ANZ Bank sends dirty data back to the source

04.04.2005
Von Julian Bajkowski

Pressure to get in line with the global financial services initiative Basel II is elevating the status of once-humble data warehouse managers to business champions.

The Australia and New Zealand Banking Group Ltd. (ANZ Bank) is implementing a raft of structural IT changes driven by the need to comply with Basel II, which requires financial companies to have a more risk-sensitive framework for the assessment of regulatory capital.

Addressing delegates at the MIP user group conference in Sydney last week, Sabyasachi Banerjee - head of ANZ"s risk technology unit - said nothing works as well as the threat of jail when it comes to compelling senior management to stump up for compliance-related IT projects.

Banerjee said the stiff penalties associated with failure to provide transparent fiduciary and risk data to regulators had now produced significant pressure on CEOs and CFOs to ensure IT systems delivered data that was clean, integrated and accurate.

"You can"t walk away and say you didn"t really know [what was in] the data and hope for the best. People go to jail," Banerjee said.

In terms of dealing with data surrounding risk management, Banerjee said his business unit opted to forge an alliance between risk modellers and IT to enable the flow of disparate data sources into the bank"s risk information store (RIS) core database.

He said this created tangible results rather than management frustration.

RIS is used to prepare heavy calculations on information and data sourced from both internal feeds such as transactional and account systems and external data sources such as credit monitors and market information providers.

"We are not going to brand people [because they are from] business or technology. They are part of the risk technology unit." The business worked very closely with IT to build the architecture, Banerjee said.

"One of the beautiful things we have achieved is that the business understands very well the capabilities of the [modelling] toolsets ... they don"t ask for things that will be difficult to implement. By marrying teams, it was easier to get people to learn from each other."

However, he cautioned that despite the success of cleansing and profiling his own unit"s data, IT managers confronted with poor data from external sources should not reward polluters by cleansing their mess for them. Rather, it should be sent back from whence it came.

"The data quality should reside with the [line of] business. We want data corrected at the source system. We don"t want to create an industry out of fixing [poor quality] source data. Responsibility must be taken by the business," he said.

Banerjee said the new and positive RIS experience had provided a "very large proof of concept" to the bank"s management, which had become wary of "large IT projects that don"t deliver and blow you out of the water" - and certainly a better option than doing time.

The deadline for Basel II compliance is end of calendar year 2007.