Mixed Estimation When the Model and/or Stochastic Restrictions are Nonlinear
The standard mixed estimation method allows the incorporation of linear stochastic constraints into the estimation of a linear regression model. The present paper shows how the method can be adapted and extended to accommodate nonlinearities in the model, in the constraints, or both. As an illustration, it shows how nonlinear constraints can be defined so as to impose strict bounds on parameters of the model, or functions of parameters.