The Role of the Population Projections for a Redefinition of the Portuguese Higher Education Institutional Network

Abstract

Population projections can be used as a tool to provide information on possible scenarios of future population and, namely, to support decision-making processes in diverse socio-economic areas, such as, higher education institutional network planning, both in public and private sectors. The dimension and the age and sex composition of future populations are influenced by mortality, fertility and migration trends. So, an accurate estimation of those future trends is crucial to evaluate how many inhabitants we will face in the future and simultaneously to prepare ourselves for their future needs. In a country like Portugal, affected by a severe economical and financial crisis, with a young population characterized by very low levels of education and qualification is fundamental to use population projections as a basis for higher education planning. The main goal of this paper is to evaluate the possible changes in the younger population size in the coming years as a tool to (re)define and (re)design, geographically the higher education institutional network in Portugal. For this purpose, we used the cohort-component method to project the Portuguese population from 2011 to 2036. For the evaluation of mortality future evolution we tested the performance, for the Portuguese case, of both the Lee-Carter (LC) method (1992) and the Booth-Maindonald-Smith (BMS) variant (2002), using data from the Human Mortality Database .Regarding the fertility projection, we used data from the Human 1 Fertility Database and applied the method proposed by Schmertmann (2003 & 2005), to model fertility rates by age. The complexity of migration flows, especially regarding its instability and the difficulties in addressing new forms of population mobility, supported the decision of include only a null migration scenario on this exercise. Considering the relevance of the projection of the number (and sex composition) of the under 18 population, we centered the discussion on the impact of different estimates of the future mortality rates for the youngest. Alongside with the main results, we will focus also on the analysis of the outcomes of LC and BMS models, performing a sensitivity analysis. We will sustain the reasons to choose one of those models as well as the use of confidence intervals to design alternative scenarios. Our findings will provide a range of reliable forecasts to support a more rational political decision contributing to an efficient and effective planning in what concerns higher education requirements adjusted to the future population.