class16-06 - Econ 444, Monday November 20, class 16 Econ...

Info iconThis preview shows pages 1–7. Sign up to view the full content.

View Full Document Right Arrow Icon
Econ 444, Monday November 20, class 16 Econ 444, Monday November 20, class 16 Robert de Jong 1 1 Department of Economics Ohio State University Robert de Jong Econ 444, Monday November 20, class 16
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Econ 444, Monday November 20, class 16 Monday November 20 1 Heteroskedasticity - repeat of last Wednesday 2 Autocorrelation 3 Probit model and fixed effect model Robert de Jong Econ 444, Monday November 20, class 16
Background image of page 2
Econ 444, Monday November 20, class 16 Heteroskedasticity is the failure of model assumption 5 The model assumptions: 1 The regression model is linear in the coefficients, is correctly specified, and has an additive error term. 2 The error term has a zero population term. 3 All explanatory variables are uncorrelated with the error term. 4 Observations of the error term are uncorrelated with each other (no serial correlation). 5 The error term has a constant variance (no heteroskedasticity). 6 No explanatory variable is a perfect linear combination of any other explanatory variable(s) (no perfect multicollinearity). 7 The error term is normally distributed (this assumption is optional but usually is invoked). Robert de Jong Econ 444, Monday November 20, class 16
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Econ 444, Monday November 20, class 16 Heteroskedasticity can take various forms Example: expenditure on housing vs. income Example: crop yield vs. world price over 24 years, technical innovation after 12 years Example: sales of VCRs in the 1970-1992 period explained from price level and income Consequence of heteroskedasticity: 1 The coefficients will still remain correct. 2 t -values, standard errors and tests will be incorrect Compare to omitted variable or endogeneity: coefficients AND standard errors and tests will be incorrect Robert de Jong Econ 444, Monday November 20, class 16
Background image of page 4
Econ 444, Monday November 20, class 16 Testing: use the White test 1 Obtain residuals 2 Run a regression of the squared residuals on the regressors and on squares and cross-products. Example for two regressors: ( e i ) 2 = α 0 + α 1 X 1 i + α 2 X 2 i + α 3 X 1 i X 2 i + α 3 X 2 1 i + α 4 X 2 2 i + u i 3 Calculate nR 2 and compare to critical value from chi-square table with appropriate number of degrees of freedom l 4 l equals the number of variables in the above regression, not counting the constant (here: 5) Robert de Jong Econ 444, Monday November 20, class 16
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Econ 444, Monday November 20, class 16 Problems with White test: 1 not “automatic" - i.e. we get no p -value that can be compared to 0.05 2 number of terms can be large if number of regressors is large Robert de Jong Econ 444, Monday November 20, class 16
Background image of page 6
Image of page 7
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 24

class16-06 - Econ 444, Monday November 20, class 16 Econ...

This preview shows document pages 1 - 7. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online