QP1 prepares technical white papers which are made available to clients of Deutsche Bank's equity businesses. In addition, members of the team regularly make presentations at academic seminars and industry conferences. Selected material may be downloaded here.
An analysis of the effects of planned tick size changes at the Tokyo Stock Exchange.
Presentation at "Mathematics at the interface between business and research", held by the Stifterverband für die Deutsche Wissenschaft, Berlin, December 2008.
White paper describing Deutsche Bank's single stock "Implementation Shortfall" algorithm.
Deutsche Bank’s Implementation Shortfall (DBIS) algorithm is a proprietary, schedule-based strategy for trading single-stock orders that aims to minimize the difference between the average price of execution and the arrival price of the order. The optimal trade schedule balances the benefits incurred by trading rapidly – lowered exposure to risk – against those gained by trading slowly – lowered costs due to market impact. Deutsche Bank’s proprietary algorithm constructs schedules based on superior models for transaction costs and optimal termination time, and is fully integrated with DB’s Smart Order Router thus adapting the execution of trades to the wide range of venues currently available.
This white paper presents the Optimal Portfolio Execution (OPX) algorithm, the latest addition to Deutsche Bank’s algorithmic trading lineup.
OPX is a 3rd generation algorithm designed to optimally execute a portfolio of stocks. The goal of the strategy is to trade a portfolio in such a way as to minimize both expected shortfall of the execution and the risk of loss due to adverse price movement. It does so by creating a joint unwinding schedule that optimally trades off the market impact that would be incurred by aggressive trading against the risk incurred by being exposed to market volatility. While the conceptual model for the strategy is based on the work developed by Robert Almgren in his seminal paper (Almgren and Chriss, 2000) several improvements have been added to the original framework. Additionally, as we will discuss in section 7, a novel proprietary approach was implemented to deal with the computational complexity that makes brute force optimization unfeasible for large portfolios. This means that OPX can perform the full optimization without having to utilize one of the strong assumptions that many competitors are forced to adopt to deal with this problem.
This white paper presents an optimal order placement strategy, which is intended to be used as a micro-trader for a higherlevel strategy, such as VWAP, Percent of Volume, Single Stock Implementation Shortfall and Optimal Portfolio Execution algorithms.
This note discusses some of the key ingredients of a robust and reliable approach to large dimensional covariance matrix forecasting over short intra-day horizons.
In this setting, the main challenge is to correctly account for market microstructure contaminations that emerge in data sampled at ultra high frequency. We contrast three competing methods, namely (i) realized covariance based on intra-day data, (ii) a statistical factor model based on intra-day data, and (iii) a fundamental factor model based on daily data. Our performance evaluation approach centers around portfolio optimality and portfolio stability. The large amount of available data, allows us to accurately discriminate amongst the performance of the alternative models and we illustrate their relative merits and weaknesses using the FTSE-100 and STOXX-50 index constituents.