EUR/USD spread patterns and history, 2003-2009

User Rating: / 4
PoorBest 
Written by Forex Automaton   
Tuesday, 08 June 2010 11:52

This article begins a series dedicated to forex spread (to be defined below as half the difference between bid and ask prices). In the series, a separate article will be dedicated to every forex pair we track. The issues of interest are evolution of spread in time, stability of spread during the day and week, and existence of predictable patterns in spread. Data from a popular non-ECN broker are used.

The algebra of profit and loss in trading on a margin (follow the link for details) has been discussed before. The same notation is used here. If total trading capital is C, its increment per time step of the system (day for Danica system) is dC, then

dC/C = dp/p × k/a (1)

where p is current price quote, dp is its increment during the time step, a is stop-loss placement as a fraction of price quote, and k is fraction of capital to be lost in the event a stop-loss is triggered. Note that cost of capital is also ignored, and leverage does not appear in Eq.1 unless one becomes concerned with the cost of capital. This formula assumes that all trades are long but I will keep it that way since the difference between long and short trades in forex is immaterial for this discussion.

Eq.1 also ignores spread (assumes it is zero). If spread s is non-zero, dC/C becomes

dC/C = (dp-s1-s2)/p × k/a (2)

Here s1 and s2 are spreads when opening and closing the trade, respectively. Obviously s is always greater than zero and therefore it always works against the trader. Note that s enters only as a ratio to price (relative spread). Therefore spread in pips has to be divided by the typical price level in order to compare profitabilities of trading strategies applied to different forex pairs under real-world conditions.

In this study, I use data from a popular non-ECN forex broker, aggregated with an hour step, for years 2003 through 2009. The spread is aggregated in the following manner.

  • The spread assigned to a tick data point is half the difference between ask and bid prices of that tick. This is consistent with price itself being though of as the middle of the ask and bid interval.

  • The spread assigned to an hour is the largest spread observed in tick data during the hour.

Tokyo 9 1011 1213 14 15 16 17 18 19 20 21 2223 0 1 2 3 4 5 6 7 8
Central Europe 123456789101112 13141516171819202122230
Greenwich 01234567891011 121314151617181920212223
Eastern US 19202122230123456 789101112131415161718

Table 1. Time zone conversion table. Seasonal time shifts, such as daylight saving time, may complicate the picture if the nations choose to enact them on different days, and are ignored.

Fig.1 shows evolution of EUR/USD spread during a day, for each of the seven years from 2003 through 2009. In the figure, the time axis is split into bins, each bin being 1-hour long. The data from a given year are averaged within the specific bin, inferring the mean and RMS.

Spread evolution during a day, absolute units, EUR/USD, 2003-2009, CET 1.1 Spread evolution during a day, relative to price, EUR/USD, 2003-2009, CET 1.2

Fig.1. Evolution of spread during a day, EUR/USD, 2003-2009. Central European time. The borders of the shaded areas indicate precision of the data, the width of the shaded band being roughly twice the precision, while the data points are located in the middle of the band. 1.1: spread; 1.2: spread/quote (relative spread).

Central Europen time is chosen for the time axis for the following reason. Forex week begins, roughly speaking (since the volume increase is gradual) on Sunday 5pm and ends Friday 6pm Eastern time. It is convenient to define this week to consist of 5 full days, from 6pm Sunday to 6pm Friday New York time. When it's 6pm in New York, it's midnight in Berlin, Paris, Madrid, Rome, Geneva and Frankfurt. These cities use Central European Time or CET. Therefore, the convenience of using CET is that one gets 5 non-interrupted, full 24-hour long trading days per week. Table 1 compares four time zones including major trading centers of the world.

Some immediate conclusions from Fig.1 are:

  • On average, EUR/USD spread is fairly stable throughout the day. The only feature occuring repeatedly year after year is a spike in spread around lunch time in New York. As you can see from our study of temporal distribution of daily low and high in EUR/USD, this artefact does not have a parallel in the daily patterns of market activity and thus its cause is most likely internal to the broker. One may hypothesize that the broker is somewhat under-stuffed during the lunch hour and imposes a spread increase to counter its reduced ability to hedge its client exposure on the external market. Apparently there is little need to do the same during the night hours, as decreasing market activity decreases risks.

  • A good news for trades is that the EUR/USD spread gradually gets lower with years, but the pattern is step-like rather than continuous. This is understandable, given that this is an administrative decision caused by competition with other brokers.

  • Another good news is that relative spread, the quantity that matters for a systematic trading strategy, seems to be drifting down more steadily, but as comparison between Fig.1.1 and Fig.1.2 indicates, this is in part a lucky result of a combination of a long-term bull trend in EUR/USD and step-like spread reduction.

Spread evolution during a day, EUR/USD, 2003-2009, CET

Fig.2. Evolution of relative spread (spread/price) during a week, EUR/USD, 2003-2009. Central European time. The borders of the shaded areas indicate precision of the data, the width of the shaded band being roughly twice the precision, while the data points are located in the middle of the band.

In much the same approach as Fig.1, Fig.2 shows spread evolution during a five-day forex week. Here a stable pattern is the abnomally high spreads in the first (and sometimes last) trading hours of the week.

Bookmark with:

Deli.cio.us    Digg    reddit    Facebook    StumbleUpon    Newsvine
Last Updated ( Monday, 14 June 2010 16:10 )