Higher Education Students’ Choices Forecasts in Portugal: Insights to Evaluate the Impact of Rural Depopulation
Lidia Patricia Tome, CIDEHUS.UE, University of Évora
Filipe Ribeiro, CIDEHUS.UE, University of Évora
Maria Filomena Mendes, CIDEHUS.UE, University of Évora
Rita Freitas, CIDEHUS.UE, University of Évora
The educational network may change dramatically and previous patterns of higher education candidates will certainly be affected. Higher education institutions need to reorganize their supply accordingly with students’ demands but also with the number of applicants they can reach. In addition, national regions need to adapt demographically to the changes resulting from the displacement of students in higher education, resulting in aged interior regions and even more depopulated than at the present moment.
We evaluate the impact of demographic changes in the Portuguese higher education network employing a combination of the cohort component projection model and coherent forecasting approaches for mortality, fertility and migration, but also to forecast population based on work made by Hyndman and Booth (2008) and Hyndman et al. (2013). Additionally, we make use of Oeppen’s (2008) compositional approach to forecast students’ first higher education choices.
Independently of the university location and fields of study, we expect that universities located at large cities and closer to the coastline reveal to be more attractive and with time, lower and lower applications completed. The obtained results will not only contribute for a better understanding of students’ choices but also to evaluate the necessity of higher education network reconfiguration.
In order to accomplish our goal, we use information on the students’ choices over last recent years (2008 - 2014), including their universities/fields of study options and specific courses. The study horizon will be 2031.
The existing empirical evidence on highereducation student’s choices and the implication on the depopulation of lessurban areas are scarce and focused on a specific region of a country (e.g.Lovén et al. 2017, Ahlin et al., 2014 and Andersson et al., 2009).
If rural regions are traditionallycharacterized by decreasing and aged population, low level of human capital andslow economic grow, combined with the impact of extremely low fertility andnegative migratory flows will result in a decline in the resident population.
Considering the total concentration ofpopulation at the coastline in Portugal, it’s important to compare all regionsand national districts. In 2015 the Portuguese ageing index register thehighest values of 144 elderly per 100 under age 15 (INE - Statistics Portugal),and approximately 80% of the population at littoral municipalities.
The student’s temporary migrationto regions other than their origin, that later can be transformed on permanentmigration, are expected to affect not only the human capital but also the labormarket transmutation of already depopulated and aged regions. If between 1940and 2011 Portuguese population census registered an increasing population,after that long period of positive population growth, current populationprojections indicate an accentuated declining pattern.
The main goal of this study is, thus, toforecast population demographic changes, evaluate its impact on highereducation applications, and to contribute for a better understanding ofPortuguese higher education students’ choices employing a Compositional DataAnalysis (CoDa) approach proposed by Oeppen in 2008 to forecast mortalitypatterns and provide essential insights for decision-makers.
Data and Methods
Data used on this study comes fromhistorical data available on the Human Mortality Database, on the StatisticsPortugal and from the Portuguese Higher Education Organization (Direção-Geraldo Ensino Superior).
Concerning the methodological approach, firstwe not only forecast mortality, fertility and migration trends but also to forecastpopulation, based on work done by Hyndman and Booth (2008) and Hyndman etal. (2013). This approach is a combination of the cohort componentprojection model and historical information about demographic subpopulationtrends, resulting in what can be designed as “coherent” forecasts. Second, weemploy a CoDa model to forecast students’ first higher education choices(Oeppen, 2008).
In 2011, the last census year in Portugal,the resident population in the country was 10 562 178 inhabitants, and it isexpected that in 2021, ignoring possible migratory movements, this figure couldbe 10 064 219 (CI95%: 9 991 349, 10 137 292). And if consider on theequation the strong negative migration impact from between 2012 and 2015, thisvalue will tend to decrease even more sharply to 10 043 008, thus presenting alarger range of variation (CI95%: 9 702 069, 10 393 018).
We identified that, despite some ficklenessover the years, the 18 years-old subpopulation began to decline even before 2011.At that age, the number of inhabitants was 114 291, and we estimate 103 547 (CI95%:103 522; 103 569) in 2021.
Although no major changes are expected inregard to the interregional flows of applications estimated for 2021, comparedto previous years, a careful observation of Figure 1 allows us to predict that,once again, it will be the Lisbon and Porto districts (themore populated ones) that most candidates register for higher education.
Figure 1: Distributionof applications (%) by district of origin in 2008, 2015 e 2021.
Although Lisbon and Porto districtscontinue to be the most likely to contribute to higher education applications(Figure 2), the truth is that our forecast points to a slight decline between2015 and 2021, 44.93% to 42.65% in this period. And an additional highlight tothe district of Setúbal (satellite district of Lisbon), which had long sincedeclined to the level of candidates for higher education between 2008 and 2015(Figure 1), is now expected to be able to "recover" and increase thenumber of candidates coming from there.
Figure 2: Prediction of interregional flowsof applications for the year 2021.
The decrease of young population inPortugal will affect the number of candidates for higher education, since eventhe proportion of young candidates increases among those who are in a positionto do so, fewer are the young population. Such condition has not only an impacton the number of candidates for higher education, but also on theirdistribution, since the most attractive institutions will concentrate themajority of students, regardless of their type of education, leaving asidethose that with less applications and traditionally at the less developedcountry areas. Take, for example, the case of the districts of Lisbon, Portoand Coimbra, which are those that in the present engage more attraction poweron the candidates, and the most likely to continue to do so in the future,increasing the depopulation effect.
Presented in Session 1235: Posters