Jarque B test of normality

[tthomas]

Can we rely on the Jarque B test of normality in Eviews for small samples of 600 observations. if the normality is getting rejected what do we do next. I am using panel data of 44 countries for 15 years with Bilateral trade as dependent variable, GDP of exporter and importer, population of exporter and importer, distance , FDI stock , Official exchange rate as independent variables.

[startz]

Why do you care about normality?

[tthomas]

Thanks for responding. According to my understanding parametric tests are relevant only if the residuals are normally distributed. Does it not apply to panel data? Request you to clarify.

[startz]

Parametric tests are generally valid if the error terms are normal or if you have a large sample. Most econometric work relies on the latter because errors are rarely exactly normal.

[tthomas]

Thanks a lot for the clarification.