An increasing life expectancy, as well as a decrease in the number of children born, resulted in an aging population. In the same period, the educational level of the population increased. These changes influenced pensions, health care, housing, employment, culture, economic growth, etc. Furthermore, for policy it is relevant to know the impact of these changes on future demographics.

A population projection describes the future population by sex, age and (sometimes) migration background. The projection is based on assumptions about fertility, migration and mortality. However, these components are influenced by education. For instance, the more educated generally have a higher life expectancy. Omitting this influence will lead to substantively different results.

For this research the Dutch population, differentiated by education, will be projected into the future (2014-2060). Specific attention is paid to assumptions about mortality patterns by educational attainment. For this, an adapted Li-Lee model is used. Because data on educational attainment for the entire Dutch population is not available, data from a new database on educational attainment is linked to register data. The new database contains 60 percent of the Dutch population and is more detailed and qualitatively better than information from a survey.

Our results show that the future population of the Netherlands will increase more in the presence of education variables than in their absence, but the age distribution is similar. More important is the educational profile of the population: in 2060 the higher educated will represent the largest share of the population. Consequently, the behaviour of the future population will primarily reflect the behaviour of the higher educated. Including education in a population projection results not only in a more accurate population forecast, but produces also more information of intrinsic and practical interest.