Serial independence tests for innovations of conditional mean and variance models

• Kilani Ghoudi ^[1] ; Bruno Rémillard ^[2]
  1. [1] United Arab Emirates University

     United Arab Emirates University

     Emiratos Árabes Unidos

     
  2. [2] HEC Montréal
• Localización: Test: An Official Journal of the Spanish Society of Statistics and Operations Research, ISSN-e 1863-8260, ISSN 1133-0686, Vol. 27, Nº. 1, 2018 (Ejemplar dedicado a: Special issue on goodness of fit (GOF)), págs. 3-26
• Idioma: inglés
• DOI: 10.1007/s11749-016-0521-3
• Texto completo no disponible (Saber más ...)
• Resumen

  • In this paper, one studies the asymptotic behavior of empirical processes based on consecutive residuals of univariate conditional mean and variance models. These processes are then used to develop tests of serial independence of the innovations. Even if the limiting distributions of the empirical processes depend on unknown parameters, it is shown that a Monte Carlo method based on the so-called multipliers can be applied to estimate the P values of the proposed test statistics. A simulation study is carried out to demonstrate the effectiveness of the proposed tests and the behavior of the statistics is also studied under contiguous alternatives