Digital twins and AI-powered predictive analytics enhance supply-chain resilience by boosting firms' anticipation and recovery capabilities, with the resilience payoff larger under greater supply-chain complexity; three distinct capability configurations — including an anticipation-led pathway — suffice for high resilience in Taiwan's electronics sector.
Digital Twin-Enabled Supply Chain Resilience: Anticipation and Recovery Capabilities in Taiwan's Electronics Manufacturing Sector
openalex
correlational
low evidence
7/10 relevance
Summary only
summary available; pdf_status=paywall
Source
PDF
Digital twins and AI-driven predictive analytics strengthen supply-chain anticipation and recovery capabilities, which together improve resilience performance, and supply-chain complexity amplifies these technology-to-capability effects, with three sufficient capability configurations identified for high resilience.
This study examines how Industry 4.0 technologies activate distinct resilience capability mechanisms in supply chains. Drawing on dynamic capabilities theory, we position digital twin capability and AI-driven predictive analytics as technology-enabled capabilities that strengthen supply chain anticipation capability and recovery capability, which together enhance supply chain resilience performance. Using survey data from 289 supply chain executives in Taiwan’s electronics sector and PLS-SEM, all six direct-effect hypotheses are supported, while supply chain complexity significantly strengthens the four technology-to-capability relationships. Complementary fsQCA based on a 98-case complete-data sub-sample identifies three sufficient configurations for high resilience performance, including anticipation-dominant and capability-driven pathways. The findings extend proactive–reactive resilience research by clarifying the temporal and functional distinction between anticipation and recovery capabilities and by showing how differentiated digital investments support resilience under supply chain complexity.
Summary
Main Finding
Digital twin capability and AI-driven predictive analytics each strengthen supply chain anticipation and recovery capabilities; those capabilities together improve supply chain resilience performance. Supply chain complexity amplifies the positive effects of the technologies on capabilities, and multiple distinct technology–capability configurations (including anticipation-dominant and capability-driven pathways) can achieve high resilience.
Key Points
- Theoretical framing: uses dynamic capabilities theory to treat Industry 4.0 technologies as technology-enabled capabilities that activate higher-level resilience capabilities.
- Technologies studied: digital twin capability and AI-driven predictive analytics.
- Resilience capabilities: distinguished temporally and functionally into anticipation capability (proactive) and recovery capability (reactive).
- Main empirical results:
- PLS-SEM on survey data (n = 289, Taiwan electronics supply chains) supports all six hypothesized direct effects (technologies → anticipation/recovery → resilience performance).
- Supply chain complexity significantly strengthens the four technology → capability relationships (i.e., complexity is a positive moderator).
- fsQCA on a 98-case complete-data sub-sample reveals three sufficient configurations for high resilience, including anticipation-dominant and capability-driven pathways (showing multiple effective combinatory routes).
- Conceptual contribution: clarifies temporal/functional distinction between anticipation and recovery capabilities and shows differentiated digital investments can be strategically aligned to those capabilities under varying complexity.
Data & Methods
- Sample: survey of 289 supply chain executives in Taiwan’s electronics sector.
- Constructs: digital twin capability; AI-driven predictive analytics; anticipation capability; recovery capability; supply chain resilience performance; supply chain complexity (moderator).
- Quantitative methods:
- PLS-SEM to test direct effects and moderation hypotheses (all six direct-effect hypotheses supported; significant positive moderation by complexity for the four tech→capability links).
- Fuzzy-set Qualitative Comparative Analysis (fsQCA) on a 98-case complete-data sub-sample to identify sufficient configurational pathways to high resilience (three solutions found).
- Limitations:
- Cross-sectional survey, single sector and country (Taiwan electronics) — limits causal inference and generalizability.
- Self-reported executive data; potential common-method bias (not detailed here).
- fsQCA sample smaller (98) and exploratory regarding pathway generality.
Implications for AI Economics
- Investment targeting and complementarities:
- Firms should treat digital twins and AI predictive analytics as complementary but distinct investments: allocate resources to align each technology with the resilience capability (anticipation vs recovery) the firm most needs.
- Returns to these technologies increase with supply chain complexity, suggesting higher willingness-to-pay (and faster adoption) for AI-enabled tools in complex supply networks.
- Strategic adoption and heterogeneity:
- Multiple effective configurations mean there is no one-size-fits-all optimal portfolio; firms can achieve resilience through different technology–capability mixes depending on context and strategic priorities.
- Economic models of AI adoption should incorporate moderating effects of task/organizational complexity when predicting diffusion and firm-level ROI.
- Risk pricing and market effects:
- As AI and digital-twin investments materially reduce supply-chain disruption costs, insurance pricing, financing terms, and supply-chain contracting may shift—firms that adopt such technologies could obtain lower risk premia.
- Competitive dynamics: early adopters in complex industries may capture outsized resilience advantages, altering market structure and entry barriers.
- Policy and ecosystem considerations:
- Public support (subsidies, standards, data-sharing platforms) may be justified in sectors where complexity magnifies social value of AI-enabled resilience.
- Interoperability and data governance lower adoption frictions and can increase the social returns of digital-twin and AI investments.
- Research directions for AI economics:
- Quantify ROI over time (longitudinal studies) and generalize across sectors/countries.
- Model how supply-chain complexity and network effects shape equilibrium adoption of AI tools and consequent macro-level productivity and resilience outcomes.
- Evaluate distributional effects (which firms/countries benefit most) and potential second-order effects on trade, insurance markets, and systemic risk.
Assessment
Paper Typecorrelational
Evidence Strengthlow — Findings are based on same-source, cross-sectional self-reported survey data and correlational methods, which leave results vulnerable to common-method bias, reverse causality and omitted-variable/endogeneity concerns; the fsQCA adds configurational insight but does not address causal identification.
Methods Rigormedium — The study uses appropriate and complementary methods (PLS-SEM for latent constructs and moderation tests; fsQCA for causal complexity/configurations) and a reasonable sample size, but rigor is limited by cross-sectional design, potential measurement/sampling biases, and no reported use of longitudinal data, instruments, or robustness checks that would strengthen causal inference.
SampleCross-sectional survey of 289 supply-chain executives/managers in Taiwan's electronics sector; analysis includes full-sample PLS-SEM and a 98-case complete-data subsample used for fuzzy-set QCA.
Themesadoption org_design
IdentificationCross-sectional survey of 289 supply-chain executives in Taiwan's electronics sector analyzed with PLS-SEM to test theorized direct and moderating effects (digital twin and AI-driven predictive analytics → anticipation and recovery capabilities → resilience) and complementary fsQCA on a 98-case complete-data subsample to identify sufficient configurations; no experimental or quasi-experimental source of exogenous variation, so causal claims rely on theory and statistical associations rather than identification from plausibly exogenous shocks or instruments.
GeneralizabilitySingle country: Taiwan — institutional and market conditions may differ elsewhere, Single industry: electronics sector — findings may not transfer to services or other manufacturing subsectors, Executive self-reports: perceptual measures may not map directly to objective performance, Cross-sectional data: cannot capture temporal dynamics or causal directionality, Potential non-random sample and moderate sample size limit population representativeness
Claims (11)
| Claim | Direction | Outcome | Confidence & Evidence | Details |
|---|---|---|---|---|
| Digital twin capability strengthens supply chain anticipation capability. Organizational Efficiency | positive | supply chain anticipation capability |
Reading fidelity
high
Study strength
medium
|
n=289
|
| AI-driven predictive analytics strengthens supply chain anticipation capability. Organizational Efficiency | positive | supply chain anticipation capability |
Reading fidelity
high
Study strength
medium
|
n=289
|
| Digital twin capability strengthens supply chain recovery capability. Organizational Efficiency | positive | supply chain recovery capability |
Reading fidelity
high
Study strength
medium
|
n=289
|
| AI-driven predictive analytics strengthens supply chain recovery capability. Organizational Efficiency | positive | supply chain recovery capability |
Reading fidelity
high
Study strength
medium
|
n=289
|
| Supply chain anticipation capability improves supply chain resilience performance. Organizational Efficiency | positive | supply chain resilience performance |
Reading fidelity
high
Study strength
medium
|
n=289
|
| Supply chain recovery capability improves supply chain resilience performance. Organizational Efficiency | positive | supply chain resilience performance |
Reading fidelity
high
Study strength
medium
|
n=289
|
| All six direct-effect hypotheses in the model were supported by the PLS-SEM analysis. Organizational Efficiency | positive | direct-effect paths in the structural model (technology→capability; capability→resilience) |
Reading fidelity
high
Study strength
medium
|
n=289
|
| Supply chain complexity significantly strengthens (i.e., positively moderates) the four technology-to-capability relationships. Organizational Efficiency | positive | strength of technology-to-capability relationships (moderation by supply chain complexity) |
Reading fidelity
high
Study strength
medium
|
n=289
|
| FsQCA on a 98-case complete-data subsample identifies three sufficient configurations for high supply chain resilience performance, including anticipation-dominant and capability-driven pathways. Organizational Efficiency | positive | high supply chain resilience performance (configurational sufficiency) |
Reading fidelity
high
Study strength
medium
|
n=98
|
| The study used survey data from 289 supply chain executives in Taiwan’s electronics sector and employed PLS-SEM as the primary quantitative analysis method. Other | null_result | methodological details (sample and analysis technique) |
Reading fidelity
high
Study strength
high
|
n=289
|
| The findings clarify the temporal and functional distinction between anticipation and recovery capabilities and show how differentiated digital investments support resilience under supply chain complexity, thereby extending proactive–reactive resilience research. Organizational Efficiency | positive | conceptual distinction between anticipation and recovery capabilities; guidance on digital investment strategies under complexity |
Reading fidelity
medium
Study strength
medium
|
n=289
|