Why is the Pearson correlation coefficient used to derive correlations between stock returns if the stock returns are not normally distributed? I was under the assumption that the data should be normally distributed or close to that but in case of stock returns, especially short time series of daily returns, they are not.