technical consequence was a change in expenditure nomenclature as he added, “Under PNPM, the funds were classified as capital expenditure, whereas under Kotaku they were classified as expenditure on goods.” The transition to pure APBN eliminated the obligation of regional contributions. FK-01 recounted, “Since Kotaku started, there has been no more DDUB, no more cost-sharing funds, so there are none left; local governments no longer feel they have a stake”. Simultaneously, new budget actors had emerged who could potentially distort the program. SF-02 observed that “council aspiration funds piggyback on the program”, while FK-02 considered its implementation “fairly extreme, as numerous legislative council (Dewan Perwakilan Rakyat, DPR) aspirations are now included”. The funding scheme has shifted from a participatory and needs-based pattern to a transactional and absorption-based one.

The cycle and implementation times of the program have narrowed drastically. SF-02 summarized, “In PNPM, it was multi-year, so it used a cycle, whereas in Kotaku, it’s one-year”. Consequently, the learning space in the community was curtailed by the pressure of the fiscal year. SF-03 described, “Ideally, the completion process that should take one year is now targeted to finish in three months”, sparking pessimism among facilitators. The time compression also extended to the preparatory stage for facilitators; SF-03 compared, “For example, the pre-service training used to be two weeks, the new model is just four days”. Furthermore, the core empowerment cycle mechanism of P2KP was eliminated, as emphasized by SF-06, “In Kotaku, there’s no cycle. The difference is, if in P2KP the cycle was strong, there wasn’t any Key Performance Indicators (KPI)”. The loss of this cycle meant that stages of community awareness were no longer a prerequisite for the sustainability of activities.

The facilitator recruitment system changed with the decentralization of authority. SF-03 explained, “recruitment [under] PNPM and Kotaku [has become] ambiguous, owing to issues of centralization and decentralization”, since the Working Unit was devolved to the district level. As a result, the selection criteria shifted from competency to formal educational qualifications, as criticized by FK-01, “recruitment is no longer based on the person’s competence or abilities, but on their diploma”. TA-01 added that in the Kotaku era, “there were people with less experience who were still accepted”. Practices of placement based on personal connections also intensified; FK-03 recounted their experience when first assigned and was immediately asked, “who’s your insider here?” This shift undermined the quality assurance of facilitators because the ability to empower communities was no longer the main requirement for their selection.

The composition of the facilitator team has decreased in terms of both the number and diversity of expertise. SF-01 noted, “Previously, one team consisted of five people, now it’s two people per team”, while FK-02 mentioned that in the final period, the pattern was “one community facilitator per urban village”. The reduction in team size has impacted on the workload. SF-07 revealed, “We have to multitask, and schedules are irregular” because a facilitator must plan, monitor, and review reports. The loss of facilitators in specific fields has lowered the quality of facilitation; TA-03 explained, “economic facilitators were not provided with UPK skills, so the new economic facilitators in Kotaku don’t know how to manage revolving funds”. A lean and less multidisciplinary team composition increased the burden on substantive facilitation functions.

The disbursement and reporting mechanism has shifted toward a system based on performance indicators and quantitative data. FK-06 explained that assessment now relied on “The assessment is based on KPIs and baseline data for each area.” This orientation shaped the facilitators’ work routines, as acknowledged by FK-03: “In PNPM, we directly learned about livelihood; the output was documents. Here [under Kotaku], the output is KPIs.” The pressure to absorb funds within a single fiscal year also gave rise to fund returns. TA-01 admitted, “It doesn’t always go smoothly. Some budgets are returned and the key is also community readiness”, differing from the old pattern described by SF-02, “In the previous program, if there was a budget shortfall, it was supplemented”.

Figure 3 summarizes the coding distribution using two measures: the number of coding references (122 in total) and the number of interview sources in which each theme appeared (51 source-level codings in total). The 51 source-level codings are the sum of the source counts across the six themes; because a single interview source can be coded under more than one theme, this total exceeds the 17 informants and does not represent 51 distinct informants or files. Source coverage is therefore reported against the 17 interview sources. On both measures, Paradigm Shift is the dominant dimension: it accounts for 39 coding references (32.0%) and was raised by most informants (12 of 17 sources). By coding references, it is followed by Recruitment System Shift (23; 18.9%), Funding Scheme Shift (18; 14.8%), Team Composition Shift (16; 13.1%), Cycle & Timeframe Shift (15; 12.3%), and Disbursement & Reporting Shift (11; 9.0%), while at the source level each of the remaining five dimensions appeared in 7–8 of 17 sources. The coding-reference percentages describe how frequently each theme appeared in the coded interview material, whereas the source counts report how many of the 17 interview sources raised each theme. This dominance indicates that for facilitators, the root of the change lies in the shift in program philosophy from empowerment to project-based, while the dimensions of funding, cycle, recruitment, team composition, as well as disbursement and reporting are structural derivatives.

This finding is consistent with the statement from FK-05, who assessed the program composition as “sixty percent project, forty percent empowerment,” as well as the complaint from SF-02 that “[Under] Kotaku, the project element became very strong. There are standards, and it becomes a source of power.” The relatively even spread of sources across the other five dimensions indicates a comprehensive and systemic transformation rather than a partial one. To complement the interview evidence, Table 3 compares the program orientation, typical interventions, selection basis, funding cycle, and implementation scope across the two program eras in Makassar. The comparison draws on program documents, BDI/BPM recipient decrees, and baseline records and is intended as an indicative rather than exhaustive overview.

This study discovered that the shift in program orientation toward physical development reduced community participation to a mere procedural formality. When success was measured by budget absorption and rapid completion of physical work, the space for awareness raising and collective decision making narrowed, leaving community involvement as nothing more than administrative attendance and signatures. FK-05 asserted that after “the physical structure is built, the community remains unchanged”, as did FK-07’s criticism of development projects that did not meet actual needs, “For the construction of the gate, I do not think there’s any benefit. For me, I have yet to see any value in it.”

The loss of mentoring knowledge, such as its philosophy, methods, and training systems, is a latent consequence of the program’s shift from empowerment to physical development. Based on the seven core themes in the coding results, this erosion of knowledge is sequential: from the fading of the philosophical foundation of mentoring, the collapse of the training infrastructure, to the breakdown of intergenerational knowledge transfer among the facilitators. The accumulation of these losses results in a decline in facilitator capacity and the weakening of the autonomy of assisted communities.

To convey how quickly this erosion unfolded, Figure 4 places the key inflection points on an actual timeline, showing that the most acute losses (the halt of facilitator training, the compression of the cycle, and the shrinking of teams) clustered after the 2016 transition to Kotaku. Under P2KP/PNPM, each facilitation team comprised 5–7 members and received 14 days of structured training, and implementation was oriented toward process—community empowerment and poverty alleviation. Under Kotaku, in both its early and later phases, teams were reduced to 5 members, dedicated training was discontinued, and the orientation shifted to project targets centered on physical slum reduction. This progression smaller teams, shorter or absent training, and a move from process to output captures the same deinstitutionalization that Figure 4 renders chronologically.

Philosophy and Mentoring Methods that form the foundation of the empowerment paradigm have eroded in line with the transition from PNPM to Kotaku. Reflective instruments that shaped the facilitators’ critical awareness were no longer recognized. SF-02 asserted, “In the PNPM days, we were equipped with training methods, facilitation techniques. In Kotaku, we aren’t. Things like coffee philosophy, the Johari window—Kotaku facilitators never hear about those anymore.” TA-03 explained that the ability to “refine feeling, refine thinking, and refine the soul” was possessed only by the first generation of facilitators and was displaced when “the program became project-oriented… We no longer need refinement of the soul or feeling, just skills, because that’s what matters now”. SF-03 added that the participatory approach became a weak point, “it [is] now more about participation models, like the focus group discussion—that is clearly the issue.” The fading of reflective tools marked the decline of core critical consciousness-raising methods for community empowerment. (Note: “Coffee philosophy” refers to an informal-immersion ethos in P2KP/PNPM facilitator training via building trust by “drinking coffee” with residents, i.e., sustained relational presence before any intervention. The “Johari window” is a self-awareness model used in that training to help facilitators understand the open, blind, hidden, and unknown areas of self in their interaction with the community. Both are reflective-method staples that Kotaku training no longer transmits.)

The Training System, as a mechanism for the reproduction of knowledge, has experienced a structural collapse, and this is the node with the highest coding intensity (ten citations). New Kotaku facilitators have received practically no preparation. FK-04 stated briefly, “In Kotaku there aren’t really any facilitator trainings”, while SF-01 added that “there have not even been any trainings for years. Since 2017, there also have been no facilitator trainings.” This void has forced facilitators to work without modules; FK-07 admitted, “The material I delivered was self-taught; there were no modules. I never saw the module”. The loss of tiered mentoring is also felt, as expressed by SF-01, “In PNPM, there was intensive coaching with the UPL”, a practice that no longer takes place in Kotaku.

The Learning by Doing mechanism, which once served as a means of developing competencies, has also disappeared. TA-03 described the now-lost multiyear learning cycle, “In the past, there was the term facilitator as learning by doing; now there is not. Back then, in the first year, we would provide support, and in the second and third years, they would come to us to share what they had done”. In addition to disrupting facilitator development, the disappearance of this mechanism has eliminated the dimension of collective learning within the community. SF-06 recalled the previous collective model, “Everyone was there BKM, UPL, whoever became Community Self-Help Group (Kelompok Swadaya Masyarakat, KSM), we learned together” and compared the underlying philosophies, “in P2KP, empowerment was enforced; in PNPM, rapid independence was enforced”. Thus, the learning space that once served as a source of knowledge reinforcement was no longer available to students.

Intergenerational knowledge transfer among facilitators has been disrupted by the loss of regeneration channels. The role of senior mentors, once crucial, is now rare; SF-01 recounted his experience being mentored by a senior, “I was continuously coached by Mr. Darwis. He always guided me, so I understood”. Formal channels have also weakened, as FK-01 explained that certification forums were no longer required, “There used to be a forum established for recruitment to issue facilitator certification. Why are people not applying? Because it’s no longer a requirement”. The departure of campus actors and NGOs who once served as conduits of empowerment knowledge has worsened the situation; according to FK-01, the impact is “clear, in terms of competence and knowledge transfer”. Consequently, the collective knowledge reservoir of the program is dwindling.

Facilitator capacity is a direct consequence of this loss. Many new-generation facilitators are unable to perform basic facilitation functions. SF-02 observed, “Many Kotaku facilitators are unable to appear or speak in front of a forum”, while FK-07 admitted their personal limitation, “I am afraid…afraid of conveying things incorrectly to the community. I never received training in facilitation techniques.” Weak communication and advocacy skills are also prominent, as FK-05 acknowledged, “I lack building communication with the community… I focus more on technical matters, rarely go to the field, and mostly work on my laptop”. As a compensation, facilitators take an unsystematic self-taught route; TA-02 stated, “My knowledge is self-taught. I also learn from senior social staff”, a survival strategy that is not equivalent to structured training.

Community self-reliance, which is the ultimate goal of the program, weakens when facilitators take over roles that should be carried out by residents. As dependence on facilitators in BKM increased, SF-03 quipped that “if only we are capable and the community is not, that’s not empowerment”. The technical capacity of the community to prepare Cost Estimate and Budget Plan (Rencana Anggaran Biaya, RAB) and Accountability Report (Laporan Pertanggungjawaban, LPJ) has also eroded. TA-03 compared the two eras, “In the P2KP/PNPM days, in the first year, they were taught; in the second year, for consultations, they came on their own; the third year was about verification. But comes the Kotaku era, it’s the facilitators who do the work”. The impactful consequence is that the behavior of the community remains unchanged; FK-05 asserted, “after the physical structure is built, the community stays the same”. While physical development is complete, awareness, which is the essence of empowerment, is not achieved.

Strengthening leadership, which was previously an explicit training agenda, is now being neglected. SF-02 stated directly, “It’s definitely different because there is no strengthening of leadership abilities there”. In fact, according to SF-01, leadership was the first competency that should be developed, “First is leadership ability. Second, communication must be built… there can be no ego within the team”. The dimension of character development has also weakened; SF-03 reminded us that “ethics and attitude are important”, to be in line with SF-06, who emphasized that the ability to speak without “ethics and respect for others” became a problem. The disappearance of leadership and character development completes the picture of the comprehensive erosion of knowledge.

The relationship between pattern changes and knowledge loss is presented as a matrix in Figure 5.

Figure 5 reads as co-occurrence between two categories: program change patterns (rows) and dimensions of knowledge loss (columns). Paradigm Shift has the widest reach, co-occurring strongly (2) with disrupted knowledge transfer and weakened community self-reliance and moderately (1) with the loss of philosophy, the training system, and learning-by-doing. Cycle-and-timeframe changes co-occur strongly (2) with collapse of the training system and moderately (1) with learning-by-doing. Recruitment system changes co-occur strongly (2) with broken knowledge transfer and moderately (1) with the training system. Funding-scheme and disbursement/reporting changes show no meaningful co-occurrence (–) with any knowledge-loss dimension. These associations suggest that knowledge erosion stems more from changes in paradigm, cycle, and recruitment than from financial-administrative mechanisms.

Together, these patterns indicate that knowledge loss was systemic and shaped the conditions to which facilitators subsequently adapted.

Facilitators are not merely passive recipients of program changes; rather, they are active agents who interpret, negotiate, and navigate pressures during the paradigm shift from empowerment to physical development. Analysis of matrix coding query from interviews with 17 informants revealed facilitators’ responses in nine interwoven thematic patterns, ranging from managing pressures, adapting roles, self-directed learning, and coaching new cadres to efforts to maintain idealism amid limited support systems. These patterns illustrated that facilitator adaptation was built through technical, relational, and ethical strategies implemented simultaneously to ensure the continuity of facilitation.

Pressure and stress are intense experiences for facilitators in the Kotaku era. SF-01 stated that the workload increased along with the large funds managed, “The stress level is high. This is because we usually receive BDI in Kotaku. I supervise three sub-districts and always receive it”. SF-05 agreed that the pressure stemmed from targets, “There’s more stress in Kotaku because it emphasizes targets more”. SF-03 described bureaucratic pressure, “Pressure from the Director General, Governor, Mayor, Sub-district Head, and so on…each brings its own pressure, which causes stress”. Facilitators made the post a safe space, supporting each other as a team; FK-03 acknowledged the closeness developed because “I’m there twenty-four hours, living at the post”, while FK-07 felt supported because “there’s a team that backs me up”. Pressure also came from community relations, as lamented by SF-07, “Very often, especially lately, we get BKM who are money-oriented”. Stress was managed more through informal collective support than through program mechanisms.

The shift in orientation from empowerment to physical development required facilitators to change their working methods. SF-01 summarized the new demands, “So you have to be multidisciplinary” and acknowledged that many roles in the community have been taken over: “In Kotaku, it is really we [facilitators] who do everything.” FK-07 described the takeover of tasks in order to meet disbursement deadlines, “I took over because the bookkeeping had to be finished before the funds could be disbursed”. This adaptation often ended up as mere formality, as emphasized by SF-03, “Now it’s just practical, a formality, just being presented”. Independent learning has emerged as a compensatory strategy for the loss of formal training for new facilitators. FK-07 recounted the lack of structured guidance, “We’re self-taught; we meet directly with the community and start talking straight away”. TA-02 had a similar experience: “As for knowledge, I am self-taught. I learned from my senior colleagues in social, including SF-01”. FK-03 even learned by imitating his senior: “I just copy and paste, I imitate my SF” and absorbed knowledge informally “from [informal] evening gatherings”. FK-06 underscored the need to be “responsive in learning something new related to the program”.

The training of new facilitators was conducted informally by senior facilitators who filled the gap left by the absence of institutional mentors in the training process. FK-04 implemented direct field mentoring: “I always accompany them to the site, so they also feel comfortable”. SF-03 emphasized the importance of building rapport before demanding performance: “Hang out for two to three months even if there are targets”. Meanwhile, SF-04 chose to foster confidence gradually by giving opportunities to take the stage: “The point is, give them the chance to start”.

The community approach remains the main asset for facilitators, even as the space for empowerment narrows. SF-05 fostered closeness through daily interactions, “Personally, I prefer to build communication or approach people by blending in”. SF-07 merged even further: “If the community drinks, I join in too, so that there’s no distance”. FK-03 highlighted the importance of linguistic diplomacy, “choosing the right language requires skill, it’s really important”, while SF-05 was sensitive to cultural characteristics that differed between regions, as in their experience in Toraja, which emphasized emotional closeness to the community. This relational resilience requires patience; SF-07 described how they rebuilt communication even after a conflict, “even if we fight with the BKM at night, the next day I’ll go to their house again to have coffee together”.

Team management has become a key skill developed by senior facilitators to maintain work cohesion. FK-04 relied on sensitivity to read people’s characters: “I can understand people’s characteristics and manage the team accordingly.” SF-01 delegated tasks based on capacity: “I assign according to colleagues’ capacities, both by location and specialization”. SF-05 maintained team members’ authority by giving personal admonishments rather than speaking to them in front of the team, while SF-03 asserted the fundamental principle of team management: “The key here is ethics and attitude”. This team management approach showed that work cohesion was maintained more through personal leadership and psychological sensitivity than through standard procedures, especially when team composition shrank owing to program changes.

Maintaining idealism appears strong through the rejection of gratuities and steadfast adherence to the principle of honesty. FK-03 declined the offer of money, citing advice from a senior, “I don’t want to take it because the amount is not significant”. SF-06 returned the envelope given by the KSM, “I still returned it” and emphasized the values he held, “Universal values: honesty and justice” as well as the principle of integrity: “One word and one deed, don’t say one thing and do another”. Facing conflict is an inseparable part of a facilitator’s work, whether conflicts with the community, within the team, or in the legal domain. SF-01 recalled a physical threat he once faced, “I even almost got stabbed, almost got attacked with a machete. The key is sincerity.” Conflict has even dragged facilitators into legal issues; FK-02 experienced problems “to the point of ending up at the police station”. SF-01 compared the two program eras: “In Kotaku, I was once summoned by the police”, as a result of targets that did not align with the quality of field implementation. Internal team conflict is also unavoidable; FK-07 admitted losing his temper with a member of his own facilitation team: “I once got so angry that I kicked [that team member] out of the group”, a rift with a fellow team member that was later resolved through a personal apology to restore working relations.

The facilitators’ hopes and reflections revolved around satisfaction, fatigue, and calls for system improvement. FK-04 took pride when the facilitators they mentored became independent: “There is pride when they are able to do it without me”. In contrast, SF-03 voiced collective exhaustion: “There is a lot of pessimism”. TA-03 hoped the system would return to its original design, “the system must go back to how the program was at the beginning”, while emphasizing the facilitator’s role as “needs to be strengthened as an agent of change”. Similarly, FK-03 placed facilitators as the “front line”, requiring adequate competence from the outset, so that expectations for improvement were directed toward restoring training and strengthening the substance of empowerment, not just physical completion.

The relationship between patterns of program change, the impact of knowledge loss, and the facilitators’ developed responses can be observed in Figure 6, which presents the flow of inter-categorical relations, along with the weighing of their coding references.

Figure 6 reads as co-occurrence between two categories: the dimensions of knowledge loss (rows) and the facilitator responses they provoke (columns). The dominant pathway runs from Loss of the Training System to Self-Learning (4 coding references), which confirms that the collapse of formal training pushed facilitators into self-directed learning, as TA-02 put it: “I learned those skills autodidactically.” Loss of the Training System also links strongly to Adaptation to the Orientation Shift (2), as does Impact on Community Independence (2), which explains why facilitators took over community roles; SF-01 said, “In Kotaku, indeed, we’re the ones who do everything.” Loss of Philosophy and Method spreads across Mentoring New Facilitators (2) and Maintaining Idealism (2), with a moderate link to Self-Learning (1), showing that the loss of philosophical inheritance is carried by informal guidance and personal commitment rather than by institutional mechanisms. The meaning of this pattern is a shift in the burden of adaptation from the system to the individual: facilitation now depends on personal initiative rather than structural program support. The coded links can be read as layered rather than simultaneous, though Figure 6 shows thematic co-occurrence, not chronology. Interview accounts suggest that self-learning arose first, as an immediate response to the absence of pre-service training; that role-takeover became more salient as fiscal-year disbursement deadlines took hold; and that informal mentoring and idealism-maintenance came later, once seniors recognized the regeneration gap. This ordering is an interpretive reading, not a demonstrated sequence, and would need longitudinal evidence to confirm.

Overall, the findings show that the burden of adaptation shifted from institutional support to individual facilitator initiative.

Taken together, the compression of time, training, and team capacity weakened the institutional conditions for substantive participation. Through Arnstein’s (1969) framework, reduced citizen involvement represents tokenism without a genuine transfer of power. Extending Cooke & Kothari’s (2001) tyranny of participation thesis, the findings show that tyranny can arise from both standardized participatory methods and the erosion of empowerment methods. Facilitators’ substitution for citizens reflects the tyranny of decision making, while bureaucratic targets overriding community needs reflect group tyranny. The deinstitutionalization of empowerment infrastructure therefore provides the structural conditions through which these forms of tyranny operate, adding an institutional dimension to the existing framework.

These findings are consistent with those of several previous studies. The decline in support for the poor due to physical orientation reinforces G​r​i​l​l​o​s​ ​(​2​0​1​7​) on budget allocation bias and S​a​g​u​i​n​ ​(​2​0​1​8​) on the failure of community-driven development to reach marginalized groups. The burden of pressure and self-taught pathways taken by facilitators echo the affective injury identified by J​a​k​i​m​o​w​ ​(​2​0​1​8​a​) and the negotiation of various program impossibilities identified by Jakimow (2018c). Value-based personal resilience aligns with the space for volunteer self-formation revealed by J​a​k​i​m​o​w​ ​&​ ​H​a​r​a​h​a​p​ ​(​2​0​1​6​). Weak collaboration and ambiguous roles in the implementation of Kotaku confirm Z​u​b​a​i​d​a​h​ ​e​t​ ​a​l​.​ ​(​2​0​2​3​), while the suboptimal practice of participatory planning is consistent with A​k​b​a​r​ ​e​t​ ​a​l​.​ ​(​2​0​2​0​).

However, there is a fundamental difference between the two. Previous studies generally captured beneficiaries, program outcomes, or policy structures at a specific time, whereas this study traced them diachronically from the perspective of facilitators as carriers of institutional memory across eras. Its new contribution lies in revealing the causal mechanism: a paradigm shift simultaneously dismantled the training system, learning-by-doing, and knowledge transfer, thereby diminishing participation. This is the “new face” of the tyranny of participation, which does not stem from oppressive participatory methods as posited by Cooke and Kothari (2001), but from the deinstitutionalization of empowerment infrastructure, a nuance that has not been captured in the previous literature. While G​r​i​l​l​o​s​ ​(​2​0​1​7​) and S​a​g​u​i​n​ ​(​2​0​1​8​) explained failure in terms of elite capture and allocation bias, this study added an explanation from the supply side, namely the collapse of institutional preconditions that previously kept participation substantive, thereby completing a picture that has thus far been fragmented. Facilitators' assessments diverged along two axes rather than uniformly favoring one era: the P2KP/PNPM period was regarded as superior on the depth-of-process axis (critical awareness, transformative agency, and community self-reliance), whereas the Kotaku period was viewed more favorably on the target-achievement axis (physical output, efficiency, and measurable performance indicators).

Read against the international CDD debate, this supply-side account complemented the predominantly demand-side explanations advanced in studies of the Philippines and India, where elite capture and allocation bias dominated the diagnosis (D​a​s​g​u​p​t​a​ ​&​ ​B​e​a​r​d​,​ ​2​0​0​7; F​r​i​t​z​e​n​,​ ​2​0​0​7; S​a​g​u​i​n​,​ ​2​0​1​8). The Indonesian case suggests, as an analytically transferable proposition, that even where capture is contained, participation could still hollow out if the institutional preconditions training, reflective method, and tiered mentoring were dismantled. As a conceptual rather than prescriptive implication, this account bears on debates among international donors and policy bodies. For institutions such as the World Bank, which seeded the original P2KP/PNPM architecture, and UN-Habitat’s slum-upgrading agenda, the case conceptually illustrates how prioritizing physical targets without protecting facilitator-development systems may shift program implementation toward procurement-oriented delivery. The Makassar case suggests that donor results frameworks may benefit from retaining process indicators (facilitator competence, learning-cycle completion, and community self-reliance) as co-equal with physical and absorption metrics and treated the empowerment infrastructure as a fundable line item rather than an assumed externality. These points are offered as analytically transferable propositions—claims about institutional mechanisms that may travel to settings sharing similar preconditions—rather than as generalizations from a single case; consistent with the Limitations noted below, the single-case Makassar design constrains statistical generalizability and invites testing in other cities.

Because the informants are facilitators and program actors, the account privileged a state-and-facilitation vantage point, and the inference about declining community self-reliance was necessarily mediated by their interpretation. The Indonesian literature on elite capture (A​s​p​i​n​a​l​l​,​ ​2​0​1​3; D​a​s​g​u​p​t​a​ ​&​ ​B​e​a​r​d​,​ ​2​0​0​7; F​r​i​t​z​e​n​,​ ​2​0​0​7) cautioned that grassroots power dynamics might diverge from facilitators’ perceptions. We therefore framed the community-self-reliance findings as situated professional assessments and flagged the incorporation of residents’, ward officials’, and local leaders’ voices as the necessary next step to test whether reduced participation was experienced as such at the grassroots.

Theoretically, this study extended the thesis of C​o​o​k​e​ ​&​ ​K​o​t​h​a​r​i​ ​(​2​0​0​1​) by adding an institutional dimension: tyranny may emerge from the erosion of empowerment methods rather than from their oppressive imposition, so that empowerment infrastructure is appropriately positioned as an explanatory variable in studies of participation. Practically, the findings called for the restoration of facilitator training systems, tiered coaching mechanisms, and multiyear learning cycles that once formed the core of P2KP, so that mentor capacity no longer relied on unequal self-taught initiative. Restoration must also protect the informal, relational spaces in which tacit facilitation knowledge and community trust were historically built. What H​u​s​n​i​y​a​h​ ​(​2​0​2​6​) theorizes as ngopi—an unhurried, trust-based practice of socializing over coffee through which Indonesians share and co-create knowledge—closely describes how P2KP/PNPM facilitators once transmitted facilitation knowledge and cultivated rapport through berbaur (integrating into community life), evening conversations, and informal coffee gatherings with residents, rather than through scheduled sessions alone. The target-driven compression of cycles foreclosed precisely this relational time, so that deinstitutionalization erased not only formal training but also the everyday, culturally grounded practices of knowledge co-creation; reinstating empowerment infrastructure therefore means safeguarding relational time, not only restoring a curriculum. At the policy level, program design needs to balance physical targets with empowerment process indicators, maintain regional contributions as an anchor of ownership, and make competence not merely diplomas or connections the primary recruitment criterion. The experience of DDUB disappearance, which made local governments “feel they have no stake”, along with the entry of aspirational funds that have piggybacked on the program, indicates that budgetary governance also determines the quality of participation. Therefore, fiscal instruments should be designed to strengthen, not displace, the logic of community needs.

The main strength of the study lies in its interpretive qualitative design and diachronic case study approach, which can capture hidden shifts behind official policy documents. Triangulation of sources across positions, method triangulation, member checking, and an audit trail strengthened trustworthiness, while deductive–inductive coding maintained the traceability of inferences from quotations to categories. The researcher’s reflexivity, given their close ties to the facilitation field, was managed transparently as a source of analytical sensitivity. However, the single case study in Makassar City limited transferability; therefore, the findings were not automatically applicable to cities with different dynamics. Dependence on facilitators’ memory reconstruction risks introducing retrospective bias and nostalgia for the PNPM era, while the absence of end-beneficiary perspectives and quantitative data on physical outcomes constrains full proof of causality. A related constraint is the evidential asymmetry across eras: because field observation could document only the contemporary Kotaku and post-Kotaku context, evidence on the earlier P2KP/PNPM period rests mainly on retrospective interviews and archival records rather than on direct observation.

This study leaves several gaps that open opportunities for further research. The voices of community beneficiaries have not yet been represented; therefore, future studies need to confirm the reduction of direct participation from the perspective of the recipients. Comparative multi-city studies could assess the analytical transferability of the identified mechanisms across comparable settings. Longitudinal research could track the impact of the loss of empowerment infrastructure on the long-term sustainability of physical outcomes. Finally, intervention studies examining the effectiveness of restoring training systems and learning cycles would bridge these findings toward empirically tested policy recommendations, while also strengthening the role of facilitators as agents of change, not merely project executors.

This study shows that the transition from P2KP/PNPM to Kotaku altered the institutional conditions of community facilitation, particularly facilitator training, reflective methods, learning-by-doing, and intergenerational knowledge transfer. These changes shifted the burden of adaptation toward individual facilitators, weakened community self-reliance, and contributed to increasingly procedural forms of participation. The study extends the tyranny-of-participation debate by identifying the erosion of empowerment infrastructure as an institutional mechanism through which substantive participation may be weakened.

For comparable urban community-driven development settings, the findings suggest that physical targets should be accompanied by process indicators and sustained investment in facilitator development. The single-case design, reliance on retrospective accounts, and absence of beneficiary perspectives limit the scope of the conclusions. Future research should incorporate beneficiary perspectives and conduct comparative multi-city studies to examine whether the identified deinstitutionalization mechanisms also operate in comparable settings. Longitudinal studies could assess the durability and longer-term effects of physical interventions, while intervention studies could evaluate whether restoring facilitator training systems and learning cycles improves program outcomes and helps translate these findings into empirically supported policy recommendations.