A human resources analyst fits a regression model: $\hat{y} = 45000 + 2500x_1 + 8000x_2$ where $y$ is annual salary (dollars), $x_1$ is years of experience, and $x_2$ is a dummy variable coded as 1 = male, 0 = female. What is the correct interpretation of the coefficient 8000?