Abstract Amid a global resurgence of industrial policy, the "manufacturing delusion"—the belief that expanding manufacturing capacity directly fosters innovation and growth—warrants rigorous empirical scrutiny. This paper utilizes a comprehensive research framework to test the nuanced relationship between manufacturing activity and innovation outcomes. Using a multi-pronged econometric strategy encompassing cross-country panel data analysis, quasi-experimental methods based on China's WTO accession, and instrumental variable approaches, we test seven core hypotheses. Our synthesized results indicate that the link between manufacturing and innovation is not automatic. The analysis shows that after controlling for institutional quality and intangible investments, the direct effect of manufacturing growth on frontier innovation is attenuated. We find that manufacturing expansion is more strongly associated with process innovation than with product innovation, and that positive innovation effects are contingent upon reaching critical thresholds in R&D intensity, supplier ecosystem depth, and the availability of skilled producer services. The findings support the view that building innovation capability is distinct from expanding production capacity, offering critical evidence for designing more effective, evidence-based industrial policies.Keywords: Manufacturing Delusion, Industrial Policy, Innovation, Econometrics, Panel Data, Causal Inference, Technology Spillovers