AMA Models

From Open Risk Manual
Revision as of 13:34, 19 November 2019 by Wiki admin (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Definition

AMA Models is the set Quantitative Models involved in the Advanced Measurement Approach (AMA), one of the possible options under the Basel[1] regulatory standard for the measurement and capitalisation of Operational Risk

Context

Under Basel II, regulated financial services firms are allowed (subject to approval) to develop own risk models to quantify required capital for operational risk.

Inputs

The AMA framework must include the use of four data elements:

  • Internal loss data (ILD)
  • External data (ED)
  • Scenario analysis (SA), and
  • Business environment and internal control factors (BEICFs)


The regulatory guidance does not specify how these data elements are to be combined

Methodologies

The heterogeneous nature of the data inputs means that there is large discretion on how to combine the elements into a coherent risk model. Some common approaches:

  • Combination of internal and external loss data into a purely historical loss driven model
  • Subjective integration with scenario analysis
  • Qualitative adjustments based on BEICF's

Integration of different risk types

The existence of widely different risk event types within operational risk poses a challenge as to how to integrate into a total risk profile

Outputs

The output of the AMA model is worst case loss at 99.9%

Issues and Challenges

  • Following unprecedented losses in the financial industry since the Financial Crisis, the AMA approach has been criticized and is now to be replaced by a standardized approach

References

  1. Basel II Accord, 2006