A multi-population evaluation of the Poisson common factor model for projecting mortality jointly for both sexes

Abstract

Mortality forecasts are critically important inputs to the consideration of a range of demographically-related policy challenges facing governments in more developed countries. While methods for jointly forecasting mortality for sub-populations offer the advantage of avoiding undesirable divergence in the forecasts of related populations, little is known about whether they improve forecast accuracy. Using mortality data from ten populations, we evaluate the data fitting and forecast performance of the Poisson common factor model (PCFM) for projecting both sexes' mortality jointly against the Poisson Lee-Carter model applied separately to each sex. We find that overall the PCFM generates the more desirable results. Firstly, the PCFM ensures that the projected male-to-female ratio of death rates at each age converges to a constant in the long run. Secondly, using out-of-sample analysis, we find that the PCFM provides more accurate projection of the sex ratios of death rates, with the advantage being greater for longer-term forecasts. Thus the PCFM offers a viable and sensible means for coherently forecasting the mortality of both sexes. There are also significant financial implications in allowing for the comovement of mortality of females and males properly.

A Double Common Factor Model for Mortality Projection Using Best-Performance Mortality Rates as Reference

“…It may be explained by the fact that the USA is a more populous country and its mortality patterns are naturally more complex. This observation agrees with the results in Li et al (2016) and Yang et al (2016). Moreover, it is not surprising to see that the DCFM (second component) tends to require more parameters compared to the Poisson common factor model (except for England and Wales and the fitting period 1970-2005).…”

“…This norm of coherence has been widely adopted in the literature. There are a range of extensions and applications of the common factor model, in terms of handling both sexes (Li, 2013;Li et al, 2016;Parr et al, 2016;Yang et al, 2016;Pitt et al, 2018;Wong et al, 2020), two hedging counterparties (Li and Hardy, 2011;Li and Luo, 2012;Li and Haberman, 2015;Lin and Tsai, 2016;Zhou and Li, 2017) or a few countries in the same geographic region (Biffis et al, 2017;Enchev et al, 2017;Chen and Millossovich, 2018) jointly.…”

The mechanism between mortality, population growth and ageing of the population in the European lower and upper middle income countries

“…But, on the other side, it decreases the proportion of the population who are old, measured by prospective concept and the population's prospective median age. The prospective continuation of increases in longevity and population ageing, and the related challenges with the budget, particularly in relation to long-term care, health care, and pensions, have become serious concerns for governments in more developed countries [22]. Ka ´čerova ´et al [4].…”

Multipopulation mortality modelling and forecasting: The multivariate functional principal component with time weightings approaches

“…Such coherent forecast models are also documented in the literature, see, for example, the earliest augmented common factor (ACF) model proposed by Li & Lee (2005), which is an extension of the Lee-Carter model with an additional common factor to capture both short-term divergence and long-term coherence among related populations. Variants and extensions of the ACF model have been subsequently developed, such as Li (2013), Li et al (2016) and Chen & Millossovich (2018). Some others like the Age-Period-Cohort (APC) model proposed by Cairns et al (2011), incorporate a mean-reverting stochastic process for two related populations and allow for different trends in mortality improvement rates in the short-run but parallel improvements in the long-run.…”