Humboldt-Universität zu Berlin - High Dimensional Nonstationary Time Series

SFB649DP2015 053

Specification Testing in Random Coefficient Models

Christoph Breunig
Stefan Hoderlein

Abstract:
In this paper, we suggest and analyze a new class of specification tests
for random coefficient models. These tests allow to assess the validity of
central structural features of the model, in particular linearity in coefficients
and generalizations of this notion like a known nonlinear functional relationship.
They also allow to test for degeneracy of the distribution of a random coefficient,
i.e., whether a coefficient is fixed or random, including whether an associated
variable can be omitted altogether. Our tests are nonparametric in nature, and use
sieve estimators of the characteristic function. We analyze their power against
both global and local alternatives in large samples and through a Monte Carlo simulation study. Finally, we apply our framework to analyze the specification
in a heterogeneous random coefficients consumer demand model.

Keywords:
Nonparametric specification testing, random coefficients, unobserved heterogeneity, sieve minimum distance, characteristic function, consumer demand

JEL Classification:
C12, C14