Assessing Southeast Asia countries' potential in the semiconductor supply chain: an objectively weighting multi-criteria decision-making approach

The paper frames semiconductors as foundational to modern electronics and global economic growth, noting the industry’s highly globalized and interdependent supply chain spanning design, fabrication, assembly, test, and distribution. It highlights the sector’s vulnerability to disruptions and the central role of leading manufacturers (e.g., Intel, Samsung, TSMC) operating capital-intensive fabs worldwide. Within this context, Southeast Asia has gained strategic relevance, with countries like Malaysia, Singapore, and the Philippines attracting investment and developing capabilities in manufacturing, design, and supporting industries. The study identifies a research gap: despite extensive work on global semiconductor supply chains, there is limited, systematic analysis of Southeast Asia’s specific readiness and roles, especially amid recent geopolitical and economic shifts. Research objective: to assess the ability and opportunity of Southeast Asian countries to integrate into the semiconductor supply chain in the near future by determining relevant factors and indicators, and evaluating country potential using an integrated, objectively weighted MCDM approach (CRITIC for objective weighting; TOPSIS for ranking by proximity to ideal solutions). The authors justify CRITIC for its data-driven, bias-minimizing weighting based on variance and inter-criteria correlation, and TOPSIS for efficiently handling multiple criteria and mixed beneficial/non-beneficial attributes to produce transparent rank-order results.

The literature review covers two strands. 1) Semiconductor supply chain studies: Prior research addresses supply chain structure, coordination among manufacturers, equipment/material suppliers, distributors, and end-users; risk management and resilience for complex, long lead-time processes; impacts of disruptions (natural disasters, geopolitical tensions, demand volatility); cross-border collaboration; sustainability (energy efficiency, waste management, recycling); and technological change (IoT, AI, blockchain) shaping operations. Empirical and conceptual works investigate shortages and bottlenecks, mitigation strategies, and the role of government policy and regional cooperation. For Southeast Asia specifically, studies examine comparative country capabilities, competitiveness, strategic location, investment climate, skilled labor, domestic market potential, technological capabilities, and policy incentives, alongside regional integration and partnerships (e.g., ASEAN initiatives, industry–government collaboration). 2) MCDM trends and applications: MCDM has evolved with AI integration, big data analytics, multi-criteria optimization, sustainability-oriented decision support, group and hierarchical decision structures, and real-time applications. Fuzzy extensions (e.g., fuzzy AHP, DEMATEL, spherical/pythagorean fuzzy methods) help model uncertainty and subjectivity. In semiconductors, MCDM has been used for facility location under geopolitical shifts, resilience assessment, supplier selection, sustainability reporting prioritization, blockchain readiness to mitigate bullwhip effect, green supplier evaluation, and mapping interrelationships for sustainable transitions. Despite broad usage, a gap exists in applying MCDM specifically to evaluate Southeast Asian countries’ semiconductor supply chain capabilities, which this study addresses by integrating CRITIC and TOPSIS for a tailored, objective regional assessment.

The study proposes a three-phase MCDM framework integrating CRITIC for objective indicator weighting and TOPSIS for ranking alternatives. Phase I (Preparation): Conduct systematic literature review across Scopus, Web of Science, Google Scholar to identify relevant indicators; collect quantitative data from recognized international sources (World Bank, UNESCO Institute for Statistics, WIPO, UNIDO, CRED/EM-DAT), using the most recent available country data; construct the m×n decision matrix of m countries (alternatives) by n indicators (criteria). Fifteen indicators are organized under Stability, R&D, Technology Infrastructure, and Economic Efficiency; some are benefit and some non-benefit. Phase II (Objective weighting via CRITIC): - Normalize raw data using min–max scaling to [0,1] to enable cross-criterion comparability, preserving relative distributions. - Compute, for each indicator j, the standard deviation σj (contrast intensity) and pairwise Pearson correlation coefficients rjk among indicators. - Derive information content Ij = Σk (1 − |rjk|); compute weights wj = Ij / Σi Ii, giving higher weight to criteria with greater variability and lower redundancy. Observed variability shows, for example, high dispersion for natural disaster occurrences; lower dispersion for government effectiveness, GDP, e-participation, ICT development, and logistics. Correlation analysis reveals strong positive correlation between logistics performance and competitive industrial performance, and strong negative correlation between patents and logistics performance. Phase III (Ranking via TOPSIS): - Apply vector normalization to construct C and a weighted normalized matrix T = [tij] with weights wj. - Identify the ideal (best) and negative-ideal (worst) solutions across benefit and non-benefit indicators. - Compute Euclidean distances D+ (to ideal) and D− (to negative ideal) for each country, and the overall score (closeness) D* = D− + D+. Higher overall scores indicate better alternatives (closer to ideal and farther from the negative ideal). The approach yields transparent, reproducible rankings, accommodates mixed indicator types, and minimizes subjective bias through objective weighting.

- Data and indicators: 10 Southeast Asian countries evaluated (Brunei Darussalam, Cambodia, Indonesia, Lao PDR, Malaysia, Myanmar, Philippines, Singapore, Thailand, Vietnam) across 15 indicators spanning Stability (Global Peace Index; Natural disaster occurrences; Government effectiveness; Rule of law), R&D (Graduates in science and engineering; GERD; Human Capital Index; Patent applications), Economic Efficiency (High-tech exports; ICT goods exports; GDP), and Technology Infrastructure (E-participation Index; ICT Development Index; Logistics Performance Index; Competitive Industrial Performance Index). - Indicator variability and correlations (CRITIC inputs): Natural disaster occurrences exhibit the highest standard deviation (≈0.490), indicating greatest dispersion. Lower variability observed for government effectiveness, GDP, e-participation, ICT development, and logistics (standard deviations ≈0.306–0.331). Strongest positive correlation: Logistics performance (114) and Competitive Industrial Performance (115), r ≈ 0.965. Strongest negative correlation: Patents (18) and Logistics performance (114), r ≈ −0.892. - Information content and weights (selected values from CRITIC): Patents applications 0.1605; Global Occurrences from Natural Disasters 0.1245; Graduates in Science and Engineering 0.0991; GDP 0.0704; ICT goods exports 0.0667; High-tech exports 0.0636; Global Peace Index 0.0611; Gross expenditure on R&D 0.0495; ICT Development Index 0.0494; Rule of Law 0.0475; Human Capital Index 0.0473; Government effectiveness 0.0416; E-participation Index 0.0426; Logistics Performance Index 0.0383; Competitive Industrial Performance Index 0.0377. High-weight drivers (>10%): Patents, Natural disasters, Graduates in S&E. Moderate (5–10%): GDP, ICT goods exports, High-tech exports, Global Peace Index. Lower (<5%): Rule of law, Human capital, Government effectiveness, E-participation, ICT development, GERD (borderline), Logistics, Competitive industrial performance. - Country ranking (TOPSIS results): Singapore achieves the highest overall score (closest to ideal), indicating strongest opportunity/readiness for semiconductor supply chain participation, supported by advanced infrastructure, skilled workforce, strong governance, and policy support. Next highest performers: Brunei Darussalam and Lao PDR. Mid-tier: Malaysia, Vietnam, Thailand (each with notable strengths—Malaysia’s established semiconductor base; Vietnam’s growth, labor cost advantage and improving infrastructure). Lower performers: Indonesia and the Philippines, reflecting challenges such as infrastructure gaps and bureaucratic hurdles, despite large markets and skilled labor potentials. - Distances (qualitative): Singapore has the smallest distance to the ideal solution; the Philippines has the largest distance to the ideal and the smallest distance to the negative ideal; Indonesia has the largest distance to the negative ideal (i.e., farthest from worst case). - Policy and investment implications: High weights on innovation- and resilience-related indicators emphasize investment in R&D, IP ecosystem, and STEM talent; disaster resilience and stability matter materially. Moderate weights on GDP and export structures indicate importance but not sufficiency without innovation capacity. Logistics and industrial performance, while necessary, are less determinant singly in this model.

The findings align with recent analyses emphasizing Singapore’s central role in the regional semiconductor ecosystem, particularly in ATP and OSAT, driven by advanced infrastructure and supportive policies. The study extends the literature by identifying Brunei Darussalam and Lao PDR as emerging potentials—countries less emphasized in prior work—suggesting opportunities from geographic positioning and evolving economic frameworks, conditional on ecosystem development (infrastructure, skills, policy incentives). Consistent with other studies, Malaysia remains attractive given its established base and incentives, while Vietnam’s increasing integration is underpinned by competitive labor, infrastructure improvements, and targeted policy support. Divergences appear for the Philippines and Indonesia: while some network analyses show increasing participation, this study’s indicator-based assessment highlights persistent infrastructural and bureaucratic constraints affecting their readiness, suggesting country-specific structural reforms are needed for fuller integration. The broader implication, including COVID-19 impacts, is the necessity of robust resilience planning and diversified regional supply chain strategies. For managers and policymakers: prioritize STEM and R&D capacity building (especially in emerging countries), continue infrastructure and digital readiness investments, and pursue collaborative regional initiatives to diversify and harden supply chains against shocks.

The study presents an objective, replicable framework integrating CRITIC and TOPSIS to assess Southeast Asian countries’ readiness and opportunities in the semiconductor supply chain. By weighting 15 indicators based on variance and inter-criteria correlation, and ranking countries by proximity to ideal solutions, the approach yields transparent, quantitative insights into national strengths and gaps. Singapore emerges as the leading performer; Brunei Darussalam and Lao PDR show promising potential; Malaysia, Vietnam, and Thailand form a competitive mid-tier; Indonesia and the Philippines face more pronounced challenges. Contributions include: a rigorous, data-driven regional assessment; actionable guidance for policymakers and investors on priority levers (innovation capacity, STEM talent, resilience, digital infrastructure); and a methodological template applicable to other regions/sectors. Future research directions: adopt dynamic MCDM models to capture evolving conditions; integrate qualitative assessments with quantitative indicators; conduct longitudinal and cross-regional comparative studies to benchmark strategies and track progress over time.

- Data timeliness and consistency: International databases have publication lags and varying country reporting standards; the study uses the most recent available data, potentially mixing years across indicators and countries. - Indicator set and weighting dependence: Results depend on the selected 15 indicators and CRITIC-derived weights; relevance may evolve, requiring periodic updates. - Static and quantitative orientation: The approach relies on static snapshots and may not capture dynamic shifts or qualitative nuances (e.g., policy changes in progress, firm-level investment pipelines). - TOPSIS simplification: Use of ideal/anti-ideal reference points may oversimplify complex decision contexts and interactions among criteria. - Regional specificity: Findings are tailored to Southeast Asia’s semiconductor context and may not generalize directly to other regions or industries without adaptation.