Ces if o Dice Report 213

7/28/2019 Ces if o Dice Report 213

1/65


2/65

CESifo DICE ReportISSN 1612-0663 (print version)ISSN 1613-6373 (electronic version)A quarterly journal for institutional comparisonsPublisher and distributor: Ifo InstitutePoschingerstr. 5, D-81679 Munich, GermanyTelephone ++49 89 9224-0, Telefax ++49 89 9224-1462, e-mail [email protected] subscription rate: 50.00Editors: Marcus Drometer, Silke FriedrichEditor of this issue: Silke Friedrich ( [email protected] )Copy editing: Lisa Giani Contini, Sabine Rumscheidt, Andrea HohenadlReproduction permitted only if source is stated and copy is sent to the Ifo Institute.

DICE Database: www.cesifo-group.org/DICE


3/65

DICE ReportVolume 11, Number 2 Summer 2013

Forum

A ccess to H igHer e ducAtion

Globalization and Access to Higher Education Policy Implications Marcel Grard and Silke Uebelmesser 3

Increasing Access to Higher Education Through Student Loans Adrian Ziderman 11

The Bologna Process and Fairness in University Education: Evidence from Italy Paolo Brunori, Vito Peragine and Laura Serlenga 19

Access to Higher Education: The Shift Towards Market-based Policies in the UK Roger Brown 2 3

Equal Opportunity and Life-long Learning: The Future of Higher Education in Germany is Only Secure with Major Policy Changes

Jo Ritzen and Cecile Hoareau 28

Research Report

Economic Freedom, Money and Happiness Why Deregulation MattersBeyond its Wealth Enhancing Effect

Bodo Knoll, Hans Pitlik and Martin Rode 35

Reform Model

The In uence Wielded by Land Developer Lobbies During the Housing Boom: Recent Evidence From Spain

Albert Sol-Oll and Elisabet Viladecans-Marsal 4 3

Database

Regulation of the Wholesale Broadband Access Market 50Evolution of Oil Taxes: an International Perspective 53Government Defence Anti-Corruption Index 2013 55Energy Ef ciency in Buildings in EU Countries 57

News

New at DICE Database, Conferences, Books 60

Launch of Visual Storytelling in the DICE Database 61


4/653 CESifo DICE Report 2/2013 (June)

g lobAlizAtion And Access to HigHer e ducAtion policy implicAtions

mArcel g rArd 1 And

s ilke u ebelmesser 2

Introduction 3

The Bologna Process was launched in 1999. Its philos-ophy is well summarized in the Bologna Declaration,which states that A Europe of Knowledge is now widelyrecognized as an irreplaceable factor for social and hu-man growth and as an indispensable component to con-solidate and enrich the European citizenship, capableof giving its citizens the necessary competences to facethe challenges of the new millennium, together with anawareness of shared values and belonging to a commonsocial and cultural space (Bologna Declaration 1999).In other words, the main aim of the Bologna Processhas been to create a European Higher Education Area(EHEA) based on international cooperation and aca-demic exchange that is attractive to European and non-European students and staff.

This article focuses on one key element that has been ne-

glected in the Bologna Process: the nancing of higher education when students and graduates, i.e. (potential)tax-payers, are mobile. More precisely, it deals with thecentral issue of who will pay for the education of mobile students. This, in turn, raises the questions of: Whatshare of higher education costs should be publicly (viz.

privately) nanced and which jurisdiction(s) should betaken into account in the nancing process?

1 Catholic University of Louvain and CESifo.2 University of Jena and CESifo.3 This article is based on research partly conducted in the framework of Project IAP 6/09 of the Belspo agency of the Belgian FederalGovernment.

Imbalanced migration

Ultimately, the Bologna Process should lead to a singlemarket for high skilled labor. Whether this also leadsto spill-over effects or externalities and endangers theef ciency of the higher education system in the EuropeUnion (EU) depends on the extent of student and gradu-ate mobility; and even more on the link between them.

Do students who have graduated in one country embark upon their professional career, earn their wage incomeand pay their taxes in that country? Or do they jointhe workforce in their country of origin or in another country? There is empirical evidence that mobility in -creases with the skill level of workers (see, for example,Ehrenberg and Smith 1994, for the US, or Uebelmesser 2006, for Germany). This would suggest that universitygraduates are comparably mobile. Besides, the mobil-ity of graduates is incentivized by mobility before andduring tertiary education (see, for example, Dreher andPoutvaara 2011; Parey and Waldinger 2011; and most re-cently Voin and Grard 2013). This means that foreignstudents are fairly likely to move on after graduation.

In terms of the nancing of higher education, graduate,and thus tax-payers, mobility would not present any

problem if migration ows were balanced. As illustratedin Table 1, however, the mobility of students is imbal-anced. A negative sign means that the country is a netimporter of young people or raw human capital and anet exporter of graduates or enriched human capital if most intra-EU migrant students do not stay after their

studies.

Those countries that face a negative balance can be sep-arated into two different groups: one group comprisingAustria, Belgium, the Czech Republic, the Netherlandsand Denmark, and the other group featuring the UnitedKingdom.

The members of the rst group all have large neighbor -ing countries France for Belgium, Germany for theother countries that are quite selective regarding ad-missions to medical and paramedical studies. Studentswho fail at the entrance examination level in their own country, either France or Germany, enroll in the

Forum

Access to HigHer e ducAtion


5/654CESifo DICE Report 2/2013 (June)

equivalent programs of their small neighboring coun-tries where language is the same or similar, access tostudies is formally easier no entrance examination and tuition fees are generally low they do not exist inAustria, Denmark and the Czech Republic (in the CzechRepublic, a EUR 1,000 per term fee is charged to stu-dents following courses taught in a language other thanCzech), and amount to EUR 830 per year in Belgiumand up to about EUR 2,500 in the Netherlands.4 As aresult of EU laws, the nationality of the degree does not preclude practice in another country. So, one could con-clude that France and Germany free ride on their small

neighbors. The latter face a net in ow of students whoare not the best in their eld and cohort. 5

The other group comprises the UK alone. In that coun-try, the language is country speci c, but used in the en -tire world aslingua franca , access to studies is intel-lectually demanding admission is very selective andtuition fees are very high by European standards, reach-ing up to GBP 9,000 a year. The UK, therefore, can be

4See http://www.studyineurope.eu/tuition-fees.5 It is worth noting that Denmark is also a net importer of studentsfrom other Nordic Countries. We will come back to this case later. Wewould like to point out that the negative sign for the Czech Republicshould be seen together with the positive one observed for the Slovak Republic.

expected to face a different, rather able, group of students.

We can conclude from the stylized

facts presented above that the cross border ows of students are imbal -anced. In a decentralized worldlike the EU, where higher educa-tion is extensively nanced by thegovernment of the jurisdiction thathosts the students, and where tui-tion fees must not discriminate be-tween EU residents on the basis of their country of origin, imbalancedtransfers, and thus externalities,

undermine the ef cient function -ing of the Bologna process. This is per se an argument for allocatingthe responsibility of higher educa-tion to the level of the EU, some-thing that is already the case whenit comes to the de nition of cursusnorms, but which seems to be im-

possible in terms of nancing.

The focus in the following istherefore on a decentralized system of nancing higher

education, whose outcome comes closest to that of acentralized one. We will, in particular, address the ques-tion of whether higher education should be nanced viataxes or fees. A two-step procedure is called for to arriveat the optimal nancing regime with open borders. In arst step, the private and social bene ts and the corre -sponding share of costs which should be borne th roughfees or taxes have to be determined. In a second step,the speci c implementation of the regime has to be con -sidered. This involves the private part and the questionof whether tuition fees should be levied at the time of

studying or after graduation (as graduate taxes or in-come-contingent loans). This also concerns the public part and how it should be allocated between the countryof education and the country or countries where the so-cial bene ts arise. In fact, with migration, the nancingwould have to be split up among four parties: the gradu-ates, according to their private returns, the country of higher education (the host country), that of previouseducation (the origin country, possibly identical to thatof birth), and the working-country or countries, which

bene t from the improved skills of the worker. 6

6For a more detailed analysis, see Grard and Uebelmesser (2013).

Table 1

Imbalanced mobility of students in the European Union

Note: Foreign students: fraction of students coming from abroad. Balance of mobility: the number of outgoing students minus the number of incomingstudents, divided by the total number of students (Erasmus students excluded).

Source: Grard (2012) based on Eurostat and own calculation.

Country Foreign students (%) Balance of mobility (%)

Austria 11.36 -8.02Belgium 6.98 -4.62

United Kingdom 4.06 -3.63Czech Republic 5.21 -3.01

Netherlands 4.17 -2.41Denmark 2.70 -1.18

Sweden 2.03 0.11

Germany 2.61 0.26

Spain 0.75 0.30France 1.60 0.33

Hungary 1.20 0.36Italy 0.54 1.06

Finland 0.74 1.37Poland 0.11 1.43Portugal 0.68 2.50

Greece 0.15 4.06Ireland 1.92 7.47Slovak Republic 1.59 9.73

Luxembourg 37.00 232.70

Forum


6/65



2006) or the riskiness of the human capital investment,what would then be the optimal nancial regime? If thedistortions on the credit market are not too large, thereis still an argument for a positive fee level (albeit smaller

than in the absence of this market imperfection andcombined with a subsidy from the public sector) in or-der to induce the optimal number of students with theoptimal ability. A pure tax- nancing regime can onlyarise if the distortions are very large. So, distortions onthe credit market justify (additional) intervention by thegovernment via tax- nancing.

Secondly, among other limitations of the data, the public bene ts displayed in Table 2 only compriseof the additional tax revenues from individuals with

tertiary education relative to those with non-tertiaryeducation, as well as saved transfer payments. Thereare other positive public bene ts of a more educated population like those related to productivity gainsand thus to economic growth. Similarly, many oth-er positive (causal) effects of education have beenestablished, for example, a reduction in crime(Lochner and Moretti 2004) and an improvement in thehealth status (Webbink, Martin and Visscher 2010).

If these two points were to be relevant, the public expenditure share would increase. Importantly, however, itis very unlikely that, even in such a case, full tax- nanc -ing of higher education would be justi ed (on ef ciencygrounds). So, in a closed economy, we would expect thetax share in the nancing of higher education to roughlyre ect the share of public bene ts in total bene ts fromhigher education (Table 2). This would point towards amixed nancing system.

The question is whether these conclusions have to be adjusted in an open economy with mobile students and / or mobile graduates.

The speci c design in an open economy

Whatever the conclusions may be for a closed economy,it is obvious that a nancial regime that relies heavilyon taxes is not sustainable in the presence of signi cantmobility of graduates, i.e. high-skilled and high-wagetax-payers (Justman and Thisse 2000; Demange, Fengeand Uebelmesser 2013). With tax- nancing, studentsreceive their education free-of-cost at the time of study-ing, as it is nanced by tax-revenues from those workingat that time. This is based, however, on the implicit un-derstanding that in the subsequent period those former-

students-and-now-workers nance the next student gen -eration with their tax-payments. But this requires that asuf cient number of those or other graduates is presentin the country at that time. If migration is unbalanced

(see Table 1), a tax-based regime is only sustainable if there are transfer agreements between the countries af-fected, in particular between the host country providinghigher education and the working-country or countries

bene tting from the graduates increased productivity.If there are no transfer agreements, the sustainability of the public budget requires that the students bear a larger share of the costs involved.8 Otherwise, the host countrywill have an incentive to underprovide higher educationto foreign students, but also in the absence of any wayof discriminating to their own native students.

Sharing of educat ion cost s among the governments of

affected countries

We now discuss the cur rent system of nancing the high -er education of cross border students in the EU and in-vestigate alternative avenues (Grard 2007). Currently,in most EU member states except for England, higher education for native and foreign students from other EUcountries is mainly publicly funded by the local govern-ment of the host country, and thus by the tax-payers of that country (Table 2). Moreover the tuition fees are ei-ther zero or equal to a very moderate amount comparedto those charged in England or the US (or in the EU tonon-EU residents). These features characterize what isreferred to here as the Host Country Principle.

A rst alternative is the Origin Country Principle.According to that mechanism, the country of origin of a student, say, the country where she/he received sec-ondary education, is responsible for her/his higher edu-cation, irrespective of whether the latter takes placeat home or abroad. In both cases, it is up to the origin

country to pay and to decide on the number of studentsit sends abroad. From the perspective of the host coun-try, the incentives to underprovide higher education arereduced.

A system based on the Origin Country Principle can be implemented through the provision of vouchers bythe government of a given country to potential studentswho are residents of that country. Those vouchers may

be used for, say, one year of studies in a particular eldin an university located at home or abroad, provided it is8Of course, nothing is said here about the relevance of special schemes,scholarships, etc. deemed to alleviate the burden and guarantee equal-ity of chances independent from individual nancial resources. See below for further d iscussion.

Forum



recognized as an institution of high quality by the issuer of the voucher. Those vouchers should either cover theactual cost of studies, or a standardized benchmark cost.It is up to the issuer of the voucher to decide whether

the vouchers are distributed upon request, or are al-located through a competition; and whether they aremade available for free or subject to a present or future(re)payment. Vouchers could also nance the cost of liv -ing during the studies or be targeted at certain socio-economic groups, or may even be used as instruments toincentivize young people to study for jobs which are notvery attractive or poorly compensated, although sociallyhighly desirable.

Provided that other countries commit to not admit stu-

dents without a voucher, this device expands the geo-graphical area of sovereignty to the set of those partici- pating countries, for example, the Bologna Area or theEU. As an example, lets imagine that Germany limitsthe number of students admitted to the rst year of medicine in German universities in order to optimizethe supply of medical services in the future; those youngGermans who go to Austria and enroll in the schoolsof medicine there try to bypass the German numerusclausus at least in cases where they plan to return toGermany after they have obtained their MD degree. If we move to the Origin Country Principle, only thoseyoung Germans with a voucher issued by the Germanauthorities will be admitted to Austrian schools of medi-cine; thus the decision to admit those students for stud-ies in medicine, even outside Germany, is in the handsof the German authorities, who are then in a positionto expand the application of their numerus clausus rulesand thus their area of sovereignty. Similarly, Austria hasthe opportunity to sell the quality of its medical schoolswhile being in accordance with German public health policy.

Although the vouchers depicted above channel thetransfer implied by the Origin Country Principlethrough students, that transfer may alternatively bedirectly operated by governments. This is already thecase in Switzerland and in the Nordic Countries. InSwitzerland, cantons that do not have universities payfor the studies of their residents in universities locatedin other cantons.

Among Nordic Countries, a similar transfer system isat work. To illustrate the Nordic system, based on a four country agreement, it is worth noting that: the previousagreement was signed in 1996, and the new agreementwill be effective from 1 January 2013 for Denmark,

Finland, Norway and Sweden () The new agreementmeans that the yearly compensation per student for Denmark will be DKK 22,000 (USD 3,800) in 2013 the same as it was in 2012 rising to DKK 30,000

(USD 5,200) in 2014. The compensation will be regu-lated according to the consumption index calculatedeach year by Statistics Denmark. Under EU regulations,Denmark is obliged to treat citizens of the EU and theEuropean Economic Area the same as Danish citizens,which means that European students are entitled to freehigher education in Denmark. The compensation agree-ment has been concluded despite this, with the cost to be carried by governments rather than by individualstudents.9 This quotation seems to indicate that directtransfers between governments are a way of bypassing

EU legislation.

Implementing transfer systems either via vouchers tostudents or via inter-governmental transfers would al-leviate the problems related to open economies and tothe sustainability of public budgets. This would makeit possible to maintain the cost sharing and the corre-sponding tax-fee mix as derived above. As far as weknow, however, the inter-cantonal transfer systems atwork in Switzerland and in the Nordic Countries areamong the very few such mechanisms implemented sofar. This means that most countries do not have a com-

pensatory system. Globalization and increased ows of students and graduates then require a shift of the costshare to the students if the public budget is to be sus-tained, and if a race to the bottom in terms of educa-tional quality or, in general, an underprovision of higher education is to be avoided.

Shift towards fee- nancing

The main question is how a larger nancial contr ibution by the students should be implemented when graduates

are potentially mobile and the negative consequences of this shift towards fee- nancing for equality of chancesare to be avoided.

A new application of the Bhagwati Tax, proposed byBhagwati (1976) and again by Wilson (2008), is one pos-sible such mechanism. Those who have studied at theexpense of one country and currently work in another country have to compensate the former for the cost of their higher education. In practice, that may take theform of a transfer by the government of the latter coun-

9 See University World News, 03 November 2012, http://www.uni-versityworldnews.com /article.php?story=20121031163939447#.UJOYFr7pNiM.email.

Forum


9/65



cussed above that we would expect a positive associa-tion between the tax-fee mix of nancing higher educa -tion and the public-private mix of bene ts on the onehand, and a negative association between that tax-fee

mix and graduate (tax-payer) mobility on the other hand.Of course, there are many additional relevant factorsthat we do not take into account here. The correlations below are, therefore, only intended to highlight some basic relations, which can be observed in a cross-coun-try perspective.

Considering rst the relation between public bene tsfrom higher education and public expenditure on high-er education (both as shares of the respective totals),we nd a positive and highly signi cant association

(Figure 1). So, a larger share of public bene ts goes handin hand with a larger share of public expenditure, al-though to a lesser degree than one-to-one.11 In Germany,for example, the public bene t share is 49 percent andthe public expenditure share is almost twice as large at85 percent, while in France, the difference is even moreimportant with a public bene t share of 34 percent and a public expenditure share of 81 percent.

As countries are affected differently by student andgraduate mobility, the second point of interest is howgraduate mobility relates to the public share of total expenditure for higher education. The hypothesis is thatwith a high level of graduate out-mobility, a system of nancing higher education that relies mostly on taxes(and not on tuition fees) is not sustainable as a (net) out-

11Remember, however, that public bene ts do not include (positive) ex -ternalities and are, therefore, underestimated.

ow of graduates also reduces the number of (potential)tax-payers.

As data about graduate mobility are not available, we

make use of data about the net brain gain (Docquier andMarfouk 2005). Figure 2 shows that there is a negative,

but insigni cant relation. This result also holds if the Nordic Countries are excluded. Following our argu-ments above, one could have expected that, for thesecountries, the mobility of students and graduates shouldnot affect the choice of the nancing mix as much asfor the other countries given their inter-governmentaltransfer system.

A negative brain gain, i.e. a net out ow of highly ed -

ucated workers is not associated, on average, with alarger share of private expenditure on total expendi-ture for higher education in the sending (host) country.Countries with comparable net out ows of between 0.4and 0.6 percent have very different nancing regimeswith only the United Kingdom relying much on privatecontributions (public share of 35 percent), while Italy,the Czech Republic and Austria have large public sharesof, respectively, 70, 79 and 85 percent.

On the other hand, the importance of public expenditurein immigration countries is relatively modest. Australia,with a net brain gain of 11.4 percent, relies on private and

public nancing in a very balanced way (public share of 48 percent). The public contributions to the nancing of higher education are slightly larger for Canada with 59

percent and a net in ow of highlyeducated migrants of 10.7 percent,while the United States, with anet in ow of 5.4 percent, has thesmallest public expenditure shareof these three immigration coun-tries with 37 percent.

Policy conclusions

The analysis conducted in thisarticle might lead to the follow-ing policy conclusions. Given theexternalities related to the pub-lic provision of higher educationwith mobile students and gradu-ates, their internalization calls for a system as close as possible to acentralized one. The system shouldinclude cross border transfers

Figure 1

0

20

40

60

80

100

0 20 40 60 80 100

Relation of public benefits to public expenditure

Share of public benefits on total benefits in %

Source: OECD (2012), Tables B3.2b, A9.3 and A9.4.

Share of public expenditure on total expenditure in %

Note for the regression results: constant 0.197 (t-value 3.357), slope coefficient 0.233 (t-value 3.122).

Forum



aimed at compensating the country that nances higher education without sharing in the bene ts mostly thehost country. Otherwise, there are strong incentives for that country to underprovide higher education if it is(mostly) publicly nanced or to shift the nancing-mixtowards fee- nancing. A compensatory system might be implemented as a (comprehensive) Bhagwati Taxwhere the transfer is from government to government.If the compensation relies on transfers by the students,income-contingent loans might be considered.

To date, compensatory transfer systems have not beenimplemented in Europe with the notable exceptionof the Swiss system of inter-cantonal transfers and the

Nordic one. Neither have the national systems of nanc -ing higher education been adjusted to increased studentand graduate mobility.

It remains to be seen when the most concerned coun-tries will put these issues on their agenda. If they plan

to do so, this contribution is meant to provide someguidelines.

References

Barr, N. (2012), Financing Higher Education: Mobility, Quality andAccess, paper presented at the conference onCross Border Mobility of Students and Researchers: Financing and Implications for Economic

Ef cienc y and Growth , CESifo Venice Summer Institute, Venice.

Bhagwati, J. (1976),Taxing the Brain Drain , vol. 1, A Proposal and the Brain Drain and Taxation , vol. 2,Theory and Empirical Analysis , North Holland, Amsterdam.

Bologna Declaration (1999),Towards the European Higher Education Area , Conference of Ministers responsible for Higher Education in 29European countries, Bologna.

Del Rey, E. and M. Racionero (2012),Choosing the Type of Income-ContingentLoan: Risk-Sharing versus Risk-Pooling, paper presented at the conference onCross Border Mobility of Students and

Researchers: Financing and Implications for Economic Ef cienc y and Growth ,CESifo Venice Summer Institute, Venice.

Demange, G., R. Fenge and S. Uebelmesser (2013), Financing Higher Education in aMobile World, Journal of Public Econo micTheory , in press.

Docquier, F. and A. Marfouk (2005), Internat ional Migration by Education al Attainment (1990-2000) - Release 1.1., TheWorld Bank, Washington.

Dreher, A. and P. Poutvaara (2011), ForeignStudents and Migration to the UnitedStates,World Development 3, 12941307.

Ehrenberg, R. G. and R. S. Smith (1994), Modern Labor Economics: Theory and Public Policy . Harper Collins, New York.

Grard, M. (2007), Financing Bologna: Which Country Will Pay for Foreign Students?, Education Econ omics 15, 44154.

Grard, M. (2012), Who Is to Pay for Mobile Stu dents?, in A. Curaj, P.Scott, L. Vlaseanu and L. Wilson, eds., European Higher Edu cation at the Crossroads: Between the Bologna Process and National Reforms ,Springer Science+Business Media, Dordrecht, 72748.

Grard, M. and S. Uebelmesser (2013), Cross-Border Mobility of Students and Researchers: Background Paper, mimeo.

Jacobs, B. and F. van der Ploeg (2006), Guide to Reform of Higher Education: A European Perspective,Economic Policy 21, 53592.

Justman, M. and J. F. Thisse (2000), Local Public Funding of Higher Education when Skilled Labor Is Imperfectly Mobile, Internati onal Tax and Public Finance 7, 24758.

Lochner, L. and E. Moretti (2004), The Effect of Education on Crime:Evidence from Prison Inmates, Ar rests, and Self-Reports, American

Economic Review 94, 15589.

OECD (2012), Educatio n at a Gl ance 2012: OECD indicators , OECD,Paris.

Parey, M. and F. Waldinger (2011), Studying Abroad and the Effect onInternational Labour Market Mobility: Evidence from the Introductionof Erasmus,The Economic Journal 121, 194222.

Uebelmesser, S. (2006), To Go or not to Go: Emigration fromGermany,German Economic Review 7, 21131.

Voin, M. and M. Grard (2013), A Contribution to the Study of GlobalCompetition for Talent: The Determinants of Student Mobility andits Consequences for the Internationalization of the Labor Market,Bruges European Economic Policy Brie ngs (BEEP) 27., College of Europe, Bruges.

Webbink, D., N. G. Martin and P. M. Visscher (2010), Does EducationReduce the Probability of Being Overweight?, Journal of Health

Economics 29, 2938.

Weizscker, R. K von and B. U. Wigger (2001), Risk, Resources andEducation, IMF Staff Papers 48, 54760.

Wilson, J. (2008), Taxing the Brain Drain: A Reassessment of theBhagwati Proposal, in E. Dinopoulos, P. Krishna, A. Panagariya andK. Wong, eds.,Trade, Globalization and Poverty, , Routledge, London,254262.

Figure 2

- 6

- 4

- 2

0

2

4

6

8

10

12

14

0 20 40 60 80 100

Relation of graduate mobility to public expenditure

Net brain gain in % of working age residents

Source: OECD (2012), Tables B3.2b; Docquier and Marfouk (2005), Tables 5.1.

Share of public expenditure on total expenditure in %

Note for the regression results: constant 4.321 (t-value 1.747), slope coefficient -4.901 (t-value -1.474);without Nordic (red) countries: constant 4.519 (t-value 1.566), slope coefficient -5.315 (t-value -1.288).

Forum


12/65

Forum

CESifo DICE Report 2/2013 (June)1111

i ncreAsing Access to HigHer e ducAtion t HrougH s tudent l oAns 1

AdriAn z idermAn 2

Four aspects of access

While an accepted, generic de nition of access isa process of enabling entry into higher education(Harvey 2004, 13), this paper approaches the issue of access more widely, from four differing perspectives. Itthen moves on to consider how the availability of gov-ernment-sponsored student loans may impinge, posi-tively, on access.

Broadening of access

Firstly, let us consider the broadening of access. Thisaspect of access relates to policies facilitating the entryof larger numbers of potential students into higher edu-cation. Many countries around the world have long-term policies aimed at increasing the percentage of the rele-vant age cohort (say 1824 years old) that enrolls in high-er education. Underlying this trend is the recognition of the importance of a larger higher-educated populationfor economic and social development, particularly in thecontext of a more competitive, global environment. The

focus is not so much on the socio-economic compositionof the student population, but rather on the relative sizeof this enrollment. Some sixty years ago, the fear that asubstantive expansion of the higher education system inthe UK would lead to falling academic standards (moremeans worse) was countered by the view that therewas a large pool of untapped ability that was not at-tached to tertiary learning at the time (Robbins Report:Committee on Higher Education 1963).

1 This paper draws heavily on the authors previous writings, particu-larly Ziderman (2013).2 Bar-Ilan University, Israel.

Deepening of access

In contrast, we may de ne the deepening of access asensuring that signi cant proportions of students fromnon-traditional areas (such as working class, ethnic mi-norities) enter higher education (Harvey 2004, 13). Herethe emphasis shifts from the need to increase the number of students in higher education, to that of changingtheir composition in order to achieve a more socially

acceptable balance amongst the various socio-economicgroups. This is achieved through reaching out to those,usually disadvantaged, groups who do not customarily pursue higher education studies. The central motiva-tion here is clearly social, and aimed at improving thelife-chances of these groups.

Retent ion and successful completion

More widely, the concept of access is understood toencompass not only entry into higher education, but alsoretention and successful completion (National Of ce

of Equity of Access to Higher Education 2008). Dropoutfrom learning is not only (or perhaps mainly) the resultof academic weakness. Unforeseen nancial dif cul -ties may play a role in many cases, particularly when brought on by such factors as tuition fee increases or adownturn in the economy (with less student employ-ment opportunities). In such circumstances, the avail-ability of student loans (for tuition fees or for living expenses, as appropriate) may be important in mitigating potential student dropout.

Maintaining freshman enrollment levels

Similarly, the advent of tuition fee increases or an eco-nomic slowdown may persuade many potential studentsto decide against enrolling in higher education. Studentloans can offset these nancial barriers.

Access and student loan schemes

What role may the availability of student loans playin facilitating these forms of access? Government-sponsored student loan schemes around the world differ in the central objective pursued. Identifying the under-


13/65

Forum

CESifo DICE Report 2/2013 (June) 12

lying objective of a particular loan scheme is thereforeimportant because this will have implications for manycentral aspects of the scheme. These include: whether loans are offered for tuition, living expenses or both; the

appropriate level of loan subsidy (if any); the need for targeting (con ning eligibility to particular student cat -egories); loan allocation and rationing procedures whereloan funding is limited. But the objective of a givenscheme, and how this effects the operation of a scheme,will also have strong implications for the effect of theloans on access.

We may distinguish three types of loan schemes: costsharing, social targeting and student independence.3

Cost-sharing model

We begin with an elaboration of the cost-sharing mod-el. In many parts of the world, university systems arefacing a nancial crisis. Resources available to univer -sities have been eroded due to a combination of a dra-matic and continuing expansion of student enrollmentsunmatched by public expenditure on higher education.Universities have attempted to alleviate these nancial pressures through the development and extension of non-government sources of funding. Cost-sharing, (or,greater cost-recovery), where a larger and signi cant

share of the costs of university education is shifted ontothe main bene ciaries of university education studentsand their families is the dominant path that is pursuedfor revenue augmentation. In particular, this has takenthe form of the introduction of tuition fees or of rais-ing them to realistic levels; in fewer cases, cost sharingtakes the form of charging for hitherto highly subsidizeddormitory and living costs.

Due to the act that substantially higher tuition ees willcause hardship or enrolled students and are thought

likely to impede university access, tuition hikes havebeen accompanied by the introduction o a state-spon-sored student loan scheme in many countries. Te dis-incentive e ects o up- ront tuition ee increases may also be o set by the availability o loans or studentsthat will cover these augmented costs. Loans enable stu-dent-borrowers to avoid up- ront payments or highereducation (whether or tuition or living expenses) by de-laying payment, which will be rendered in manageableinstallments out o enhanced earnings a er graduation.State intervention is necessary because banks are loath-to make commercial loans to students to nance tuition

3See Ziderman (2002) for a fuller taxonomy of loan scheme objectives.

costs, given the higher risk, lack of collateral and thenebulous nature of the human capital asset that the loanwill generate.

The availability of student loans helps to make tuitionfee increases more acceptable, both politically and so-cially. In Singapore, the 1988 university tuition feerises were accompanied by subsidized loans equiva-lent to about half the value of the new tuition fees. Themuch-discussed Australian loan scheme was introducedin tandem with the imposition of university tuition feesin 1989. In the early 1980s, large tuition fee increases inChile were accompanied by the introduction of studentloans administered by the universities.

Social targeting mode

Student loan schemes may serve the more deliberaterole of increasing the accessibility of the poor and of other marginal groups to higher education. When tar-geted speci cally at such disadvantaged groups, loanschemes (particularly in cases where they are substan-tially subsidized), may lead to greater access of the poor to university education, thus contributing to socialequity. In many countries the relatively low enrolmentof poor and disadvantaged youth in tertiary education(and also in non-compulsory secondary education) is acause of social concern. Increasing the access to univer-sity education among these segments of the populationhas become a major element in educational and social policy. While the cause of low access of the poor is mul-ti-faceted (and a full discussion is beyond the scope of this paper), nancial constraints evidently play a role.There is now a broad consensus on the need to offer clear nancial incentives to poor, potential students, notonly to overcome the burden of fee payment and livingexpenses, but also to offset both parental resistance toreductions in family income and the fear that the bene-

ts of the educational process may not be sizeable. The provision of nancial aid therefore may be regarded as anecessary, though not suf cient condition for achievinggreater participation of the poor.

But what form should this nancial assist ance take? Thetraditional, and most effective, method of enhancingthe educational access of the poor has been through the provision of means-tested grants to cover tuition fees(where schooling is not free) and, often, to cover livingexpenses as well. However, a widespread grants schemeis likely to be expensive. The use of loans, rather thangrants, proactively targeted at the poor, offers a methodthat may both increase access for the poor and reduce, or


14/65

Forum


at least contain, public expenditure on student supportover the longer term, as loan repayments build up. To beeffective in increasing the higher education access of the poor, loans may need to be made available under soft

lending conditions.

Subsidized loan policies can have a limited effect onraising access of the poor; but this role needs to be com- plemented by appropriate action far earlier on in the ed-ucation process. Insuf cient academic preparation andthe lack of willingness of large numbers of the poor toenroll in higher education have their roots much further upstream in the education system.

Student independence model

Even when tuition fees are minimal, students (both themore af uent and the disadvantaged) may face consid -erable nancial burdens: potential earnings are foregonewhile studying, and living expenses may be sizeable,especially when the student does not attend a univer-sity near home. Financial pressures may have negativeeffects on a students academic performance (and thuscompromise the process of human capital investment);the fear of such pressures will act negatively on deci-sions to enrol in tertiary education. These pressures can be mitigated by the broad availability of student loansfor living expenses. While such burdens may fall rel-atively heavily on the poor, in principle loans for this

purpose could be made broadly available, to more af u -ent as well as poorer (current and potential) students, aslong as these loans are not unduly subsidized.

Eligibility, and the extent of loan support, is determined by parental income in many loan schemes. The conceptof parental support is a central element in loan schemesin many European countries, including Austria,

Germany, France, Italy, Portugal, Spain, and the UK.However, parents are not legally required to make thedesignated parental contribution. Thus many stu-dents, including those from non-poor backgrounds, mayface nancial dif culties during study, while potentialstudents may not enroll if they feel that the parentalcontribution will not be forthcoming. A very differentapproach is taken in a number of countries, which basestudent support on the concept of student nancial inde - pendence; student entitlement to loan support is basedon student, not parental, income. Such arrangements ex-

ist in Denmark, Finland, Norway and Sweden; in these Nordic countries, where tuition is free, grants and loansare made for living expenses only.

Table 1 provides a matrix of selected national loanschemes, in which loan scheme coverage is mappedagainst loan scheme purpose (i.e., the three types of loan models, outlined above). The cost-sharing mo-del is illustrated, principally, from experiences of howthe Australian scheme and the new scheme in Englandwork. Examples of the social targeting model are drawnfrom ve UNESCO-supported Asian case studies. 4 Thestudent independence model is illustrated mainly from acluster of European countries where typically no tuitionfees are charged and loans cover living expenses only.

4 The regional study, a joint endeavor of UNESCO-Bangkok and theInternational Institute for Educational Planning in Paris, consisted of ve in-depth studies on the functioning of student loan schemes inAsia: China, Hong Kong, the Republic of Korea, the Philippines, andThailand. A synthesis study is provided in Ziderman (2004).

Type of access and loan scheme objective / coverage, country examples

Loan schemeobjective

Loan scheme coverage

Tuition fees only Living expenses only Tuition and living expenses

Cost-sharingAustralia England, New Zealand

Student independenceHong Kong (NLS) a)

Korea (GECP) b) Denmark, Finland, Hungary,

Norway, SwedenCanada

Social targeting The PhilippinesHong Kong (LSFS) c)

Korea (MOE) d) China, South Africa

Thailand e)

a) Hong Kong: non-subsidized scheme (NLS), b) Korea Government Employees scheme, c) Hong Kong: subsidized scheme(LSFS), d) Korea Ministry of Education scheme, e) Thailand Student Loan Scheme (SLS)

Source: Ziderman (2013).

Table 1


15/65

Forum


Loan scheme objective and expec tations

A clear distinction may be drawn between the cost-shar-ing model and the other two models (Table 2). While

loan schemes conforming to all three objectives willhave implications for access, it is only the social tar-geting model that focuses predominantly on increasingaccess. Cost-sharing is concerned mainly with facilitat-ing tuition fees increases and generating funding for theuniversity sector; it has constituted the major rationalefor the spread of student loan schemes in industrializedcountries. The other two objectives are not concernedwith augmenting university funding as such, but arewider in scope, with a clear social perspective. Loanschemes aimed at cost recovery would be restricted to

universities in the public sector, while in meeting theother two objectives loans should, in principle, be avail-able to students enrolled in public and private univer-sities, on an equal basis. Most loan schemes are highlysubsidized, mainly because they are offered at belowmarket interest rates; however, in most cases, such sub-sidization cannot be justi ed (loan subsidizat ion is dealtwith subsequently). While the aim should be near fullloan recovery, loan schemes targeted at the poor mayconstitute an acceptable exception.

In Table 3, type of access (as discussed in section Four aspects of access ) is mapped against loan scheme objective. Cost sharing and student independence loanscheme affect positively the broadening of access, whilesocial targeting schemes may lead to the deepening of access. All three loan scheme categories affect studentretention positively, while cost sharing loan scheme may

Student loan schemes: objectives and expectations

Expectations

Loan scheme objective

Cost-sharing model Social targeting model Student independence model

Loans will facilitateincreased tuition fees Yes No No

Loans will generateadditional university funding Yes No No

Loans are restrictedto public universities Yes No No

Loans are highly subsidized No Probably No

Loans are confinedto a target group No Yes No

Source: Ziderman (2013).

Table 2

also help maintain new student enrolment levels whenfees increase.

Financial viability

Almost all government-sponsored student loansschemes are highly subsidized by governments. Tismeans that, unlike commercial loans, a sizeable pro-portion o the total loans outlay by the loans body, be itgovernment department, loan scheme authority or com-mercial bank, will not be received back in repayment.A large and sustained gap between disbursements andrecovery implies continuing governmental nancialsupport or the scheme. Given pressures on govern-

ment budgets, the continuation o these subsidies may not be assured, thus compromising the viability o theseschemes over the longer term.

Loan repayment and loan recovery

But why (unlike commercial loans) do govern-ment-sponsored student loans schemes ail to recoverthe sums loaned out through the scheme? A number o actors militate against ull recovery o loans.

Firstly, there are built-in subsidies, incorporated intothe design o the loan scheme. While these lighten theburden o repayment alling on the individual stu-dent-borrower, they reduce repayment income accruingto the lending body. Tese so loan conditions includezero or below-market interest rates on the loan, periodsin which no interest is levied on outstanding debt (both


16/65

Forum


during study and in grace periods a er study comple-tion) and repayments not linked to the rate o in ation.

Te e ect o these built-in subsidies is ampli ed whereamortization periods are long. Te loan repayment ra-tio indicates how much o the loan an average borroweris required to repay. It is measured by the ratio o total(discounted) required payments or each borrower, tototal (discounted) loan sums received. 5

Secondly, there are inefciencies in running the schemeas a whole, in terms o substantial repayment de ault andhigh administration costs that are not passed on to theborrower. Te repayment ratio relates to the typica l bor-rower; it ails to show the ull extent o recovery to theloan und, rom the overall perspective o the scheme asa whole. Even i student loans were not subsidized, andthe individual student was required to repay in ull, notall o the sums loaned would be recouped by the loanauthorities.

Thus, overall loan recovery depends not only on the to-tal of all individual cash repayments. It also takes ac-count of administrative costs that are not passed on tothe student borrowers and of the extent of non-repay-ment, including both default and loan forgiveness,

for reasons like low graduate earnings that fall belowa designated income threshold, disability, death, ac-ademic performance (South African, Norwegian andDutch schemes) and the encouragement of graduatesto enter skills-shortage occupations (Korean Ministryof Education Scheme for Engineering graduates) or to practice in underserved geographical areas (teachersand doctors in the US).

The loan recovery ratio is measured by the ratio of to-tal (discounted) repayments to total (discounted) loanscheme outlays. Clearly, the recovery ratio is always

5 Both measured in terms of present values.

lower than the repayment ratio, because the latter takesno account of the probability of repayment default and

does not include general administration costs.

Repayment and recovery ratios: interna tional

comparisons

How large are these gaps in practice? A joint paper bythe author probes this issue for 44 loan schemes in 39countries (Shen and Ziderman 2009, updating Zidermanand Albrecht 1995). The analysis shows considerablevariation in the size of the repayment and recovery ra-tios across schemes. Many loan schemes exhibit size-able built-in subsidies accruing to student borrowers.The average repayment ratio is 61 percent (so that, onaverage, borrowers are required to repay only about 60 percent of the total loan received).

The distribution of repayment ratios across the 44schemes is shown in Table 4. 13 schemes (about 30 percent of the sample) have relatively high repaymentratios, in excess of 80 percent. However, most schemescontain large built-in subsidies: the repayment ratio in18 schemes (over 40 percent of the loan schemes exam-ined) is less than 60 percent.

Overall loan recovery is considerably lower. No schemehas a loan recovery ratio exceeding 80 percent. Onlyve programs (above 20 percent of the sample) displayrecovery ratios higher than 60 percent, for the most partloan recovery is not high; 80 percent of the schemes dis- play recovery ratios of 60 percent or less. In a third of the cases, loan recovery does not rise above 20 percent.Overall, the average recovery ratio is 39 percent.

Two noteworthy points emerge from the results report-ed in Table 4; both dispel prevalent myths about the -nancing of loan schemes. The rst relates to the shortfallfrom full recovery in almost all government-sponsored

Type of access and loan scheme objective

Loan schemeobjective

Type of access

Broadening access Deepening access Student retention Maintaining freshmenenrollment levels

Cost-sharing X X X

Student independence X X

Social targeting X X

Source: The author.

Table 3


17/65

Forum


loans (and the very heavy losses in some). The impli-cation of this is that government subsidization is to beseen as an enduring feature of these schemes; the wide-ly-held view that loan schemes can act as a revolvingfund which, once capitalized, will nance themselvesthrough repayments of earlier loans, is not consonantwith the facts in almost all cases. The second relates tothe supposed dominant role played by repayment defaultand high administrative costs in accounting for low loanrecovery. As shown in the bottom section of Table 4,the major factor, by far, accounting for recovery loss isthe large, built-in, interest rate subsidy element in mostschemes. Excluding default and administration charges,recovery from the amount that graduates are required torepay (i.e., the repayment ratio) is surprisingly low onaverage. The average repayment ratio is 61 percent, rep-resenting a hidden grant to the student and a loss to thescheme of some 40 percent. The addition of default andadministration costs reduces recovery by only a further 20 percentage points (ten percent in each case). The fol-

lowing section looks at the issue of whether such largeinterest rate subsidies can be justi ed.

Can student loan subsidies be justi ed?

In loan schemes where either cost recovery or studentindependence constitutes the central objective, the casefor heavy built-in student loan subsidies is not strong.For current students, the intended effect of student loansin both of these cases is to reduce the nancial bur -den on students during study and to delay fee payment(through borrowing) until after graduation, when pay-ment is more readily made from the expected enhance-

ment of earnings. For potentialstudents, the availability of a loan programme will encourage access,under the concept of study now,

pay later.6 Since the aim should be near-full loan recovery in thesetwo cases, the level of built-in sub-sidy is often excessive in practice.

It is only where loan schemes areaimed directly at social targetingthat a clearer case for sizeable built-in subsidies can be made.However, such subsidies, as wehave noted, will entail considera-

ble budgetary costs. Since a grantoffers a stronger and more directincentive for access than a (par-

tially) repayable loan, the apparent advantage of loansover grants is less clear-cut. This highlights a centralconundrum in loan policy: at what level of in-built loansubsidy does a grant become a more cost-effective in-strument for helping the poor than a subsidised loan(with hidden grants)? This suggests that, in country set -tings where state budgets are constrained, a more ap-

propriate nancial aid program to encourage access of the poor is likely to involve a combination of both loansand grants, with a relatively larger overt grant elementfor the very poor. This is common practice in the LFSHong Kong scheme, in England and in many other loanschemes.

In the comparative study of loan schemes in South EastAsia, most of the schemes studied were shown to con-form to the social targeting model (Ziderman 2004). Yetthe evidence did not indicate any high degree of successin increasing the university access of the poor. A number of essential conditions for success were lacking. These

included a suf ciently high level of individual supportto cover necessary expenses; a broad coverage of poor students to achieve national impact, and careful anddeliberate loans targeting so that loans do indeed reachthe poor and other disadvantaged groups, otherwise thecentral objective of the scheme is compromised. Loanschemes aimed at greater participation of the poor areoften not effective because these ingredients for successare missing.

6This is the name of the Philippines national loans scheme, a social-tar-geting scheme that is so limited in coverage that it has little effect in in-creasing access of the poor. Little attempt is made at collecting due loanrepayments, so recovery is minimal; thus the scheme is often (and ap- propriately) dubbed the Study Now, Repay Never scheme.

Loan repayment and recovery ratios: international comparisons

Ratio

Number of schemes

Repayment ratio Recovery ratio (with defaultand administration costs)

Above 80 percent 13 0

6180 percent 13 5

4160 percent 8 11

2140 percent 7 2

20 percent or less 3 8

Total number of schemes

44 26

Average repayment ratio: 61%, average recovery ratio, overall 39%

Source: Shen and Ziderman (2009).

Table 4


18/65

Forum


The upshot of this discussion is that the levels of built-insubsidies, resulting in low repayment ratios, are oftenexcessive. High subsidies may be either unnecessary(cost sharing and student independence models) or not

very effective in pract ice in achieving objectives (socialtargeting). Since the level of built-in subsidy is xed bygovernment, these subsidies may be reduced, as appropriate, by government decision. However, vested inter-ests may militate against these desirable changes.

The repayment burden

Finally, we consider the concern that many student loansschemes saddle graduates with an inordinately heavy

level of debt. In particular, it is argued that the fear of entering into student-loan debt acts as a disincentive for prospective students from lower socio-economic back-grounds to apply for university studies.7 However, theapproach adopted here is not concerned with the totalsize of the debt facing a student on graduation (nor with prospective students perception of this debt), but rather with the extent to which repayment of the loan does, in

practice, constitute a nancial burden. The repayment burden falling on the borrower each year may be meas-ured by the required annual loan repayment expressedas a percentage of annual income.

In the case of loan schemes where repayment is a xed percentage of income (income-contingent repaymentschemes), this percentage is de ned by the conditionsof the loan. In the case of the Hungarian loan scheme,this stands at six percent; in the South African schemerepayment varies from between three to eight percentdepending on annual income, in New Zealand at ten percent and in the current scheme in England at nine percent. Since the repayment percentage out of incomeis built into income-contingent schemes, the repayment

burden may, by design, be kept within acceptable lim-its and is the same for all borrowers. Furthermore, lowincome earners and the unemployed are protected by aminimum income threshold for repayment.

With mortgage-type loan schemes (where the periodic sum to be repaid is xed), the situation is very different.While all borrowers repay the same annual amount, therepayment burden falls over time, as incomes increase.The size of the repayment burden will depend not only

7 This is strongly argued by Callender and Jackson (2005) in the Englishcontext; but is not supported from studies in Australia (Andrews 1999)and the Netherlands (Vossensteyn 2005).

on graduate annual incomes and loan size, but also onthe size of loan subsidies and the length of the repay-ment horizon.

Chinese loans schemes do not carry high subsidies, re-sulting in heavy repayment burdens. Subsequent loanscheme reforms, notably through increasing the number of years over which loans must be repaid, have ledto lower repayment burdens. In the early years of loanscheme operation, the short four-year repayment hori-zon resulted in a heavy repayment burden of 24 percenton average over each repayment year and high repay-ment default. The number of repayment periods wassubsequently increased to six in 2004, together with atwo-year grace period; currently the repayment horizon

is ten years, resulting in a more acceptable repayment burden of 8.8 percent in the rst year and falling steadilyto 2.6 percent.

High levels of state subsidy in the Thai Student Loanscheme (SLS) aimed at increasing access of the poor imply a low repayment ratio of only 21 percent andmoderate annual repayment obligations. Consequently,the repayment burden is very light: some 2.5 percent for males and, because of their lower earnings, around 3.5 percent for females. However, a recent paper (Chapmanet al. 2010) argues that such average estimates are mis-leading because they do not show the considerably high-er repayment burdens borne by low-earning graduates.The repayment burden for graduate borrowers falling inthe lowest decile of earners is shown to be nine percentfor males and 13.9 percent for females. This may notonly enhance repayment default, but may also act as adisincentive for access of those potential students whoare pessimistic about their future earnings. The policyresponse to these ndings is to incorporate measuresinto mortgage-type loan design, to protect low earn-ers from excessive repayment burdens; this may be

achieved by the introduction of suf ciently high incomerepayment thresholds, as is common in income-contin-gent schemes.

A concluding comment

Student loans can have a positive, yet limited, role inaugmenting access, as de ned in this paper; but the gen -eral case for heavy loan subsidisation is weak. Whenthe central loan scheme objective is access deepening reaching out to the poor and other disadvantaged groups student loans (probably subsidised) can constitute animportant element in the available policy toolbox for


19/65

Forum


increasing access. However, loan schemes need to bewell-designed to avoid an excessive repayment burdenand default.

References

Andrews, L. (1999), Does HECS De ter? Factors Affecting University Participat ion of Low SES Groups, Department of Employment,Education, Training and Youth Affairs, Canberra.

Callender, C. and J. Jackson (2005), Does the Fear of Debt Deter Students from Higher Education?, Journal of Social Policy 34 (4),50940.

Chapman, B., K. Lounkaew, P. Polsiri, R. Sarachitti and T.Sitthipongpanich (2010), Thailands Student Loan Fund: Interest RateSubsidies and Repayment Burdens, Economics of Ed ucation Review29 (5), 68594.

Harvey, L. (2004-13), Analytic Quality Glossary, Quality Research

International, http://www.qualityresearchinternational.com/glossary(accessed 15 Apr il 2013).

National Of ce of Equity of Access to Higher Education (2008), National Plan for Equity of Access to Higher Education , 2008-2013 ,Higher Education Authority, Dublin.

Robbins Report: Committee on Higher Education (1963), Higher Education: Report of the Commit tee Appoin ted by the Pri me Minister Under the Chairmanship of Lord Robbins 1961-63, Cmnd. 2154,HMSO, 23 September 1963, London.

Shen, H. and A. Ziderman (2009), Student Loans Repayment andRecovery: International Comparisons, Higher Education 57, 31533.

Vossensteyn, J. J. (2005), Perception s of Student Price-Re sponsivene ss: A Behavioura l Economics Exploration of the Relationshi ps betweenSocio-Economic Status, Perceptions of Financial Incentives and Student Choice , Ph. D Dissertation, Center for Higher Education PolicyStudies (CHEPS), University of Twente, Brno: VUTIUM Press.

Ziderman, A. (2002), Alternative Objectives of National Student LoanSchemes: Implications for Design, Evaluation and Policy,The Welsh

Journal of Educat ion 11 (1), 3747.

Ziderman, A. (2004), Policy Options for Student Loan Schemes: Lessons from Five Asian Case Studies , International Institute for Educational Planning, UNESCO, Paris.

Ziderman, A. (2013), Student Loans Schemes in Practice: GlobalPerspectives, in D. E. Heller and C. Callender, eds.,Student Financing of Higher Education: A Comparative Perspective, InternationalStudies in Higher Education series, Routledge.

Ziderman, A. and D. Albrecht (1995), Financing Universitie s in Developing Count ries , Stanford Series on Education and Public Policy,Falmer Press, London.


20/65


21/65

Forum


Below we propose an analysis of the reform from the point of view of equity: more speci cal ly, we estimateinequality of opportunity in access to university in Italy between 1995 and 2007 and show that inequality of opportunity declined immediately after the reform, butsubsequently increased again.4

Inequality of education opportunity

We propose to track access to university before and af-ter the reform, with a focus on inequality of educationalopportunity (IEOp hereafter). Firstly, let us de ne our concept of the equality of educational opportunities.Following on from the recent economic literature onequality of opportunity (Fleurbaey 2008 and Roemer 1998), we model access probabilities as a function of two kinds of variables: variables beyond individualcontrol (called circumstances) and variables of respon-sibility (called effort). We de ne IEOp as that portion

of inequality in the probability that can be att ributed tocircumstances beyond individual control and we look atthe change in IEOp over time. We are aware that it is not possible to control for all possible sources of change inIEOp; hence, our exercise may not be strictly considereda policy evaluation of the reform. It should be seen in-stead as an attempt to understand whether the positiveeffect of the reform vanished in a decade.

To operationalise the concept of IEOp, the rst step con -sists of dividing the possible factors in uencing univer -sity access into circumstances and effort. In principle,4 Hence we update our previous analysis (Brunori et al. 2012) by ex- ploiting a new wave of the Istat dataset on upper secondary graduates .

circumstances are all variablesthat affect individual outcome, butare not direct or indirect effects of choices. Different characteristics

may be considered as fair or un-fair sources of inequality depend-ing on the subjective or collectivenormative beliefs. Race and soci-oeconomic background are largelyagreed sources of unfair inequali-ty; many believe that inequalitiesdue to innate ability, on the other hand, which is de nitely a charac -teristic beyond individual control,are morally legitimate.

De ning a domain for responsi - bility is particularly complex in

the case of education for two main reasons: i) educationis a fundamental way of obtaining labour market opportunities, ii) most education takes place in the earlyyears, when individuals have a limited understanding of the consequence of their choices. Supporters of the rstargument underline the role of education as an instru-mental good: skills acquired in school and university produce income opportunity in the future. Equality of opportunity in the labour market should therefore re-quire full equality in education (Howe 1989). However,if education is clearly a source of opportunities on theone hand, it is itself the result of circumstances andchoices on the other, so this paper considers access totertiary education as an end, and not a means. The sec-ond issue challenges the idea that young adult studentscan be considered responsible for the choice they make.While there is a general consensus that children cannot be held responsible for their choices, it seems instead plausible to hold individuals responsible for their choic-es at the age of around 19 years.5

Model and data

We measure equality of opportunity in access to tertiaryeducation by looking at the conditional probabilities of access to university for individuals with different cir-cumstances. As discussed above, the outcome is deter-mined by two kinds of variables: circumstances andresponsibility variables, where circumstances are allobservable variables beyond individual control like gen-der or socioeconomic background. Responsibility vari-

5 See Trannoy (1999) and Brunori et al. (2012) for a discussion.

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1 9 8 1 / 8 2

1 9 8 3 / 8 4

1 9 8 5 / 8 6

1 9 8 7 / 8 8

1 9 8 9 / 9 0

1 9 9 1 / 9 2

1 9 9 3 / 9 4

1 9 9 5 / 9 6

1 9 9 7 / 9 8

1 9 9 9 / 0 0

2 0 0 1 / 0 2

2 0 0 3 / 0 4

2 0 0 5 / 0 6

2 0 0 7 / 0 8

2 0 0 9 / 1 0

enroll/coll grad

enroll/19 cohort

col grad/19 cohort

Enrollment ratios%

Source: CNVSU (2011).

Figure 1


22/65

Forum


ables are summarised by a proxy that we will call ef-fort. IEOp represents total inequality as a share of the probability of accessing tertiary education due to cir-cumstances. The population of students is divided intogroups characterised by identical circumstances (types),and each type is sub-divided into groups of students thatexerted the same degree of effort (tranches). To meas-

ure IEOp, we build a matrix of probabilities, where the probability of accessing university for students in thesame type and same tranche is reported in each elementof the matrix. There are also at least two approaches tomeasuring IEOp in distribution: ex-ante and ex-post.The former focuses on the idea that all inequality due toeffort is unproblematic. Ex-ante IEOp is obtained resid-ually: in a rst step all inequality due to effort is elim -inated (within type), and the residual inequality is sub-sequently measured. Ex-post IEOp is obtained directly, by measuring for all degrees of effort (hence tranche bytranche) the difference in probabilities due to circum-stances. As discussed in Fleurbaey and Peragine (2013),these two approaches differ and although they general-ly return consistent estimates, they could, in principle,move in opposite directions over time.

To measure IEOp in tertiary education we rst de nean outcome of access to tertiary education for an upper secondary school graduate. We then identify the variables beyond individual control (circumstances): gender,family socioeconomic background (based on parentaleducation), region of residence (Centre/North, South),

and educational attainment at the age of 15 (high, lowgrades).6

Our proxy for effort is related to the grade of the upper secondary nal exam. We recognise that this grade can -not be considered a proxy for how hard a student tried,as grades are affected by a students circumstances.Therefore, like Roemer (1998), we believe that such ameasure of effort is only suitable to compare individu-als belonging to the same type, as they are all subject to

6 We consider the educational attainment at the age of 15 a very rele-vant predictor of the future success in education. We also consider it asa circumstance beyond individual control because due to circumstancesand individual choices made at an age in which pupils cannot be heldresponsible for their decisions; see Brunori et al. (2012) for a discussion.

the same circumstances. In order to make it comparablefor individuals of different types, we de ne effort as therank in the type speci c distribution of observed effortas an ordinal and inter-type comparable measure of ef-fort. Hence, two individuals are declared to have exertedthe same degree of effort if they sit at the same positionin their respective type speci c grade distribution.

In order to implement our measures we use datafrom Indagine sullInserimento Professionale dei

Diplomati (IIPD), a survey published every three years by the Italian National Bureau of Stat istics (Istat). Thesurvey focuses on the transition from upper secondaryschool to work and university of a representative sample of Italian students, who completed upper secondaryschool. We estimate IEOp in ve waves: 1998, 2001,2004, 2007, and 2010, each containing information onstudents that completed upper secondary school threeyears previously. The survey data includes informationon students socioeconomic background, school curric-ula and access to both university and labour market after upper secondary school.

Discussion

Our results are in line with what we found in our pre-vious analysis.7 Table 1 shows the ex-ante and ex-postIEOp measures. The two measures quantify inequalityin the probability of enrolling in university, the former

between types and the latter within tranches. In bothcases inequality is measured by the Gini index.

The ex ante and ex post IEOp show similar trends: theysigni cantly drop in 2004, slightly decrease from 2004to 2007, and increase somewhat from 2007 to 2010. Infact, both the ex ante and the ex post measures returnedto their 1998 level in 2010. Hence, our evidence showsthat the 2001 university reform had only a short-termeffect in terms of IEOp in the access to tertiary educa-tion. This may be due to the fact that the reduction of

7 Estimates differ in absolute terms from the measures presented inBrunori et al. (2012) because in that case we were controlling for demo-graphic change across time.

Ex-ante and ex-post IEOp measures

1998 2001 2004 2007

Ex ante 0,289 0,333 0,275 0,266Ex post 0,311 0,356 0,304 0,287

Source: Authors elaboration on Istat IIPD.

Table 1


23/65

Forum


inequality might be linked to a number of side-effectsof the 2001 university reform such as the sharp rise inthe number of university degrees awarded (the number of courses offered by Italian university totalled 2,444 in

2000 and 3,234 in 2001), the spread of university loca-tions across Italy (the number of cities with an universitygrew from 93 in 1995 to 146 in 2001) and the reductionin the workload required to obtain a degree (Bratti et al.2007). Given that enrolled rates among students fromwell-off social background was already very high prior to the reform, the effect of an increase in university de-grees and/or locations might have acted as an incentivefor students from less advantaged social backgrounds,lowering IEOp in 2004.

Why did the reform have such limited effects? One pos -sible explanation is that, at the time that the reform wasintroduced, there were high expectations of the oppor-tunities that the new system may bring. Many studentswho had completed upper secondary school decided toenrol in higher education expecting high returns for ashorter investment in human capital. However, withina few years students and parents learnt that there werelower returns from the new shorter degrees and the en-rolment rate declined as a result, especially for less ad-vantaged students. Two facts are in part icular consistentwith our interpretations: (i) the large percentage of grad-uates of 3 year courses that enrolled in two-year master degrees varies across universities, but is generally above60 percent; (ii) growth in the rate of youth unemploy-ment is shown to be even higher for university graduatesthan for students that completed upper secondary schoolamong individuals aged under 35 years old.

References

Bratti, M., C. Broccolini and S. Staffolani (2007), Mass TertiaryEducation, Higher Education Standard and University Reform: A

Theoretical Analysis,Quaderni di ricerca 227 , Universit degli studidi Ancona, Diparti- mento di Economia.

Brunor i, P., V. Peragine and L. Serlenga (2012), Fairness in Education:The Italian University Before and After the Reform, Economics of

Education Revi ew 31 (5), 76477.

Cappellari, L. and C. Lucifora (2009), The Bologna Process andCollege Enrollment Decisions, Labour Economic s 16 (6), 63847.

CNVSU (2011),Undicesimo Rapporto sullo Stato del SistemaUniversitario , Ministero dellIstruzione e della Ricerca Universitaria,Comitato Nazionale per la Valutazione del Sistema Universitario,January.

DHombres, B. (2007),The Impact of University Reforms on Dropout Rates and Student s Status: Evidence from Italy , Joint ResearchCentre Institute for the Protection and Security of the Citizen.

Di Pietro, G. and A. Cutillo (2008), Degree Flexibility and UniversityDrop-Out: The Italian Experience, Economics of Education Review27, 54655.

Fleurbaey, M. (2008), Fairness, Responsibilit y and Welfare , 1st ed.,Oxford University Press, New York.

Fleurbaey, M. and V. Peragine (2013), Ex Ante Versus Ex PostEquality of Opportunity, Economica 80 (317), 11830.

Howe, K. R. (1989), In Defense of Outcome-Based Conceptions of Equal Educational Opportunity, Education al Theory 39, 31736.

Istat (2006), Universit e lavoro: orientarsi con la statistica 2006. Tech.rep., ISTAT.

Roemer, J. E. (1998), Equality of Opportun ity , Harvard UniversityPress, Cambridge, MA.

Trannoy A. (1999), Lgalisation des Savoirs de Base: Lclairagedes Thories conomiques de la Responsabilit et des Contrats, inD. Meuret and M. Duru. Bellat, eds., Justice et Educ ation , De Boeck,Bruxelles, 5576.


24/65


25/65

Forum


The Government has summarised the case for thechanges in the following terms: Our reforms tacklethree challenges. First, putting higher education on asustainable footing. We inherited the largest budget de -

cit in post-war history, requiring spending cuts acrossgovernment. By shifting public spending away fromteaching grants and towards repayable tuition loans, wehave ensured that higher education receives the fundingit needs even as substantial savings are made to publicexpenditure. Second, institutions must deliver a better student experience; improving teaching, assessment,feedback and preparation for the world of work. Third,they must take more responsibility for increasing socialmobility (Department for Business, Innovation andSkills 2011, Executive Summary, paragraph 3).

As a result of these reforms, the English system of high-er education now conforms quite closely to the marketmodel:

Universities are legally private institutions with con-siderable nancial and operational autonomy.

Whilst most providers are not for pro t institutions,for pro t providers are beginning to emerge, withconsiderable of cial encouragement.

There is erce competition for students, with studentshaving a wide choice of subjects, programmes, pro-viders and modes of study, together with a consid-erable amount of information on which to base their choices.

Tuition fees now represent all of the costs of teachingfor most subjects.

Students must meet tuition and living costs fromtheir own or their families resources, although in-come contingent loans remain available to full-timestudents to cover fees and living costs together witha system of non-repayable maintenance grants and bursaries for students from households with incomes

of up to GBP 42,000 (part-time students are entitledonly to fee loans).

The historical background

These reforms can be seen as the latest stage in a pro-gressive process of marketisation that began with theThatcher Governments decision in November 1979 toend the subsidy for overseas students fees. Other major steps have included: The separation of public funding for teaching and

research, and the introduction of selective research

funding, from 1986. The increase in the level of the still-subsidised Home

and EU Student Tuition Fee in 1989, and the corre-sponding reduction in the institutional grant for

teaching (partially reversed in 1993). The introduction of top-up loans for student support

from 1990. The abolition of the binary line between universities

and polytechnics in 1992. The introduction of top-up fees of GBP 1,000

in 1998. The changes in the rules for university titles in 2004

to enable institutions without powers to award de-grees to obtain a university title.

The introduction of variable fees of GPB 3,000

in 2006.We should also note the corporatisation of universitygovernance and the introduction of sector-wide institu-tional indicators and benchmarks (Brown 2012a and b).

What has been the impact of these reforms?

There can be little doubt that, through research selec-tivity, cost-sharing between taxpayers and students/graduates, and the increase in the number of providers,UK higher education is much more ef cient in its use

of resources than it was even ten years ago. This hasenabled public and private funding to go much further.UK universities are also much more responsive to stu-dents and other external stakeholders, and much moreentrepreneurial (as anyone who has encountered their representatives at overseas student recruitment fairs cancon rm). This is re ected in the extent to which theyhave diversi ed their revenues away from reliance onthe taxpayer, to an extent that was simply unimaginablethirty years ago.

But there has also been a downside. There has been areduction in institutional diversity and an increase ininstitutional strati cation and inequality. This, in turn,has limited educational innovation, reduced oppor-tunities, and damaged social mobility. The effect onquality has been mixed. In both research and teaching,improvements became increasingly marginal as institu-tions devoted more resources to compliance and learn-ing to play the game. Research selectivity has becomethe pursuit of status, quality assurance has becomereputation management, admissions has become mar-keting, Vice-Chancellors have become entrepreneurs.Above all, marketisation threatens the implicit contractthat higher education has with society whereby univer-


26/65

Forum


sities enjoy certain privileges in return for the publicgoods that they provide: The classic justi cation for the non-pro t status of educational institutions is that itredresses information asymmetry between buyers and

sellers. Because consumers cannot adequately monitor the quality of educational services, they prefer dealingwith institutions they can trust not to take advantageof them to make a pro t. But maximising revenue nowlooks a good deal like making a pro t. Private universi -ties now engage in such deceptive practices as award-ing less aid to early admission students or front-loadingthe rst year of aid packages (McPherson and Schapiro1998). Students in the aggregate may gain greater wagesthrough these arrangements, but each student must fendfor themselves. Trust in this relationship can no longer

be assumed (Geiger 2004, 171).

It is clearly too soon to be able to evaluate the impactof the current reforms (Brown, submitted for review).The rest of this article looks at the main claims madefor market-based policies and considers how far suchclaims are justi ed, drawing not only on UK experi -ence, but also on the experience of other systems thathave gone down the market route.

Marketisation of higher education: pros and cons

Five main claims are usually made for introducing or increasing market competition in higher education:

Increased ef ciency. Increased quality. Increased innovation. Increased diversity. Increased student choice.

Increased ef ciency

The argument is that the introduction or intensi cationof competition will increase the ef ciency of resourceuse (as can reduced funding, either in aggregate or per student). This is positive because it makes resourc-es go further, and reduces claims on the taxpayer andthe economy, thus freeing up scarce capital for other potentially more valuable or important uses. However,increased competition can also lead to signi cant waste,for example, the amounts spent on marketing and oth-er activities unrelated to the quality of education, thetransaction costs of bidding competitions, the additionalresources needed for collecting fees, chasing debts, etc.This is, of course, ironic given the rationale for marke-

tisation in the rst place, namely to obtain better valuefor money and make resources go further. Of course,regulation in a non-market system can also consumeand divert resources. All this points to the need for ef-ciency to be measured and the effects of the changesmonitored.

Increased quality

Quality of service can be increased if institutions haveto take greater account of consumers views, enhanceservice responsiveness, give quicker and better feedback to students, handle complaints more expeditiously,etc. But quality may also be damaged if consumers or consumer pressures have too much purchase (grade in-

ation, grade grubbing); and especially if quality judg -ments are taken out of academic hands by the marketor by managers acting in response to market conditions/signals (Brown 2009). There are also clear dangers fromstudents adopting a more instrumental attitude to their studies (commodi cation: see Naidoo and Jamieson2005). This points to the need for quality to be mon-itored independently of market conditions/controls, but taking account of resourcing levels and uses, and by an agency independent both of the institutions andgovernment.

Increased innovation

Innovation can increase ef ciency, raise quality, ex - pand choice and increase consumer satisfaction, etc.Innovation, however, can also be damaged if institutionsdont wish to be seen to be departing too far from rec-ognised academic models. Innovation is closely linkedto differentiation (see below). A lot depends on whichinstitutions are doing the innovating, and it will rare-ly be the high prestige ones. In any case, innovation onits own is not necessarily good: it can be at the expense

of quality. All this points to the need for innovation to be de ned and monitored, for example, for its effect onquality

Increased diversity

Differentiation can occur to extend demand for an in-stitutions product by distinguishing or emphasisingspecial characteristics, creating and occupying specialniches, etc. But it can also be reduced if a single insti-tutional model the large, multi-faculty, research-in-tensive, selective university becomes dominant. Thisis linked to the risk that marketisation will strengthenthe dominant institutions, which, in turn, increases the


27/65

Forum


risk that their needs/interests will distort the system(as, arguably, will research due to the costs of build-ing up research capability). This is almost certainlywhat has happened in the UK: reputational hierarchy

has replaced functional diversity. Diversity can also bereduced through the rationalisation of institutions (ab-sorption of specialist institutions) and/or of programmes(as low demand subjects are curtailed or withdrawn be-cause of cost pressures on providers, cherry picking bynew entrants, etc). This points to the need for diversi-ty to be de ned, monitored and protected, if necessarythrough state intervention to constrain possibilities for institutional development.

Increased student choice

Choice can be increased as the number of institutionsgrows, especially if entrant institutions offer some-thing different (innovation) or existing institutionswiden the programmes/subjects/modes on offer (diver-sity). But it can also be reduced through rationalisationand reduction of differentiation at institutional leveland/or rationalisation at a programme level (reductionof cost subsidies). This points to the need for the de -nition, monitoring and costing of all aspects of studentchoice.

Equity

Not even the present British government claims thatmarketisation improves equity; it is, in fact, far morelikely to exacerbate, rather than to reduce inequalities between different socioeconomic or ethnic groups.This can happen in at least three ways: through theincreased importance of economic factors in studentchoice (including the increasing use of merit or non-needs based aid), through the unwillingness of prestig-ious universities to expand places in line with increases

in demand, and through strati cation (of universitiesand the social groups they serve). We should also notethe (mutually reinforcing) interaction between marke-tisation at university level and strati cation at second -ary education levels, not to mention the links betweenthe marketisation of higher education and other policiesto introduce market or quasi-market competition intoother parts of what used to be called the public sec-tor. This points to the need for state interventions to protect equity against market competition. Even if the basic causes of differential participation in higher ed-ucation lie much further back in the education system,we should

Documents

Ces if o Dice Report 213