G-2003-52
On the Asymptotic Distribution of the Residual Autocovariance Matrices in the Autoregressive Conditional Multinomial Model
référence BibTeX
The autoregressive conditional multinomial model describes vector time series of conditional probabilities, where the variable of interest is multinomial. In financial applications, such time series could consist of price changes that are assumed to have a finite support. These models have been introduced by Russell and Engle (2002). In this note, we derive the asymptotic distribution of the residual autocovariance matrices in this class of models. As an application of this result, a new test statistic for diagnostic checking is presented and its asymptotic distribution is established.
Paru en août 2003 , 11 pages
Publication
jan. 2004
On the asymptotic distribution of the residual autocovariance matrices in the autoregressive conditional multinomial model
ECONOMICS LETTERS, 83(2), 193–197, 2004
référence BibTeX