All Outputs (4)

The role of information in nonstationary regression (2019)
Journal Article
Marsh, P. (2019). The role of information in nonstationary regression. Statistics, 53(3), 656-672. https://doi.org/10.1080/02331888.2019.1605516

The role of standard likelihood based measures of information and efficiency is unclear when regressions involve nonstationary data. Typically the standardized score is not asymptotically Gaussian and the standardized Hessian has a stochastic, rather... Read More about The role of information in nonstationary regression.

Properties of the Power Envelope for Tests Against Both Stationary and Explosive Alternatives: The Effect of Trends (2019)
Journal Article
Marsh, P. (2020). Properties of the Power Envelope for Tests Against Both Stationary and Explosive Alternatives: The Effect of Trends. Journal of Time Series Analysis, 41(1), 146-153. https://doi.org/10.1111/jtsa.12458

This paper details a precise analytic e¤ect that inclusion of a linear trend has on the power of Neyman-Pearson point optimal unit root tests and thence the power envelope. Both stationary and explosive alternatives are considered. The envelope can b... Read More about Properties of the Power Envelope for Tests Against Both Stationary and Explosive Alternatives: The Effect of Trends.

Nonparametric series density estimation and testing (2018)
Journal Article
Marsh, P. (2019). Nonparametric series density estimation and testing. Statistical Methods and Applications, 28(1), 77–99. https://doi.org/10.1007/s10260-018-00432-y

This paper .rst establishes consistency of the exponential series density estimator when nuisance parameters are estimated as a preliminary step. Convergence in relative entropy of the density estimator is preserved, which in turn implies that the qu... Read More about Nonparametric series density estimation and testing.

Approximation for the conditional distribution of the MLE with application to autoregression (2014)
Journal Article
Marsh, P. (in press). Approximation for the conditional distribution of the MLE with application to autoregression. Advances and applications in statistics, 42(1),

The famous p* formula provides a higher-order approximation for the conditional distribution of the maximum likelihood estimator given an exact ancillary. By collating well known existing results including application of a formal higher-order expansi... Read More about Approximation for the conditional distribution of the MLE with application to autoregression.