Hello, everyone. When I am research the stock index data of different country, I have a question: I set the time is from 2006-01-01 to 2013-12-31, and I got the historical data of FTSE100, SP100, HANG SENG INDEX, SSE Composite Index from Google Finance(daily data). But I found that, even if the time width is the same, but the sample size is different, for example, there are just 2194 sample for SSE Composite Index, but there are 2278 for FTSE100. I want to use this data for Copula and Monte Carlo simulation for VaR estimation. Does anyone can give me some hints about how to deal with this?

Not answering your question as such, but just adding that Cholesky decomposition may need to be used when random sampling from joint multi-variate distributions. Others, please feel free to correct me -or- comment further. Thanks.

I don't know what you're going to do with it, but one suggestion is to just define a 'good pair'.Say a good pair is where you have both returns on the same day.Of course, the trading hrs are different; so by 'same day" I really mean aligned so the disparity in trading hrs is as smallas possible. Given any two series, this may involve shifting one of them by a day -- I hope that is clear. Anyway, just keep the good pairs and throw out the rest.

work with with weekly returns... it will make your life much easier and I suspect you will not lose much information.