London Quant Group - Annual Spring Seminar
Monday 11th May - Tuesday 12th May 2009
Monday 11th May 2009
09:00 ROM Simulation (Random Orthogonal Matrices)
Professor Carol Alexander, University of Reading
ROM simulation is a new approach to multivariate simulation based on random, orthogonal matrices. Unlike Monte Carlo methods there is no simulation error in the multivariate sample moments of the simulations. The approach is based on a new theoretical result from linear algebra that provides necessary and sufficient conditions for a matrix to be a covariance matrix. We realise a sample from a given covariance matrix using an L-Matrix, which is a new class of orthogonal matrix, and explain how to change the features of the sample (e.g. its higher moments, and volatilty clustering) using different types of random orthogonal matrices. Applications to finance are described.
10:00 The small cap effect in ex post vs. ex ante returns
Mathijs A. van Dijk, the Rotterdam School of Management, Erasmus University
We show that the disappearance of the size effect in U.S. stock returns after the early 1980s can be attributed to unexpected shocks to the profitability of small and big firms. After accounting for the price impact of these profitability shocks, we find a robust size effect of close to 10% per annum over the period 1963-2005. Our results have a number of implications for investment practitioners. First, our analysis indicates that tilting equity portfolios toward small caps produces systematically higher expected returns. Second, the cross-sectional model that we use to construct profitability shock estimates can serve as a powerful tool to forecast the profitability of individual firms. Third, our methods of adjusting realized returns for the impact of profitability shocks can be readily applied to other investment strategies. Studying the role of profitability shocks will not only strengthen our understanding of these strategies, but will also help enhance and fine-tune their implementation and performance.
11:00 Coffee
11:30 Madoff : Market Mayhem
Dan di Bartolomeo, Northfield Information Systems Inc
At the request of a client in early 1999, we analyzed the track record of a hedge fund then known to Northfield only as “Manager B. This manager was subsequently revealed to be the Bernard Madoff organization. Within a few hours it was concluded that the Madoff returns were either fictitious, or had arisen from a strategy other than was being represented to investors wherein returns were probably being enhanced by illegal means. Our client, Harry Markopolos, repeatedly reported his urgent concerns about Madoff to the US Securities and Exchange Commission starting in 2000, but his warnings went unheeded. In this presentation, we will review the key analytical methods used in this matter (derived from di Bartolomeo and Witkowski, FAJ, 1997), and the specific related findings. We will illustrate how our analysis was able to identify obvious risks that others charged with due diligence on the Madoff fund apparently did not perceive.
12:30 Lunch
14:00 My Experiences in Quant Investing
Olivier Ledoit - University of Zurich
Over the past 10 years, quantitative techniques have taken an important place in proprietary trading at major investment banks. This presentation will review the recent evolution of quant trading and highlight some of its important features.
What is perhaps most significant about quant trading is that it has been able to generate considerable profits. Before the current crisis, quant traders believed that such profits were fairly earned in the competitive marketplace. Big bonuses represented a just reward for their hard work, market savvy and superior mathematical skills.
By contrast, since the crisis, most people – even in financial circles – feel that these gains were somehow ill-gotten. For example Jim Rogers, who co-founded the Quantum Fund with George Soros, blamed the crisis on “29-year olds driving Maseratis.” This is not just idle talk: such arguments are used to justify the maximum wage of $500,000 per year announced by the US government for traders working in banks that have accepted bail-out money.
We will provide an economically rigorous investigation of this question in order to find out who is right and who is wrong.
15:00 Deconstructing the Fear Gauge: A Better Understanding of the VIX with Kurtosis Adjusted Volatility Estimators
Taher Khan - Millenium Global Investments
This method is a poor man's estimator that provides a simple implied volatility for a kurtosis adjusted Black-Scholes model, such as found in Corrado and Su (1996) and Brown and Robinson (2002). It also provides an easy way to forecast volatilities with extreme values, other examples being Chou's CARR (2005) model and Range-based volatility estimators a la Rogers and Satchell (1991).
16:30 Robust Performance Hypothesis Testing with the Sharpe Ratio
Michael Wolf - University of Zurich
Applied researchers often test for the difference of the Sharpe ratios of two investment strategies. A very popular tool to this end is the test of Jobson and Korkie from the 1980s, which has been recently corrected by Memmel. Unfortunately, this test is not valid when returns have tails heavier than the normal distribution or are of time series nature. In real-life financial applications, typically at least one of these `violations' occur.
As a result the Jobson-Korkie-Memmel test tends to overstate the significance of the observed difference between the two sample Sharpe ratios.
Instead, we propose the use of robust inference methods, in particular a studentized bootstrap approach, to obtain valid inference under realistic conditions.
A simulation study demonstrates the improved finite sample performance compared to existing methods.
In addition, two applications to real data are provided.
19:30 Optional Dinner
This will be at the Dispensary Pub and Dining Room, 19a Leman Street, London, E1 8EN. East London Pub Of The Year, The Dispensary has been chosen as the 2009 Pub of the year by the East London and City (ELAC) Branch of the Campaign For Real Ale (Camra)
Tuesday 12th May 2008
9:00 Optimisation Issues and Portfolio Construction
Richard Louth, University of Cambridge
This paper seeks to advance the methodological basis for the use of non-Gaussian alternatives to traditional mean-variance analysis for large dimension portfolio optimisation problems. Through the application of a threshold acceptance algorithm and an appropriately chosen distribution from the Johnson family, we show how portfolio weights for an otherwise computationally intensive asset allocation problem can be obtained in a quick and efficient manner. The inherent flexibility and generality of this approach is further illustrated by two practical extensions. Firstly, we introduce the idea of data re-weighting as a simple and, most importantly, computationally tractable method for improving the robustness of our optimisation algorithm. And secondly, we demonstrate how return forecasts ('alphas') can be seamlessly incorporated into the estimation procedure via Bayesian updating. In this case, we utilise an important property of the Johnson family to build a model of the joint dependence between returns and their forecasts using the class of meta-elliptical distributions.
10:00 Coffee
10:30 Is Quant Dead? – A Debate
Edward Fishwick, BlackRock
Slides are not available
Jason MacQueen, R-Squared
Quantitative investment strategies have been having a tough time. The “quant crisis” in 2007 highlighted the potential for crowding in “quant” stocks, the related issue of high model correlation or equivalence, and generally focused attention on the potential for poor returns from conventional quant strategies. In addition, the extreme volatility of the past 18 months and the rapid transition from one environment to another inevitably calls into question any approach driven off backward-looking, codified relationships. Here then, we will debate the implications of these issues, and attempt to understand the ways in which quant investing might evolve.
12:30 End of Seminar