1) Use plot of the stock return and consider the Autocorrelation Function to determine the auto-regressive structure of the data and explain why you think the return is stationary.
2) Use the information in (1) to estimate a univariate auto-regressive (AR) model of the return series (at the minimum estimate an AR(1) or AR(2) model).
3) Save the residuals from the model in (1) and use their square to run an auxiliary regression to test for ARCH.
4) Re-estimate the time series model either correcting for ARCH (using the GLS method in the lectures) or by augmenting the model by one or two dummy variables to correct for non-normality (any large shocks in the stock return).