Forex Automaton

This second article in a series dedicated to patterns and stability of forex spread deals with USD/JPY. In this pair, one of the most liquid in forex,  the spread is seen to be quantitatively very close to EUR/USD; its evolution during the day and week is also similar.

The algebra of profit and loss in trading on a margin (follow the link for details) has been discussed before and used in the USD/JPY spread article to justify the conclusion that not the spread itself but its ratio to the price is of direct iterest to the trader and strategy developer intersted in comparing profitabilities of trading strategies applied to different forex pairs under real-world conditions.

In these studies, spread is defined as half the difference between bid and ask prices. I use tick 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 difference.

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

Fig.1 shows evolution of USD/JPY 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.

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, USD/JPY 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 USD/JPY, 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.

  Note that the artefact itself and its discussion in the above paragraph are fully analogous to that for EUR/USD.

• A good news for trades is that the USD/JPY spread gradually gets lower with years, but the pattern is step-like rather than continuous. 2009 brought about a significant reduction in USD/JPY spread charged by this broker, which may reflect the brokerage market conditions at large.

• Unfortunately, the downward trend of spread looks more systematic in the absolute, not relative spread. This may be the result of the fact that it is the absolute spread that the brokers compete on among themselves, based on their perception of (unsophisticated) client demand.

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 abnormally high spreads in the first (and sometimes last) trading hours of the week.

Bookmark with: