Abstract: In this paper, we propose a new nonparametric method to impose time varying restrictions in econometric models. The method relies on the assumption that the structural parameters evolve smoothly over time and they are required to satisfy some pre-specified constraints. Smoothing techniques are employed in order to estimate the parameter sequences. In particular, we propose a local constrained least squares method. The asymptotic properties are derived and a test procedure for the validity of the time varying constraints is derived. The methodology is applied to estimate a time varying demand system and a simulation study is performed to illustrate the estimation method as well as the test procedure.
Key words and phrases: Constrained estimators, nonparametric estimation, time varying coefficients, time varying cross restrictions.