A Participatory and Performance-Based Evaluation Model for Adaptive Local Governance

Regional development planning in decentralized systems increasingly demands not only institutional autonomy but also strategic coherence, data integration, and citizen participation [1, 2]. In Indonesia, local governments are legally empowered to design their own development strategies (RPJMD), annual programs (RKPD), and budgets (APBD). However, persistent misalignments among these instruments and weak participatory mechanisms hinder the realization of Sustainable Development Goals (SDGs) [3-5].

Among existing evaluation tools, the CIPP (Context, Input, Process, Product) model is widely used for its structured and formative evaluation framework [6, 7]. Introduced by Stufflebeam in the 1970s, it evaluates not only outcomes but also the contextual and procedural factors shaping program success [8-10]. However, critics argue that in dynamic policy environments—particularly decentralized governance—the CIPP model often lacks integration with digital platforms, real-time feedback systems, and participatory processes [11-13]. Its linear and static structure limits applicability in fragmented planning environments, where coordination, responsiveness, and inclusivity are essential [14, 15].

In addition to CIPP, several other evaluation frameworks—such as the CIPPO model, the Kirkpatrick model, and Logic Models—have been widely applied across education, training, and development sectors. CIPPO extends CIPP by adding an “Outcomes” component but remains limited in addressing participatory governance and digital adaptability. The Kirkpatrick model evaluates impact through four levels (reaction, learning, behavior, and results), yet it lacks policy integration and stakeholder feedback mechanisms essential in public sector planning. Logic Models, though visually intuitive, typically follow linear logic and offer limited flexibility for real-time corrections or feedback loops in multi-stakeholder environments. These limitations reveal a gap in models capable of integrating adaptive feedback, data-driven planning, and multi-stakeholder participation in decentralized governance contexts [16-18].

These limitations are evident in regional contexts such as Dairi Regency, where program fragmentation, underutilization of digital tools like SIPD, and procedural citizen engagement continue to challenge effective planning and evaluation [16-18]. Recent studies suggest that modern governance requires evaluation frameworks that are not only evidence-based but also adaptable and participatory—able to integrate stakeholder feedback and support plan–budget coherence throughout the policy cycle [19-21].

As digital platforms proliferate, systems like Indonesia’s SIPD provide real-time tracking of financial and development indicators. These tools have proven effective in increasing transparency and coordination when supported by institutional capacity and participatory governance [22, 23]. Simultaneously, participatory planning mechanisms—such as citizen scorecards, community forums, and multi-actor coalitions—have gained traction for enhancing legitimacy, responsiveness, and policy relevance [6, 24]. Yet, few evaluation models successfully merge digital systems and participatory structures into a cohesive planning framework [25, 26].

To address this gap, this study introduces the EPABKD framework (Evaluative, Participatory, Adaptive Model Based on Performance and Data), which restructures the foundational logic of CIPP by embedding five dynamic components. In its operationalized form within Dairi Regency, EPABKD takes shape as the D-SMART model (Dairi – Sustainable Monitoring, Alignment, and Result-Based Transformation)—a planning-evaluation hybrid designed for scalability in decentralized, digitally diverse, and SDG-oriented governance environments [27-29].

To guide this inquiry, the study is structured around the following research questions:

1. What are the institutional, participatory, and technological challenges that limit the effectiveness of regional development planning in decentralized settings such as Dairi Regency?
2. How can the classical CIPP evaluation framework be expanded to integrate participatory governance, digital monitoring, and SDG-oriented planning?
3. What are the practical impacts and theoretical contributions of the D-SMART model in enhancing planning coherence, performance accountability, and citizen participation at the local level?

3.1 Data sources and analytical basis

Despite the RPJMD’s stated commitment to digital transformation, inclusive infrastructure, and participatory development, APBD execution revealed structural incoherence. A stark example lies in the mismatch between fiscal targets and actual revenue realization (Figure 2 and Table 3).

This reactive revenue pattern illustrates the absence of a fiscal-policy alignment mechanism, reinforcing the need for Smart Alignment, a core component of the D-SMART model. Moreover, capital expenditures (Belanja Modal) remained marginal relative to operational spending, undermining infrastructure and innovation ambitions.

Table 3. Realization of local revenues against RPJMD targets

2.png

Figure 2. Revenue composition of APBD Dairi 2024 regional retribution far exceeded its target, while core revenue components underperformed

3.2 Deficits in participatory and adaptive planning

Forums such as Musrenbang remained largely procedural, with minimal conversion of community proposals into actionable budget allocations. The TEPD (Regional Evaluation Team) existed informally and lacked legal mandate or resources, while mid-year evaluations remained budget-centric rather than policy-responsive. SIPD tools were underutilized, primarily functioning as a reporting interface rather than a dynamic planning or monitoring platform as shown in Table 4.

Table 4. Governance challenges identified in stakeholder interviews

3.3 Component-level validation of the D-SMART model

1. Data-driven planning

Sectoral data remained siloed, non-standardized, and inaccessible for planning. Most departments continued to rely on static formats such as spreadsheets or PDFs, while SIPD dashboards remained significantly underutilized (Figure 3).

3.png

Figure 3. Simulated SIPD utilization by OPD in Dairi Regency planning agency showed moderate usage; most line departments lagged

D-SMART addresses this through integrated sector dashboards and centralized access via SIPD-linked tools.

2. Smart alignment

Using the Program Alignment Index (PAI):

$\begin{gathered} { Program\ Alignment\ Index\ (PAI) }= \left(\frac{ { Number\ of\ Programs\ funded\ in\ APBD }}{ { Total\ Strategic\ Programs\ in\ RPJMD }}\right) \times 100\end{gathered}$

Example: 6 of 10 RPJMD programs were funded in 2024 PAI = 60% (Table 5).

Table 5. Sample alignment between RPJMD and APBD

40% of RPJMD strategic priorities lacked budget execution—highlighting urgent need for D-SMART’s alignment dashboard.

3. Multi-stakeholder participation

Participation lacked continuity and authority. D-SMART introduces a Stakeholder Evaluation Board with formal quarterly reporting and input rights.

4. Adaptive feedback loop

No structured process exists for intra-year program revision. D-SMART embeds Triannual Digital Review Cycles via SIPD dashboards, enabling data-based mid-cycle corrections.

5. Result-based management

Current indicators focus on activity counts (e.g., “number of trainings”). D-SMART emphasizes outcome-linked SDG indicators such as reduction in maternal mortality or improvements in youth employment.

3.4 Validation outcomes of the D-SMART model

The D-SMART model underwent multi-method validation as described in Section 2.5, with each method contributing distinct evidence of the model’s utility and robustness.

1. Comparative assessment

In the head-to-head comparison with the classical CIPP model, D-SMART scored higher across all six evaluative dimensions (Table 6):

Table 6. Comparative performance: D-SMART vs. classical CIPP

These results validate D-SMART’s superiority in addressing real-world governance challenges through structured innovation.

2. Expert panel evaluation

Five evaluators (three local policy practitioners, two national planning experts) rated the model's:

• Feasibility: 4.6 / 5
• Clarity: 4.8 / 5
• Scalability: 4.4 / 5

Qualitative feedback highlighted the model’s ease of integration into SIPD and its potential for broader use across Indonesian districts.

3. Simulation Exercises

Simulation of Dairi’s 2022–2023 planning cycle using D-SMART demonstrated:

• 14% improvement in RPJMD–APBD alignment,
• Integration of 9 previously excluded community proposals,
• Dynamic mid-year adjustment to 6 underperforming programs.

4. Sensitivity checks

Cross-checking budget execution data revealed that D-SMART’s adaptive review cycles could have identified and corrected policy lags at least one quarter earlier than current mechanisms.

3.5 Discussion

The classical CIPP (Context, Input, Process, Product) model, though foundational in evaluation theory, struggles to address the complexities of decentralized, SDG-oriented governance. Its limitations—static sequencing, weak stakeholder integration, and lack of digital feedback—are increasingly incompatible with the adaptive demands of modern policy environments [28, 33, 34].

This study positions the D-SMART model as a strategic upgrade, retaining CIPP’s structured logic while embedding digital tools, participatory mechanisms, and real-time performance tracking. Each CIPP element is re-engineered: Context integrates live sectoral data; Input aligns strategic and fiscal codes; Process introduces digital review cycles; Product emphasizes SDG-linked outcomes; and a fifth domain—Governance—institutionalizes stakeholder inclusion (Table 7).

These innovations position D-SMART as a hybrid governance tool that bridges evaluation, strategic planning, and participatory oversight. Unlike prior adaptations (e.g., CIPPO), D-SMART fully integrates data, decentralization, and democracy, making it highly relevant for SDG localization and responsive governance [30, 31, 35].

Table 7. Comparative framework: classical CIPP vs. D-SMART innovations

3.6 Comparative review of evaluation models

While the classical CIPP framework remains a widely used foundation for program evaluation in governance contexts, alternative models such as CIPPO, the Kirkpatrick Model, and Logic Models also offer unique evaluative logics that merit discussion.

The CIPPO Model, an extension of CIPP, introduces an additional "Outcomes" component to better reflect long-term impact. However, it remains largely linear and does not explicitly embed digital adaptability or participatory structures, which are central to decentralized governance.

The Kirkpatrick Model, originally designed for training evaluation, focuses on four levels: Reaction, Learning, Behavior, and Results. Although robust for capacity-building programs, it lacks policy alignment tools and does not emphasize citizen participation or data integration—key elements in the D-SMART model.

The Logic Model visualizes causal chains between inputs, activities, outputs, and outcomes. It is widely used in development evaluation but is often criticized for its static structure and limited applicability in complex, multi-stakeholder environments. D-SMART, by contrast, emphasizes adaptive feedback loops and cyclical responsiveness.

Thus, D-SMART fills a crucial gap by integrating participatory governance, digital platforms, and SDG alignment within an adaptive and performance-oriented framework (Table 8).

Table 8. Summarizes how D-SMART compares across key criteria