Tuesday, May 10, 2011

Time-of-day effects in FX trading

As I mentioned in a previous post, one of the main ingredients of success in constructing a profitable momentum trading strategy in Forex (and futures) is to pay attention to the entry and exit times. I haven't seen any good momentum strategy that has "time-translation invariance", i.e. works without reference to a fixed time of the day. The fixed time can refer to a benchmark level of the market (e.g. the previous close), or it can be the entry or exit time.  (This is in contrast to mean-reverting strategies where the reference price can often be just a moving average.) A recent research paper (Hat tip: William) points to another example of such time-of-day effects in FX markets: a currency typically depreciates during its local trading hours.


Anonymous said...

did anyone try algo-mentor.com working with the UK market ? I know they cover the US market but not sure about the euromarket.

Shaun Overton said...

These guys post a ton of time based analysis on forex markets. I found this last week and am still browsing the material:


Bernd said...

Hi Ernie!

What I really don't know is if Momentum strategies are as appealing as they seem to be.

If you do it in the very short term and you use a stop your broker might "gun" your stop.

Another thing to take in account is that, for knowing when a Momentum strategy has stopped working, you probably need to lose more times than when you do pair trading.

Anyway, most of the strategies I have seen on the net rely on price ranges, breakouts + Fibonacci levels to predict the maximum reach of the momentum, and the usage of certain moving averages as support.

Other ppl claim to be really profitable trading NFP and such.

Do you believe that Momentum trading can beat pair trading consistently?

My guess is that, these kind of strategies are really very time-dependant, being then also more difficult to predict.


Ernie Chan said...

Hi Bernd,
Most momentum strategies have lower Sharpe ratios than mean-reverting strategies, and good momentum strategies are generally harder to find also. Furthermore, when more and more people catch on to a trend, we will need to enter into a position earlier and earlier.

But I want to add momentum strategies to my portfolio not because they are better than mean-reverting strategies, but because their risk-return profile is quite uncorrelated, or maybe even anti-correlated, to that of mean-reverting strategies, hence the overall Sharpe ratio of the portfolio will be higher. In other words, it acts like a hedge.


Paul said...

Hi Erine

I had read the reference reseach paper, it mentions that their FX test data is from EBS pool, I knew bloomberg has EBS but it is very expensive for retail trader to get a bloomberg professional terminal, do you know where I can source the EBS FX data at reasonable cost. Thank you

Ernie Chan said...

While the research was done using EBS data, it is worthwhile to test it on other data. For e.g. you can get half a year of free FX data, even for 1s bars, from Interactive Brokers via their API.

Do let us know if you find that the results are different on IB!

Michael said...

I agree 100% Ernie with your portfolio comment about MR and MOM.

This concept is lost on so many people. Most people only target one or the other.

It is comforting to know that others have come to the same conclusion.

Anonymous said...

Hi Ernest, very nice blog and book of yours, it speaks highly about your work!

I have one question regarding the UIP (Statistical Arb) strategy described by Irene Aldridge in her book; High frequency Trading on page # 191.

Essentially I cannot figure out the first term of equation (13.7) mainly because it looks like it’s a price that will take place in the future i.e. St+1,CHF/USD

In such equation, there is no confusion regarding the terms that appear in the right side of the equality, since I could OLS the alphas and betas, plus I could add the difference between the interest rates of each country.

As a matter of fact I could even know the Spot rate that is the second term from left to right of the equality ( i.e. the first one to the left of the equal sign)….. So I’m stuck trying to solve for the unknown St+1,CHF/USD…. I wonder if you have any idea what symbol or financial instrument should I use?..... because I don’t understand what spread do I buy or sell on the left hand side of the equation!



Ernie Chan said...

Hi Brian,
Thanks for your kind comment on my book. I actually don't have Aldridge's book, so I would have to purchase a copy before I can answer your question!

J. Monson said...

Fx microstructure studies show that there are clear time of day effects. The London and Tokyo trading "sessions" are U-shaped when it comes to trading activity: there are many trades for the first couple of hours, then things die down, then there are many trades made during the last 2-3 hours.

There is no "Monday morning" effect--nothing out of the ordinary happens because trading starts after the weekend. Regarding what currencies are traded on, "home" currencies are always traded much more heavily than others; i.e., */USD currency pairs are traded hevily during the US session (7AM EST in the US until 5PM EST).

Regarding "home" currencies depreciating, this has been seen in a study or two, but that's all. Most studies do not note this, they simply note heavy trading on them.

All this comes from Economics micromarket/microstructure studies.

Lenore said...

Quite useful material, thanks so much for the post.