Integrated Stress Testing Architecture amazonaws com
White Paper
Integrated Stress Testing Architecture The need for a consolidated, future-proof framework
WE PUT THE BANKING INTO BUSINESS INTELLIGENCE www.icreate.in
Integrated Stress Testing Architecture
Integrated Stress Testing Architecture The need for a consolidated, future-proof framework
Risk-control assessment in financial institutions should not just be limited to identifying failures in processes and exposure levels. It must also perform routine ‘what-if’ analyses to stress test the changes in capital requirements. An ideal stress testing solution requires integrated risk architecture along with excellent data handling capabilities and business Intelligence (BI) capabilities. Stress testing primarily focuses on 1. Establishing the interdependency of variables 2. Cascading downward from macro (economic, political scenarios) to country to sector (Banking) to exposure levels 3. Arrest the degrees of freedom in a plane to fix the event (mathematically this may mean introducing more constraints for variables) 4. An empirical methodology to predict the next crisis or default And with financial institutions increasingly gaining the capability to predict most economic scenarios that can impact their balance sheets, an interesting experiment would be to reverse the stress test to find out the vulnerabilities in the business models itself.
Framework and Data Quality Informed business decisions are taken in conjunction with the stressed test results. The results could be enhanced capital buffer, product-pricing, limit-setting and many other qualitative aspects. Hence the results of the stress test are extremely important to the financial institutions from a sustenance as well as an earnings perspective. The quality of the results is directly attributed to the quality of the data, framework of the stress test, the approach and assumptions and internal models. Moreover, the framework of stress testing should be implementable and repeatable. It should detail the procedures to find out the inter-relationship of different functional units within a bank and also at an institutional level. This is most often guided by the regulators in the respective countries. It can apply to the banks basis their maturity levels and level of preparedness. On the other hand, quality in this context refers to the historical availability of data, the frequency at which the data is made available, the depth, integrity and completeness of the data. Absence of any of the above will impact the quality of results. For example, if assumptions are made for want of historical data, drill down capabilities of the output results may be affected.
Regulatory Mode of Stress Testing In several countries, FIs produce the stress tested results under a common framework, using a ‘bottom-up’ approach using internal data modelling. The assumptions that the FIs make here are very relevant to their business scenarios and operational philosophy. Hence more often this gives a view of ‘risk budgeting’ to the FI. From a supervisory perspective this helps the regulator assess whether the FI will be adequately capitalized under the defined circumstances (the framework being common across FIs). This approach is quite ‘forward looking’ and does not take any economic considerations in to the picture. Regulatory authorities typically use a ‘top-down’ approach to gauge the health of the financial system arising from economic conditions. This gives an idea of undetected risks or losses if only historical data is used. As we see, when the supervisor applies this effect on the individual FI’s portfolio, it applies it at an income / balance sheet level. Standard econometric modelling may be applicable and techniques such as multivariate regressions (as applicable) are familiar in modelling this.
An iCreate White Paper
2
Integrated Stress Testing Architecture
Integrated Risk Management Not limiting to the above, FIs are gradually moving towards an integrated approach to risk management. a. The concept here is to stress test the various components of risk such as market, credit, liquidity, solvency, strategic and operational. This provides the complete breadth of the capital adequacy to the FI. Irrespective of the regulatory needs, this helps the bank maintain economic capital, structure the lines of business and stay in touch with market dynamics. b. A ‘top down’ approach (within the FI) especially around earnings will help steer clear of downturns. This simple logic when extended could mean establishing the correlation between the economic capital and volatility of earnings. We can extend this logic either ways by inclusion of more economic variables one side and drill down to exposure level impact on the other side. In the same manner, FIs may also would like to build custom-build scenarios and assign periodic losses and probability on an on-going basis and hence callable in future for appropriate calculations. To add to the above, provision for reverse stress testing using standard or custom defined algorithm is the path ahead. For all of the above, the Risk Architecture plays a prime role in covering all of the above functionalities.
Data analytics is key To have the ability to store granular level data of all deals along with the cash flows. Frequent updates of data help ensure accuracy. For example, the user might want to take the mark-to-market value of one of the most illiquid instruments, which can be a substantial holding in his portfolio. Hence an external feeder for ‘bid ‘and ‘ask’ data might be required on a daily basis, specific to the exposure. For complex transactions, data might be required on an intraday basis with varying frequencies. More than just making the data available, analysis of the same with varying dimensions and capability to build models over it is important.
Integrated view To have an integrated view of all risk types (including liquidity and solvency) and also risk characteristics by product type as well. Each of these components can be user configurable and can have detailed drill down capabilities.
Capabilities Apart from the regulatory defined and other pre-defined scenarios, the user should be able to define custom defined scenarios (ex: defining yield curves and mapping scenarios to it). Business Intelligence (BI) to produce user-defined dashboards, reports, workflows and audits is feather on the cap. While an ‘ideal / fits-all’ option may not be available, a good stress testing solution must certainly have the above mentioned capabilities and should be modular / configurable to suit local regulatory requirements as well as the individual needs of the FI
Integrated Stress Testing Architecture The need for a consolidated, future-proof framework
WE PUT THE BANKING INTO BUSINESS INTELLIGENCE www.icreate.in
Integrated Stress Testing Architecture
Integrated Stress Testing Architecture The need for a consolidated, future-proof framework
Risk-control assessment in financial institutions should not just be limited to identifying failures in processes and exposure levels. It must also perform routine ‘what-if’ analyses to stress test the changes in capital requirements. An ideal stress testing solution requires integrated risk architecture along with excellent data handling capabilities and business Intelligence (BI) capabilities. Stress testing primarily focuses on 1. Establishing the interdependency of variables 2. Cascading downward from macro (economic, political scenarios) to country to sector (Banking) to exposure levels 3. Arrest the degrees of freedom in a plane to fix the event (mathematically this may mean introducing more constraints for variables) 4. An empirical methodology to predict the next crisis or default And with financial institutions increasingly gaining the capability to predict most economic scenarios that can impact their balance sheets, an interesting experiment would be to reverse the stress test to find out the vulnerabilities in the business models itself.
Framework and Data Quality Informed business decisions are taken in conjunction with the stressed test results. The results could be enhanced capital buffer, product-pricing, limit-setting and many other qualitative aspects. Hence the results of the stress test are extremely important to the financial institutions from a sustenance as well as an earnings perspective. The quality of the results is directly attributed to the quality of the data, framework of the stress test, the approach and assumptions and internal models. Moreover, the framework of stress testing should be implementable and repeatable. It should detail the procedures to find out the inter-relationship of different functional units within a bank and also at an institutional level. This is most often guided by the regulators in the respective countries. It can apply to the banks basis their maturity levels and level of preparedness. On the other hand, quality in this context refers to the historical availability of data, the frequency at which the data is made available, the depth, integrity and completeness of the data. Absence of any of the above will impact the quality of results. For example, if assumptions are made for want of historical data, drill down capabilities of the output results may be affected.
Regulatory Mode of Stress Testing In several countries, FIs produce the stress tested results under a common framework, using a ‘bottom-up’ approach using internal data modelling. The assumptions that the FIs make here are very relevant to their business scenarios and operational philosophy. Hence more often this gives a view of ‘risk budgeting’ to the FI. From a supervisory perspective this helps the regulator assess whether the FI will be adequately capitalized under the defined circumstances (the framework being common across FIs). This approach is quite ‘forward looking’ and does not take any economic considerations in to the picture. Regulatory authorities typically use a ‘top-down’ approach to gauge the health of the financial system arising from economic conditions. This gives an idea of undetected risks or losses if only historical data is used. As we see, when the supervisor applies this effect on the individual FI’s portfolio, it applies it at an income / balance sheet level. Standard econometric modelling may be applicable and techniques such as multivariate regressions (as applicable) are familiar in modelling this.
An iCreate White Paper
2
Integrated Stress Testing Architecture
Integrated Risk Management Not limiting to the above, FIs are gradually moving towards an integrated approach to risk management. a. The concept here is to stress test the various components of risk such as market, credit, liquidity, solvency, strategic and operational. This provides the complete breadth of the capital adequacy to the FI. Irrespective of the regulatory needs, this helps the bank maintain economic capital, structure the lines of business and stay in touch with market dynamics. b. A ‘top down’ approach (within the FI) especially around earnings will help steer clear of downturns. This simple logic when extended could mean establishing the correlation between the economic capital and volatility of earnings. We can extend this logic either ways by inclusion of more economic variables one side and drill down to exposure level impact on the other side. In the same manner, FIs may also would like to build custom-build scenarios and assign periodic losses and probability on an on-going basis and hence callable in future for appropriate calculations. To add to the above, provision for reverse stress testing using standard or custom defined algorithm is the path ahead. For all of the above, the Risk Architecture plays a prime role in covering all of the above functionalities.
Data analytics is key To have the ability to store granular level data of all deals along with the cash flows. Frequent updates of data help ensure accuracy. For example, the user might want to take the mark-to-market value of one of the most illiquid instruments, which can be a substantial holding in his portfolio. Hence an external feeder for ‘bid ‘and ‘ask’ data might be required on a daily basis, specific to the exposure. For complex transactions, data might be required on an intraday basis with varying frequencies. More than just making the data available, analysis of the same with varying dimensions and capability to build models over it is important.
Integrated view To have an integrated view of all risk types (including liquidity and solvency) and also risk characteristics by product type as well. Each of these components can be user configurable and can have detailed drill down capabilities.
Capabilities Apart from the regulatory defined and other pre-defined scenarios, the user should be able to define custom defined scenarios (ex: defining yield curves and mapping scenarios to it). Business Intelligence (BI) to produce user-defined dashboards, reports, workflows and audits is feather on the cap. While an ‘ideal / fits-all’ option may not be available, a good stress testing solution must certainly have the above mentioned capabilities and should be modular / configurable to suit local regulatory requirements as well as the individual needs of the FI
