Hi am trying to plot co-integration of TCS and INFY over 100 days periods using historical data of several years.
The problem is I am getting different results when running on 10 year data, 5 year data and 2 year data even though I am only running it for 100 days at a time. This must mean that each file gives a different result for similar date range.
What do I not understand here?
Thanks for your query. It seems like you are using the first 100 days for each of the datasets and checking for cointegration which is giving you different results on a different amount of historical data. If the ending point is the same then you can use the last 100 days of data to check for cointegration. This would give you same results.
Also, the idea is cointegration is associated with the long-term property of a time-series, so 100 days are good for checking for cointegration. 