Why assuming the normality of the errors is important in lin

Why assuming the normality of the errors is important in linear regression? What happenes if it does not hold? Ho will the values of b₀, b₁, ..., b_k be affected, if at all?

Solution

not necessary to assume all time normality assumption. You can take any other continuous distribution whose support is real line. but in general error should be this  which behaves normally in nature. The estimate of parameters are not differ but while you are supposed to test this paremeter then there will be difficulty in it.

So for simplysity we assume nomality assumption.