Quant Congress 2015:  Risk Management, Model Validation and Regulations

Quant Congress 2015: Risk Management, Model Validation and Regulations

Jim Jockle (Host): Welcome to Numerix Video
Blog, your expert source for Derivative trends and challenges. I am your host Jim Jockle.
Quant Congress in its 13th year was just held recently this week and joining me today is
SVP of financial engineering of Numerix, Dan Li, to discuss some of his thoughts and insights
from the conference. Welcome Dan. Dan Li (Guest): Hi. Thank you. Jockle: So Dan, to Quant or not to Quant seems
to be kind of the theme as we’re looking ahead to risk management and there seems to
be a little bit of a controversy as it related to what is the future role of quantitative
finance in some institutions. So, maybe we just kind of jump right in there and give
us a little bit of the issues of the debate that happened at the event. Li: Yeah. Yeah, actually I attended a panel
discussion actually specifically about this market trend and it was very interesting and
we can also clearly see that after Lehman collapsed and after the financial crisis,
we can see the market trends shift really for the Quant roles really from the front
office role. Really focus more on the main office role, risk management, model validation
and focus on the regulations. There are a lot of regulations really taking on that risk
management to Retilaty, from liquidity to counterparty risk, to market risk, so it covers
everything including the capital. So those have really been a trend we have seen in the
market. Jockle: So how is that changing the profile
of the Quant, or the day to day? You know, you think of back into the days of the boom,
or innovations of products coming to market. There are still so many quantitative challenges
in meeting all of these new regulatory requirements or changes in valuation. How is that changing
the skill set required? Li: Yeah, that’s a very good question. And
actually, there are also some debates related to these topics. And we can see the trend,
I mean in terms of the skill set, that for front office quants that used to be very mathematical
oriented, meaning, they typically would be PhD’s in physics, or PhD’s in mathematics
or statistics, they were very attractive back to 1990’s and 80s. And now we can see that
those type of skill sets are very useful and we still need all those types of skill sets.
In terms of risk management, and in terms of underpriced solutions, and usually we’ve
see more trends in terms of the skill sets, we may also need to see good probabilities
and knowledge in the probabilities and statistics and also some of the econometrics, portfolio
optimization type of the skill set. So those we can see the trend on that area. Also, in
terms of market risk, counterparty risk, that actually needs even higher skill sets in terms
of the model understanding because you really need to understand more than like one asset
class. It’s like that really pushes, in terms of the skill set, it’s not just a
simple shift, and people usually assume you have a PhD background, but ideally also have
a mathematical finance background as well. So, that is sometimes where we see the mix.
Some of the banks, actually they have a slightly decreasing trend in terms of hiring the fresh
Phd’s, but they prefer some of the mathematical, finance degree, either fresh from school or
maybe with a few years of experience. That is what we can see.
Jockle: So, looking ahead, I think one of the things when you think about CCAR or some
of the other regulations that are coming down, it’s also the evolution of projections into
better understanding the future. So, no longer just Monte Carlo and backwards looking into
historical data, but where are you seeing the role of, you know, simulation technology
and being played going forward. Li: Yeah, that’s actually really true, especially,
the whole world, not only finance, actually moved to a big data world, so from that reason
especially in the risk management world or model validation world, that we really see
the trend and the demand on the scalable, like underpriced solution. So, for that reason,
in terms of the skill sets they really actually shift; we need a good programmer, we need
really an architect, on all those scalable solutions side and in terms of Monte Carlo,
also we need there’s market needs that really demand on the high computing, really efficient
computing, in terms of the convergence, in terms of to really handle the large dimension
of the whole portfolio. It’s really, we move to the very complex world, even more
exotics than before, even most of the portfolios tends to be more vanilla, compared to several
years ago before the crisis. Jockle: Other trends I’d love to get your
perspective on is in the insurance side of the business we’ve seen actuaries embracing
more typical capital markets modeling and get a wider understanding of different risk
factors as it relates to securities performance. But in the typical quant world we’re starting
in, and you even mentioned it a little bit before, more of the introduction of econometrics
into the through process. You know, how do you see that evolving in realms of capital
markets and banking and finance? Li: Yeah, that’s actually back to the old
80’s. Those type of the trends it’s like a risk neutral role with all the modeling
vs econometrics or statistics or like a historical estimate and do the forecasting. So those
are like two types of world, running simultaneously: one is really focused on the portfolio, focused
on portfolio management or focused on like hedge fund hunting whatever the profit that
they can get from the market and from the market maker starts dealer specs they typically
really within the same consistent risk neutral framework not for all asset class and using
that for the risk management or al the risk measures. And now we can see actually the
two worlds, seems like, not merged, but definitely we can see the converging to the degree, meaning
people need to think especially like the insurance industry that you mentioned, insurance used
to use like the historical estimate, everything for the future simulations could be under
the real-world measure. And, but in the capital market, actually, people when they try to
do the hatch, try to do pricing, and get the sensitivities of the value at risk or maybe
counterparty risk on top of those portfolios those are all in the risk neutral. So in other
worlds it’s really the real-world versus risk neutral but we can see actually the trend
bridge them together, that’s also ideal, actually it would also need to be within the
same framework but maybe with resampling techniques or Monte Carlo techniques that really can
bridge those two together; and from the real-world, the modeling part that could involve like
statistical, econometric, and other advanced metrics. And for the risk neutral that would
still be the market standard and calibrate to the market and then bridge them together.
So that’s what we can see also the trend. Jockle: And also, I’m sure being benefactors
of better data, more robust data, and longer data sets, is only going to help advance that.
So, I want to turn to the other side of the debate. So I think it’s like 12 or 13 currencies
we have negative rates, low-yield environment, the US is poised to raise rates which puts
everybody in some sort of concern, even though its priced into the market. So, as investors,
we need to make our return. So, finding and seeking alpha, and what is the debate on the
other side of not just the risk management change, but what are some innovations in terms
of quantitative finance that people are thinking about in terms of getting return on their
risk? Li: Yeah, on the one side, for negative rates
or others, we can see that some of the after crisis, some of the market behaviors or market
trends really push the market quotes break the fundamental. Like the Russian … [9.32]
or even push some of the existing standard models to the limit, in this case as what
you just mentioned the negative rate, that people traditionally assumes all those arbitrage
free conditions, then all the curves need to be, especially the forward rate, needs
to be positive. That’s really the foundation for the arbitrage free. But after the crisis,
you could see the Euro, you could see the Swiss and that’s all the negative rates.
This is what we can observe in the market, but in terms of the typical, traditional,
modeling, that actually breaks all the modeling assumptions there, that’s what we can see
from that point of view; actually we do see the many innovations I mean in terms of the
shift to SABR, or even like Free Boundary SABR, like a more advanced approach to really
tackle the heart of these problems in the market. On the other side of the world, that
the seeking alpha we can see, I mean most of the portfolio management or buy-side, they
would still adapt and have the deep understanding of the existing model framework and including
the limitation, and then from their side what they need is really a strategy to seeking
alpha or reduce various in terms of the more optimization and maybe how you really control
the frequency of rebalancing and also in terms of a better approach basically for the portfolio
management and also for that reason because usually we are in the big data world and then
they typically would also seeking to tackle the big data challenges in terms of they typically
need a more scalable solution and maybe a more flexible framework can help them with
a more different strategy and to enhance their returns.
Jockle: Well Dan, I want to thank you so much sharing your thoughts on Quant Congress and
of course we want to talk about all the trends that you want to talk about, so please follow
us on LinkedIn or on Twitter @nxanalytics. Thank you, I’m Jim Jockle.

Leave a Response

Your email address will not be published. Required fields are marked *