[PF1]Point Forecasts

The method used to compute the forecasts of each country was based on conventional cohort-component book-keeping:

(i) A starting, or jump-off population was established.

(ii) Survival probabilities were calculated by sex for each forecast year t = 2003, 2004,..., 2049, for each age x = 0, 1,..., 99, to survive the cohorts to the next age. These were based on "projective" age-specific mortality rates computed annually from the parallelograms of the Lexis diagram. Survival from birth to age 0 was based on direct estimates of average survival probability. Within age-group 100+ the fraction of survivors was inferred from the rates of ages 98 and 99.

(iii) Births were generated for years t = 2003, 2004,..., 2049, using age-specific fertility rates for women in ages x = 15, 16,..., 49.

(iv) Migration during years t = 2003, 2004,..., 2049 was handled in terms of additive net migration numbers that already incorporate the effect of mortality within the year of entry or exit.

For each step, a point forecast, i.e., the most likely value, was established for the variables of interest.

Jump-off Population

Population as of January 1, 2003 was taken from the New Cronos database, at the end of May, 2004. For Denmark, Finland, Iceland, the Netherlands, Norway, and Sweden the figures come from a population register. For Austria, Belgium, Italy, Luxembourg, Spain and Switzerland the estimates were based on the 2000 (conventional) census, and updated by data from the population register. Germany currently relies on the 1987 census and population register updates. France, Greece, Ireland, Portugal and the U.K. use the census of 2000 and updates from vital registration and survey data. Special estimates for the oldest ages had to be made for Austria (ages 95+), Luxembourg (ages 95+), Portugal (ages 85+) and the U.K. (ages 90+).

More recent population values were available for several countries. These were not used in the current exercise for technical reasons. Incorporating them would have complicated the uncertainty analysis of EU/EEA jointly.


A jump-off value of age- and sex-specific mortality was established by smoothing the observed values of years 1998-2002 and adjusting for increase during the period to match the level of 2002. For all countries, rates in ages 95+ were imputed using information in younger ages. For Germany, Greece, Portugal and the U.K. additional special estimations were made to establish starting values for the highest ages.

The point forecast for age- and sex-specific mortality was calculated by starting from the jump-off value and applying an age-specific rate of decline during years t = 2002-2003, 2003-2004,...., 2048-2049, to the value obtained until then. A country-specific initial rate of decline was estimated and a linear change towards an eventual rate of decline was assumed to occur by year 2030.

The initial rate of decline was empirically estimated from years 1993-2002. The values were constrained to be non-negative and smoothed separately for males and females before use. In some cases extremely high rates of improvement were observed in the highest ages. These were thought implausibly high for forecasting purposes, and judgmentally reduced to the level of immediately younger ages.

The eventual rate of decline was empirically estimated from the data of Austria, Denmark, Finland, France, Germany, Italy, the Netherlands, Norway, Sweden, Switzerland and the U.K., during the latest 30-year period ending between 1997-2001, for which the data were available. The estimates were smoothed separately for males and females. The same eventual rate was used for all countries.


Recorded age-specific fertility in ages x = 15, 16,..., 49, in t = 2002, was used to provide jump-off values. The rates were smoothed before use. For some countries for missing values were obtained by using linear interpolation and extrapolation.

Ultimate value for the total fertility rate was specified for groups of countries, and the total fertility rates of the intermediate forecast years were obtained by linear interpolation.

Mean age at childbearing (defined as mean age of age-specific fertility schedule) was assumed to increase from the value at t = 2002, to age 31.0, to be reached by year 2017. A log-linear model was used to implement the change.

Net Migration

Net migration was defined as the difference in reported population growth and natural increase. (This may differ from estimates obtained by subtracting estimates of out-migration from estimates of in-migration.)

Jump-off values at t = 2002 were judgmentally specified based on the most recent officially reported values, levels estimated from time-series models, and reported values in countries deemed similar with respect to migration. The jump-off value was assumed to persist for 10 years. Then, a linear change to the ultimate level was assumed. The ultimate level of net migration was specified based on time series models and judgment. For the year t = 2003 a correction was made to conform to the published estimates.

The age-structure of net migration was assumed to start from a national pattern estimated from data in 1990-2000, and to change linearly to an average pattern, estimated from Austria, Belgium, Denmark, Iceland, Netherlands, Norway, Sweden and Switzerland, after ten years and then held constant for the rest of the forecast period. For Greece the average pattern was used from the start.

Last updated 24.9.2004