http://www.stocktradingsites.org/uncategorized/finance-research-arch-and-garch-coefficients-in-stock-returns Stock Trading Sites and Tips Tue, 15 Nov 2011 00:52:25 +0000 hourly 1 http://wordpress.org/?v=3.2.1 http://www.stocktradingsites.org/uncategorized/finance-research-arch-and-garch-coefficients-in-stock-returns#comment-1655 Michael K Fri, 23 Jul 2010 20:08:18 +0000 http://www.stocktradingsites.org/uncategorized/finance-research-arch-and-garch-coefficients-in-stock-returns#comment-1655 (General) auto-regressive conditional heteroskedasticity measures the concentration of variance from the mean trend in a sample. It is useful to determine whether the excess returns observed in a sample are the result of one or two great picks or whether the entire sample exhibits out-performance. If you remove the one or two outliers (which may be errors, or the result of industry or country skewness) and then the sample loses most of its alpha, you have a dangerous model. On the other hand, if there are no general outliers, and the returns are normally distributed, then you have a good stock selection model, which should perform well as long as the underlying assumptions remain valid. (General) auto-regressive conditional heteroskedasticity measures the concentration of variance from the mean trend in a sample. It is useful to determine whether the excess returns observed in a sample are the result of one or two great picks or whether the entire sample exhibits out-performance. If you remove the one or two outliers (which may be errors, or the result of industry or country skewness) and then the sample loses most of its alpha, you have a dangerous model. On the other hand, if there are no general outliers, and the returns are normally distributed, then you have a good stock selection model, which should perform well as long as the underlying assumptions remain valid.

]]>