# regression wth just an intercept

$Y=\beta_0+U_i$
Now, if by doing a sample of 500 observations, I find that , $\beta_{hat} = 2.5$, how can I build a 95% Confidence interval if I don't have the std. error?