Seemingly unrelated regression equations models
Seemingly unrelated regression equations models
Computational Statistics & Data Analysis
Seemingly unrelated regression model with unequal size observations: computational aspects
Computational Statistics & Data Analysis
Efficient algorithms for block downdating of least squares solutions
Applied Numerical Mathematics - Numerical algorithms, parallelism and applications
Bootstrap-based tests for deterministic time-varying coefficients in regression models
Computational Statistics & Data Analysis
Preface: Second Special issue on Computational Econometrics
Computational Statistics & Data Analysis
| Hi-index | 0.00 |
A nonparametric method to estimate time-varying coefficients in seemingly unrelated regression equations models is proposed. The procedure presents two main advantages with respect to other proposals in literature. First, it allows to incorporate both cross and time-varying restrictions into the parameters. Second, the estimator is obtained in a closed form, and there is no need of an iterative method to compute its value. However, this computation requires to solve a linear system where the number of equations and coefficients increases with the sample size. This problem is overcome by using an algorithm that reduces the computation cost. The algorithm enables to solve the linear system in a recursive manner where (in each step) a lower dimensional linear system is solved and there is no increase with sample size.