In reviewing the 2011 PM mock, I came across questions 26 and 27 which bugged me.
The answer to 26 explains how the use of appraisal data (less-frequent valuations) causes correlations with other assets to be biased downward in absolute value. I get this, I agree with this, I remember reading this.
Then, moving onto 27, the answer explains how high-frequency data is more sensitive to asynchronism and, thus, tends to produce lower correlation estimates.
I’m obviously missing something fundamental here? Where is the line drawn?