Serving the Quantitative Finance Community
How good is the assumption of normal distributions for financial returns?
I think any 'linear' process driven by normal shocks will have unconditional gaussian distributions with a broad range of static conditional correlations (i.e. non-IID)...i.e.y_n = sum<1-K> c_k y_{n-k} + epsilon_nIn continuous time, I think O-U processes must have unconditional gaussian distributions and negative autocorr b/c of the mean-revert.