Blog

Get the latest updates and news from FINCAD. Subscribe and never miss a post! 

Subscribe

Blog
With AD, Everything is Illuminated
By Russell Goyder PhD | August 15, 2017

For most financial organizations, using traditional finite difference methods (also known as “bump and grind” or “bumping”) to calculate greeks and other sensitivities is the status quo. Unfortunately, the downfall with bumping is that typically firms cannot afford to bump every quote in their portfolio. Thus they are forced to rely on intuition when deciding which limited set of bumps to make. From a risk management perspective, determining this is like navigating a dark and dangerous landscape with only a small flashlight.

Using Algorithmic Differentiation (AD), on the other hand, you don’t have to guess about which quotes to calculate portfolio sensitivities. Everything is illuminated. 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.

Below are five resources that explain more about the many advantages of using AD for risk calculations, and why every firm should be utilizing this proven method.  

1. Video: Algorithmic Differentiation and the Buy Side

In this video interview, John Hull, PhD and Alan White, PhD discuss adjoint algorithmic differentiation (AAD) and the impact it is having on the buy-side. “AAD is very consistent with the evolution of the derivatives market, which has moved from exotic products to now exotic risk management. There are now incredibly computationally-intensive procedures firms have to use to calculate the appropriate risk measures,” commented White. He went on, “One might say, regulators are pushing institutions in the direction of AAD. It is essentially impossible to measure risk of OTC derivatives without using Monte Carlo simulation, and so anything that can speed up computations is very helpful.” Watch the video.

2. eBook: The Case for Algorithmic Differentiation

In a world where markets — and risk — move at a remarkable pace, traditional risk measurement strategies no longer cut it. Finite difference methods, otherwise known as bumping, provide a slow and limited view of risk. AD offers a fully comprehensive analysis of portfolio risk in real-time. This technique empowers financial organizations to proactively hedge and manage exposure, resulting in: more robust market risk management, optimal deployment of capital and greater potential for profit. Download the eBook to learn more.

3. Blog Post: How to Speed Up Risk Calculations by 1000X

For firms using AD, it’s not uncommon to get a 1000x improvement in calculation speed when compared against finite difference methods. So why isn’t every firm using this method? 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. Read the blog post for more information.

4. Webinar: Improve Trading Performance with Algorithmic Differentiation

Speed was once a roadblock to intra-day risk analysis. But today the latest technology has broken through, delivering fast, accurate valuation and risk calculations to traders and portfolio managers who are using algorithmic differentiation to drive better trading decisions. View this on-demand webinar to discover the many benefits that real-time risk can deliver to the front office, including more accurate hedging of complex portfolios and superior risk management.

5. Blog Post- Algorithmic Differentiation: Shattering the Myths  

In this blog post, I set the record straight on some inaccuracies I’ve read about concerning AD.  One misunderstanding pertains to the role of AD in hedging. Greeks generated by AD are exact derivatives, not the differences that many firms rely on to hedge. As such, some believe that the Greeks produced by AD need to undergo additional transformations before they can be used for proper hedging. For this reason, their recommendation is to stick with the conventional bumping method that produces differences out of the gate. In the blog, I counter this erroneous line of thought, explaining why there are no additional transformations needed to use AD sensitivities in hedging, over bumped ones.

For more on AD and other related topics, check back here and in our resources section, where we post new content regularly. We have a full calendar of timely and educational topics lined up for the remainder of this year.   

About the author
Russell Goyder PhD
Russell Goyder PhD
Director of Quantitative Research and Development | FINCAD

Russell Goyder, PhD, is the Director of Quantitative Research and Development at FINCAD. Before joining FINCAD’s quant team in 2006, he worked as a consultant at The MathWorks, solving a wide range of problems in various industries, particularly in the financial industry. In his current role, Russell manages FINCAD’s quant team and oversees the delivery of analytics functionality in FINCAD’s products, from initial research to the deployment of production code. Russell holds a PhD in Physics from the University of Cambridge.