If a regression model is calculated as $\hat{y} = 2x + 5$, and you add a non-significant predictor $x_2$ to the model, what happens to the Adjusted $R^2$?