Comparing strategies for matching mortality forecasts to the most recently observed data: exploring the trade-off between accuracy and robustness

Genus

Journal of Population Sciences

Table 1 Mean absolute forecast error (MAFE) of remaining life expectancy at age 65 for the first and fifth year in the forecasting period, for six different choices of the jump-off rates applied to a Lee-Carter model, for eight Western European countries, men and women combined, fitting periods 1960–2005, 1960–2006, …, 1960–2014

Jump-off rates	FRA	ESP	SWE	BEL	UK	NOR	FIN	NLD	Av¹
Mean absolute forecast error in the first year of the forecasting period
Model values	0.16	0.16	0.24	0.10	0.56	0.24	0.41	0.39	0.28
Last observed values	0.13	0.17	0.08	0.14	0.13	0.12	0.10	0.13	0.12
Average 2 years observed	0.13	0.15	0.07	0.15	0.16	0.11	0.08	0.16	0.13
Average 3 years observed	0.13	0.16	0.09	0.14	0.19	0.12	0.09	0.21	0.14
Average 4 years observed	0.14	0.17	0.10	0.15	0.24	0.13	0.12	0.26	0.16
Average 5 years observed	0.16	0.19	0.11	0.17	0.28	0.16	0.16	0.32	0.19
Mean absolute forecast error in the fifth year of the forecasting period
Model values	0.20	0.17	0.31	0.18	0.76	0.45	0.41	0.55	0.38
Last observed values	0.16	0.30	0.10	0.17	0.38	0.20	0.13	0.31	0.22
Average 2 years observed	0.17	0.32	0.11	0.18	0.41	0.21	0.14	0.35	0.24
Average 3 years observed	0.18	0.35	0.13	0.19	0.44	0.24	0.15	0.41	0.26
Average 4 years observed	0.20	0.38	0.16	0.20	0.46	0.29	0.18	0.47	0.29
Average 5 years observed	0.21	0.40	0.18	0.21	0.49	0.34	0.19	0.55	0.32