I have a time series dataset of say e.g car sales from Jan 2008 to Dec 2017. And I have to find the correlation between Car sales in March 2009 and March 2012, and between April 2009 and June 2009. So i open up Gretl and I see that my data is non-stationary in terms of seasonality so it means there's definitely seasonality involved but i don't know how to find the correlation between sales from different months.
So what should I do here? Do i need to look at the ACF and PACF of the values? Or is there some calculation involving them that i'm missing. Or do i need to find out if i have multiplicative or additive seasonality and how do i do that?
Thanks
I tried to just compare the ACF between differences for the corrected time series. E.g my time series needed to be written in log terms because of the variance and it had seasonality non-stationary so i applied a seasonal diff and got sd_l_cars and then i checked correlogram and saw ACF for the amount of difference. E.g March 2009-12 has a difference of 36 months so i checked for ACF for 36. I followed this formula i found: rho_j/rho_0 where j is the number of periods that separate the two values. So rho_0 is then the variance of the value itself and that must be 1 but i don't understand that