Scotiabank Creates a Shared View of Risk Across Trading Desks and Loan Departments Scotia Capital, a division of Scotiabank, has been developing technology that lets credit risk officers and traders at the bank share counterparty risk assessments of potential trades. The $575 billion-asset, Toronto bank has already created a real-time view of counterparty risk for its energy, equity and base-metal derivatives trading desks; rates derivatives are next. More recently, it developed a way to turn that information into loan-equivalent data that can be shared with credit officers as well as with the traders.
"We would like the most accurate measurement of risk we can get and we're looking for an overall comprehensive exposure," says Mark Engel, managing director and head of the global analytics and financial engineering group at Scotia Capital. "We'd also like to be able to drill down to the details and really understand what's happening."
The bank's primary reason for using this technology is classic, prudent risk management, he says. "As much as there's new technology and new, clever things we can do, this is a core risk banks have had ever since the [over-the-counter] market started. So we're doing the same thing we've been doing all along, but better."
The software Scotia Capital is using, Algo real-time credit engine from Algorithmics, calculates credit-value adjustment, capital exposures, potential future exposure and limit use to help with pricing of trades. It produces a close-to-worst-case scenario of exposure. "If we know at deal time the credit valuation adjustment and impact on capital exposure, that lets us have sharper pricing, manage capital and limits and make risk-reducing trades," Engel says. The software helps with the pretrade estimation of capital under Basel rules. The real-time view of risk stops a trader from making a decision without knowing how much capital would be required, from making capital-inefficient transactions or mispricing the transaction.
"We've added more tools to the software so the person executing the trade has the ability to make a more informed decision on the value of that trade," says Alyson Bailey, director of global analytics and financial engineering for Scotia Capital. "We're enabling people to evaluate trades across a multitude of scenarios and through time."
In the broader picture at the banking group, to provide a comprehensive view of risk across the trading and lending areas, the bank aggregates the output from the Algorithmics software with information about its loan exposures. The capital markets group assesses the counterparty and determines the potential exposure it would have from making a trade. It provides the exposure information for new trades to the banking group in the form of a "loan equivalent."
Historically, credit risk officers would get a single risk assessment number for each counterparty from the capital markets group and they would line those up with their loans. Now the credit risk officers receive two numbers for each company, one that lines up more directly with their loans, the other the 95th percentile worst case that the capital markets group uses to limit exposures.
As part of the Algo software implementation, the bank is building tools that will identify significant trades, such as those that are larger, longer dated or have a material credit value adjustment, so they're highly visible and receive special attention. The flagging of these transactions, again, will help traders and management make more informed decisions.
The bank has made some technology changes to accelerate the performance of this software, which processes large quantities of information. "We run many scenarios through time on about 10,000 risk factors for a whole lot of counterparties," Engel says. "To have that come back with a very fast response intraday, there are some clever things you need to do." One thing the bank has done is create a large compute grid of commoditized servers using software from Data Synapse. "We're looking to open up the whole grid to all applications, which means that at peak times there's that much more bandwidth available," Engel says.
The company is still building an enterprisewide view of risk and making the needed connections between existing risk data feeds and the Algorithmics system.
