Tackling the over-dispersion of operational risk: Implications on capital adequacy requirements

• The over-dispersed versus the equi-dispersed nature of operational risk is tested.
• The impact of the extra-Poison variance into Capital adequacy requirements (CARs) is evaluated.
• Compound Poisson–Gamma models perform better than the traditional Poisson model.
• CARs are even more sensitive to extra-Poisson distributions under heavy-tailed scenarios.
• Ignoring over-dispersion leads to unreliable regulatory capital charges for operational risk.

Having proved Basel II to be ineffective to prevent the global financial crisis, Basel III seeks to enhance the resilience of individual financial institutions by strengthening their capital buffer and by building counterbalancing capacity to absorb liquidity shocks. Under this new regulatory framework the increase of capital adequacy ratios is a matter of utmost importance for promoting the soundness and stability of the financial system. With regard to the operational risk, Basel III suggests a greater convergence in the measurement methodologies as well as a higher supervision. To this extent, the objective of this paper is threefold: (i) to test the over-dispersed nature of operational losses; (ii) to capture the extra-Poison variance into the Loss Distribution Approach (LDA); (iii) to assess its potential impact on the capital adequacy requirements (CARs) for operational risk. Our findings point out a higher capital charge associated to the alternative extra-Poisson distributions; even more significant under heavy-tailed scenarios. In consequence, the over-dispersion phenomenon should be addressed very carefully not only by the financial institutions when designing their internal measurement approaches, but also, by the supervisors when validating such models, both ensuring the appropriate specifications to provide with a more realistic capital charges.