The document discusses heteroscedasticity, including its nature, sources, consequences, detection, and remedial measures. Heteroscedasticity occurs when the variance of the error term differs across observations, violating the assumption of homoscedasticity. It can arise from sample heterogeneity or outliers. While OLS estimates remain unbiased, their standard errors are invalid under heteroscedasticity.
The document discusses heteroscedasticity, including its nature, sources, consequences, detection, and remedial measures. Heteroscedasticity occurs when the variance of the error term differs across observations, violating the assumption of homoscedasticity. It can arise from sample heterogeneity or outliers. While OLS estimates remain unbiased, their standard errors are invalid under heteroscedasticity.
The document discusses heteroscedasticity, including its nature, sources, consequences, detection, and remedial measures. Heteroscedasticity occurs when the variance of the error term differs across observations, violating the assumption of homoscedasticity. It can arise from sample heterogeneity or outliers. While OLS estimates remain unbiased, their standard errors are invalid under heteroscedasticity.
The document discusses heteroscedasticity, including its nature, sources, consequences, detection, and remedial measures. Heteroscedasticity occurs when the variance of the error term differs across observations, violating the assumption of homoscedasticity. It can arise from sample heterogeneity or outliers. While OLS estimates remain unbiased, their standard errors are invalid under heteroscedasticity.
6 Sources of heteroscedasticity • 1- heterogeneity of the studied sample ( for example a sample of income of a large number of developed and underdeveloped countries) Heteroscedasticity is more common in cross-sectional data than in time series data. As the cross-sectional data deals with different individual observations at a given point of time, While in time series data, one generally collects data for the same object over certain period of time. • 2- Heteroscedasticity can also arise as a result of the presence of outliers. outlier, is an observation that is much different (either very small or very large) in relation to the observations in the sample. 7 The Consequences of Heteroscedasticity 1- OLS estimator is still linear unbiased and consistent, even if we do not assume homoscedasticity • The estimator is consistent means that it converges upon the true value as the sample size increases. 2- The OLS estimator is no longer best. There is another estimator with a smaller variance 3- The standard errors of the estimates are biased if we have heteroscedasticity 4- If the standard errors are biased, we can not use the usual t statistics or f statistics for drawing inferences .so confidence interval, f and t tests are invalid 8 Detection or Testing for heteroscedasticity • Informal method – Graphical method • Formal method – Park test – Glejser test – Goldfeld-Quandt test – White test
9 The Detection of Heteroscedasticity
Informal testing for Heteroscedasticity
Graphical method – Run the regression assuming no heteroscedasity • Then plot the squared estimated residuals with respect to predicted Yi or any Xi and examine whether there is a systematic pattern or not • We note that there is no systematic pattern in the first graph (a) so there is no problem while in other graphs there are systematic pattern of the relationship between the squared estimated residuals and any Xi so there is heteroscedasity problem but we must do formal test to be sure that there is a problem or not. 10 11
