Stress-Test Your Portfolio Without Risking Your Data

As the US economy enters headwinds and credit stays tight, the risk of bank failures continues to mount. The 2008 financial crisis taught us asset crashes and bank failures can cascade into each other across all scales of the financial system, and today, regulators are increasingly paying attention to small banks. It is increasingly likely that small banks will eventually have to do the same stress-test simulations that the larger players do to minimize risk and ensure their asset portfolios can survive any unwelcome economic surprises.

Monte Carlo simulations are one of today’s most important tools for assessing risk and estimating the capital requirements needed to ensure solvency. But to accurately assess the impact of long-tail events, thousands, possibly tens of thousands of iterations must be performed, which has a high computational cost. This kind of computation requires a high-performance machine, one with hundreds of hardware threads and rock-solid reliability, the kind of performance and stability only IBM Power can provide.

IBM Power10 CPUs deliver massive floating-point throughput with up to 192 hardware threads per socket, each running as fast as 4.0 GHz. This kind of compute power brings down the total cost of computation and enables a fast turnaround time on decisions. Even with processors this powerful, some simulations can run all night, and that’s where IBM’s industry-leading reliability and zero-cost security ensure that simulations can run as long as needed without any worry about a server going down or malicious attackers compromising data.