I have a dataset with three variables: 1) mutual fund returns (MF), 2) stock index returns (SI), 3) oil price returns (OP).

I have computed a rolling (overlapping) window of correlation coefficients between i) MF and OP, and between ii) SI and OP. I get two vectors of N correlation coefficients each.

The eyeball-metric seems to indicate that case i) has a much higher correlation coefficient than case ii).

However, I want to test the difference in correlation statistically. How should this be done?

  • $\begingroup$ What is rolling (overlapping) window of correlation ... $\endgroup$ – Subhash C. Davar Oct 3 '17 at 16:25

1) It should be clarified first whether you made a proper data preprocessing to remove unstationary patterns from all of your data.

2) You can use the Fisher's transform of the Pearson correlation coefficients to get unbiased normal statistics and standard errors, and then make a two sample T-test.

  • $\begingroup$ Hi Alexey, thanks for your comment and suggestions! 1) No, the data is not stationary, but correlation coefficients are computed from returns. 2) Thanks, I will do that! $\endgroup$ – Thomas.LRV Oct 4 '17 at 17:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.