Dear community,

i am currently trying to create min. volatility portfolios including several indexes. My issue is that i can't just cut the historic stock returns at the last point where all indexes have values as this would drop over 30 years of data.

I am sure that i can somehow predict / estimate the non-existing past returns of the shorter time series.

Can i use AR models on each Time series to predict missing data?

Any piece of help is appreciated!

Best,

Johannes