Using intraday data, the authors demonstrate that contrary to widely held perceptions about the liquidity of the foreign exchange market, the market could actually be extremely illiquid. Furthermore, they find a strong commonality in liquidity trends across various currencies, with equity and bond markets limiting the diversification potential. As for carry trade returns, it is possible that this liquidity risk is to some extent priced into their returns.
The authors provide the first systematic study of liquidity in the foreign exchange (FX) market using a new comprehensive intraday dataset from Electronic Broking Services (EBS). They begin with a brief review of related literature and an introduction to the dataset and measures of liquidity used. They then present analysis and findings on liquidity in the FX market, commonality in liquidity across FX rates, relationship of FX liquidity to funding liquidity, and liquidity of the US equity and bond markets as well as an analysis of the impact of liquidity risk on carry trade returns.
How Is This Research Useful to Practitioners?
Given its size, the FX market is generally believed to be highly liquid. But because of its opaque and fragmented nature, the liquidity in this market is not well understood. From a portfolio manager’s perspective, liquidity in the FX market is critical because it is closely tied to arbitrage strategies and funding costs.
The authors find that there are significant temporal and cross-sectional variations in currency liquidities. Over time, FX rate liquidities show large co-movements, which suggests that they are driven largely by shocks that affect the FX market as a whole rather than individual rates. More liquid FX rates tend to have lower liquidity sensitivities than less liquid ones. The liquidity-based rankings of various currencies remain stable across periods. There are also strong contemporaneous co-movements among FX, US equity, and US bond market liquidities, which suggests potential limitations of international and cross-asset diversification benefits during shocks.
The authors extend earlier work on an explanation of high carry trade returns through liquidity frictions/spirals. They provide supporting evidence that when traders’ funding liquidity deteriorates, they are forced to liquidate positions, which further reduces marketwide liquidity and eventually triggers large price drops.
Low interest rate/investment currencies (e.g., JPY, CHF) tend to be more liquid, exhibit lower liquidity sensitivities, and have negative liquidity betas (i.e., an increase in prices is in the opposite direction of liquidity). High interest rate currencies (e.g., AUD, NZD) have the opposite attributes.
The authors find EUR/USD to be the most liquid exchange rate and USD/CAD and AUD/USD to be the least liquid. The high liquidity they find for EUR/CHF and USD/CHF is potentially related to investors’ “flight to quality” during the financial crisis. Relatively poor liquidity measures for GBP/USD from the EBS dataset are likely the result of Reuters being its main venue of trade.
From a central bank perspective, an implication of these findings is that providing liquidity for a specific FX rate may have positive spillover effects by alleviating liquidity strains on other investment currencies. But injection of abundant liquidity into one currency may spread liquidity to other currencies and could lead to increased speculative trading.
How Did the Authors Conduct This Research?
One of the reasons for the lack of a systematic understanding of FX liquidity has been the deficiency of comprehensive data as a result of the inherently fragmented and opaque nature of the FX market. The authors use comprehensive and high-frequency data from EBS, the leading platform for spot FX interdealer trading with more than 60% market share. EBS provides best bid and ask quotes, volume indicators, and the direction of trades for spot rates continuously on a one-second basis. All EBS quotes are transactable and virtually free of counterparty risk.
Using second-by-second data from January 2007 to December 2009, the authors calculate liquidity measures for nine currency pairs: AUD/USD, EUR/CHF, EUR/GBP, EUR/JPY, EUR/USD, GBP/USD, USD/CAD, USD/CHF, and USD/JPY. The liquidity measures they consider are price impact, return reversal, bid–ask spread, effective cost, price dispersion, and principal component. The authors use statistical techniques to assess for liquidity and commonalities in various FX rates and equity and bond markets. To quantify the economic relevance of liquidity in the FX market, the authors analyze the impact of illiquidity costs on carry trades. To compute liquidity sensitivities or liquidity betas, they introduce a new tradable liquidity risk factor and show that it has a strong impact on carry trade returns.
Analyzing a time period when various central banks implemented significant policy measures to revive economies through monetary policy, this article provides insight to the mechanics of the potential externalities of these actions on global currencies and other financial markets. As has been observed, financial markets tend to move together during a financial crisis. The period of this study (2007–2009) has a large overlap with the run-up to and eventual global financial crisis. It would be insightful and add to the robustness of the findings if the period of study were extended to analyze whether similar trends are observable in other time periods.