Joint mortality modeling based on Wang transform

Tandy Xu

Research output: Contribution to journalMeeting abstract


Purpose: This research analyses the joint behaviour of mortality in different populations, and aims to model their difference and similarities using a joint model. Originality: Existing mortality studies typically analyse single populations with each population analysed separately and without reference to other populations. Such an approach limits the pooling of common trend and cross sectional information across populations and limits the ability to discern and understand differences between populations. The multiple population framework permits detailed analyses of differences and co-integration behaviour. Key literature/Theoretical Perspective: Wang Transform, which was proposed as a new method to project the mortality by Jong and Marshall (2007), is basic for this research, and an extension of the application of mortality forecasting based on Wang Transform, the linear framework of Wang Transform, is deduced in the research. Then the results will be compared with some other similar studies, Li and Lee (2005), Debón et al. (2010), and Cairns et al. (2010). Design/methodology/approach: Wang Transform has been proposed to forecast mortality. This research extends this application to a multiple population framework, and then analyses data from different and similar populations. Therefore, is will focus on theoretical deduction and quantitative analysis. Findings: Similarities across populations are specified as restrictions and are tested. Combining populations makes for more efficient forecasting and the analysis and understanding of divergent trends in different populations. Research limitations/implications: Restrictions are not easy to establish, and reasons for differences may be complex and not simply quantified. Practical and Social implications: This research may lead to theoretical developments that will derive a new method to describe the mortality of a group of populations. Furthermore, if a population is too small to be estimated, or its data is not reliable, it can be forecast by reference to a highly similar population.
Original languageEnglish
Pages (from-to)106-107
Number of pages2
JournalExpo 2011 Higher Degree Research : book of abstracts
Publication statusPublished - 2011
EventHigher Degree Research Expo (7th : 2011) - Sydney
Duration: 10 Oct 201111 Oct 2011


  • Joint mortality model
  • Wang Transform
  • Linear Framework
  • Combined Population

Fingerprint Dive into the research topics of 'Joint mortality modeling based on Wang transform'. Together they form a unique fingerprint.

Cite this