economicsstudent wrote:Carrying out tests of efficiency on stock market returns - if i have concluded that the majority of tests support the returns being stationary, is it a contradiction to then say that the same data displays evidence of ARCH effects?
You don't want them to be stationary, you want them to be uncorrelated. Those aren't even close to being the same thing.
ARCH effects means that the returns are dependent in second moments, so they aren't serially independent. If you're testing a version of the EMH that requires independence, then the presence of ARCH would reject that. If you use a weaker version which requires only a lack of correlation, then the presence of ARCH wouldn't reject that.
If you are confused about this, you need to discuss it with your teacher. Most of your questions have nothing to do with using the RATS program.