Survey_FL_2008

Abstract

We begin our analysis by identifying 100 underlying instruments with liquid futures contracts, such as physical commodities, financial assets or indices. In initial investigations, we studied a period from 23 November to 11 December 2006. However, to reduce risks of over-fitting, following Biais, Hillion, and Spatt (1995) we select a fresh period to report our results. We find little change. This period is from 16th to 20th April 2007. For each of the 100 instruments we select the most actively traded delivery date over our observed period. For this purpose we proxy market activity by the number of changes to depths or prices at the best quotes on the electronic limit order book per minute, calling this measure ‘updates per minute’. We confine our attention to trading on limit order books, rather than on trading floors or pits.