Internship Subject: Noise Microstructure Analysis in Financial Data

A new field in quantitative finance is emerging : high-frequency trading. This area of derivatives trading is very different from traditional trading (say volatility trading or convert arbitrage) because of the massive amount of data that needs to be processed and because of a relative lack of academic work that has been done on the subject. One big issue that modern finance has not managed to solve yet, is the impact of microstructure noise. Indeed, a stock price tends to bounce between the bid and the ask price and this move, invisible at the daily level, becomes a problem at high frequency. This noise at the tick level pollutes traditional computations of volatility or correlation. In addition, the fact that data under the 2-minute interval may be sparse and unevenly spaced in time makes traditional statistical analysis obsolete: at a high-frequency level, we are dealing with what we call \textit {stochastic time}. We propose an internship research subject that aims at understanding how to handle microstructure noise. Who should apply ? Students from ENS Ulm, Polytechnique, Centrale Paris, Ecole des Ponts, Mines de Paris, Telecom Paris. Students from DEA of P6 and P7 may only apply if they graduated from one of the previous schools. Internship dates Dates are flexible starting from January 2010 or later, for 3 months to 6 months. Email: stages@arbitragis.com

microstructure_noise.pdf — PDF document, 68Kb