The article explores how accurate the least squares estimator is for the autoregressive parameter in a first-order autoregressive model. The researchers found that a simple approximation method works well for this distribution, accurate to the second decimal place in most cases. This method is as good as or better than other existing approximations and is easier for numerical calculations. The study shows that the leading term of the approximation is highly effective in estimating the distribution.