Assume that the assumptions of a classical linear regression model apply but the true value of the constant is zero, so that:
$$y=X\beta+u$$
where $X$ does not contain a column of ones in its first column.
How can I compare the expected value and variance of the least squares slope estimator computed without a constant term compared to estimator computed with constant term using the model:
$$y=\beta_{0}+X\beta+u$$
Much Thanks in advance.