Innovation ecosystem and corruption: evidence from the global innovation index

Purpose This study investigates the complex relationship between a country's national innovation system and levels of corruption, focusing on both corruption control (CC) and corruption perception. Design/methodology/approach Using Global Innovation Index data (2013–2022) and machine learning algorithms (random forest, partial dependence plots), the study evaluates the role of innovation-related factors. Findings Political stability and regulatory quality are the strongest predictors of CC and perception. R&D investments are particularly important in upper-middle-income countries, while infrastructure plays a key role in low-income countries. Practical implications Policymakers should foster balanced innovation systems, strengthen governance, and adapt anti-corruption strategies to national income levels. Originality/value This is among the first studies applying machine learning to explore the innovation–corruption nexus across income contexts, offering novel insights for governance and sustainable development.