y=b.x + e

where b is the vector of coefficients, x are the explanotory variables and e follows a normal distribution, mean zero with variance s^2(e).

As part of the estimation, the standard deviation is also part of a regression.

s(e)=d.z + v

where d are the coefficients, z is the explanotory variables and v is the residual with mean zero and variance s^2(v).

What is the name of this model? Once i know the name i assume i should be able to find plenty of information on how to estimate it.

Cheers,

TF