Title: Varying random coefficient models
Authors: Christoph Breunig - Humboldt-Universitat zu Berlin (Germany) [presenting]
Abstract: A linear model with varying random coefficients (VRC) is considered. VRCs are modeled additively separable with an unknown nonlinear function of covariates and an unobservable part. The VRC model allows for heterogeneous marginal effects which might vary with covariates. Identification of the distribution of marginal effects is established. The estimator is based on weighted sieve minimum distance. Its $L_2$ rate of convergence is derived. Pointwise and uniform limit theory of functionals is derived. Our estimator is easy to implement and performs well in finite sample.