Pitfalls and merits of cointegration-based mortality models

Jarner, Søren Fiig; Jallbjørn, Snorre

doi:10.1016/j.insmatheco.2019.10.005

Cited by 7 publications

(7 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, coherence in itself does not guarantee the type of ongoing change in improvement rates that we advocate. For a more detailed discussion of coherent models and their pros and cons, see Jarner and Jallbjørn (2020) and the references therein.…”

Section: Motivating Examplementioning

confidence: 99%

“…Moreover, the parameters of the Lee-Carter model are not identifiable without additional constraints which precludes the use of more flexible time dynamics such as error-correction models, a limitation the parametric (identifiable) structure does not have, see for example Hunt and Blake (2017a) and Jarner and Jallbjørn (2020) for a detailed discussion.…”

Section: A Lee-carter Baseline?mentioning

confidence: 99%

“…Even though formal testing could be done using the comprehensive statistical framework developed by Johansen (1995), it was deemed problematic that the underlying time dynamics could change annually following data updates as this would destabilize projections and potentially damage the model's credibility. Cointegration is therefore used merely as a modelling tool to achieve reasonable projections, rather than to gain insights into the joint behaviour of the time-varying mortality indices, see also Jarner and Jallbjørn (2020) for further discussion on this point.…”

Section: Cointegrating Gender Dynamicsmentioning

confidence: 99%

“…An explicit representation of the (median) forecast for a horizon h ∈ N 0 can be discerned via the Granger representation theorem, see for example Jarner and Jallbjørn (2020). We have…”

Section: An Error-correction Modelmentioning

confidence: 99%

See 3 more Smart Citations

The Saint Model: A Decade Later

Jarner¹,

Jallbjørn²

2022

ASTIN Bull.

Self Cite

View full text Add to dashboard Cite

While many of the prevalent stochastic mortality models provide adequate short- to medium-term forecasts, only few provide biologically plausible descriptions of mortality on longer horizons and are sufficiently stable to be of practical use in smaller populations. Among the very first to address the issue of modelling adult mortality in small populations was the SAINT model, which has been used for pricing, reserving and longevity risk management by the Danish Labour Market Supplementary Pension Fund (ATP) for more than a decade. The lessons learned have broadened our understanding of desirable model properties from the practitioner’s point of view and have led to a revision of model components to address accuracy, stability, flexibility, explainability and credibility concerns. This paper serves as an update to the original version published 10 years ago and presents the SAINT model with its modifications and the rationale behind them. The main improvement is the generalization of frailty models from deterministic structures to a flexible class of stochastic models. We show by example how the SAINT framework is used for modelling mortality at ATP and make comparisons to the Lee-Carter model.

show abstract

Section: Motivating Examplementioning

confidence: 99%

Section: A Lee-carter Baseline?mentioning

confidence: 99%

Section: Cointegrating Gender Dynamicsmentioning

confidence: 99%

“…An explicit representation of the (median) forecast for a horizon h ∈ N 0 can be discerned via the Granger representation theorem, see for example Jarner and Jallbjørn (2020). We have…”

Section: An Error-correction Modelmentioning

confidence: 99%

See 2 more Smart Citations

The Saint Model: A Decade Later

Jarner¹,

Jallbjørn²

2022

ASTIN Bull.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Several papers extended this set-up to account for issues such as (semi)-coherence (i.e. long-run convergence of the mortality rates of several populations, see Li et al 2017), Bayesian estimation (see Antonio et al 2015), cointegration (see Yang and Wang 2013;Jarner and Jallbjørn 2020), factor-based approaches (see Chen et al 2015). To our knowledge, only Zhou et al (2013) and Özen and Şahin (2021) considered two-population models with jumps in the discrete-time framework.…”

Section: Stochastic Mortality Models and Pandemics: Multi-populationmentioning

confidence: 99%

Stochastic Mortality Models and Pandemic Shocks

Regis

Jevtic

2021

Springer Actuarial

View full text Add to dashboard Cite

After decades of worldwide steady improvements in life expectancy, the COVID-19 pandemic produced a shock that had an extraordinary immediate impact on mortality rates globally. This shock had largely heterogeneous effects across cohorts, socio-economic groups, and nations. It represents a remarkable departure from the secular trends that most of the mortality models have been constructed to capture. Thus, this chapter aims to review the existing literature on stochastic mortality, discussing the features that these models should have in order to be able to incorporate the behaviour of mortality rates following shocks such as the one produced by the COVID-19 pandemic. Multi-population models are needed to describe the heterogeneous impact of pandemic shocks across cohorts of individuals. However, very few of them so far have included jumps. We contribute to the literature by describing a general framework for multi-population models with jumps in continuous-time, using affine jump-diffusive processes.

show abstract