Given two confidence bands, it seems natural to consider the confidence band of the product of these two distributions. Does this have a name? Are there any reference that might be helpful regarding this notion?
Edit: Assume that the errors of the first, say, regression curve are not correlated with the errors of the second.