Key word in Part 1 is 'frustration', can't argue with that ))

Cointegration = stable short term correlation, so for your strategy to generate sustainable profits in a long run you still have to take correlation as a basis in the first place.

Validation quiestion - age vs. blood pressure are correlated or cointegrated?

Key word in Part 1 is 'frustration', can't argue with that ))

Cointegration = stable short term correlation, so for your cointegration strategy to generate sustainable profits in a long run you still have to take correlation as a basis in the first place.

Validation quiestion - age vs. blood pressure are correlated or cointegrated?

Great article, thought provoking . Thanks for sharing Martyn.

Thanks Martin. I hope you enjoy parts 2 and 3

Hi Oleksandr. Yes I understand and agree with what you are saying. I have a slightly different way of achieving this longer term assuramce without using correlation though. I will explain this in parts 2 and 3. I would value your feedback once you have read those in terms of whether I have addressed your concerns. many thanks for the comment. Martyn

Maybe from different time prespsective every correlation is also a cointegration. I doubt that the cointegration for any cause is naturally given by all the time. It is only a temporary case.

The best explanation I ever heard of cointegration is the very funny analogy of a drunk man walking a dog with a leash….the man will zigzag all over the road but stay, somehow, in his way home and the dog will always stay in a distance limited by the leash. Each one, man and dog, will deviate and reunite during the walk but always stay close and arrive home safe (let’s hope :) ). That’s cointegration to me!!! :)

Now more seriously, Martyn, do you still find value in cointegration trades? My perception is that it got very much squeezed out of the market.

A very good explanation of cointegration, with a fair amount of mathematics but very directed to trading (including matlab codes) is Ernest Chan’s book “Algorithmic Trading – Winning Strategies and Their Rationale”. A must-read….

Looking forward to read Part 2 where he will go over the math.

Thanks Martyn, for shining the light on an important topic. The Co-integration framework is indeed useful when constructing tactical portfolios, managing risk or indeed pair-trading. A great introduction to the subject, starting with Granger in the 80-ies, is found it the first chapter of Banerjee et al "Co-integration, Error-Correction and the Econometrics analysis of non-stationary data".

Thanks Johan. I'll take a look and maybe incorporate some of it onto my next posts if relevant. Thanks for the recommendation

I'm just going to put this out here: cointegration is most useful in trading when it breaks down.

Hi Timothy. Interesting thought. I assume you mean as part of some kind of a pairs breakout system? Is this something that you actively trade?

Yes it's all I trade. There's more to it than just that, but it is the major element for me.

Thanks Timothy. That's certainly given me something to think about. I'll put it on my list of future studies. Thank you for your input. By the way, I don't suppose you've written about this on a blog or similar so that i could get more of an understanding?

I say pairs trading is only interesting if its a true pairs trading, and you want to be market neutral. Else, pairs trading is fundamentally wrong, as Timothy indicate.

Completely agree with Soren... Cointegration is a dream... Just because the systemic risk of the first leg stock will never match continuously the systemic risk of the second leg stock. If you wish stable cointegration look at basket of stocks, but then it become not tradable because of the slippage incurred by the sum of the bid ask spread of each legs....so the best is to find pseudo cointegrated pairs and mix them

To insure a cointegrated book

SIBI Alexandre D. Timothy Spears, and Søren Lanng - Thank you all for your feedback. It is very useful to me because I am still in the process of working out what strategies will work best for cointegrated instruments. By the way, I wasn't trying to suggest which trading systems were viable in my post. The series is not focused on this - it is merely aimed at explaining how to determine cointegration and explain the calculations behind it. Nevertheless your comments have provided me something to think about. Especially yours Timothy. I think your technique of building a system around cointegration breaking down is really interesting and is certainly something I will investigate. Thanks again. It's great to be a part of such a vibrant community with people who are willing to share. Cheers

Quick message to all readers of'Cointegration for Pairs Trading Part 1' - Part 2 is now available here http://hub.tradesignalmachine.com/2016/06/02/cointegration-for-pairs-trading-part-2/

Thanks