Front-Running and Collusion in Forex Trading

This paper examines the market-wide effects of front-running and information-sharing by dealers in a quantitative microstructure model of Forex trading. Recent investigations by government regulators and court proceedings reveal that there has been widespread sharing of information among Forex dealers working at major banks, as well as the regular front-running of large customer orders. I use the model to study the effects of unilateral front-running, where individual dealers trade ahead of their own customer orders; and collusive front-running where individual dealers trade ahead of another dealer's customer order based on information that was shared among a group of dealers. I find that both forms of front-running create an information externality that significantly affects order flows and Forex prices by slowing down the process through which inter-dealer trading aggregates information from across the market. Font-running reduces dealers' liquidity provision costs by raising the price customers pay to purchase Forex, and lowering the price they receive when selling Forex. These cost reductions are substantial; they lower costs by more than 90 percent. Front-running also affects other market participants that are not directly involved in front-running trades. The information externality makes these participants less willing to speculate on their private information when trading with dealers. This indirect effect of front-running can reduce participants' expected returns by as much as 10 percent. My analysis also shows that collusive front-running has larger effects on order flows than unilateral front-running because information-sharing reduces the risks dealers face when trading ahead of customer orders. However, in other respects, the effects of collusive and unilateral front-running are quite similar. Greater collusion lowers the costs of providing liquidity and it reduces other participants' expected returns, but the effects are small.