Service Navigation

Interview: Dr Yves J. Hilpisch on VSTOXX Advanced Services

Release date: 14 Aug 2013 | Eurex Exchange

Interview: Dr Yves J. Hilpisch on VSTOXX Advanced Services

Q: Given their properties, volatility derivatives are considered as 'derivatives of derivatives' – in fact, many investors believe these products are highly complex. With its web-based VSTOXX® Advanced Services application, Eurex Group has embarked upon a new way to support trading in these products using an innovative web tool. Which user group did you particularly have in mind when you conceived the tool?

Dr Yves J. Hilpisch: In principle, we developed the VSTOXX® Advanced Services for everyone interested in volatility derivatives. There are, however, two main target groups: For professional investors, volatility indexes such as the VSTOXX® and the corresponding derivatives contracts are increasingly gaining importance for areas such as asset management, trading or risk management. Given that the derivatives usually have a strongly negative correlation with their base index (the EURO STOXX 50® Index), this product category offers attractive opportunities to diversify European equity portfolios. At the same time, they facilitate the cost-efficient execution of pure volatility trading strategies.
For users with an academic interest, volatility, volatility indexes and derivatives based thereon remain an active field of financial markets research, both theoretical and empirical. For researchers, university students and teachers alike, the new offer provides a good entry point, whether they're looking for theoretical basics or more hands-on aspects.

Q: What are the specific benefits of the new platform for users, in terms of content and methodology? Where do you see the highest added value?

Dr. Hilpisch: VSTOXX® Advanced Services address key issues which arise upon an in-depth analysis of volatility derivatives:

  • Data analysis: How can you efficiently analyze historical data as well as different strategies using VSTOXX® and the corresponding derivatives?
  • Index calculation: How are the VSTOXX® index and the related sub-indexes calculated? How can you implement the calculation(s) yourself – technically speaking – and how can you reconcile the numerical results?
  • Valuation of derivatives: How can you value futures and options on the VSTOXX®, and how can you implement such calculations from a technical perspective?
  • Calibrating models: What is the fundamental approach to calibrate a valuation model for options on the VSTOXX®?

Nowadays, research generally – and financial markets research in particular – is characterized by a trilogy of languages: the written language, the descriptive mathematical language and a suitable programming language. VSTOXX® Advanced Services offer a series of tutorials based on this trilogy; besides theoretical concepts, these tutorials also illustrate the technical implementation using scripts written in Python, a programming language. This permits users to directly understand the background information and formal basics provided.

For the more technically savvy users, the Python scripts provide an entry point for mapping this attractive product category in their in-house IT systems, for trading and/or risk management purposes. Finally, academic users may want to take the scripts as a starting point for their own analyses and research activities.

Q: The Python programming language is a key element for replicating VSTOXX® data. Why did you choose this language, and what makes Python a candidate for a benchmark programming language used in the financial sector?

Dr Hilpisch: Indeed Python is about to become the benchmark programming language in the financial sector, due to the following features (among other factors):

  • Open Source: Python itself (as well as the majority of supporting libraries and tools available) is open-source software, which can be used for free. Of course, this is a major benefit in an academic environment in particular, as banks and other financial institutions may draw upon professional support and other services offered by duly qualified experts.
  • Syntax: Python's syntax is quite compact. In many cases, it is very similar to a descriptive mathematical language. Python is usually easy to understand, even for those users having little or no experience programming in another language. From a corporate perspective, the compact code offers major benefits – both in terms of development and maintaining larger financial applications.
  • Performance: Even though Python itself is an 'interpreted' language, the majority of libraries required in the financial space (such as NumPy) have been implemented in a high-performance manner, using C or Fortran. In cases where it is impossible to implement any performance-critical function in this way, Python code may be compiled directly, e.g. using libraries such as Numba or Cython. Moreover, Python also facilitates the efficient use of parallel program execution, either using multi-core CPUs distributed via clusters, or using highly-parallel processing via GPGPUs (General Purpose Graphical Processing Units).
  • Multi Purpose: Python's growing popularity in the financial sector is not only due to the fact that the language facilitates very efficient interactive data analysis, or the implementation of complex financial algorithms. Python in fact supports all issues which are relevant to software programming in a production environment. In many areas, this avoids having any breaks between different programming languages or the duplication of effort.

Q: VSTOXX® Advanced Services was developed, above all, to facilitate understanding a complex derivative contract – the VSTOXX®-Future – and trading the contract. Having said that, the structure and concept might be transferable to other complex derivatives. Which contracts do you consider to be eligible in this context?

Dr. Hilpisch: The approach can, of course, be applied to virtually any product category – naturally, more complex products tend to benefit most. Ultimately, what is important to any market participant is being able to understand those financial products traded and held in their portfolio in detail, as well as to map them in their IT systems without any problems. The easier both objectives can be achieved, the better for market participants.

VSTOXX® Advanced Services precisely aim to cover both these aspects – on the one hand, having a central point of contact where I can find virtually everything concerning the product category involved. What is just as important, on the other hand, is being able to draw upon directly-usable Python scripts that allow users to put the knowledge imparted into practice, both technically and numerically, without having to devote major efforts.

A product category for which the tool is very suitable indeed are Eurex Group's dividend futures. But of course, the tool would also work for those products considered 'benchmarks' today – such as options on the DAX® or the EURO STOXX 50®. The focus for those products would be on advanced topics such as implied volatility surfaces, or on more complex option pricing models and their numerically-efficient calibration.

Q: The content of the tool box also shows a clear focus on the valuation of volatility options – based on the model published by Andreas Grünbichler and Francis A. Longstaff in 1996. What was the reason why you chose exactly this model?

Dr Hilpisch: In their current version 1.0, VSTOXX® Advanced Services cover a broad range of issues concerning the VSTOXX® index and the derivatives offered by Eurex Group based on the index. We selected the Grünbichler/Longstaff option pricing model because it is relatively easy to handle. Since it maps the dynamics of the volatility index, the model only requires one stochastic differential equation. For European-style volatility options, the model provides a valuation formula which can be implemented efficiently in terms of numerical efforts.

We have focused on topics such as a Monte Carlo simulation of the model, where we show how Python can be used to easily implement automated tests and analyses. Likewise, we provide a tutorial to illustrate how to calibrate the model to market prices of volatility options.

Certainly, the Valuation area of VSTOXX® Advanced Services offers itself to the presentation of additional models used in market practice today, illustrating them using Python scripts. For instance, state-of-the art variants of the model simultaneously map a base index like the EURO STOXX 50® and a volatility index such as the VSTOXX®, in order to achieve a uniform valuation basis for options on the base index – as well as on the volatility index.

In this context, the Grünbichler/Longstaff model should be seen as a possible entry point for addressing and finding answers to the key valuation issues for volatility futures and options.

Q: Of course, back-testing is a core element of this web-based application. Do you have any plans to expand the back-testing module, to include other asset classes, in order to approach portfolio managers' diversification needs?

Dr. Hilpisch: That is correct – the back-testing application is a key component of the platform provided. It allows users to analyze various investment strategies based on volatility derivatives in an extremely efficient manner – both numerically and graphically, with up to four different assets in any portfolio. Furthermore, it allows the free determination of certain parameters such as transaction costs.

In its current version, the application provides a choice of two base indexes – the EURO STOXX 50® and the DAX® – and the VSTOXX® itself (as a hypothetical investment), the VSTOXX® Future and calls and puts on the VSTOXX®.

Volatility product data for back-testing purposes is available from the launch date of the respective product, whereby the analysis period can be chosen freely. As a special feature, a yield-maximizing or risk-minimizing portfolio composition can be determined for a given period and portfolio, with a simple click of the mouse.
The architecture of this web-based back-testing application (whose core has also been implemented using Python) is flexible enough to facilitate expansion by any additional assets (or asset classes) in the future. For instance, it is conceivable to include the EURO STOXX 50® and/or the DAX®. Likewise, numerous other asset classes might be offered – such as bonds (via the REX index and/or the Euro Bund Future) or commodities, using corresponding indexes.

We have already expanded the application – for internal demonstration purposes – by adding more than fifty different assets into operation. Hence, there is sufficient potential for expanding the core functionality as well as the overall offer.

Yves J. Hilpisch – biographical details

• Graduate in business administration (banking and financial markets)
• Dr rer.pol. (option pricing theory)
• Lecturer in mathematical financial theory at Saarland University
• Managing Director of Visixion GmbH – The Python Quants (
• Managing Director of Continuum Analytics Europe GmbH – Python Data Exploration & Visualization (
• Private web site: