Hi all, I am looking at intraday FX spot anomalies / patterns. I have obtained minute by minute data for the last 1 year for a variety of currency pairs, but have only looked at USDEUR. One particular anomaly I have found is that during normal london trading hours (0830-1730 GMT/BST), the EUR tends to weaken against the USD. The statistical results are highly significant, at all levels. However I am having a tough time finding a rational economic reason for this anomaly. Any ideas? Also if anyone has any idea at what other intraday anomalies I can look at, please let me know.In return I am happy to give you the data for USDEUR (377720 Observations).Thanks very much.

I heard about a similar pattern in USDJPY : over the past 20 years or so, the USD strengthens significantly vs the JPY on Fridays (whereas USDJPY overall weakened).I thought it might be because important US statistics are realeased on Friday (NFP, michigan), so if the US economy has been in good shape overall, the USD should strengthen when these stats are released, whatever the shape of Japan economy might be.I'm not sure whether the same reasoning can be applied here,as Eurozone stats are released early in the morning (London time) whereas US figures are released between 13.30 and 15.00, so both should have roughly the same impact (although US stats are far more looked at than EUR stats by traders and investors). Also, US stats haven't been that good in the past year.

Never mind. Upon further reflection, I was probably wrong. I'm not big on hasty "ideas" when it comes to trading...Too much wine with my Communion...

QuoteOriginally posted by: farmer Yeah. Next time, go to a college where "economics" isn't just some screed about how "pollution" is an "externality" which needs to be taxed, or about "monopoly" or some other pure garbage.Ouch, farmer that hurt.I appreciate economics is far from being an exact science. Maybe I phrased my question badly, maybe I meant a "rational" explanation rather than "economic" explanation.QuoteOriginally posted by: farmer The answer is so obviousSir, do you care to share ?

Actually, you did your tests on only one year, so this is just on 252 data points, on a period where EURUSD is roughly at the same level where it was 1 yr ago.It might be just a coincidence (statistical significance is not always significant )By how much does EUR weaken against USD in average during those hours?

I don't find any significance in the last 3.5 months (the hourly average change over the period is -0.24 pips (ie EURUSD went down by a quarter of a pip) , and during london hours it is around -0.4 pips).Can you send me your data if you want. Ideally hourly data would make for smaller filesize...or just send some summary statistics to see if we are in the same ballpark.Ta

For doubletouch and quantstudent:Based on minute data over the past year:The daily change during london trading hours given by the regression is -0.32 bpsThe t statistic given by eviews is -7.54 which corresponds to a p-value of 0.000000From an academic point of view,t he most interesting thing is the pretty huge tstat. I am beginning to see this as an example of intraday geographical herding amongst london traders, rather than anything else.doubletouch, what's your data source?

- 0.3 pips... this is nothing.take into account transaction costs, you're 100% sure to lose moneythe huge tstat is still weird though

i am using reuters (hourly bids) and I can only get hourly data for the last 3.5 months, thats why i did not check the whole of last year.to be clear:-0.4 pips during london hours i find is average hourly move in eurusd, so that is almost 4 pips a day. BUT it is not significant.Quote--------------------------------------------------------------------------------Originally posted by: davidh96The daily change during london trading hours given by the regression is -0.32 bps---------------------------------------------------------------------------------0.32 bps: you mean a third of a pip? 0.000032 in EURUSD... that's really not important at all (bid offer spread is 1 pip at best), but the t-stat is weird (unless you are comparing bids minus offers!).Can you send the data?

hi davidh96since it is "0 sum game", EUR has to strengthen during the US trading hours. Does your data confirm this?

hi david,sounds interesting but i am skeptical that it is not spurioussend me the data and i will take a look: jamesh83@gmail.com