Unit root and White/Newey-west

[torc]

Hi guys,

Hope you can help me. I have a couple of questions all in relation to time series data.

1) What is a unit root and why is it important to test for in relation to time series data and OLS? What will happen if you used OLS with the data containing a unit root?
2) What is the difference between the White test and Newey-West (time series data)? When will you use White and when will you use the Newey-West?

I'm currently working with a assignment where we have to investigate the oil price influence on stock returns in specific industries. As our "benchmark" oil we are using London Brent and WTI respectively. The two oils are almost perfectly correlated with eachother, but when we use the two oils in the same regression model with the same dependent and independent variables, we get very different results. Someone have an explanation to this??

[Kran]

This is a pretty late reply. but I have had the same question. regarding White and Newey-West.

First of all the White Test is a test for heteroscedasticity.
The white estimator makes your regression heteroscedasticity-consistent
The Newey-West estimator makes your regression heteroscedasticity and autocorrelation consistent.

I am no stat-genious, But this is what I can infeer from "Introductory Econometrics for Finance" by Chris Brooks page 152-153.

By the way an amazing book!

-Kris