CEFAGE-UE, Universidade de Évora, Palácio do Vimioso, Lg. Marquês de Marialva, 8, 7000-809 Évora, Portugal Telf: +351 266 706 581 - E-mail: [email protected]- Web: www.cefage.uevora.pt CEFAGE-UE Working Paper 2011/13 Higher Education ‘Market’ in Portugal: a diagnosis Conceição Rego 1, António Caleiro 2 1 Departamento de Economia & CEFAGE-UE, Universidade de Évora 2 Departamento de Economia & CEFAGE-UE, Universidade de Évora
25
Embed
Higher Education ‘Market’ in Portugal: a diagnosis · Marquês de Marialva, 8, 7000-809 Évora, Portugal Telf: +351 266 706 581 - E-mail: cefage@uevora ... HEI’s, we intend
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
CEFAGE-UE, Universidade de Évora, Palácio do Vimioso, Lg. Marquês de Marialva, 8, 7000-809 Évora, Portugal Telf: +351 266 706 581 - E-mail: [email protected] - Web: www.cefage.uevora.pt
CEFAGE-UE Working Paper2011/13
Higher Education ‘Market’ in Portugal: a diagnosis
Conceição Rego 1, António Caleiro 2
1 Departamento de Economia & CEFAGE-UE, Universidade de Évora
2 Departamento de Economia & CEFAGE-UE, Universidade de Évora
When proceeding as describe above, we find a strong dominance of higher education
provision concentrated on the coast (about 79% of places available in the academic
year 2008/09). Public and private sectors behave, according to this criterion, very
differently: while in the inner‐land the places predominate in the public sub‐system
(78%), on the coast the distribution between public and private is quite even (the
public system only holds 51% of supply). Naturally, these features can only be linked
to how the population is distributed throughout the country, particularly the young
population in the age group entering higher education, but also the working
population, as we also know that this part of the population is mainly settled in
coastal regions north of the Peninsula of Setúbal.
3.b) The 'demand' for higher education
In the academic year 2008/09 tertiary education in Portugal had attained about
373000 students, which represents a value 4.5 times higher than the one recorded in
the late '70s and means 17% of the global set of Portuguese students (figure 2). This
reflects the strong change that higher education registered in the past 30 years, rising
to reach a wider pool of students, within which differences in terms of economic,
social and geographic origin started to emerge.
Figure 2: Number of students enrolled in higher education in Portugal
Figure 3: Evolution of the rate of growth in the number of students enrolled in higher education in Portugal
10
Source: Pordata.
Despite the systematic increase in the number of students enrolled in higher
education, we can’t fail to notice that the rate at which these developments took
place was highly irregular (Figure 3). Two distinct periods can be identified: first, until
1991, the rate at which students in higher education was growing increased; since
then the number of students grew at decreasing rates, negative rates having been
recorded, as a matter of fact, in 2004 and 2007. This behavior depends fundamentally
on the 'demand determinants', i.e., the evolution of the age population attending
higher education, and the evolution of the number of students able to access this
education system. Indeed, the positive developments that occurred during the 80's
and early 1990 reflect the extension of schooling that allowed a greater number of
young people to access higher education and the positive behavior of the population
in the 'natural' age group (15‐19 years) of applicants to higher education (in this
respect see Table 1 in Annex). The opposite behavior, seen in the following years,
besides reflecting the behavior of demography, also results from the stabilization of
the proportion of young people completing secondary education (see figures in
Annex) and continuing their studies in higher education. In this area, the data revealed
by the OECD Education at a Glance document, published in 2009, confirm that in
Portugal since 1990 the rate of completion of secondary education is of about 70%,
making the country one of the OECD members most poorly ranked on this indicator
and far from Germany , for example, where all students complete secondary
education. Furthermore, the proportion of young people who access higher
education, according to the OECD, does not exceed 64% on average (57% for male
11
and 72% for female). This factor has been, obviously, conditioning demand in higher
education.
Students who enrolled in higher education in Portugal do it mainly in public
institutions (Figure 4), whose evolution follow and reflect the behavior of students in
higher education. The private higher education institutions experienced a more
significant momentum during 1990, when this subsystem swiftly adapted to the
changing demand.
Figure 4: Evolution of students in higher education in Portugal, for public and private subsystem
Figure 5: Regional structure for students enrolled in higher education in Portugal in 2008/2009 academic
year
0
50000
100000
150000
200000
250000
300000
350000
400000
450000
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Public Private Total
Source: Figure 4: Pordata; Figure 5: GPEARI.
Carrying on with the analysis in order to verify the geographic location of the
Portuguese higher education students, we have found (Figure 5) that, as recorded
with 'supply' the majority of students enrolled is concentrated in the Lisbon area and
soon , in the north (this behavior is identical to the variable on the first students
enrolled in higher education).
3.c) The higher education market
Before proceeding to data analysis, we propose a simple comparison between some
of the key variables to characterize the 'market' for higher education. Thus, by
12
comparing (figure 6), at NUTS II level, the places available (variable supply) with
students placed in the 1st year for the first time (variable demand), we conclude that,
with the exception of the North (where supply exceeds demand), there is a balance
between supply and demand in the 'market' for higher education. This is somehow
natural if we assume that demand moves freely to locations where it is known to be
available supply.
Figure 6: Comparison of some structural variables for the behavior of the 'market' for higher
education in Portugal
Legend: A: VacancHEI’s available for NUTS II (% of national total) B: Students placed for the 1st time in 1st year by NUTS II (% of national total) C: Residents in the age group of 15‐19 years, by NUTS II in 2008 (% of national total)
Source: GPEARI
In this study we intend to address the issue of the location of higher education, in
terms of territories and of the 'balance' in the whole country. In our comparison we
added the proportion of population in the age group to which candidates to
‘traditional’ higher education belong. Comparing the three variables in the analysis we
noticed the existence of two regions where the 'divergence' between the variable on
population and the variables on higher education is more manifest: the highest
proportion of population can be found in the North and in Lisbon the variables on
higher education exceed the population of the age group to which the candidates to
higher education belong. At first glance, and assuming that throughout the country,
i.e. for all NUTS II considered, the proportion of young people seeking higher
education does not display significant differences, we would be led to conclude from
13
figure 6, that, except for the Center, all other NUTS II and particularly the North, there
is a demand that apparently is attracted to the Lisbon region. However, a more careful
analysis of the data, which will be done in the next section, allows us to verify that it is
not the case.
For the time being, let us stick to establishing two hypothetical scenarios for the
determination of a (possible) equilibrium in the 'market' for higher education: first, we
consider that distance does not matter, in other words that for the applicants to
higher education distance between residence and the location of the HEI chosen for
their higher education is absolutely irrelevant at the moment of decision, and second,
we try to test the opposite hypothesis, i.e., only the distance matters in the decision,
meaning that candidates to higher education select schools located closest to the
residence of the household.
In the first case, assuming that candidates to higher education have no "sensitivity" to
distance, what matters is whether the number of places available in the various HEI’S
located on the national territory is sufficient to welcome all who those who apply. For
example, from data available through GPEARI we found that in 2009, as far as public
higher education is concerned, there were approximately 52,552 applications (first
choice) but only 45,295 students were placed. This imbalance in the "component" of
public higher education market opens the way for a private supply response to fill this
gap. In this reasoning we assume that all students interested in attending higher
education are trying to find a place in the public sub‐system, which is obviously not
true. Factors of various natures, such as proximity or economic and financial
availability, may lead students to look for a more pragmatic approach in order to
ensure entry into a higher education institution, which makes that only a few students
apply for private establishments as first choice.
Figure 7: Comparison between the proportions of population in age group 15‐19 years with the students enrolled in 1st year, for the 1st time, by NUTS III
14
Source: INE and GEAPRI.
The opposite case which we intend to verify concerns the hypothesis of distance be
the determining variable in the selection process of the school. In this case, we chose
to use as a geographic reference unit NUTS III. We conducted our comparison, within
the various NUTS III, between the proportion of individuals in the age group of 15‐19
years [weighted by the rate of further education to higher education (64%)] with the
proportion of students enrolled in higher education, in the 1st year for the first time in
institutions located in those territories. ur intention, here, is to verify the existence (or
not) of a balance between supply and demand in the various NUTS III (see figure 7).
Despite the fact that this analysis is still at a rough stage (particularly in that we use
the age group 17‐19 years, where he can truly leave the candidates to higher
education, unlike what we consider to use the group 15 ‐ 19 years) it already allows us
to confirm the ability of major universities in Lisbon, Porto and Coimbra to attract
students from other regions. In addition, we also found, without surprise, that where
the Universities of Évora (Alentejo Central NUTS), Beira Interior (Cova da Beira NUTS)
and Minho (Cávado NUTS) holds the attraction of students from other regions to
these regions.
15
In this approach we limited ourselves to consider as 'decisive' the demand for higher
education that comes from young people insecondary schools. However, higher
education today is being sought (and attended) by growing proportion of students in
working age, employees in general, who already have some professional experience,
and who go back to the first degree of higher education as a way to improve their
training and skills with the purpose of upgrading their professional qualification and
get progression in their careers.
4. The perspective of the multidimensional scaling analysis
4.1. Methodological considerations
Multidimensional scaling (MS) is a technique that allows performing the inverse
operation when determining the distance between points represented as coordinates
in a figure, such as a map (Kruskal & Wish, 1978; Cox & Cox, 1994; Borg & Groenen,
2005).
For example, it is obviously easy to calculate the distances between the municipalities
head towns, measured in terms of kilometers needed to travel from one place to
another. It is also possible to estimate, although results are more questionable, for
example, the distance in time that is needed to travel those routes. To put it simple,
what multidimensional scaling does is the inverse operation, i.e. to determine a
graphical representation (usually in two dimensions) that plausibly has generated
those distances, given that these may be different in nature, as the above examples
show. Thus, when the structure underlying the data generation process is complex,
MS thus provides a useful figure representation of data, insofar as, for example, allows
viewing, in the first place, how far/dissimilar are the "objects" and, secondly, to verify
which of them are similar, despite being relatively far apart, according to some
concept of distance (e.g., geographic).
16
In the application that follows the districts of mainland Portugal and the autonomous
regions of Azores and Madeira will be considered as the territorial unit. In general, the
first phase of MS consists on obtaining the di/similarity data, which is built on a
particular concept of distance (e.g. Euclidean) between "objects". Using this
information on di/similarity, in a second phase a solution (configuration) is obtained,
which is the location of objects in the space of a few dimensions (usually two or at
most three), where the distances between points in space are as close as possible as
the di/similarities between data.
4.2. Application of the methodology
Taking into account our objectives, we considered the data corresponding to the
national appliance to public higher education (in 2009) in terms of home versus
placement (for districts and autonomous regions) of the students that have applied to
the numerous public universities in Portugal.
Given the obvious differences in the population levels of the 20 geographical units in
question, we considered the data regarding the percentage of students (by
geographical unit of provenience).1 Figure 8 shows these rates.
Figure 8: The placement rates for the place of origin.
1 The data source is the Direcção Geral do Ensino Superior.
17
Figure 8 shows a pattern that clearly points to a preponderance of the district of origin
itself. Thus, retention rates ‐ visible in the main diagonal of Figure 8 ‐ show that only
the districts of Guarda, Santarém and Setúbal, most particularly the first one ‐ have
not displayed the highest rate of student placement in the district itself. Coimbra and
Lisbon are quite remarkable in this respect, with retention rates of about 79% while
attracting many students from multiple geographical origins.
The pattern shown in Figure 8 suggests that geographical distance is of crucial
importance. If true, one should try to understand the extent to which districts and
autonomous regions are distant in terms of student placement in public HEI’s. The
distances between the placement rates2 obtained by applying multidimensional
scaling are shown in Figure 9.3
2 The calculation of distances was based on a freely available add‐in for Excel http://www.cse.csiro.au/poptools/index.htm, while the methodology was applied using a Matlab routine. The multidimensional scaling technique could have been applied using the macro OOs Statistics (David Hitchcock) for the free OpenOffice program, available at http://sourceforge.net/project/showfiles.php?group_id=87718&package_id=106652, or the routine for the program R free, available at http://www.r‐project.org/ (see, for example, http://www.statmethods.net/advstats/mds.html) or yet). The use of this technique to the level of another free program, PSPSS, style SPSS, it is also possible (see http://www.gnu.org/software/pspp/). 3 It should be noted that the need for a convergence in the 'objective function' led to a decrease in the critical level of convergence.
18
Figure 9: The results of multidimensional scaling
In these results major universities, including Coimbra, Lisbon and Oporto, stand out ,
each of them in a particular way, by being located at the edge of in distinct quadrants.
5. Concluding remarks
This paper aimed at understanding the behavior of the higher education 'market'
education in Portugal. In it we have confirmed that HEI’s are scattered throughout the
territory, despite the pre‐eminence of the 'coastline', which concentrates the majority
of the population and of the economic activity. Consequently it is also the region
where the supply of higher education is concentrated.
In general, institutions located in the different regions examined here exert significant
attraction and influence on candidates living in the same region.
In examining the relationship between supply and demand for higher education, we
found that big universities such as those of Lisbon, Coimbra and Oporto exert
considerable attraction not only on young people living there but also on those living
19
elsewhere in the country. Likewise, more recently founded university centers, located
in medium‐sized cities such as Braga, Covilhã, and Évora, show some ability to attract
young students.
In summary, we found that, first, higher education students in Portugal are not
indifferent to distance when choosing the place to attend higher education and,
second, large and medium sized university centers also exert some attraction on
students from other locations. This is an expected outcome of the 'unbalances'
between supply and demand for higher education throughout the country.
References
BEESON, R.J.; & MONTGOMERY, E. (1993), “The effect of college and universities on
local labour markets”, Review of Economics & Statistics, 75 (4), 735‐761
BLEANEY, M.; BINKS, M.; GREENWAY, D.; REED, G.; & WHYNES, D. (1992), “What
does university add to its local economy?”, Applied Economics, 24, 305‐311.
BORG, Ingwer; e GROENEN, Patrick J.F. (2005), Modern Multidimensional Scaling:
Theory and Applications, 2.ª edição, New York: Springer.
BROWN R. H.; HEANEY, M.T. (1997), “A note on measuring the economic impact of
institutions of higher education”, Research in Higher Education, 38 (2), 229‐240.
CALEIRO, António (2010), "Como se pode distinguir Évora do resto do Alentejo?: Uma
abordagem de estatística espacial", MPRA Paper 22057, University Library of Munich,