The ability to swiftly, efficiently and accurately analyze large datasets, exposure to various market and credit risks, and identify trading opportunities has emerged as one of the key differentiators for large numbers of capital markets firms in recent years, given the complexities of the analysis underlying such undertakings.
One firm that appears to be leading that charge is Swiss investment bank Credit Suisse, which wins this year’s AFTA for the best analytics initiative, thanks to its capital stress-testing, financial planning and forecasting platform, dubbed ProPL, an abbreviation of its official name: Projections Platform. As part of the bank’s comprehensive capital analysis and review (CCAR) stress-testing program, it needed to develop a platform for managing the front-to-back execution of capital projection models.
Early in 2017, Credit Suisse successfully filed for a US patent for the platform, an offering that seeks to improve the bank’s ability to forecast pre-provision net revenue (PPNR), balance sheet, risk-weighted assets (RWA), losses, and other key financial metrics for each area of the business. According to the bank, “these forward-looking capital projections are based on data-driven models that consider the impact of multiple financial and economic variables over an extended period of time. Projection results inform strategic business planning by providing a capital adequacy assessment using defendable, repeatable and quantitative processes.”
ProPL sources over 100 data feeds from across the organization and tracks and tags all incoming data assets, which helps with BCBS 239 compliance and provides data quality checks and reports. It also houses and conducts the execution of the firm’s capital projection models, automatically resolves the dependencies between models and sequences their execution accordingly. The architecture that ProPL is built on is scalable so that any calculation produced can be traced back to the exact set of models used and their precise configuration at the time, along with every data input, model parameter, output adjustment, and overlay applied.
The platform has over 100 models representing over 90 percent of Credit Suisse’s revenue within the Intermediate Holding Company (IHC). CCAR submissions, which typically take two to three weeks to run, can be executed in under 30 minutes, according to the bank.
Last year this category was split into two separate categories, with Morgan Stanley taking the award for the sell side, while JPMorgan Asset Management won the buy-side award.
Bryan Cross, who heads UBS Asset Management's QED group, joins to discuss alternative data and AI.Subscribe to Weekly Wrap emails