This is an excellent chapter by two of the best. In a field where expedient compromise is the rule rather than the exception, for some time now, Lee and Tuljapurkar have been doing the most thoughtful basic work on population forecasting. While forecasting practitioners in government agencies are sometimes constrained by what are viewed as politically acceptable ranges, the more basic constraint is that they rarely have the time or resources to conduct inquiries challenging the theoretical and statistical pillars upon which population forecasting rests. That job has been given to the likes of Ron Lee and Shripad Tuljapurkar, and it is in good hands indeed.

Demographic forecasting is an extremely useful analytical tool. Instead of relying on summary aggregate statistics, such as the total size or growth rate of the population, the value of the demographic approach is that it builds up its projections from demographically disaggregated vital rates. We know with reasonable certainty that the size of the population will not continue to grow (as aggregate growth rates might suggest), but rather will soon begin to decline and in fact will do so in a narrowly specified time span. We also know a good deal about the future age and sex structure of populations far into the future. The “no change status quo” scenario built up from disaggregated vital rates is often quite at odds with projections based on population aggregates.