Columbia Home
Practitioners’ Seminar 2021

The seminar takes place in the Spring of 2021, Tuesdays and Thursdays 7:40pm — 8:55 pm.

Organizer: Lars Tyge Nielsen

In some cases, the speaker will not make copies of the presentation available — to protect intellectual property or comply with company rules.

Registered participants, please join via Zoom in Courseworks.

Schedule of Presentations

Click here for the Schedule of Past Presentations.

Tuesday, March 2, 2021 — Spring Recess, no seminar
Thursday, March 4, 2021 — Spring Recess, no seminar
Tuesday, March 9, 2021

Speaker: Julien Guyon, Bloomberg L.P. and Columbia University
Julien Guyon is a senior quantitative analyst in the Quantitative Research group at Bloomberg L.P., New York. He is also an adjunct professor in the Department of Mathematics at Columbia University and at the Courant Institute of Mathematical Sciences, NYU. Before joining Bloomberg, Julien worked in the Global Markets Quantitative Research team at Societe Generale in Paris for six years, and was an adjunct professor at Universite Paris Diderot and Ecole des ponts ParisTech. He co-authored the book Nonlinear Option Pricing (Chapman & Hall, 2014) with Pierre Henry-Labordere. His main research interests include nonlinear option pricing, volatility and correlation modeling, and numerical probabilistic methods. A big soccer fan, Julien has also developed a strong interest in sports analytics, and has published several articles on the FIFA World Cup, the UEFA Champions League, and the UEFA Euro in top-tier newspapers such as The New York Times, Le Monde, and El Pais, including a new, fairer draw method for the FIFA World Cup.

Title: The Joint S&P 500/VIX Smile Calibration Puzzle Solved: a Dispersion-Constrained Martingale Schrödinger Problem

The very high liquidity of S&P 500 (SPX) and VIX derivatives requires that financial institutions price, hedge, and risk-manage their SPX and VIX options portfolios using models that perfectly fit market prices of both SPX and VIX futures and options, jointly. This is known to be a very difficult problem. Since VIX options started trading in 2006, many practitioners and researchers have tried to build such a model. So far the best attempts, which used parametric continuous-time jump-diffusion models on the SPX, could only produce approximate fits. In this talk we solve this long standing puzzle for the first time using a completely different approach: a nonparametric discrete-time model. Given a VIX future maturity T1, we build a joint probability measure on the SPX at T1, the VIX at T1, and the SPX at T2 = T1 + 30 days which is perfectly calibrated to the SPX smiles at T1 and T2, and the VIX future and VIX smile at T1. Our model satisfies the martingality constraint on the SPX as well as the requirement that the VIX at T1 is the implied volatility of the 30-day log-contract on the SPX.

The model is cast as the unique solution of what we call a Dispersion-Constrained Martingale Schrödinger Problem which is solved by duality using an extension of the Sinkhorn algorithm, in the spirit of (De March and Henry-Labordere, Building arbitrage-free implied volatility: Sinkhorn’s algorithm and variants, 2019). We prove that the existence of such a model means that the SPX and VIX markets are jointly arbitrage-free. The algorithm identifies joint SPX/VIX arbitrages should they arise. Our numerical experiments show that the algorithm performs very well in both low and high volatility environments. Finally, we discuss how our technique extends to continuous-time stochastic volatility models, via what we dub VIX-Constrained Martingale Schrödinger Bridges, inspired by the classical Schrödinger bridge of statistical mechanics. The resulting stochastic volatility model is numerically implemented and is shown to achieve joint calibration with very high accuracy.

Thursday, March 11, 2021

Speaker: Sanne de Boer, Voya Investment Management
Sanne de Boer is director of Quantitative Equity Research at Voya Investment Management responsible for overseeing the firm’s quantitative equity research agenda. Prior to joining the firm, he was a senior research analyst for quantitative strategies for Invesco. Previously, he was a research analyst for global quantitative equities at QS Investors as well as ING Investment Management, Voya’s predecessor firm. Sanne’s research has been published in the Journal of Asset Management, the Journal of Index Investing, and the Journal of Investing. He received a Ph.D. in Operations Research from the Massachusetts Institute of Technology and an M.S. in Mathematics and an M.A. in Econometrics cum laude from the Vrije Universiteit Amsterdam. He holds the Chartered Financial Analyst designation.

Title: Intangible ironies: investor mispricing of company assets on and off its balance sheet

We examine how investors evaluate the mix of company assets both on and off its balance sheet. On aggregate, they appear to correctly value tangible assets but misprice intangible assets, which have increased in economic importance. In particular, investments in stakeholder capital such as innovation, brand, and employees often go unrecognized both on the balance sheet and by investors. In contrast, the premium paid for past acquisitions which is included on the financial statements as goodwill generally fails to deliver on expectations, being written down too slowly by management and shareholders alike. Corroborating a recent surge of papers on this topic, we find that adjusting valuation metrics for the actual benefit of such intangibles leads to better performance thereof in global equity markets. More impactfully, investors can diversify value exposure by targeting companies with latent such growth assets. Our findings suggest market efficiency would be served by better accounting standards for intangible assets, allowing more flexibility on what types of investment may be capitalized while simultaneously tightening rules around impairment and amortization.
Tuesday, March 16, 2021

Speaker: Samim Ghamami, Financial Services Forum


Thursday, March 18, 2021

Speakers: Andrew Brenner, NatAlliance Securities, and Leon Tatevossian, NYU Tandon and NYU Courant

Andrew Brenner
Andrew Brenner is a senior partner and head of international fixed income at NatAlliance Securities (an Austin, Texas-based regional dealer). Andrew’s prior experience includes senior fixed-income roles at Nomura, Societe Generale-FIMAT, and Guggenheim Securities. He is a regular commentator on CNBC with Rick Santelli, and also appears with Liz Claman on Fox Business network. Andrew is quoted frequently in Reuters, the Financial Times, and The Wall Street Journal. He is on the board of Oak Hill Capital (a hedge fund), treasurer of the Maine Jewish Museum, and on the boards of two children’s pediatric cancer charities (Children’s Cancer Research Fund of New York Medical College and the Maine Children’s Cancer Research Fund). Andrew graduated from The Wharton School of the University of Pennsylvania with a bachelor’s degree in accounting and an MBA. He was also awarded a degree in international economics from Penn’s College of Arts and Sciences. He holds the CPA designation.

Leon Tatevossian
Leon Tatevossian is an adjunct professor in the Finance and Risk Engineering Department at NYU Tandon and in the Mathematics in Finance Program at NYU Courant. From 2009-16 Leon was a director in Group Risk Management at RBC Capital Markets, where he covered securitized-products market risk in secondary-trading, origination, and proprietary-trading businesses. He has thirty-one years of experience in the fixed-income capital markets, including roles as trader, quantitative strategist, derivatives modeler, and market-risk analyst. Leon’s product background includes US Treasury securities, US agency securities, interest-rate derivatives, MBSs/ABSs, and credit derivatives. He graduated from MIT (SB; mathematics) and was a graduate student in mathematics at Brown University.

Risk and Reward in the Fixed-Income Market – Where are We Now?

Dislocations caused by the Covid-19 crisis extended across all parts of the economy (corporate, consumer, and the public sector) and the capital markets. In response we saw aggressive fiscal intervention, most notably the $2.2 trillion CARES Act (signed into law by President Trump in March 2020). Additional support is expected with Congress currently debating the $1.9 trillion stimulus package proposed by President Biden. Steps taken on the monetary front include the multi-faceted bond-buying and liquidity programs initiated by the Fed. This combination of actions to: [1] underpin the consumer, business, and corporate sectors; and [2] stabilize financial-asset valuations has been unprecedented in its size.

This seminar talk (in a question-and-answer form with supporting slides) will explain some of the basic valuation parameters in the fixed-income markets, how these metrics got upended by the swift arrival of the pandemic, and how market sentiment evolved as the Treasury and Fed actions worked through the system.

Of particular interest: How reliably have the “usual suspects” guided the valuation of “risk assets,” “structure,” and “credit”? The investment performance of these sectors has always exhibited connections to the “real economy” (consumer and business) and to the decision-making and risk appetites of institutional investors. What’s the scorecard so far on the effectiveness of the policy steps? Which valuation relationships and economic metrics do we interrogate for a clear perspective on where we’ve been and where we might be headed?

Tuesday, March 23, 2021

Speaker: TBA, Graham Capital


Thursday, March 25, 2021

Speaker: Rosanna Pezzo-Brizio, New York Life Investment Management and Columbia University


Tuesday, March 30, 2021

Speaker: Kelly Ye


Thursday, April 1, 2021

Speaker: David Lando, Copenhagen Business School


Tuesday, April 6, 2021

Speaker: Andres Jaime Martinez


Thursday, April 8, 2021

Speaker: Gordon Ritter, Ritter Alpha, LP, and NYU


Tuesday, April 13, 2021

Speaker: Asset Tarabayev


Thursday, April 15, 2021

Speaker: Peter Carr, NYU
Dr. Peter Carr is the Chair of the Finance and Risk Engineering Department at NYU Tandon School of Engineering. He has headed various quant groups in the financial industry for the last twenty years. He also presently serves as a trustee for the National Museum of Mathematics and WorldQuant University. Prior to joining the financial industry, Dr. Carr was a finance professor for 8 years at Cornell University, after obtaining his Ph.D. from UCLA in 1989. He has over 85 publications in academic and industry-oriented journals and serves as an associate editor for 8 journals related to mathematical finance. He was selected as Quant of the Year by Risk Magazine in 2003 and Financial Engineer of the Year by IAQF/Sungard in 2010. From 2011 to 2014, Dr. Carr was included in Institutional Investor’s Tech 50, an annual listing of the 50 most influential people in financial technology.

Title: Multiplication, Addition and Options

Multiplication and addition are commutative and associative binary operations with the former distributing over the latter.We treat optionality as a third binary operation operating between a fixed strike price and the current known price of a risky underlying security.

We show that when all prices are real valued, then demanding commutativity and associativity of the optionality operation implies that the risky underlying security price at maturity must be logistically distributed. In this case, ordinary addition distributes over the optionality operation. We discuss applications and variations to eg non-negative domains.

Past Presentations

Tuesday, January 12, 2021

Speaker: Alexander Fleiss, Rebellion Research’s CEO Alexander Fleiss has spoken about Artificial Intelligence Investing in the Wall Street Journal, New York Times, Fox News, BusinessWeek, Bloomberg News, MIT Technology Review, Yomiuri Shimbun, Wired, Geo Magazine, The Economist and Institutional Investor. Chapter 24 of Wall Street Journal Reporter Scott Patterson’s book Dark Pools is on Mr. Fleiss. Mr. Fleiss teaches at Cornell Financial Engineering, has guest taught at Amherst College for over a decade and Yale School of Management for 4 years. Prior to co-founding, Mr. Fleiss was a Principal at KMF Partners LP, a long-short US equity hedge fund co-managed by John Merriwether of Michael Lewis’ Liars Poker. Mr. Fleiss began his career managing an Amherst College-funded Ai research project, then as an analyst & programmer for Sloate, Weisman, Murray & Co which was acquired by Neuberger Berman. Mr. Fleiss developed investment algorithms with the firm’s CEO, Laura Sloate who is now a partner at Neuberger Berman and one of the investors featured in Peter Tanous’ book Investment Gurus. Mr. Fleiss received a BA Degree from Amherst College

Title: Deep Reinforcement Learning and Nowcasting’s Fatal Flaw

Thursday, January 14, 2021

Speaker: Antoine Savine and Brian Norsk Huge, Danske Bank
Antoine Savine is a French mathematician and chief quantitative analyst at Danske Bank in Copenhagen (in-house system of the year 2015 Risk award, excellence in risk management and modeling RiskMinds award 2019). He has held multiple leadership roles in the derivatives industry in the past 25 years, including head of research at BNP-Paribas, and also teaches Volatility, Computational Finance and Machine Learning in Finance at Copenhagen University. Antoine holds a PhD in Mathematics from Copenhagen University and he is the author of the books ‘Modern Computational Finance’ with John Wiley and sons. He published ‘Differential Machine Learning: the shape of things to come’ with Brian Huge in Risk Magazine in October 2020:

Brian Norsk Huge works for Danske Bank, where he has worked since 2001. He is working in the Quant group as Chief Analyst in Copenhagen (in-house system of the year 2015 Risk award, excellence in risk management and modeling RiskMinds award 2019) with focus on FX and equity derivatives. Brian has a Ph.D. in Mathematical Finance from Copenhagen University. The thesis title is “On defaultable claims and credit derivatives”. Brian was awarded Quant of the Year 2012.

Title: Differential Machine Learning: the shape of things to come (Risk, Oct 2020)

Differential machine learning combines automatic adjoint differentiation (AAD) with modern machine learning (ML) in the context of risk management of financial Derivatives. We introduce novel algorithms for training fast, accurate pricing and risk approximations, online, in real-time, with convergence guarantees. Our machinery is applicable to arbitrary Derivatives instruments or trading books, under arbitrary stochastic models of the underlying market variables. It effectively resolves computational bottlenecks of Derivatives risk reports and capital calculations.Differential ML is a general extension of supervised learning, where ML models are trained on examples of not only inputs and labels but also differentials of labels wrt inputs. It is also applicable in many situations outside finance, where high quality first-order derivatives wrt training inputs are available. Applications in Physics, for example, may leverage differentials known from first principles to learn function approximations more effectively.

In finance, AAD computes pathwise differentials with remarkable efficacy so differential ML algorithms provide extremely effective pricing and risk approximations. We can produce fast analytics in models too complex for closed form solutions, extract the risk factors of complex transactions and trading books, and effectively compute risk management metrics like reports across a large number of scenarios, backtesting and simulation of hedge strategies, or regulations like XVA, CCR, FRTB or SIMM-MVA.

Tuesday, January 19, 2021

Speaker: Irene Aldridge, AbleMarkets
Irene Aldridge is a co-author of “Big Data Science in Finance”, (co-authored with Marco Avellaneda, Wiley, 2020), an internationally-recognized quantitative and Big Data Finance researcher, Adjunct Professor at Cornell University and President and Managing Director, Research, of AbleMarkets, a Big Data for Capital Markets company. Previously, Aldridge was, in reverse order, a quant on a trading floor; in charge of risk quantification of commercial loans; Basel regulation team lead; technology equities researcher; lead systems architect on large integration projects, including web security and trading floor globalization. Aldridge started her career as software engineer in financial services.Aldridge holds a BE in Electrical Engineering from Cooper Union, and MS in Financial Engineering from Columbia University, and an MBA from INSEAD. In addition, Aldridge studied in two PhD programs: Operations Research at Columbia University (ABD) and FInance (ABD). Aldridge is the author of multiple academic papers and several books. Most notable titles include “Real-Time Risk: What Investors Should Know About Fintech, High-Frequency Trading, Flash Crashes” (co-authored with Steve Krawciw, Wiley, 2017), “High-Frequency Trading: A Practical Guide to Algorithmic Strategies and Trading Systems” (2nd edition, translated into Chinese, Wiley 2013), and “The Quant Investor’s Almanac 2011: A Road Map to Investing” (Wiley, 2010). Her recent academic publications include “Neural Networks in Finance: Design and Performance” (with Marco Avellaneda in the Journal of Financial Data Science, 2019), “Big Data in Portfolio Management” (Journal of Financial Data Science, 2019), “ETFs, High-Frequency Trading and Flash Crashes” (Journal of Portfolio Management, 2016), and “High-Frequency Runs and Flash Crash Predictability” (Journal of Portfolio Management, 2014). Aldridge presently serves on the Editorial Advisory Board for the Journal of Applied Data Science to Finance.

Title: Big Data as a New Paradigm in Finance

Big Data shows promise delivering unbiased optimal factorization even in a noisy environment with substantial missing observations. This talk will discuss applications in Finance, including optimal driver identification and future prediction.
Thursday, January 21, 2021

Speaker: Mikhail Smirnov, Columbia University

Title: Leveraged ETFs and Dynamic Portfolio Management of Classic Portfolios.

Leveraged ETFs provide a convenient mechanism to dynamically change portfolio exposure and can be successfully used to construct robust portfolios that perform well during equity market drops. Instead of leveraging with borrowing at the portfolio level, we can use a portfolio of leveraged ETFs. We consider several classic balanced stocks/bonds portfolios created with leveraged ETFs but without borrowing money at the portfolio level and show that they present a very attractive risk-adjusted alternative to just stock index and classical stocks/bonds portfolios without leverage. The performance of these types of portfolios first proposed in 2017, in recent years including 2020, was surprisingly robust. A classical portfolio insurance strategy of Black-Jones-Perold can be easily implemented with leveraged ETFs. We consider more complex dynamic portfolio strategies and how they can also be implemented using leveraged ETFs as well as consider a Dynamic Leverage a risk measure extending classical VAR-based risk measures.
Tuesday, January 26, 2021

Speaker: Praveen Kolli
Praveen Kolli is a Senior Machine Learning Engineer working in the Marketplace Ads team at Houzz Inc. At his current position, he built machine learning models for predicting bid prices for the 10M+ products that Houzz advertises on exchanges such as Google and Microsoft Bing. He also developed ML algorithms for reranking products on various landing pages. Prior to that, he obtained his PhD in Mathematics from Carnegie Mellon University and Masters in Mathematics of Finance from Columbia University.

Title: Machine Learning in Ads

We will present an overview of the Ads systems in social media companies. We will address some of the challenges that arise due to petabyte scale data and how machine learning is used to tackle these challenges.
Thursday, January 28, 2021

Speaker: Robert Chang, FI Consulting
Robert is a Manager of Model Development and Model Validation at FI Consulting in Washington D.C., leading projects for Commercial and Federal financial institutions, including Capital One Bank, Northern Trust Bank, Freddie Mac, the Federal Housing Administration (FHA), and the U.S. Department of Treasury. Before joining FI Consulting, Robert was a Derivatives Trader for 10 years. Robert is a 2007 graduate of the Columbia MAFN Program, holds a Master’s degree in Information Management Systems from the Harvard Extension School, and is a certified Financial Risk Manager (FRM).

Title: Using Knowledge Graphs for Anti-Money Laundering and Transaction Monitoring

Today’s anti-money laundering (AML) and transaction monitoring systems need to be quicker and more agile to identify increasingly complex fraudulent transactions. Due to rapid evolution of fraudulent behavior, often layered behind seemingly innocuous transactions, AML models require greater sophistication to remain effective. Graph-based approaches that utilize advanced computational techniques are needed to adapt to changing fraud patterns and to create effective rules for detection.
Tuesday, February 2, 2021

Speaker: Natalia Zvereva, JP Morgan Asset Management
Natalia Zvereva, CAIA, is an Executive Director, Investment Risk, at J.P. Morgan Asset Management. Natalia and her team focus on risk management of the Alternative Investment Funds, including Hedge Funds, Private Equity, and Liquid Alternatives. Natalia has been working in risk management since 2009, and held a number of roles in JP Morgan within market risk and credit risk space. Prior to her current position, she covered counterparty risk, credit and funding pricing of cross-asset derivatives portfolios. Prior to that, Natalia was a market risk manager in OTC Derivatives Clearing, where she helped to facilitate the launch of the OTC Clearing business. Natalia is an analyst and associates champion for JPM Asset Management Risk. She created a training curriculum and organized global technical training and networking events for analysts, associates, and interns in Asset and Wealth Management. Natalia holds a Master’s Degree in Financial Mathematics from Columbia University (2014), BBA in Finance & Investments from Baruch College (2009), and is a member of Chartered Alternative Investment Analyst Association.

Title: Derivatives Uses and Risk Management

In this talk, we will discuss the use of derivatives across asset classes and client types, and will go over various derivatives risk management techniques used in the industry.
Thursday, February 4, 2021

Speaker: Graham Giller, Giller Investments (New Jersey), LLC
Graham Giller is one of Wall Street’s original data scientists. Starting his career at Morgan Stanley in the UK, he was an early member of Peter Muller’s famous PDT group and went on to run his own investment firm. He was Bloomberg LP’s original data science hire and set up the data science team in the Global Data division there. He them moved to JP Morgan to take the role of Chief Data Scientist, New Product Development, and was subsequently Head of Data Science Research at JP Morgan and Head of Primary Research at Deutsche Bank.

Title: Financial Data is not Normal

The Normal distribution is omnipresent in social sciences and physical sciences due to it’s unique properties as a stable distribution with a finite variance. It is a key part of stochastic calculus and a large amount of research in finance, both empirical and theoretical, is built upon it. However, financial data is not Normal! Examining real data in many markets leads to an unambiguous rejection of the hypothesis of homoskedastic Normal distributions in favour of heteroskedastic and fundamentally leptokurtotic data and the consequences of this for active managers are significant.
Tuesday, February 9, 2021

Speaker: David Fournie, Bank of America
David Fournie is Head of US Equity Exotics Trading at Bank of America. He got his Ph.D from Columbia’s mathematics department, 2010.

Title: Contingent claim replication with funded assets

We provide a self-contained framework for a contingent claims replication theory that accounts correctly for the daily collateralization exchange and different funding rates of derivatives and underlying assets. We apply it to the example of a normal variance swap.
Thursday, February 11, 2021

Speaker: Samvel P. Gevorkyan, Freepoint Commodities
Samvel Gevorkyan is a Quantitative Researcher at Freepoint Commodities. He manages a crude oil and natural gas portfolio with strategies based on probabilistic machine learning: from data processing to alpha factor engineering to testing and deployment. He holds a Master’s Degree in Financial Mathematics from Columbia University (2018), B.S. in Mathematics/Economics from UCLA (2016).

Title: Deep Learning in Systematic Investing

Deep Learning research in the past decade has seen monumental success in areas like NLP, Computer Vision, etc, with efficiency gains allowing to completely transform entire industries. Applications to financial trading, however, have been relatively stagnant. We will attempt to touch on various important ideas in machine learning and observe how a probabilistic framework can contribute to alpha generation in systematic portfolios.
Tuesday, February 16, 2021

Speaker: Alberto Botter
Alberto Botter is a Managing Director within the Portfolio Management department at AQR Capital Management. In this role, he oversees the construction, optimization and management of AQR’s Equities Long-Short products. Previously, Mr. Botter was a Quant in the Wealth Strategies group at Morgan Stanley.

Mr. Botter earned a B.S. and an M.S. in economics from the University of Bologna, as well as an M.A. in financial mathematics from Columbia University

Title: Fundamentals of Active Investing

Thursday, February 18, 2021

Speaker: David Skovmand, University of Copenhagen
David Skovmand is an associate professor of financial mathematics at the University of Copenhagen.

Title: Term Rates, Multicurve Term Structures and Overnight Rate Benchmarks: a Roll-Over Risk Approach

In this paper we model the risk that a financial institution may not be able to roll over its debt at the market reference rate, the so–called “roll–over risk”. We construct a model framework for the dynamics of reference term rates (e.g., LIBOR) and their spread vis–à–vis benchmarks based on overnight reference rates, e.g., rates implied by overnight index swaps (OIS). In this framework, different interest rate term structures are endogenously generated for each tenor, that is, a different term structure for each choice of the length of the interest rate accrual period, be it overnight (e.g., OIS), three–month LIBOR, six–month LIBOR, etc. A concrete model instance in this framework can be calibrated simultaneously to available market instruments at a particular point in time, but more importantly, we explicitly obtain dynamics of term rates such as LIBOR. Thus models in our framework are amenable to econometric estimation. For a model class based on affine dynamics, we conduct an empirical analysis on EUR data for OIS, interest–rate swaps, basis swaps and credit default swaps. Our model achieves a better fit to time series data than other models proposed in prior literature. We find that credit risk typically contributes only about 30% of the IBOR/OIS spread, with the balance of the spread due to the funding liquidity component of roll–over risk. Looking ahead, we show that, even if credit risk is entirely mitigated by repo transactions, the presence of roll–over risk confounds attempts to obtain term rates from overnight rate benchmarks. As various jurisdictions transition away from panel–based term rate benchmarks towards transaction–based overnight ones (such as SOFR in the United States), the framework presented in this paper thus provides important insights into some of the consequences of this transition.
Tuesday, February 23, 2021

Speaker: Miquel Noguer Alonso

Title: Deep Learning for Equity Time Series Prediction

We examine the performance of Deep Learning methods applied to equity financial time series. Predicting equity time series is a crucial topic in Finance. To form equity portfolios and do asset allocation, we need to predict returns, compute their risk, and optimize market impact. One of the modeling benefits of Deep Learning architectures is the ability to model non-linear highly dimensional problems. The lack of transparency and a rigorous mathematical theory could be considered less positive sides. The fact that most progress in Deep Learning has been made by trial and error is also cumbersome. Equity financial time series is a challenging domain with some stylized facts: weak stationarity, fat tails in return distributions, small data sets compared to other areas of Artificial Intelligence (AI), slow decay of autocorrelation in returns, and volatility clustering, to name the most important ones
Thursday, February 25, 2021

Speaker: Michael Pykhtin, Federal Reserve Board
Michael Pykhtin is a manager in the Policy Research and Analytics function at the U.S. Federal Reserve Board. Prior to joining the Board in 2009 as a senior economist, he had a successful nine-year career as a quantitative researcher at Bank of America and KeyCorp. Michael has edited “Margin in Derivatives Trading” (Risk Books 2018), “Counterparty Risk Management” (Risk Books, 2014) and “Counterparty Credit Risk Modelling” (Risk Books, 2005); he is also a contributing author to several recent edited collections. Michael has published extensively in the leading industry journals; he has been an Associate Editor of the Journal of Credit Risk since 2007. Michael is a two-time recipient of Risk Magazine’s Quant of the Year award (for 2014 and 2018). Michael holds a Ph.D. degree in Physics from the University of Pennsylvania and an M.S. degree in Physics and Applied Mathematics from Moscow Institute of Physics and Technology.s

Title: Local Gaussian approximation for modeling collateralized exposure

In this presentation, we are going to review the local Gaussian approximation and its applications to modeling collateralized credit exposure. These applications include: time grid interpolation that allows one to obtain exposure on a daily grid without daily revaluation of the portfolio; calculation of the future initial margin requirements along exposure simulation paths; eliminating excessive simulation noise via replacing exposure realizations along the simulation paths with conditional expectations of these realizations. In addition to these applications, which rely on estimation of the local volatility of the portfolio market value along each simulation path, we will review a powerful scaling technique that allows one to calculate the impact of initial margin on the expected exposure profile without any knowledge of the future initial margin requirements, so that estimates of the future local volatility are not needed.
Print this page