Liquidity in the FX Market: Empirical Evidence from an Aggregator

Abstract

In foreign exchange (FX) trading, an aggregator is used to connect traders with liquidity providers (LPs). In an aggregator, a trader receives a continuous stream of bid and ask quotes from a predefined set of LPs, and the difference between the best bid and ask prices over a set of liquidity streams is called an inside spread. In this paper, we empirically study liquidity in an FX aggregator and show that, on average, traders obtain a relatively tight spread with four or five streams; the use of more streams yields a marginal benefit only. For given numbers of liquidity streams, we determine the optimal combinations of streams minimizing the spread. Our findings indicate that most of the traders could — at least in theory reduce the average spread by more than half with the optimal combination of streams, and a trader could save significantly, even up to 0.18 basis points in dollars per €1 traded. However, traders may not be able to fully exploit improvements in spreads because, in practice, the liquidity streams are set by LPs and not by the trader. In addition, we find that Oomen's [Quantitative Finance, 17, 3, (2017)] model for the liquidity dynamics and contract formation process in the aggregator, which can be used to characterize the observed inside spread, fits our empirical data accurately, even under simplistic assumptions.