This study investigates the relationship between Enterprise Risk Management (ERM) implementation, agency costs, and financial performance volatility among publicly listed companies on the Indonesia Stock Exchange (IDX). Using a balanced panel dataset of 57 IDX-listed companies during 2019–2023, resulting in 285 firm-year observations, this study employs fixed-effects panel regression with firm and year fixed effects and clustered standard errors at the firm level. ERM implementation is measured using a composite ERM Index developed based on the COSO 2017 framework, while financial performance volatility is assessed through the three-year rolling standard deviation of Return on Assets (ROA), Return on Equity (ROE), and Tobin’s Q. Agency costs are measured using the Agency Cost Index (ACI) as a mediating variable. The findings show that ERM implementation is negatively associated with agency costs (β = −0.374, p < .001), indicating that stronger ERM practices are related to lower agency-related inefficiencies. Furthermore, agency costs are positively associated with ROA volatility (β = 0.491, p < .001) and ROE volatility (β = 0.428, p < .001). The mediation analysis indicates that agency costs statistically mediate the relationship between ERM implementation and accounting-based financial performance volatility, as the direct effects of ERM on ROA and ROE volatility become insignificant after including ACI in the models. In addition, ERM implementation shows a direct negative association with Tobin’s Q volatility (β = −0.218, p < .01), although the agency-cost-mediated pathway for Tobin’s Q volatility is not examined in this study. These findings extend ERM literature by highlighting agency cost reduction as a potential mechanism through which ERM relates to organizational stability, particularly in an emerging-market context. The findings also suggest that firms should view ERM not only as a compliance practice but as a governance capability that supports monitoring, accountability, and financial stability. However, the observational panel design limits causal interpretation, and future research should employ stronger identification strategies and alternative agency cost measures to further examine these relationships.