Is there a way to find the most cointegrated pair between multiple time series? (Pair trading)
1 view (last 30 days)
Show older comments
Hi,
I wondered if there exists a function which takes as input 1) a N x M matrix (e.g. M stocks with daily data for N days) 2) whether to include nonzero mean and/or a trend, and returns the cointegrated pairs, preferably sorted from "most-likely-cointegrated" pair to "least-likely" (Maybe even testing the series for being I(1) beforehand...?).
Since I want to use the output for two-stock pair-trading, I am not looking for a Johansen-test output, where I'll have to take positions based on the eigenvectors.
I think I remember some webinar where someone had done it by returning a two dimensional color output, such that you could see the most - and least cointegrated pairs.
I hope someone can help. Thanks a lot in advance
0 Comments
Answers (0)
See Also
Categories
Find more on Cointegration Analysis in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!