Which of the following is not a requirement for the GaussMar

Which of the following is not a requirement for the Gauss-Markov theorem to be true The model has the form Y=beta_0+beta^1 X 1+beta_2X_2...+beta^pxP+e The error terms must have a normal distribution. The mean of the error terms, E(e_1,) equals 0 Any two error terms e_1,and e_1, are uncorrelated The standard deviation of the error terms is sigma, which does not change with the values of the x variables

Solution

the mean of erros terms equals 0