Canada’s Sabean on tackling a data governance overhaul

Establishing good data governance at a central bank requires effort and investment, but the Bank of Canada is starting to reap the rewards, says Trevor Sabean, senior director at the central bank.

Sabean has been at the heart of efforts to reform the Bank of Canada’s data governance framework, allowing more agility and innovation, and paving the way for central bank economists to harness new big data and machine learning techniques.

“Good data governance is all about how you establish your IT systems, people, processes, policies to best exploit your investments in data and data support in pursuit of your business objectives,” he says, speaking on the sidelines of Central Banking’s recent training seminars at Christ’s College, Cambridge.

It is important that the governance framework is well understood across the organisation – people should be clear on both their roles and responsibilities and the benefits the new system offers, Sabean says.

Changes should align with corporate objectives. For instance, if a central bank is looking to make more innovative use of data, a locked-down structure where access is limited is “unlikely to be successful”, he says.

The Bank of Canada has been making “significant investments” in data, expanding its oversight committees, funding greater computing power and hiring data scientists. It is also considering the role of cloud technology – an area most central banks are still cautious about.

“What we’re also seeing is a tremendous amount of work from the ground up where staff are using these new tools and techniques to really do some cool things related to how we analyse and visualise data, or quality assess it – things we weren’t dreaming of five, 10 years ago,” Sabean says.

Central banks looking to overhaul their governance should recognise that the project will take time and there will be a series of problems to overcome along the way, he says. The organisation will need “tenacity and patience”.

But central bankers should also be willing to take on a little risk. Data science is a moving target, Sabean says, so if the organisation waits for all the answers, it may end up waiting too long, he says. Central banks that want to be agile have to accept “a certain level of risk”.