This paper aims to examine how institutional voids, Environmental, Social and Governance (ESG) pressures and sustainability paradoxes intersect in Latin American global value chains (GVCs). It develops a multilevel framework to explain how institutional void configurations filter global ESG demands and generate paradoxical outcome patterns.
The study adopts a sequential explanatory mixed-methods design. The quantitative component uses an unbalanced firm-level panel of listed companies operating in Brazil, Mexico, Chile, Colombia and Argentina for 2021–2024 (n = 792 firm-year observations). It combines Refinitiv ESG indicators with World Bank governance measures and UNCTAD trade data. Multilevel regression models and robustness checks are employed to estimate the effects of institutional voids and ESG pressures on the intensity of the sustainability paradox.
Institutional voids significantly intensify the sustainability paradox, whereas ESG pressures reduce it. However, this mitigating effect of ESG pressures is substantially weakened in high-void contexts, particularly in mining-related value chains. The results support the idea of “institutional filtering”: weak institutional environments filter and partially neutralize externally imposed ESG standards, leading to persistent tensions between sustainability claims and everyday practice.
The evidence is limited to five Latin American countries and to listed firms, which limits generalizability to the vast informal economy that characterizes much of Latin America. Since non-listed SMEs face even steeper resource constraints and opaque institutional pressures, the paradox intensity reported here likely represents a conservative lower-bound estimate.
The paper advances paradox theory and critical international business by conceptualizing institutional voids as filtering mechanisms that condition the impact of ESG pressures on sustainability outcomes in emerging markets. It integrates paradox theory, institutional voids and GVC governance in a Latin American context, using a mixed-methods, multilevel research design.
