Financial Engineering Practitioners Seminar: Tim Klassen
Date: November 16, 2015 from 6:00 pm to 7:30 pm EST
Location: 501 NWC
Contact: For further information regarding this event, please contact IEOR Department by sending email to info@ieor.columbia.edu .
Info: Click Here to Visit Website.
Bookmark and Share

Title: Equity Implied Vols For All: Part 1, Pricing with Cash Dividends

Abstract: The pricing of (even) vanilla options on underliers with cash dividends is a surprisingly active research subject, for both European and American exercise style. Neither on the listed options side (calls and puts) nor on the flow/structured side of custom vanillas or light exotics are market participants in agreement on what model to use, nor on what an efficient practical implementation of the chosen model would be. The modeling problem boils down to the question of what a proper generalization of the Black-Scholes model to the case of cash dividends is, i.e. what should replace simple geometric Brownian motion (GBM).

We discuss this question with the aim of taking a first step towards a rationalization and normalization of the equity volatility market. We compare the two main classes of models in use, namely the "spot model"(piecewise GBM) and several "hybrid models" (shifted GBM). We are interested in consistency, simplicity, speed, and generality (covering all traded vanilla options, dividend and borrow rate assumptions, etc). We show that: (i) all hybrid models are closely related on a mathematical level -- despite qualitatively different financial properties -- with simple and accurate relationships between calibrated parameters (borrow costs and volatilities) for both European and American options with cash dividends; (ii) all hybrid models allow accurate and very fast pricing of vanilla options using fine-tuned tree methods; (iii) some hybrid models have essentially all the desired properties outlined; in particular, we describe a hybrid model closely related to the spot model, motivated by the spot-strike adjustment idea of Bos and Vandermark.

Along the way we discuss the calibration problem that vanilla market participants face.

Bio: Tim Klassen is an equity derivatives quant that has built and headed modeling teams in both options market making and flow/exotic derivatives. Besides discovering the first classes of modern, arbitrage-free implied volatility surfaces, he is best known for designing the (new) VIX, the widely quoted fear gauge. He is an expert in fast and robust pricing and calibration methods, as well as volatility arbitrage and automated risk management.

He is now starting a venture to provide equity options market participants with robust real-time and historical implied borrow curves and arbitrage-free volatility surfaces, as well as associated services like valuation, risk, and PnL decomposition.

Tim started his career in finance as an equity derivative quant in Emanuel Derman's group at Goldman Sachs in 2000. There he solved the long-standing problem of coming up with robust local volatilities by designing and calibrating the first modern implied volatility curves. He also designed the "new VIX" that the CBOE started disseminating in Sep. 2003.

In 2008 he joined Getco to build the options market making analytics and quant team.The pricing and fitting framework was designed from the ground up using a modern Bayesian approach, where all inputs and outputs have error bars, using priors as necessary. With careful attention to the modeling of dividends, borrows, vol time, and vol curves he designed algorithms that allow for very fast and robust calibration of any (reasonable) market: SPY and other ETFs with dividends or high borrows, single stocks like AAPL with W-shaped vol curves around earnings, and robust curves for illiquid names. Getco became one of very few successful new options market makers since the penny pilot began; it was competitive even in super-liquid names like SPY, AAPL, KOSPI.

He received his Ph.D. in particle physics from the University of Chicago. As a postdoctoral research scientist at Cornell and Columbia University he played a leading role in the simulation of QCD using improved lattice schemes, leading to orders of magnitude speed-ups. He has authored over 27 articles with 1500+ citations in particle and statistical physics, as well as finance.