GAMING IN FUTURES MARKET SUGGESTS MODEL COULD BE IMPROVED?

Evidence from UK interest rate futures market reveals that traders have been “drowning” the market with oversized orders, increasing their allocation under a pure pro-rata matching algorithm. The 2007 introduction of a time element to the order matching mechanism has modified the behaviour of traders. No longer is the order book drowned with orders which try and force maximum allocation, instead traders now strategically gain priority through the submission of single share orders, around their overall trade size. The matching algorithm used for short-term interest rate futures contracts (STIRs) at LIFFE has undergone many changes throughout the years. Most recently, in 2007 the mechanism changed to time pro-rata. In this system the size of an order and its relative position in the order book are considered in determining what proportion of a market order will be allocated to each resting limit order. This mechanism is used in many futures markets worldwide, including the LIFFE Euribor, Short Sterling and Euroswiss contracts, and the CME Eurodollar contracts. New evidence from the CMCRC identifies that in the wake of this change to a time pro-rata algorithm, market depth at the best bid and offer reduced from over £10 million to less than £500,000. Similarly, the average notional size of submitted and cancelled orders dropped from over £4 billion to less than £1 billion. While these reductions may sound like they are bad for the market, the reason that depth was so high was that participants were submitting limit orders many times larger than their desired allocation. This “over-sizing” of orders was a strategy used to increase the proportional fill orders. The reduction in the incentive to oversize orders appears to have improved the quality of liquidity, allowing traders to enter orders which reflect their true desire to trade. The removal of over-sized orders also has another benefit – traders no longer need to cancel the unexecuted portion of their oversized orders. This reduces the cost of managing these unexecuted orders, as can be seen in the significant reduction in the order to trade ratio, from 25 to 1 prior to the change down to 12 to 1 after. No change was evident in transactions costs, measured using effective and quoted spreads.