This
** preview**
has intentionally

**sections.**

*blurred***to view the full version.**

*Sign up*
**Unformatted text preview: **the VIFs. Based on VIF values ﬁnd out which variables are most aﬀected by multicollinearity. 3. Consider the following regression equation, Y i = β + β 1 X i + ± i (2) Now suppose, V ar ( ± i ) = σ 2 X 3 i (3) (3.a) How would you transform the model to achieve homoskedastic ( or constant) error variance? Explain. (3.b) Is the OLS estimator of the transformed regression BLUE? 4. Download RD.xls from Carmen. The ﬁle contains cross section data on Research and Development expenditure in US in 1988 for diﬀerent industries: Sales i- sales in the industry i. RDexp i- R&D expenditure in industry i. Now suppose you want to estimate, RDexp i = β + β 1 Sales i + ± i (4) 4.a Estimate (4) using Eviews. Submit your output. 4.b Using Park’s test, test whether there is heteroskedasticity in (4) at 5% level of signiﬁcance. Report your eviews output for this test and state clearly whether you ﬁnd heteroskedasticity in the data. 2...

View
Full Document