Question about cross correlation(intepreting the result)

[Panzer]

Hi im working on a statistics problem for my college that requires me to use cross correlation. Im having trouble figuring out the result eviews is displaying .

I checked the manuals but they are not clear on this subject(if there are some that are, pls do point me to it)

How do i interpret results from the output (cross sorrelogram) after running it in the software? that is, if a coefficient is 0.0123 what does it mean? if the highest value is at lag 0 what does it mean? like that.........

Does the cross correlation coefficients (or are they called correlograms?) provide conclusive information as to the strength of the relationship? if not, what other tests do i have to perform to see their statistical significance? Granger causality test? T statistic?

any help would be much appreciated.

[theologos]

Hello,

What you are estimating is the Pearson correlation coefficient which measures the strength of linear association. The value of 0.0123 means that there is negligible positive linear association, which obviously is non-significant. Whenever the lag is zero you receive the contemporaneous correlation. E-views provide cross-correlations for different lags and leads of the series. It depends on you to decide the extent of lags to examine. To establish association between the variables I am afraid that just reporting the correlation coefficient is not enough piece of evidence. Further testing is needed!


Regards