|
Abstract:
|
This paper considers a simple linear regression model with long memory design and nonparametric heteroscedastic long memory moving average errors. In some cases, the first-order asymptotic distribution of the least square estimator of the slope parameter is degenerate. In these cases, under mild conditions, the second-order asymptotic distributions of this estimator are provided. In addition, an estimator of H, based on pseudo-residuals, is shown to be \log (n)-consistent. All these estimators and findings are then used to carry out a lack-of-fit test of a simple linear regression model using a marked empirical process. The proposed lack-of-fit test is then applied to fit a simple linear regression model to a currency exchange dataset.
|
