Modelling risk – how safe is your scorecard?

These are testing days for risk managers. The increasing stringency of the regulatory environment is creating new and complex demands – not least through the requirements of Basel II which are now coming into full force. In addition to these pressures, risk managers face a constant balancing act as they seek to reconcile the drive for profitability against the need to lend responsibly, all in a highly competitive market.

Clearly, effective risk management is more important than ever. However, a recent survey of financial services executives conducted by PricewaterhouseCoopers financial services group, in cooperation with the Economist Intelligence Unit, found that despite the focus created by regulatory reform, risk management is still not delivering the value that it should 1. For risk managers, the demand is not just to deliver risk models to comply with regulation – it is also vital that they are able to provide evidence that the risk function is being properly undertaken and that the risk models are being properly monitored.

At times like these, it makes sense to look towards the tools that technology can provide to support effective decision making and provide the evidence that best practice is being adhered to. There is nothing new in saying that the key lies in scorecards. But if scorecards are the answer, why is risk management still having difficulty delivering and proving its value?

Part of the problem lies not in the “what” but in the “how”. We need to take a closer look at way we approach scorecard monitoring.

The data required to monitor scorecards can come from disparate sources all around the business. Usually, the analytics systems used by the analysts give the development data while one or more operational systems give the actuals. As a result, risk analysts, risk managers and portfolio managers usually end up with bespoke reporting systems using a variety of tools, with the output often ending up in spreadsheets.

At first glance, tailor-made spreadsheets might appear to be the ideal way to reflect the individuality of the business and provide exactly the right kind of reports for that business’ needs. But spreadsheets are not an efficient tool for collating data from different sources and there is always the risk that important segments of data are either uploaded or input incorrectly. As a result, a tool that begins by looking efficient and relevant soon requires a great deal of manual processing and tinkering. Information and knowledge lies not so much in the tool as in the individuals who are working with it. If the people go, the knowledge goes with them.

Along with this complexity comes delay. It can take anything up to three or four months to create the scorecard monitoring reports and then several hours each month to generate them. Time, energy and resource are all tied up unnecessarily.

If risk managers are to meet the challenges of today’s environment, they need to challenge the tools that they are using. Current software options, even beyond self-created spreadsheets, do not serve risk managers well. However, the measures that are used to monitor scorecards are relatively standard throughout the industry so there must be an opportunity for a standard solution.

If we challenge current approaches to scorecard monitoring and ask the question “what should the tool really deliver?”, I believe the list of requirements looks like this:

• The tool should easily pull and combine data from a wide range of sources and systems throughout the business
• It should cover all the industry standard measures of scorecard strength and stability and characteristics strength
• It should be able to refresh the information as often as required and operate as close to real time as required
• Information should be displayed in a clear, easy to understand format and should reflect key performance indicators, thresholds and trends
• The tool should allow users to drill down into specific pieces of information in order to discover what data has led to a particular result
• As well as providing transparent, useful reporting, the tool should be able to model “what if” scenarios and measure the effects of tweaking scorecards

By taking this kind of new approach to scorecard monitoring and asking more of the tools that they use, risk analysts can devote more time to delivering value and less to wrestling with data.

By Peter Constance, managing director
pancredit

1 Creating value: effective risk management in financial services – PricewaterhouseCoopers, March 2007.

ENDS