FRTB and Basel III, like many of the regulatory reforms of recent years, have added new layers of complexity to the banking world. However, the structural shifts and technological demands of profitability and reforms have started to draw up contemplation and conversation about the future of quantitative finance – a key focus of this year’s 2016 Quant Summit (Formerly Quant Congress). On the heels of the conference, Risk Magazine Quant of the Year Dr. Alexander Antonov, of Numerix, joins CMO James Jockle to discuss how banks are evolving in this next phase of banking and addresses the role technological advancements will play in the evolution of quants themselves and how they’ll need to adapt to this new environment.
Jim Jockle (Host): Hi welcome to Numerix video blog, I’m your host Jim Jockle. It is a casual summer Friday here in July and we’re going to sit down and have a nice little casual conversation with our Quant of the Year, Dr. Alexander Antonov, Alexander welcome.
Alexander Antonov (Guest): Thank you, thank you Jim.
Jockle: So you were just at Quant Summit this week here in New York City. One of the trends that I’m starting to see on the conference circuit right now is looking at “What is the future of quantitative finance?” Arguably, there’s been a shift in terms of or merger if you will in terms of front office and risk management. What are some of the comments and concerns that are going on in these conversations?
Antonov: Yeah, as you said you have quite a big merge between the front offices, the back offices and the middle offices where we calculate the risk metrics like XVA for different aspects of the risk. So we don’t only need to avoid the price, but also all the valuation adjustments, some of them are for the banks, some the regulators organize, and so on.o we have more and more new adjustment. So this calculation becomes really bespoke. For example, if you have initial margin we should come up with the initial margin valuation adjustment, which changes your prices in the future. A lot of regulatory new things coming from FRTB and other Basel III and so on. So a lot of things are done around the portfolios, big portfolios, and so this changes the nature of quantitative finance.
Jockle: So, arguably some of the Basel, FRTB calculations themselves are pretty straight forward. They’re prescriptive but yet there are still levels of complexity that cannot be dismissed within these calculations, is that accurate?
Antonov: That’s right, that’s absolutely right. So if you know how to relate the sensitivities of our products or portfolios we can come up with the quantities required by FRTB or Basel. However, if you want to calculate the valuation adjustment in the future, this will require some more complexities from our modeling side. That’s the difference for the current calculation required by the regulators and the future which banks may be interested to see how much money they will spend to raise this capital, this required capital by this FRTB and Basel III regulation.
Jockle: Now one of the other changes that we’re continuing to see is, in really in terms of quantitative finance. You know, obviously there’s some changes and evolutions still within the pricing model space but more technology advancements in terms of performance and AAD seems to be gracing the pages of risk magazine.
Antonov: That’s right, exactly right. Yes, as far as we need more and more information about the current and future prices including the sensitivities and we need to calculate them rapidly and almost at the same time as pricing – maybe less frequently, not in real time, but still we need them quite frequently. So that is why we need the speed of calculation and the speed of calculation can come either with new hardware things like GPU or new software things like AAD, which is a generalization which permits us to get the derivative of whatever price of XVA revenue.
Jockle: And that’s also coming in the back work of some of the work you’ve done in terms of American Monte Carlo as well, isn’t that correct?
Antonov: This is true, these American Monte Carlo methods can help us to calculate the sensitives and also the XVAs rapidly, so this is one of the mechanisms which permits us to do these things properly, almost at the real time.
Jockle: But I think one of the common misperceptions in the market, with whether it be GPUs or American Monte Carlo or AAD is that a one-size fits all approach is not necessarily the most efficient and practical approach in terms of kind of meeting that scalability and performance requirements.
Antonov: This is true all of these methods can be somehow bespokely used for different levels and instruments or portfolios and different levels of the XVAs. So once you of course use these methods properly, American Monte Carlo will underlie all the complicated instruments involved in XVA. But for example, for less complicated instruments we can get an analytical means, we can also sometimes we know just an analytical price is a function of the parameters so we don’t use AAD. The wisdom is of course to apply the things which you know they need for concrete tasks.
Jockle: Also just kind of looking at the trades there’s been a lot of notable movements in the market of different quant teams. Other areas kind of changing their asset class focus and different banks ramping up their focus but also starting to see a little bit of a migration of quantitative resources to the buy side even more. Any insights and thoughts on kind of what’s going on in terms of the shift and priorities of institutions now as it relates to quantitative teams.
Antonov: I would say that now quants are not necessarily writing their formulas with a pen and a sheet of paper, but we should also be hands on for the programing, hands on for the technology like the GPU and other hardware accelerators. So the quants should understand all the fine details of FRTB and Basel agreements. So a lot of extra information is needed for quant work. Having said that, they will probably need less knowledge in mathematics, but more…it’s always the compensation. If a quant goes more in the technology end of the details of the capital and other agreements, then he doesn’t have the time and maybe even the desire or whatever to try and write things with a pen and paper. The models are less complicated but the instruments and the whole aggregation of the results are more complicated.
Jockle: So you know, I think arguably in every career path and discipline as this point, we’re all challenged to know more about different things and it seems like the same for the quant world at this point.
Antonov: For sure. Yeah.
Jockle: Well Alexander I want to thank you so much for your insights and I hope you’ll get some rest this summer after your negative rates world tour and I’m sure we’ll l look forward to some new, interesting papers from you in the fall.
Antonov: Thank you very much, Jim. Thank you.
Jockle: Thank you. That’ll conclude todays video blog. And of course we want to talk about the things that you want to talk about so follow us on LinkedIn and on Twitter @nxanalytics. Thanks, Alexander.
Antonov: Thank you, Jim. Thank you.