Changes, from anonymous

May 27, 2015 |

I am very interested in changes in relationships between tradeable financial markets. The key words here are changes and tradeable.

Using a period of, say, three years (totally arbitrary choice I know) most relationships between tradeable assets and other tradeable assets with a lag of, say, 1 day usually present R-squared readings circa 0.00. ( maybe a -0.01 to a + 0.01 range). You might be tempted to call this random and that is probably a reasonable categorisation.

What is of interest is when seemingly out of the blue, the two tradeable assets start exhibiting a strong positive or negative relationship. So what is of interest is the 'inflection point' from 'random' to something potentially predictive.

Studying this phenomenon is truly a meal for several lifetimes. These times should be co-incident with strong periods of either trend following or countertrend performance depending on the resultant positive or negative correlation to previous moves.

What is also very true is this. The market allows these departures from 'randomness' for varying periods of time but almost always takes them back to R-squared = 0 territory again. The lifecycle of the move from a horizontal line of best fit to a positive or negative slope and back again is very interesting.

One example would be the DAX futures contract and the EUR USD spot FX rate. Now over the last few years the movement of the DAX today had little to do with yesterday's movement EURUSD. Of late (looking at the moves in a more complex fashion) the movement in the DAX today has been negatively correlated with the movement in the EURUSD yesterday.

I wonder what caused this change. The proximate reason would be the QE program of the European Central Bank and that is a reasonable assertion.

What we should look at are magnitudes, runs & durations in relationships to see if the departures from randomness and back again are predictive.

My own analysis found one of the more incredible relationships I have ever seen between a weak EUR in a previous period to a strong DAX in the next producing some 2000 DAX points in short order. In subsequent testing this has flatlined and indeed started losing.

I strongly believe that most relationships are pulled back to 'randomness' by ever changing cycles.

Here is a prediction: the easily derived DAX bullish signals from EUR weakness will revert back to a win some, lose some proposition from recent 'Rosetta Stone' significance.

I hope this example helps us clarify our thinking further on these important issues.


WordPress database error: [Table './dailyspeculations_com_@002d_dailywordpress/wp_comments' is marked as crashed and last (automatic?) repair failed]
SELECT * FROM wp_comments WHERE comment_post_ID = '10330' AND comment_approved = '1' ORDER BY comment_date




Speak your mind


Resources & Links