Deriving Business Value from RAROC
In the past six months, banks have taken a pounding on their bottom-line, with approximately US$380bn sub-prime assets being written off. Sub-prime losses in just six banks account for approximately US$154bn (Citi, UBS, Merrill Lynch, HSBC, Bank of America, Morgan Stanley).1 With more news coming on breakdowns of banks’ capital structure, the need for an effective capital management is gaining more attention from various stakeholders in the bank’s operations: shareholders, employees, regulators and the government. Before the banks reach the finish line, there will be more changes in the industry such as bank consolidations, government financial support for struggling banks, employee layoffs and severe cost cutting. The finish line of this financial crunch seems to reach a new low with every dispirited quarter results and once this crunch is over, banks will then be faced with an environment of lower returns, higher interest rates and stringent regulations. Banks must equip themselves with all the necessary tools for more effective capital management and wise business decision making in the post credit crunch scenario, as they may not get the same amount of financial comfort, they enjoyed with all the leverage before.
Contriving Basel II for banks and Solvency II for insurance companies was one step taken by regulators to mitigate various risks and provide some stability in operations. However, Basel II could not avert sub-prime losses as there was no guideline to validate external ratings for the mortgage backed securities. Further, as Basel II focuses on capital adequacy and does not directly influence business decision making, banks have started focusing on a more developed approach to risk management.
Monitoring risk adjusted return on capital (RAROC) is one such approach employed by banks to appropriately account for risk while providing products and services. RAROC does not take into consideration the external ratings, but depends on self-evaluated risk parameters. RAROC was first used by the Bankers Trust in the late 1970s to limit exposure of the bank’s credit portfolio. It is a part of the family of risk-adjusted performance measures. It helps in precise estimation of profits, an important determinant of decision-making. ‘Risk versus reward’; something which is inculcated in every decision-making is more explicitly covered in RAROC than in any other measure. RAROC is different from Basel II as it considers risk capital evaluated under one-factor value at risk (VaR) model as against capital as evaluated under standardised or advanced measurements approach. Risk capital is not the same as Basel II capital. Although the gap between capital as evaluated under RAROC and that evaluated under Basel I has reduced after the introduction of Basel II, Jaime Caruana, Chairman of the Basel committee opines that economic capital is the model of best practice to which banks, particularly international ones, should aspire.2
More banks have started considering RAROC as an important tool in their armoury for making decisions. Banks have started formalising RAROC figures not only in small deal decisions, but also in strategic decisions at organisation level. Banks that are already RAROC-enabled have an edge over banks that are yet to realise the benefits of RAROC.
Every division of bank works in its own risk environment; however when it comes to evaluating its performance, profitability is considered as the sole determinant. Risks are unavoidable in business and it is only rational to consider quantum of risk into performance evaluation for business lines. RAROC calculates profit generated per unit of risk (in percentage) borne by the division. This measure acts as a benchmark across business lines and this is a valuable input for management to decide which business lines to expand and which ones to discontinue. RAROC also helps to measure precisely the incremental effect on economic capital of the business that is in the process of getting finalised.
Forecasting profitability for divisions is important while estimating capital requirements. Banks use various methods such as trend analysis, economic factor modeling, etc. In RAROC methodology, risk capital is measured considering the risk (volatility) evaluated on a longer period. Hence RAROC, encapsulating futuristic risk estimation, is useful in estimating cash flows. If contribution of individual factors to the net revenue is analysed further, banks will have valuable insights into its operations.
Most of the banks remunerate their traders, division leads, managers as a fixed percentage of profit generated, thereby jeopardising those who have generated same or more profit for lower risk. Trader’s compensation can be adjusted for the riskiness of activities undertaken under RAROC methodology. This will now drive these personnel to focus on risk minimisation as well.
Although banks take into consideration the risk profile of customers before pricing their services, this pricing methodology does not take into consideration the organisation’s risk profile and the marginal risk added by the products provided by specific divisions. From activity based pricing of products, banks can now price products and accounts more precisely using RAROC, which takes into consideration not only the risk of the division in isolation, but also the marginal risk added by the division. Product pricing can now provide for remuneration to the capital as well. Using RAROC as a ‘supplementary’ tool, banks can bring consistency in approving their products and services. Past deal decisions can also be analysed for their appropriateness using this methodology.
In an environment of credit crunch, banks are forced to borrow external funds at increased cost. RAROC figures evaluate the proposal’s precise risk and cash contribution to the organisation. This is an important factor in capital budgeting decisions. Further, if RAROC is implemented at enterprise level, capital allocation can be greatly optimised thereby making more capital available for revenue generating activities.
On the numerator, economic value added is considered after accounting for overhead costs, expected loss and taxes. Expected loss is evaluated the same way as evaluated under Basel II rules. For instance, for channel finance operations, the overhead costs would be the administration expenses, staff salary overheads, etc. The total revenue would include fee based as well as interest income expected from a deal. Risk capital is considered in the denominator representing risk assumed by the operations for generating the profits. RAROC is measured in percentage revenue generated per unit of risk capital.
Banks should consider the following before implementing RAROC:
Although directionally correct, RAROC has its limitations. RAROC models are based on assumptions that can skew results, if they are not complied. For getting the best out of the model, participation from the line managers coupled with motivation and support from the top leadership is necessary. When RAROC is based on simple risk ratios, comprehensiveness of the framework is sacrificed and the output is then just another financial figure with no added relevance. Some banks are of the opinion that RAROC’s impact on pricing is more direct with some products than with others. For instance, in the case of HNI customers, the services are less driven by RAROC figures, but more by strategic benefits, something which is not captured in the model. Banks have not been able to successfully use RAROC model, because they attach less importance to the hidden assumptions of the model. Following are some of the challenges in designing a RAROC model:
Banks have to accept that there is no single approach to business decision-making. This holds true for RAROC as well. RAROC is by no means a substitute for effective leadership. It merely captures return per unit of risk. Strategic decision-making on the other hand includes factors such as financial, market, regulatory etc. Although RAROC can be used as a guiding tool, these factors cannot be objectively captured in this model. Dependency on RAROC as a single tool can lead to strategic opportunity losses.
Perhaps the biggest challenge for banks in RAROC implementation is to identify a consistent risk framework across the enterprise. Definitions of risk, probabilities, and time horizons should be consistent across operations. However, in some business lines, estimating these parameters can be extremely different and there is no best way to do it. For instance, while it is easy to evaluate parameters for a credit swap, evaluating capital requirements for private banking operations is extremely difficult, as it is less governed by rational measures of credit and market risk. If risk appraisal is not uniform across business lines, then consistency, which is the value proposition of RAROC methodology, is jeopardised. It is extremely difficult to get a ‘buy-in’ from divisions, if non-standardised models are used. Division risk managers will not consider the models effective, if the models are incapable of capturing risks to their satisfaction.
Risk faced by a specific business line is influenced not only by the inherent risks in its operations, but also by risks from other business lines. Calculating correlation and co-variance for risk interdependencies between different business lines is extremely difficult and is possible only in case of banks that have a perfectly compatible technology and information architecture. While designing a RAROC model for a bank, analysts should focus not only on the transaction level risks, but also on enterprise level risks. Further due to the dynamic nature of market risk and unpredictability of operational risk, integrating these risks into RAROC models is a challenge to risk managers across the world. Banks in such cases can opt to implement RAROC model initially by considering only one risk and can thereafter enrich the models by adding more risk factors, once the model has been tested.
Unlike capital adequacy, RAROC is yet to be standardised. Hence there is no standardisation in the way banks select risks and assign values to confidence level, volatilities or time horizon. Banks have the freehand to identify risk metrics relevant to its operations. Consequentially, RAROC figures reported by banks cannot be benchmarked without adjusting them for the inconsistencies. This obscures the benefits of external reporting as well as benchmarking.
A well-designed RAROC methodology can provide a strong boost for managing capital precisely, if it is well supported by sound integrated information architecture. The primary goal of this exercise is not to change the existing risk management operations, but to enrich it with RAROC. Needless to say, the solution has to be scalable for handling increased product complexity. Providing flexible component based architecture that can handle complex user queries are critical enablers to having a sophisticated RAROC model. It has to be borne in the mind, that data consolidation is absolutely necessary. Banks can adopt a bottom-up approach and handle consolidation at a later stage. If these pointers are managed, we would have a comprehensive and pragmatic RAROC model.
I would recommend a five-phased approach to RAROC implementation:
Banks should be aware that risk profiles are unique and hence the model has to be tailor made for its operations. A RAROC framework is based on certain assumptions and banks should identify the limitations of these assumptions before implementing the model. Further, these models should not be the sole decision determinants. Rather they should be considered as a supplementary tool for decision-making. It is imperative to refine the models regularly in order to maintain their relevance. Volatilities used in these models should undergo timely updates (It is recommended to have dynamic updates) depending on the business environment.
A major challenge in RAROC implementation is varied risk profiles of business lines and the inflexibility of line managers to reach a consensus on the model. Inadequate communication about RAROC benefits is the main hindrance for getting ‘buy-in’ from stakeholders. Each business line’s unique risk characteristics should be taken into consideration in the process of designing a RAROC framework. Implementing RAROC into the bank’s reward system can motivate active participation in the design process.
Designing model parameters is driven by a bank’s desire to reach a benchmark risk rating. Instead of focusing all the possible risks that a business line is subject to, an analyst may opt to identify major risks that account for 80% (say) of the total risk. Banks may bucket the identified risks into four categories such as credit risk, market risk, country risk and business risk.
Time horizon for the model should be decided on the basis of implicit risks and default rates. Considering the reliability of data, banks can opt for a 1-year time frame for the model. However, most of the business cycles last for five to 10 years. In order to mitigate any term bias, banks should give appropriate weights to the data projected for longer time frames. The data used for model designing can be a mix of historical and forward-looking data. As RAROC is also used as a profitability estimation tool, it is recommended to give more weight-age to the forward-looking data.
The confidence level for calculating risk capital must be aligned to the bank’s risk appetite. The quantum of protection required against unexpected losses is what determines the confidence level. If the bank is risk averse, a confidence level of 99.99% is recommended.
Volatility can be based on number of factors such as inherent risks in operations and volatility in market prices. Identifying volatility for only the major risks identified by the bank rather than thousands of factors is a wise decision.
RAROC output should be compared with a ‘hurdle rate’ in order to make business sense from the model outcome. The hurdle rate is the business unit’s price plus risk premium. Having differentiated hurdle rate for all business lines is recommended as this improves the utility of RAROC in decision-making.
RAROC models should be compatible with the information and technology architecture. Quality data required by the model should be provided in optimum quantity as required by the model. Instead of having information operating at division level, it is recommended to have a common data hub that captures organisation wide data for model operation. It is recommended to have data at account level rather than at product level, as this would provide account level view to the banks. Considering the ticket size, banks can use consolidated data for input into the model. Risk managers should chalk out information segments in granular order before proceeding with model parameters.
Backtesting of model output is vitally important before the roll out. Stress testing PD, LGD, EAD for extreme values will help identify output reliability. Testing model output for combinations of these inputs for expected results has to be a part of the overall review process. For this purpose, rejected deals or turned down decisions should also be considered. Similarly RAROC model output, such as expected profit, expected risk, expected cash flow, should be tested with actual figures. In case of joint exercise between a group of banks, FIs can run a sample of representative deals through their model and can share the results anonymously for benchmarking purpose. Banks should test other models, if they are able to reduce the gap between model expected output and actual results.
In the absence of any industry standard, it will take a long time, before the industry can agree on a standardised framework for RAROC. Meanwhile, with every RAROC implementation experience, the industry would have adequate understanding of the bank’s benefits and challenges. With every RAROC implementation, there would be increased knowledge among stakeholders and this in the end strengthens shareholder confidence. The motivation and drive for RAROC-enablement is not in considering it as compliance, but treating it as a worthwhile management tool.
1Wall Street Journal
2‘Basel II – Back to the future’, Jaime Caruana, Governor of the Bank of Spain and Chairman of Basel Committee on Banking supervision: 7th HKMA Distinguished Lecture, February 2005.