To estimate standard error, this article uses a nonparametric method, bootstrapping, which does not assume homoskedasticity
and normality of errors.
Keywords: regression model, homoskedasticity
, testing for heteroskedasticity, software environment MATLAB
(2) To test the Homoskedasticity
of the residuals, the Breusch--Pagan/Cook-Weisberg test is performed.
Here both tests strongly reject the null hypothesis of homoskedasticity
; across all our models, the chi-square statistic is highly significant (the p value is zero to four decimal places), indicating that the identification assumption of Lewbel's method about heteroskedasticity holds.
These two tests are executed where the null hypothesis is of constant variance (No heteroskedasticity/ Homoskedasticity
).The results of White's and Breusch-Pagan's test are given in the following Table 3.
assumption is lifted by considering different weighting matrices.
[[mu].sub.it] denotes that homoskedasticity
is assumed and not correlated over time [[alpha].sub.i] is time variant and homoskedasticity
is assumed across firms.
The white test results for Model I and Model II depicted (probability value> Chi2) 0.2518 and 0.5118 respectively which indicated the acceptance of null hypothesis at 5 percent level of significance and favoured Homoskedasticity
Our null hypothesis is data homoskedasticity
. The results from the White Test render F-test p-values much below the 5-percent level, indicating that the data are highly heteroskedastistic.