Over the last several years, leading quant teams have been working to modernize and optimize risk management practices. They are doing this by applying Algorithmic Differentiation (AD) to accelerate solving complex pricing and analytics problems.
For those unfamiliar with the concept of AD, it is a mathematical technique that radically speeds up the calculation of greeks. In fact, for typical portfolios it is not uncommon to get a 1000x improvement in calculation speed when compared against finite difference methods.
For most financial organizations, relying on traditional finite difference methods (also known less formally as “bump and grind” or “bumping”) to calculate greeks and other sensitivities is the status quo. The biggest problem with bumping is that it is notoriously slow. As slow as a turtle one might analogize. An entire portfolio valuation is required for every sensitivity calculation. What this means is that firms are forced to sacrifice intra-day risk reporting and pre-trade risk while relying on overnight snapshots of their exposure.
In an effort to speed up bumping, many firms have thrown considerable hardware at the issue. However, investment in improvements like graphics processing units (GPU) have still not fully solved the problem. At the end of the day, even if you strap a rocket onto a turtle, it’s still a turtle. It may be a faster turtle—but a turtle nonetheless. Using AD, however, the speed up is in the realm of 100x – 1000x. Managing exposure is no longer an overnight activity, but a pre-trade one.
For teams with complex multi-asset, multi-currency portfolios that need to bump, corners are often cut to reduce run-time, including not bumping every quote or bumping curves altogether with a parallel shift, twist, or other aggregate bump. Another major issue with bumping is that typically firms cannot afford to bump every quote. So then, how can one properly determine what to bump? From a risk management perspective, determining this is akin to trying to navigate a dark and dangerous landscape with only a small flashlight.
However, using AD, you don’t have to make guesses on which quotes to calculate portfolio sensitivities. You have crystal-clear visibility into everything. Essentially, you trade your flashlight in for a floodlight, yielding a complete view of the risk landscape. Sensitivities to every relevant quote – including intermediate ones – are available for a fraction of the cost.
So taking into account all of the associated benefits of AD, you may be wondering, why isn’t every firm using it? Well historically the biggest challenge to adopting AD has been the high implementation cost, which has put it beyond reach for many firms. Fortunately today there are more affordable tools available in the marketplace. FINCAD F3 incorporates its own patented implementation of the method we call Universal Algorithm Differentiation (UAD). UAD offers financial institutions comprehensive, real-time measurement of the sensitivities of a portfolio, trading book or fund. This helps you take advantage of more trading opportunities by allowing rapid assessment of the impact of a new trade’s exposure on your portfolio.
In speaking with clients using F3, many say the biggest advantage of FINCAD’s UAD is that it is truly universal and future-proofing. You can be confident that whatever you trade—whether it is exotic or vanilla—and whenever you trade—whether it be today, tomorrow or next week – it will be covered and handled in a consistent manner. Furthermore, by using UAD to extend your models you can cover new markets and take advantage of new opportunities all while safely managing your exposure throughout.
For more information on how AD can help you advance risk management at your firm, download our eBook: Improve Trading Performance with Algorithmic Differentiation.
