Banks, as well as insurance companies, face regulatory and economic challenges when choosing to implement an Internal Model Approach (IMA), which would allow them to calculate the capital requirements in a preferred way. For instance, to stay eligible for using an IMA, an institution needs to demonstrate its model’s ability to continuously replicate the actual profit and loss (P&L) behaviour with sufficiently high accuracy. Performing P&L attribution (PLA) tests is one way to do that. This article discusses challenges related to PLA as well as similarities between the two regulations specifying the requirements for the tests: Fundamental Review of Trading Book (FRTB) for banks and Solvency II for insurance companies.

FRTB

In 2018, the Bank for International Settlements revised the IMA as of the previous January 2016 standard (Basel Committee on Banking Supervision, 2016) by releasing a consultation paper (Basel Committee on Banking Supervision, 2018) for the FRTB. The revised document revealed that banks must perform the PLA tests continuously to ensure that their trading desks are eligible for the IMA over time. The purpose of the PLA is to ensure the alignment between the P&L estimates calculated by the risk management models and the front office.

The accuracy of the trading desks’ P&L estimates is evaluated using a traffic light approach, consisting of three zones (green, amber, red). If the performance of an evaluated trading desk falls into the red zone, the firm would have to revert to using the standardised approach often characterised by a higher capital requirement.  Alternatively, if the performance of a trading desk is within the thresholds of the amber zone, its capital requirement would be specified between the levels prescribed by the IMA and standardised approach. Hence, the introduction of the amber zone lowers the impact of an enforced switch from an IMA to a standardised approach in case of performance failure. However, as discussed in (Risk.Net, 2018a), industry participants argue that the amber zone is too narrow and that it is too easy for the trading desks to fall into the red zone. In connection to this, (Risk.Net, 2018b) argues that the red zone needs to be reduced and simultaneously support a widening of the amber zone. Under FRTB, a trading desk needs to pass all PLA tests independently to avoid falling into the red zone and if the amber zone is too narrow, too many good desks are deemed ineligible for an IMA.

Solvency II

Banks, insurance and reinsurance undertakings are obliged to continuously perform satisfactorily on PLA tests to be permitted to use an internal model under Solvency II. According to the Solvency II directive, insurance and reinsurance undertakings shall “demonstrate how the categorisation of risk chosen in the internal model explains the causes and sources of profits and losses. The categorisation of risk and attribution of profits and losses shall reflect the risk profile of the insurance and reinsurance undertakings” (Official Journal of the European Union, 2009). According to the Commissions Delegated Regulation, “The categorisation of risks chosen in the internal model shall be adequate, and sufficiently granular, for the purpose of risk-management and decision-making…” (Official Journal of the European Union, 2015).  Hence, the procedure of conducting a PLA within Solvency II shares the same overall goal as the FRTB test, ensuring that the risk management model can replicate actual P&L behaviour sufficiently well.

The Solvency II PLA assesses the ability of the internal model’s risk factors to describe a yearly change in basic own funds (BoF), using the same risk factors and assumptions as for calculating the capital requirements. The change in BoF would be allocated between the internal model risk factors and the risk not covered by the internal model. The conclusion on the internal model’s appropriateness is based on an overall assessment combining different measures. The main tasks of this assessment involve;

  • Evaluating the size and constituents of any unexplained yearly change in BoF;
  • The yearly outcomes’ percentiles with regards to the internal model probability distribution forecasts;
  • Basis risk of the risk factors.

Challenges

One challenge related to PLA within the scope of FRTB is the elimination of valuation differences between the front-office and the risk management model. Similar challenges arise from other potential reasons for divergence such as; the systems use different market data sources; or the P&L is calculated at separate times. It is difficult and time-consuming to manually monitor and manage the integration between multiple systems, data sources as well as any timing differences between the systems. This results in high key personnel risk and discrepancies impacting the accuracy and reliability of capital requirement and PLA calculations. Hence, having unaligned risk systems, inconsistent valuation methodologies and timing differences can amount to significant complications hindering PLA automation and depleting the chance of staying eligible for internal model usage.

Therefore, it is crucial for banks operating with different front-office and risk management models to ensure an appropriate alignment between them, facilitate communication and preserve a mutual representation and pricing of assets. The reference (Risk.Net, 2017) discusses the advantages and disadvantages of having a centralised or decentralised way of implementing FRTB. Here a centralised approach implies establishing one central risk management model, which mirrors all the asset positions the company has at a specific point in time. The decentralised approach refers to a process of the front-office performing calculations that are later aggregated into the risk management model. Implementing a decentralised method simplifies the process of achieving consistency across the systems but might lead to difficulties in operating and correcting issues on a daily basis (Risk.Net, 2017). A centralised method is better suited for automation of processes since the same calculations do not have to be performed twice.

Similar arguments as above are applicable for the insurance companies performing PLA under Solvency II. Our experience suggests that a significant challenge faced by insurers is achieving the alignment between the risk management model, asset management as well as the Solvency II balance sheet. Different data sources and risk factor granularity may result in timing and valuation differences. This, in turn, leads to a suboptimal PLA testing process that induces ambiguities among employees.

Another challenge regarding PLA is to replicate changes in the portfolio composition during a specific holding period for the risk management model. This is often based on a static portfolio that is exposed to specific price shocks, despite that significant changes in the portfolio composition are common in large trading institutions. This challenge is most apparent for the PLA test within the Solvency II framework since it is not based on daily P&L measures, inducing significant difficulties for the risk management model to replicate the trading that has occurred during the year.

To conclude, challenges related to PLA for banks and insurance companies can arise due to several underlying issues that do not necessarily have to be related to the modelling accuracy. Hence, PLA is not only a process consisting of technical analysis but also a process that is highly dependent on the abundance of data and system quality assessment.

Summary

In this article we have discussed some of the challenges related to the PLA test and how they can make it daunting to consider implementing an internal model approach under the Solvency II or the FRTB framework. However, if approached correctly, they can not only ensure compliance but also improve firms’ internal processes. We have summarised our expertise and experiences of resolving these challenges in the document attached below.

Risk Process Automation - FRTB

Risk Process Automation - Solvency II

References

1. Basel Committee on Banking Supervision, 2016. Minimum capital requirements for market risk 

2. Basel Committee on Banking Supervision, 2018. Revisions to the minimum capital requirements for market risk

3. Risk.Net, 2018a. Amber zone in new P&L almost useless, say banks 

4. Risk.Net, 2018b. The revised P&L attribution test and the suitability of newly proposed thresholds

5. Official Journal of the European Union, 2009. Directive 2009/138/EC of the European Parliament and of the Council of 25  November 2009 on the taking-up and pursuit of the business of Insurance and Reinsurance (Solvency II)

6. Official Journal of the European Union, 2015. Commission Delegated Regulation (EU) 2015/35 of 10 October 2014 supplementing Directive 2009/138/EC of the European Parliament and of the Council on the taking-up and pursuit of the business of Insurance and Reinsurance (Solvency II)

7. Risk.Net, 2017. The right strategy for a head start on FRTB