Demographic Forecasting

The authors have given an excellent statement of the current state of demographic forecasting. The main points of their paper are as follows:

• A key element in demographic forecasts is projection of fertility rates. Volatility in rates, partly induced by timing choices, makes fertility projections uncertain. Over a horizon of two or three generations, this has a major impact on the reliability of projections of total population and age distribution.

• Another key element in demographic forecasts is projection of survival curves. Life expectancies are projected to continue to rise, but it is uncertain whether trends over the past century will be sustained.

• Current Census projection methods are flawed because (1) too much attention is paid to recent fertility and mortality rates, (2) current fertility projections appear to be too high, (3) current mortality rate projections appear to be too high, and (4) the high/medium/low scenarios do a poor job of representing uncertainty in the forecasts.

• The elderly dependency ratio (65+ population/20–64 population) will peak around 41% to 45% in the decade 2030 to 2040, depending on the forecast. (Census projections made in 1966 were more optimistic, predicting a maximum elderly dependency ratio around 35% to 36%.)

Forecasting Methodology and Stochastic Simulation

The engine that drives most demographic forecasts is a cohortcomponent accounting relationship (see Figure 2-1.1).