Im creating a model for volatility of shares prices using EGARCH(1,1) in Eviews.My mean equation is: Share Price = C + CDS Price (cds price as a proxy for %sool data - as my investigation relates to the short selling ban)When running tests for statistical significance most of my results are good. However when I run the ARCH LM test with 1 lag my Obs*R-squared figure is generally much lower than critical value. I appreciate this means my model leaves ARCH factors unexplained. Is there an effective way to fix this and ensure I get good results for my LM test? Many thanks in advance.

Solution: On the off chance another student comes across this same problem - be careful with historic CDS data, some may be rather illiquid therefore price constant, thus not good in a model for volatility.