In a study of salary discrimination in a company, data are collected on the monthly salary and the characteristics of the employees. A regression model is specified with monthly salary Y as the dependent variable and gender as an independent variable. Beside gender, the only other independent variable is whether the employee does clerical work or not. Hence the independent variables in the regression model are D1, a dummy variable that is 1 if the employee is female and 0 otherwise and D2, a dummy variable that is 1 if the employee does clerical work and 0 otherwise. The regression model has also an intercept. It is suggested that the treatment of male and female employees in the regression model is asymmetrical. To remedy this, the researcher includes a third dummy variable D3 that is 1 if the employee is male and 0 otherwise. The researcher does not succeed in estimating the model that also contains D3. Why?