Forecasting Human Capital: Using Demographic Multi-State Projection Methods to Show the Long-Term Effects of near-Term Investments in Education

Wolfgang Lutz, International Institute for Applied Systems Analysis (IIASA)
Anne Goujon, Vienna Institute of Demography
Annababette Wils, Academy for Educational Development (AED)

This paper argues that forecasting human capital is important for several reasons and that the most appropriate methods for doing so are demographic multi-state population projections. Under this method the population of a country or region is cross-classified by age, sex and different categories (states) of educational attainment. The population is projected into the future based on assumed education-specific fertility, mortality and migration rates as well as age- and sex-specific transition rates from one educational status to another. This explicit consideration of such dynamic features makes this method more appropriate for the projections of human capital than other methods previously suggested. Such multi-state models can also apply to the detailed analysis of education flows, modeling intake and enrollment as well as completion rates. The paper demonstrates the feasibility of the method through three detailed education projection studies for Guinea, Zambia and Nicaragua.