Given $y = \beta_0 + \beta_1 x + \epsilon$, if the error term $\epsilon$ has a mean of $5$ instead of $0$, what is the effect on the least-squares estimator $\hat{\beta}_0$?