FIRM SOCIAL CAPITAL AND CORPORATE RISK-TAKING: CAUSAL MACHINE LEARNING EVIDENCE FROM VIETNAM
Anh Nguyen H. , Thanh Ngo P., Truc To T. T., Dong Nguyen N.A.
University of Economics and Law and Vietnam National University,
Ho Chi Minh City, Vietnam
Abstract: In transition economies where, formal institutions remain underdeveloped and information asymmetry is pervasive, relational networks substitute for market mechanisms in allocating resources and managing uncertainty. This study investigates the impact of firm-level social capital on corporate risk-taking behavior among 338 non-financial companies listed on the Ho Chi Minh Stock Exchange (HOSE) from 2019 to 2024. Our contribution to the literature is twofold. First, we move beyond traditional binary metrics to build a comprehensive, four-dimensional formative index of corporate social capital. This index integrates an automated CSR disclosure score aligned with GRI standards, a biographical locus measure reflecting the CEO’s social network, political connections, and bank connectivity. Second, we address the challenge of causal identification by deploying a dual strategy. Specifically, we pair Two-way Fixed Effects with the Panel Double Machine Learning method to eliminate nonlinear confounding and ensure the robustness of our causal inferences. The empirical evidence suggests that social capital functions as a strategic risk management toolkit with dual roles. While political and bank ties act as institutional buffers that dampen earnings volatility, they simultaneously act as a catalyst for financial expansion. These ties increase leverage capacity by 14.6 percentage points for every standard deviation increase in connectivity. Furthermore, CEO networks and political ties drive a shift toward precautionary cash holdings. Notably, the Panel Double Machine Learning approach uncovers significant effects overlooked by linear models, confirming the material importance of nonlinearity in relational networks. By dismantling the binary view of social capital as either a risk-booster or a risk-reducer, this study positions it as a sophisticated portfolio that stabilizes performance while broadening financial flexibility.
Keywords: Firm social capital, corporate risk-taking, Panel Double Machine Learning, automated content analysis, Vietnam, transition economy
VOLUME 10 ISSUE 04 2026: 1 – 20