how to make loop for Johansen cointegration test for multivarite time series and get all possible combinations in Rstudio?

Question

I am very new in R and statistics. I would like to perform the Johansen cointegration test for 75 columns by using a loop. And then to have the result as a data frame or a list sort out trace test and confidence levels 90% 95% 99% into data frame.

I have a time series in format data.frame. As I said, 75 columns with different names (company with daily data) and 495 rows.

According to ADF test time series are stationary.

So basically my data looks like: Name Name2 Name3 Name4 Name5 Name6 Name7 Name8 ... Name75
(all names are different)

I know that to get all possible combinations I have to use combn(), so I presume that my code should look like:

 combn(seq_len(ncol(myData)), 2,
         FUN=function(x) ca.jo(Farm2[, x[1]], Farm2[, x[2]]), simplify=FALSE)

Or may be something like :

for (col in colnames(myData)) {
  print(col)
  data <- myData[, col]
  print(data)
  Jo <- ca.jo(data, type="trace", K=2, ecdet="none", spec="longrun")
  summary(Jo)
}

But nothing is working.

Can someone help me understand and correct my loop?

score 0 · Answer 1 · answered Oct 13 '20 at 12:58

0

Keep in mind that critical values are only reported for systems with less than 11 variables and are taken from Osterwald-Lenum, maybe you should consider restructuring your input data if you want to use Johansen test for cointegration in urca package.

how to make loop for Johansen cointegration test for multivarite time series and get all possible combinations in Rstudio?

1 Answers1