In a recent article in Risk.net, a leading publication for news and analysis on risk management, derivatives and regulation, Scotiabank was recognized for its delivery of the Bank’s new risk engine in Global Banking & Markets, a system built for valuation adjustment (XVA) calculations that has resulted in both speed and accuracy improvements for the Bank’s clients.
Scotiabank’s new system combines four technologies: cloud, graphic processing units (GPUs), adjunct algorithmic differentiation (AAD) and deep neural networks (DNN), a branch of artificial intelligence (AI). Today, Scotiabank has industry-leading compute power, which allows traders to gain greater insight into the risk in the derivatives market, and ultimately provide customers with more timely, accurate pricing.
Karin Bergeron, Managing Director and Head of XVA Trading at Scotiabank, described how during the onset of the pandemic when markets became stressed, Scotiabank’s new risk engine was able to run multiple scenario analyses on-demand to provide a rich understanding of market sensitivities. “We tested different flavours of correlation and used these to guide our hedging strategy as we saw which scenario unfolded in the markets on any given day. The migration to this new platform has given us greater visibility into our risk,” Bergeron told Risk.net.
In the article, Andrew Green, Lead XVA Quant at Scotiabank, talked about Scotiabank’s work with deep neural networks. “We were first to market using AAD and GPUs for XVA. And the first to use deep neural networks, for XVAs in production full stop.”
Read the full article from Risk.net here.