We’re now in “Moving Operational Risk Forward” or “Getting Value from ORX Data and Tying Operational Risk into Each Business Unit” with Joe Sabatini, JP Morgan, and Simon Wills, ORX. The introduction is being given by David Millar, PRMIA, who opened the session with a statement on the fire evacuation procedures. Some will remember that a fire alarm during an operational risk conference is not unheard of.
Sabatini started out by echoing comments from a previous session: namely, that the increased regulatory pressure will increase the challenges of managing operational risk at regulated entities.
Loss data, according to Sabatini, has been one of main drivers for change within the operational risk field. Before loss data was collected, no one really knew how much money was being lost on operational risk. With the collection of loss data, business lines understood how critical operational risk was.
With regard to capital calculation, the Enron/Worldcom data points included in the traditional LDA approach for capital would suggest for JP Morgan that they need $50 billion in capital driven somewhat by investment banking underwriting risk. Sabatini discussed an approach similar to that in the credit world where you calculate the probability of default, loss and investors winning a suit. This approach produces a more realistic capital number.
Sabatini also discussed some of the challenges and opportunities with regard to risk management, including business unit benchmarking, trend analysis, correlation with business metrics, and dynamic reporting. He also suggested that a significant advance would to have a real time dashboard that would allow what-if analysis discussion between market, credit and operational risk functions.
Simon Wills then gave an overview of ORX, our customer and partner. He said that they will be up to 54 member institutions when they announce their newest member tomorrow. Wills noted that ORX follows the Basel II categorization, with an additional category for corporate losses (ransom paid for a kidnapping of the chairman, for instance).
ORX also collects data on the product (e.g. equity derivative) and process (sales and marketing) associated with the losses, which provides a greater degree of granularity to the loss. ORX also collects additional information on large losses (over €10 million).
Wills shared some recent data on operational risk losses, and noted that sales and trading have been the driver of the large number of losses in 2008, whose aggregate severity rivals that of the Enron/Worldcom losses of 2002.
ORX is interested in a better visualization of the data to improve the communication and engagement of operational risk with the business. Corporate finance, for instance, tends to have low frequency and high severity losses, the opposite of losses in the retail business. Wills showed a 3D graph of the two different loss data sets, with dramatic spikes in the corporate finance business.
Wills talked about ORX sector services that will provide insight for different business units to benchmark against their peers, and, in this way, provide real business performance value to operational risk managers and their business line colleagues.