1 / 9

Econometrics

Econometrics. Chapter 13 Heteroskedasticity. Heteroskedasticity - Chapter 13. Gauss-Markov Theorem – A3 Errors have a constant variance from observation to observation If this is Not True:

wynter-graves
Download Presentation

Econometrics

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Econometrics Chapter 13 Heteroskedasticity

  2. Heteroskedasticity - Chapter 13 • Gauss-Markov Theorem – A3 • Errors have a constant variance from observation to observation • If this is Not True: • Some data points will be more accurate measures of the true relationships between the variables than others • This is True because:

  3. Heteroskedasticity – continuedWhy Some Data More Accurately Reflect Relationships Among Variables • Larger random numbers have larger variances • Averages have smaller variances as the number of observations that produced the averages rises • Quality of data measurement varies • Inherent variances in the unobserved factors can cause Error Terms in our models

  4. Heteroskedasticity – continuedConsequences of Heteroskedasticity • OLS Estimates of the Betas are Unbiased • OLS Estimates of the Betas are Consistent • OLS Estimates of the Betas become Inconsistent • OLS Estimates of the Standard Errors of the Betas become Biased and Inconsistent

  5. Heteroskedasticity – continuedDetecting Heteroskedasticity • Graphically • “Trumpet-Shaped” Residuals • Goldfeld-Quandt Test • Different “Grouped” Regressions • GQ = SSR1 / SSR2 • If No Hetero, Then GQ = 1 • Use F-test • Breusch-Pagan-Godfrey Test • Squared Residuals are Regressed on Squared LHS Variables • BPG = N x R2 should = 0 if Homoskedastic • Use Chi-Squared Test

  6. Heteroskedasticity – continuedCorrecting For Heteroskedasticity • Transform the variables using weighted least squares method • Known form of Heteroskedasticity • White Test • Then use WLS • Used for Unknown Forms • Best Guess as To Weights • No “Good” Solution

  7. Heteroskedasticity – continuedSummary – Chapter 13 • Heteroskedasticity: Some data points are more reliable than others – they have a smaller variance, lie closer to the true relationship between the variables in a regression, and are better indicators of what the true data generating process is

  8. Heteroskedasticity – continuedSummary – Chapter 13 • Scale, Averaging, and Data Quality may cause Heteroskedasticity • OLS Is Unbiased and Consistent in the presence of Heteroskedasticity • OLS becomes Inefficient in the presence of Heteroskedasticity • The Estimated Standard Errors of the parameters are Biased – can’t assess the precision of estimates or test hypothesis about them

  9. Heteroskedasticity – continuedSummary – Chapter 13 • Weighted Least Squares can be used to de-emphasize high variance observations and produce efficient estimates when Heteroskedasticity is of known form • Goldfeldt-Quandt and Breusch-Pagan-Godfrey Tests can detect Heteroskedasticity • White Test can detect Heteroskedasticity of unknown form

More Related