Risk Disclosure Quality and Financial Resilience in Sri Lankan Listed Commercial Banks

AbstractRisk disclosure is often treated as a reporting requirement, but it can also affect how investors, depositors,regulators and managers understand a bank's ability to absorb shocks. This paper examines whether the qualityof publicly reported risk information is associated with financial resilience in Sri Lanka. I hand-code an eightdimensionRisk Disclosure Quality Index (RDQI), scored from 0 to 32, from 42 official annual or integratedreports issued by six Sri Lankan-incorporated, Colombo Stock Exchange-listed private commercial banksbetween 2018 and 2024. The resilience measures combine reported total capital ratios, return on assets and theinverse of the gross-loans-to-customer-deposits ratio; a separate extended measure adds the inverse of thereported gross non-performing-loan or Stage 3 ratio where available.The descriptive evidence shows a clear upward reporting trend. Mean RDQI rises from 24.83 in 2018 to 32.00 in2024. However, the positive association in the bank-fixed-effects-only model becomes small and statisticallyimprecise after year fixed effects are added. The two-way fixed-effects coefficient is 0.049 (HC3 p = 0.544) forthe core resilience index, 0.088 (p = 0.233) for the extended index and 0.005 (p = 0.935) when current resilienceis related to one-year-lagged RDQI. The same caution appears in component models for capital, reported ROA,asset quality and liquidity. The results are therefore consistent with common time effects and crisis-responsereporting playing an important role in the raw relationship. They do not show that disclosure caused resilience.The paper contributes a transparent Sri Lankan dataset, a page-traceable coding framework and a reproducibleempirical baseline. It is a complete empirical manuscript prepared for journal submission; it has not been peerreviewed. The manuscript reports its remaining limitations directly: the RDQI is single-coded, outcomedefinitions are not perfectly harmonised and the sample is purposive rather than a census of all licensedcommercial banks.