A General Semiparametric Approach to Inference with Marker-Dependent Hazard Rate Models

Gerard J. van den Berg; Enno Mammen; Lena Janys; Jens Perch Nielsen

doi:10.1016/j.jeconom.2019.05.025

A General Semiparametric Approach to Inference with Marker-Dependent Hazard Rate Models

Gerard J. van den Berg^*, Enno Mammen, Lena Janys, Jens Perch Nielsen

^*Corresponding author for this work

Research output: Contribution to journal › Article (Academic Journal) › peer-review

Abstract

We examine a new general class of hazard rate models for duration
data, containing a parametric and a nonparametric component. Both
can be a mix of a time effect and possibly time-dependent covariate effects. A number of well-known models are special cases. In a counting process framework, a general profile likelihood estimator is developed and the parametric component of the model is shown to be asymptotically normal and efficient. Finite sample properties are investigated in simulations. The estimator is applied to investigate the long-run relationship between birth weight and later-life mortality.

Original language	English
Number of pages	25
Journal	Journal of Econometrics
Early online date	5 Mar 2020
DOIs	https://doi.org/10.1016/j.jeconom.2019.05.025
Publication status	E-pub ahead of print - 5 Mar 2020

Structured keywords

ECON Econometrics
ECON CEPS Data

Keywords

Covariate effects; duration analysis; kernel estimation; mortality; semiparametric estimation

Access to Document

10.1016/j.jeconom.2019.05.025

Embargoed Document

Full-text PDF (author accepted manuscript)
Accepted author manuscript, 2.2 MB
Licence: CC BY-NC-ND
Embargo ends: 5/03/22
Request copy

Professor Gerard J van den Berg
- gerard.vandenberg@bristol.ac.uk
- School of Economics - Honorary Professor
- Bristol Population Health Science Institute
Person: Member, Honorary and Visiting Academic

Cite this

@article{0ccb3df912fe453b97aadf64d0181b21,

title = "A General Semiparametric Approach to Inference with Marker-Dependent Hazard Rate Models",

abstract = "We examine a new general class of hazard rate models for durationdata, containing a parametric and a nonparametric component. Bothcan be a mix of a time effect and possibly time-dependent covariate effects. A number of well-known models are special cases. In a counting process framework, a general profile likelihood estimator is developed and the parametric component of the model is shown to be asymptotically normal and efficient. Finite sample properties are investigated in simulations. The estimator is applied to investigate the long-run relationship between birth weight and later-life mortality.",

keywords = "Covariate effects; duration analysis; kernel estimation; mortality; semiparametric estimation",

author = "{van den Berg}, {Gerard J.} and Enno Mammen and Lena Janys and Nielsen, {Jens Perch}",

year = "2020",

month = mar,

day = "5",

doi = "10.1016/j.jeconom.2019.05.025",

language = "English",

journal = "Journal of Econometrics",

issn = "0304-4076",

publisher = "North-Holland Publishing Company",

}

TY - JOUR

T1 - A General Semiparametric Approach to Inference with Marker-Dependent Hazard Rate Models

AU - van den Berg, Gerard J.

AU - Mammen, Enno

AU - Janys, Lena

AU - Nielsen, Jens Perch

PY - 2020/3/5

Y1 - 2020/3/5

N2 - We examine a new general class of hazard rate models for durationdata, containing a parametric and a nonparametric component. Bothcan be a mix of a time effect and possibly time-dependent covariate effects. A number of well-known models are special cases. In a counting process framework, a general profile likelihood estimator is developed and the parametric component of the model is shown to be asymptotically normal and efficient. Finite sample properties are investigated in simulations. The estimator is applied to investigate the long-run relationship between birth weight and later-life mortality.

AB - We examine a new general class of hazard rate models for durationdata, containing a parametric and a nonparametric component. Bothcan be a mix of a time effect and possibly time-dependent covariate effects. A number of well-known models are special cases. In a counting process framework, a general profile likelihood estimator is developed and the parametric component of the model is shown to be asymptotically normal and efficient. Finite sample properties are investigated in simulations. The estimator is applied to investigate the long-run relationship between birth weight and later-life mortality.

KW - Covariate effects; duration analysis; kernel estimation; mortality; semiparametric estimation

U2 - 10.1016/j.jeconom.2019.05.025

DO - 10.1016/j.jeconom.2019.05.025

M3 - Article (Academic Journal)

JO - Journal of Econometrics

JF - Journal of Econometrics

SN - 0304-4076

ER -

A General Semiparametric Approach to Inference with Marker-Dependent Hazard Rate Models

Abstract

Structured keywords

Keywords

Access to Document

Embargoed Document

Fingerprint

Profiles

Professor Gerard J van den Berg

Cite this