فصل روزانه در معاملات بازار ارز

This paper examines the intra-day seasonality of transacted limit and market orders in the DEM/USD foreign exchange market. Empirical analysis of completed transactions data based on the Dealing 2000-2 electronic inter-dealer broking system indicates significant evidence of intra-day seasonality in returns and return volatilities under usual market conditions. Moreover, analysis of realised tail outcomes supports seasonality for extraordinary market conditions across the trading day.

In recent times much has been made of the trading revolution in currency markets brought about through the screen-based electronic trading and broking systems that have come to dominate foreign exchange trading activities. The importance of these systems is illustrated by the fact that, according to the Bank for International Settlements (BIS) survey on foreign exchange, electronic trading makes up 48% of activity in the largest market, the UK (Williams, 2005). This study examines the intra-day seasonalities of returns and volatilities in FX transactions. More specifically, it examines the intra-day return and volatility seasonalities for both limit and market orders involving the DEM/USD exchange rate from the D2000-2 electronic FX broking system. The analysis is based on actual transaction data rather than the more common (but less reliable) use of indicative quotes in which there is no firm commitment to transact on the stated terms. The dataset also represents the transaction outcomes of a multiple-dealer market. Limit orders represent an order to buy or sell at some prespecified price, whereas market orders are orders for immediate execution at whatever price can be obtained.1 Previous microstructure studies have demonstrated the importance of order type return and volatility characteristics. For example, Hasbrouck and Saar (2004) find evidence of a market order certainty effect where increased limit order volatility is associated with a reduction in the proportion of limit orders in incoming order flow.2 Relatedly, Harris and Hasbrouck (1996) compare the execution performance of market and limit order and show that limit order trading strategies generally perform best across various spreads, order sizes and position (buy or sell). However, neither study examined the role of intra-day seasonality of limit and market order realisations.3 There is also reason to believe that traders will alter their trading strategies and associated order mechanisms during periods of extreme market movements (Goldstein & Kavajecz, 2004). We therefore believe that it is also important to distinguish between normal and extraordinary market conditions (Longin, 2000). To do so, we also examine the intra-day seasonality of the tail behaviour of market and limit orders to both determine if seasonalities and trading patterns differ between normal and extreme market conditions.4 We model the tail behaviour using Extreme Value Theory (EVT). Many previous studies have documented the presence of heavy tails for the distribution of exchange rate price changes over different frequencies (e.g. Dacorogna, Gencay, Muller, Pictet, & Olsen, 2001). Recently, in comparing limit order versus market order realised returns, Cotter and Dowd (2007) find fat-tails are associated with limit and market order transactions and that on average the extreme returns of limit orders are found to have fatter tails than market orders. However, they do not examine whether there is any seasonality in these findings across the trading day. The rest of the paper proceeds as follows. Section 2 outlines the foreign exchange market and limit and market order features of the D2000-2 dataset on which the study is based. Section 3 presents EVT and explains its properties and the measurement techniques used to examine market and limit order tail behaviour in the market considered. Section 4 presents the empirical findings. Some conclusions are given in section 5.

This paper has examined risk and return realisations for FX market and limit orders across a trading day. We analyse a unique data set of transacted exchange rates overcoming the common use of indicative quotes that may or may not be binding. We find that both limit and market orders exhibit seasonality through the variation of their returns and volatility across the trading day, and this is so (although in different ways) for both ordinary and extraordinary market conditions. The results have implications for the timing of trading strategies using limit and market orders. Our work presents preliminary findings about the behaviour of intra-day seasonalities in FX markets, but much more work will be needed to determine how typical our findings might be of seasonalities in other periods or other markets, and to determine whether, and if so how, seasonalities might be changing over time.