The inefficiency of Reuters foreign exchange quotes

https://doi.org/10.1016/S0378-4266(98)00004-1Get rights and content

Abstract

Reuters foreign exchange (FXFX) page is the world wide predominant information source to foreign exchange traders. In this study we compare the indicative spot exchange rate quotes from Reuters with their matching futures exchange rates from the Chicago Mercantile Exchange. We find that the indicative quotes on Reuters FXFX page are inefficient and could be improved by incorporating information from the futures market. This casts doubt on the way banks determine these quotes, as well as on the informational content of these quotes as an indicator of the current exchange rate.

Introduction

The spot foreign exchange market is a 24 hours electronic market with brokers and traders around the world. Brokers display quotes to their customers. These quotes are the best bid and ask price provided by a limited number of banks regularly contacted by the broker. Since there are many brokers each contacting their own circle of banks to obtain quotes and having their own customers, the natural question arises whether this market is informationally efficient.

To date, no data sets have been available allowing for a direct test of the efficiency of the spot market in foreign exchange. Goodhart et al., 1996, Goodhart et al., 1997study 7 hours of the Reuters-2000 electronic trading system, which at the time of sampling was still a relatively small broker in the spot market. Lyons (1995)studies one week of all transactions of a New York broker.

Obviously, any data set on quotes from brokers will only reflect a part of the spot market. In fact, the only information source available to all traders around the world consists of indicative quotes, as provided by Reuters foreign exchange (FXFX) page, and those provided by its competitors Knight Ridder and Telerate. As such these quotes play an important role in the spot market, indicating the current foreign exchange rate. Though the quotes are only `indicative', studies using the quotes claim banks have reputation considerations and will most likely trade against their quotes if called within a short time after appearance on the Reuters FXFX page. This assumption is crucial for studies like De Jong et al. (1995)who study triangular arbitrage, Bollerslev and Domowitz (1993), and Dacorogna et al. (1993)who study the trading intensity and volatility patterns in the spot market, and Bollerslev and Melvin (1994)who study the relationship between the spreads and volatility. Similarly, Olsen and Associates who use these quotes to forecast the foreign exchange rate, started a boom in empirical research by releasing 1 year of Reuters quotes in 1994.

In this study we further investigate the assumption that one can actually trade against the Reuters quotes. We compare these spot exchange rate quotes with their matching futures exchange rates traded at the Chicago Mercantile Exchange (CME). The futures market is a highly liquid market, but in value terms relatively small as compared to the spot market. Nevertheless, we find that the futures market is leading the `quoted' spot market for up to 3 minutes. The results of a simple trading strategy show that profits can be made from the futures lead, unless trading against the Reuters quotes is not (always) possible. This could (partly) explain our results, which we therefore interpret conditional on the possibility of trading against the Reuters quotes:

(i) If one can trade against the Reuters quotes, then our results show that gains can be made and hence the spot market is inefficient.

(ii) If one cannot (always) trade against the Reuters quotes, our results can be (partly) explained by that fact. However, this implies that studies which made the assumption that one can trade against Reuters quotes were incorrect to do so. Their results will then have to be taken with care.

In both cases banks displaying quotes on Reuters FXFX page can improve upon these quotes by paying more attention to the futures market. The futures price provides a more adequate reflection of the (true) current spot exchange rate than the Reuters quotes.

The remaining part of this study is organised as follows. Section 2discusses in further detail the functioning of the spot and futures markets. In Section 3the data set is discussed. Section 4describes the methodology, and Section 5elaborates upon the results. Section 6focuses on the effect of prescheduled news announcements and high volatility periods. In Section 7the profitability of a simple trading strategy is tested. Finally, Section 8will conclude.

Section snippets

Microstructure of foreign exchange markets

The major difference between the futures and the spot market in foreign exchange is the trading system. While the futures contracts at the CME are traded in an open outcry (OOC) market, the spot market is an electronic market with brokers and traders around the world. In addition to brokers setting quotes, market participants can also offer or obtain quotes via Reuters, Telerate or Knight Ridder.

We will use Reuters FXFX data for the spot rates. These data consist of, mainly indicative, bid and

Data

The data set consists of intraday Reuters quotations of the DM/$ exchange rate from the Olsen and Associates data set, and of futures prices of the DM/$ exchange rate for the September 1993 contract from the CME. The sample period covers June, July and August 1993. We choose to analyse the DM/$ exchange rate since it is the most liquid exchange rate in terms of number of contracts (futures market) and number of bid–ask quotes (spot market).

The futures data contain the transaction price, the

Methodology

Having discussed the univariate data series, we can now proceed with their joint analysis based on the covered interest rate parity (CIRP). For the necessary US and German interest rates we obtained daily Eurocurrency rates from Datastream. We refrain from using intraday interest rates. First, they are not readily available. Second, foreign exchange transactions are not settled within the trading day but at the end of the trading day when the final position in foreign currency is deposited at a

Results

In this section we will first test for cointegration between the market observed futures prices and theoretical (spot-induced) futures prices based on the CIRP. Next, we will estimate the VECM specified in Eq. (4). From this we will calculate impulse–response functions and the information share of both the futures and spot market. In Section 5.2we investigate simple cross-correlations between the spot and futures returns to determine the lead–lag structure between the two markets.

The effect of high volatility periods

In this section we split the sample into two parts: one where volatility is relatively high and one where volatility is at its normal level. In cases of high volatility or prescheduled news announcements, traders prefer first to trade and only then update their Reuters quotes. However, traders will then also need to closely follow the current developments in the market.

A simple trading strategy

Having established a statistically significant lead of the futures market over the spot quotes, it is interesting to investigate its economic significance. Of course the underlying test would be more reliable if we would have quotes from brokers at which we could certainly trade. On the other hand, Reuters FXFX page quotes usually include the broker's quotes, and thus the underlying strategy overvalues the trading costs. The spread of the broker can reduce to 1 tick, while the minimal spread on

Conclusion

This study compares the DM/$ futures prices in Chicago with Reuters foreign exchange page displaying spot exchange rate quotes of many different banks. Even though Reuters' FXFX page is mainly indicative, reputation considerations might induce banks to trade at their quotes when asked to within reasonable time after the appearance on the screen. Also, broker's quotes are usually within the Reuter's quotes. Interestingly, we find that the futures market is leading the quotes on Reuters FXFX page

Acknowledgements

The authors would like to thank Yuan-chen Chang, Michel Dacorogna, Theo Nijman, Antoon Pelsser, Piet Sercu, Ton Vorst, Siegfried Trautmann, Casper de Vries, two anonymous referees, participants of the 13th International Conference of the French Finance Association in Geneva (1996), and participants of the 23rd European Finance Association meeting in Oslo (1996) for useful comments. We are also grateful to the ABN-AMRO bank for allowing us to visit their dealing-room in Amsterdam and speak with

References (18)

There are more references available in the full text version of this article.

Cited by (19)

  • Which market dominates the price discovery in currency futures? The case of the Chicago Mercantile Exchange and the Intercontinental Exchange

    2022, Global Finance Journal
    Citation Excerpt :

    In summary, the information revealed in one market is positively related to trading volume and the number of shares traded, but is negatively related to transaction costs and price volatility. Several researchers argue that the currency futures market is more informative than the currency spot market because informed traders prefer to trade in the futures market (Chatrath & Song, 1998; Chou & Chung, 2006; Chu, Hsieh, & Tse, 1999; Crain & Lee, 1995; Garbade & Silber, 1983; Martens & Kofman, 1998). However, other studies show that the spot market is more informative than the futures market because the spot market is more liquid and transparent (Cabrera, Wang, & Yang, 2009; Chen & Gau, 2010; Ivanov, Jones, & Zaima, 2013; Poskitt, 2009; Rosenberg & Traub, 2009).

  • Information shares in the US Treasury market

    2008, Journal of Banking and Finance
    Citation Excerpt :

    Hutcheson (2003) finds that the highly liquid spot market leads the futures prices. Martens and Kofman (1998) find that indicative Reuter’s FXFX spot market quotes do not subsume the futures market quotes and Rosenberg and Traub (2005) report that futures order flow seems to dominate in price discovery. Ex ante, it is not obvious whether spot or derivatives markets should dominate in bond price discovery.

  • The truth about interest rate futures and forwards: Evidence from high frequency data

    2008, Global Finance Journal
    Citation Excerpt :

    There is some debate over the appropriateness of using indicative quotes as proxies for transaction prices. While reputation considerations constrain OTC dealers from abusing the indicative quote convention by displaying misleading or “off-market” quotes (Goodhart & Figliuoli, 1991), indicative quotes might tend to lag behind transaction prices in a volatile market as dealers will be more focused on trading than on maintaining up to date quotes (Martens & Kofman, 1998). Moreover the fundamentally different architecture of a decentralised interbank market has implications for the properties of a time series of indicative quotes.

View all citing articles on Scopus
View full text