So reading 25 says that returns correlated across time will artificially lower the standard deviation. But from our normal standard deviation formula (below), if correlation increases, it increases standard deviation. Why are these contradicting? variance = w1^2xvar1^2 + w2^2xvar2^2 + 2xw1xw2xcorrelation x SD1xSD2.
The second formula looks at the variance coming from two factors (e.g. two firms in the portfolio). If the two factors move in sync they will reinforce each other’s volatiltiy.
The first statement is just about 1 factor (the firm’s return). If a fund’s return is likely to be correlated with its previous return (e.g. 10,11,11,12) then volatility looks low. This is artificial as if it then drops low, it will likely stay there (e.g. 2,2,3,3).