Margin of Conservatism

From Open Risk Manual

Definition

Margin of Conservatism (MoC) is a term used primarily in the context of the Internal Ratings-Based Approach option of the Basel III accord for the calculation of Pillar I capital requirements. It denotes the application of a conservative (erring on the side of caution) adjustment to remedy any shortcomings of quantitative estimates of Risk Parameters.

The margin of conservatism aims to mitigate two forms of Model Risk, namely issues with Data Quality and potentially Intrinsic Model Risk associated with the selection of methods and quantitative methods underpinning the model.

Regulatory Guidance

As a general concept institutions are required to address the identified deficiencies in data or methods via appropriate adjustments and margin of conservatism. Comprehensive guidance on setting up a framework for assessing Margin of Conservatism in given in [1]

An appropriate adjustment consists in rectifying the identified errors, for instance missing data points are filled in with the most probable information or the inaccuracies in data are corrected. The objective of the appropriate adjustment is to achieve the possibly most accurate estimates. However, as the appropriate adjustment is an estimate due to data deficiency, additional MoC has to be added to address the uncertainty related to this estimation. Moreover, MoC aims at addressing all errors that cannot be rectified through appropriate adjustment and any other uncertainties related to the estimation of risk parameters.

Identification of deficiencies

Institutions should have a robust process for identifying all deficiencies, including data errors and any uncertainties that lead to estimation errors, and for classifying them in the following categories:

  1. Category A: Expected range of estimation errors due to data or methodological deficiencies;
  2. Category B: Expected range of estimation errors due to diminished representativeness of historical observations


In addition, they should quantify a general estimation error and present it in a separate category (Category C: General Estimation Error)

For the purposes of applying the MoC during the phases of model development, estimation and calibration institutions should consider:

  • for the errors classified under Category A (Identified data and methodological deficiencies) at least the following triggers:
    • missing or materially changed default triggers in historical observations including changed criteria for recognition of materially past due credit obligations;
    • missing estimated date of default, leading to late default detection;
    • missing or outdated rating information used for the purpose of calculation of default rate per grade or pool; or average realised LGDs per grade or pool;
    • missing or inaccurate information on the source of cash flows;
    • missing, inaccurate or outdated data on risk drivers and rating criteria;
    • missing or inaccurate data for the calculation of economic loss;
    • missing or inaccurate information used for the estimation of future recoveries;
    • limited representativeness of the historical observations due to the use of external data;
    • potential bias stemming from the choice of the approach to calculating the average of observed one year default rates;
    • necessity of adjusting the average of observed one-year default rates
    • missing information for the purpose of estimating loss rates or for the purpose of reflecting economic downturn in LGD estimates;
    • diminished representativeness of the historical observations due to the changes in the definition of default
    • the rank order estimation error ;
    • estimation error in the calibration;
    • estimation error in the long-run averages due to necessary adjustments
  • for the errors classified under Category B, at least the following triggers:
    • changes to underwriting standards, collection or recovery policies, risk appetite or other relevant internal processes;
    • unjustified deviations in the ranges of values of the key risk characteristics of the application portfolio compared with those of the dataset used for risk quantification;
    • changes to market or legal environment;
    • forward-looking expectations regarding potential changes in the structure of the portfolio or the level of risk, especially based on actions or decisions that have already been taken but which are not reflected in the observed data.

Quantification of estimation errors

In order to overcome estimation errors in PD and LGD estimates stemming from the categories of deficiencies A, B, institutions should apply adequate methodologies for correcting the identified errors (‘appropriate adjustment’). Institutions should ensure that the appropriate adjustment results in a more accurate estimate of the risk parameter, where this adjustment can have both positive and negative effect on the risk parameter.

Where such appropriate adjustments are used institutions should apply a MoC to account for the additional estimation error associated with these adjustments. The MoC related to the economic adjustment should be proportionate to the impact of the adjustment on the risk parameter.

Institutions should also apply a MoC to address any errors that have not been corrected via appropriate adjustment and any identified uncertainties. Institutions should ensure that the impact of the MoC does not ever result in lowering PDs or LGDs.

The Guidelines do not prescribe any specific method for the quantification of MoC as the appropriate approach will depend on the character of the deficiency and the available data. However, institutions should keep in mind that model aspects that appear conservative in one model may not be truly conservative compared with alternative methods. For example, simply picking an extreme point on a given modelled distribution may not be conservative if the distribution was misestimated or misspecified in the first place. Furthermore, initially conservative assumptions may not remain conservative over time. Therefore, it is expected that the methods applied to derive the MoC will be regularly revised in order to ensure that the effect on the risk parameters is adequate and proportionate to the estimation error related to the identified deficiencies.

Institutions should assess the MoC at the level it is identified but they should reflect and report it with respect to the final risk parameter estimate used for own funds requirements.

Any occurrence of any of the triggers should result in the application of a MoC. Where more than one trigger occurs, a higher aggregate MoC should be applied. The MoC related to each trigger should be proportionate to the estimation error in the estimated parameter that results from the identified deficiency. Institutions should quantify the estimation error that results from the identified deficiency in order to justify the level of MoC. Institutions should quantify the appropriate adjustment and MoC at least for every calibration segment.

Institutions should provide for a customable IT implementation solution, which ensures that MoC can be implemented in a timely manner

Institutions should consider the overall impact of the identified deficiencies and the resulting MoC on the soundness of the model and ensure that capital requirements are not distorted due to the necessity for excessive adjustments

Monitoring

Institutions should regularly monitor the levels of the appropriate adjustments and MoC. The adoption of a MoC by institutions should not replace the need to address the causes of errors or uncertainties and to correct the models to ensure their full compliance with the requirements of the Regulation (EU) No 575/2013. Following its assessment, institutions should develop a plan to rectify the data and methodological deficiencies and reduce the estimation errors within a reasonable timeline, taking into consideration the materiality of the estimation error and the materiality of the rating system.

When reviewing the levels of MoC institutions should ensure all of the following:

  • that the MoC stemming from Category A is reduced over time;
  • that the MoC stemming from Categories C is eliminated after the error is rectified in all parts of the rating system that were affected

Documentation

For each rating system, the MoC applied should be documented in the relevant model documentation and methodology manuals. The documentation should at least contain:

  • a complete list of all potential and identified deficiencies and the potentially affected model components or risk parameters,
  • a description of the methods used to apply appropriate adjustments to rectify the data and methodological errors, where relevant;
  • a description of the methods of addressing the deficiencies, including errors and uncertainties, via the application of an MoC;
  • the category under which these errors and uncertainties are classified

References

  1. EBA, Guidelines on PD estimation, LGD estimation and the treatment of defaulted exposures, 2017