# Autocorrelation

#### Luigi D'Amato

##### New Member
Hi everyone,
I have a small theoretical doubt about autocorrelation. I am running a montecarlo simulation of the inflation rate in 3 different years (2019,2020,2021) using Crystal Ball and I now have to specify the correlation between the base values. In order to do that I have computed the lag 1 and lag 2 autocorrelation of a historical time series with inflation data.

So is that correct to use the autocorrelation (lag1) coefficient to specify the correlation between 2019 and 2020, while the autocorrelation (lag 2) coefficient to specify the correlation between 2019 and 2021? Instead, if I have to specify the correlation between 2020 and the other two years, should I only use the autocorrelation (lag 1) coefficient?

#### hlsmith

##### Less is more. Stay pure. Stay poor.
Is your data months or years? Perhaps provide a data sample for us. If it was months, that does seem right.

#### Luigi D'Amato

##### New Member
The values to be correlated are those in screen one. The input data is found in screen 3 and the values are calculated on a yearly basis. In order to come up with the autocorrelation factor for years that are contiguous (e.g. '19-'20.'20-'21) I have used a lag of 1 (screen 2). Instead, to calculate the autocorrelation factor for years that are not contiguous (e.g. '19-'21) I have used a lag of 2. Is this all correct?

1)

2)

3)