Latency in Electronic Securities Trading: A Proposal for Systematic Measurement

Miroslav Budimir 1. is a strategy developer and market microstructure expert at Deutsche Börse Group in Frankfurt, Germany 2. Uwe Schweickert 1. is a senior vice president at Deutsche Börse Group in Frankfurt and a postgraduate student at E-Finance Lab Frankfurt am Main e.V. House of Finance, Goethe University, Frankfurt, Germany. (uwe.schweickert{at}deutscheboerse.com) <!-- --> 1. To order reprints of this article, please contact Dewey Palmieri at dpalmieri{at}iijournals.com or 212-224-3675. Latency is one of the major issues in today’s electronic securities trading. The demand for low latency services has increased tremendously since the advent of algorithmic trading. Milliseconds are a competitive edge, both for the demand and the provision of electronic execution services. Speed is important to sophisticated market participants because it impacts the profitability of their innovative trading strategies. It is equally important to marketplace operators because algorithmic trading is a major source of order flow. The sound understanding of latency is important to all concerned parties. However, the term latency still lacks a common definition and a transparent methodology for measurement in securities trading.After reviewing the literature, we define latency and develop a benchmarking methodology to comprehensively assess latency. We analyse the observed latency of 18.4 million order actions in the electronic trading system Xetra. Our results show that trading activity, time of day and distance are the main drivers of latency. The descriptive section gives an in depth view on trading at light speed. TOPICS: [Statistical methods][1], [equity portfolio management][2], [security analysis and valuation][3] [1]: https://www.pm-research.com/topic/statistical-methods [2]: https://www.pm-research.com/topic/equity-portfolio-management [3]: https://www.pm-research.com/topic/security-analysis-and-valuation-0