Nonparametric spatial regression under near-epoch dependence
2011; Elsevier BV; Volume: 167; Issue: 1 Linguagem: Inglês
10.1016/j.jeconom.2011.11.008
ISSN1872-6895
Autores Tópico(s)Economic and Environmental Valuation
ResumoThis paper establishes asymptotic normality and uniform consistency with convergence rates of the local linear estimator for spatial near-epoch dependent (NED) processes. The class of the NED spatial processes covers important spatial processes, including nonlinear autoregressive and infinite moving average random fields, which generally do not satisfy mixing conditions. Apart from accommodating a larger class of dependent processes, the proposed asymptotic theory allows for triangular arrays of heterogeneous random fields located on unevenly spaced lattices and sampled over regions of arbitrary configuration. All these features make the results applicable in a wide range of empirical settings.
Referência(s)