The supplemental archive contains one R file. The file presents the functions and data allowing the replicability of the analysis of the article.
The data can directly be extracted from the Human Mortality Database (HMD) using the R program. 


The steps to replicate the results are :

1) Open the R file,
2) Install and open the libraries listed at the beginning of the R file (forecast, demography, compositions, MortalitySmooth, ggplot2),
3) To have access to the data, the user has to be registered on the Human Mortality Database (HMD) website (mortality.org),
4) Enter your HMD username and password at the beginning of the R code (step 2 in "Start" section),
5) Select the countries to be forecast in countries<-c() using the "demography" library country code, e.g. Denmark=DNK (step 3 in "Start" section),
6) Run the rest of the code.


Information for further analysis:

1) The code output is for females only. To obtain the males results, change load.data[[i]]$pop$female for load.data[[i]]$pop$male; and load.data[[i]]$rate$female for load.data[[i]]$rate$male in the "Load data"section.
2) In the "Information" section, the time indexes for the CoDa and CoDa-coherent models are set to be forecast with an Arima model of order (0,1,1) with drift and (1,1,0) without drift, respectively, as mentioned in the paper. The order can be changed.
3) In the "Information" section, to reduce the time requires to run the functions, the number of simulation for each procedure defined in Appendix B of the paper are set at 10. For more reliable prediction intervals, this number should be increased.