Cointegration is often used in statistical abitrage as a way to identify how to combine some tradable instrument to obtain a *mean reverting* one.
Before asking any question about cointegration, please read this paper: A Drunk and Her Dog: An Illustration of Cointegration and Error Correction by Michael P. Murray in The American Statistician, Vol. 48, No. 1 (Feb., 1994), pp. 37-39.
From the viewpoint of the users of quant.stackexchange, questions on cointegration can be clustered in several categories:
how to put in place arbitrage once cointegration is identified?
how to build a cointegrated relatinship between stochastic processes?
- general statistical questions
- questions related to the time frame / window
- generic presentation / questions
- A gentle introduction to cointegration
- How are correlation and cointegration related?
- What is the intuition behind cointegration?
- Is there a measure for the 'degree' of cointegration
- How to interpret the eigenmatrix from a Johansen cointegration test?
- practical and specific questions
Before asking a new question you should check that it has not been already asked, and at least look at the associated wikipedia article. Please feel free to enrich the wikipedia page thanks to the answer found at stackexchange.
For more bibliographic reference, see the associated citeulike search.