I am trying to construct a portfolio based on 5 assets and initially i used their monthly returns over past 7 years to find out the global minimum variance portfolio (GMVP). Then i used their daily returns of the same period to determine the GMVP and interestingly its sharpe ratio increased dramatically. I am wondering what caused this discrepancy, and if i am going to invest in these assets in a long term perspective, let say 5 years, would you recommend monthly returns or daily returns for analysis and why? Thanks in advance!
Daily returns have too much noise. At the other extreme, annual data have very little noise, but also much less meaningful data (12 data points vs. 260 data points).
Weekly or monthly returns have pretty much the best compromise between signal-to-noise ratio (high) and loss of meaningful data (low).