How do I quantify the time-lag in a time series data set? I “know” there is time-lagged correlation, but I don’t know how “long” or “short” the time lag is. How should I approach this?

Answer 1

What you need to do is calculate the correlation coefficient between the x's x's and the (x-1)'s, (x-2)'s,

(x-3')'s, ..................., (x-n)'s..

The value of n which gives you the highest value of autocorrelation is your lag period.