In my research study, I'm trying to model financial series using two series as data bases: the first series is monthly, and the second series is daily.
I've tried to render the second series monthly as well, by taking the instantaneous average of each month, but I've noticed that the wide fluctuations that make my research so interesting have disappeared.
how can we render a daily time series into a monthly series while retaining the effect of extreme values in the series?
what is the appropriate measure or statistical technique in this case?