Network Structure and Pricing in the FX Market
59 Pages Posted: 12 Feb 2020
Date Written: January 17, 2020
Foreign exchange (FX) settlement data define a network, for which we may construct centrality measures and profit attributions. Our sample of settlement data from CLS Bank spans diverse currency pairs, participants and execution platforms over the Aprils of 2013 and 2016. We assign settlement members to (five) groups ranked by unweighted degree centrality. We define an average centrality differential as the return to the more-central counterparty in the trade, and model this as a function of the two counterparties’ centrality groups. Estimates of the average centrality differential are generally positive: the more-central counterparty receives a higher return. Additionally, the differential generally increases as the counterparties’ centralities diverge. These two results are consistent with a pervasive centrality premium. The estimates are robust to the choice of pre- or post-settlement benchmarks, to inclusion of settlement size interactions, and to grouping on volume-weighted degree centrality. Across currency pairs the centrality profit varies considerably, and typically amounts to about one-third of bid-ask half-spread. The centrality premium is consistent with the hypothesis that central agents exercise bargaining power. We also find, however, evidence suggesting that the premium is partially offset by losses that central agents incur in supplying liquidity.
Keywords: foreign exchange, CLS Bank, centrality, settlement
JEL Classification: F31, G12, G15, G23
Suggested Citation: Suggested Citation