Economics of Education Review 21 (2002) 367–380 www.elsevier.com/locate/econedurev

Financing student loans in Thailand: revolving fund or openended commitment? Adrian Ziderman Economics Department, Bar-Ilan University, Ramat Gan 52900, Israel Received 30 May 2000; accepted 3 November 2000

Abstract A student loans scheme (SLS) came into effect recently in Thailand, covering both upper secondary and tertiary level schooling. The central objective of the scheme is social — to increase access of poor students and to prevent student dropout. The loans scheme is highly subsidized owing to the extremely favorable repayment conditions, which in turn calls into question the longer-term financial viability of the scheme. The average repayment ratio on loans is only about 20%, while an overall loan recovery, taking into account repayment default and administration costs, is 10% or less. It is recommended that the loans scheme for the upper secondary schooling be converted to a grants scheme; alternative reforms are suggested to raise the average repayment and recovery ratios on loans for tertiary level students, closer in line with international experience.  2002 Elsevier Science Ltd. All rights reserved. JEL classification: I-21; H-52 Keywords: Educational finance; Student financial aid; Student loans; University finance; Cost-sharing; Thailand

1. Introduction A student loans scheme (SLS) has been established recently in Thailand, starting its operation in the academic year 1996/1997. As in the other countries where student loans schemes have been established, the Thai scheme covers tertiary education (comprising public and private universities, Rajabhat teacher training colleges and technical institutes); unusually, the Thai scheme also covers the upper secondary schooling. Throughout the world, student loans schemes almost exclusively relate to tertiary education; though there are a few notable exceptions.1 Yet while loans schemes for upper secondE-mail address: [email protected] (A. Ziderman). One is the Swedish loans scheme, but even here the loans element in the total aid package available to students is very small; another is the Japanese Scholarship Foundation, the most important source of financial assistance for students in the country, which covers both secondary school and higher education students. 1

ary schooling are unusual, it is at this level of non-compulsory schooling that enrollment rates fall off drastically and the risk of dropout increases. Poor students are particularly at risk, because of the high opportunity costs of studying rather than working. For this reason, government subsidy of the private costs of upper secondary education, particularly for the poor, is seen as an important element of social policy in Thailand. A central question posed in this paper is whether subsidized loans targeted at the poor (rather than grants) constitute the best instrument for achieving the desired outcome. Repayment conditions in the Thai SLS are extremely generous, implying an extensive loan subsidy. The introduction of soft terms of repayment may have resulted from the conception of the loan scheme as one that has essentially social objectives, with considerably less weight given to the financial aspects of the scheme. The open issue, however, is the resulting size of the loan subsidy; an overlarge subsidy brings into question the financial viability, and hence the long-term sustainability, of the scheme as a whole. Surprisingly, little attention

0272-7757/02/$ - see front matter  2002 Elsevier Science Ltd. All rights reserved. PII: S 0 2 7 2 - 7 7 5 7 ( 0 1 ) 0 0 0 2 4 - 3

368

A. Ziderman / Economics of Education Review 21 (2002) 367–380

has been paid to this issue in Thailand. Yet the sums involved are very large; annual budgetary allocations for student loans now exceed 20 billion Baht (a sum equivalent to some 10% of the national education budget). The plan of the paper is as follows. Sections 2 and 3 are contextual and comparative; Section 2 provides comparative data on the financial viability of loans schemes in other countries while Section 3 presents alternative objectives and policies for student loans schemes, based on international experience of student loans and contrasts this with the Thai scheme. Section 4 examines the size, growth and coverage of the scheme, in terms of financial commitments and number of borrowers. Section 5, focusing on the individual borrower’s loan account, poses the question: how much of a loan does a typical student repay to the fund? Section 6 examines the financial viability of the scheme. It estimates the proportion of the loans budget that is likely to be recovered by the loans fund after taking account of the probability of repayment default and administration costs. Building on these findings, Sections 7 and 8 suggest reform measures that might be adopted for student loans at the upper secondary and tertiary levels, respectively.2

2. Financial viability of student loans schemes: international experience Student loan programs have been developed in various forms in over 50 countries throughout the world.3 Table 1 provides some comparative figures on the financial viability of loans schemes in about half of these countries. Estimates are provided on the average repayment on a typical loan, as expressed by the repayment ratio (present value of total repayments in relation to loan size) and of the recovery ratio (present value of net repayments, including repayment default and administration costs, in relation to loan size). From the viewpoint of the individual borrower, the size of the subsidy received on the loan depends on the extent to which repayment conditions (interest rates, repayment linkage to inflation) depart from those on commercial loans. Following Johnstone (1986), we may regard this subsidy as a hidden grant to the student.4 The lower the repayment ratio on an individual loan, the 2 A broader discussion of many of the issues presented in this paper is given by Ziderman (1999), a paper prepared for UNESCO-Bangkok as part of the Asian Development Bank Social Sector Program for Thailand. 3 For a concise overview of student loans in an international context, see Woodhall (1994). Also see Albrecht and Ziderman (1991, 1992, 1993) for applications to developing countries. 4 Johnstone’s classic discussion of student loans in the context of ‘cost sharing’ appears to be the first to define and calculate the hidden grant element in subsidized loans schemes.

Table 1 Recovery from student loans schemes, selected countriesa Country

Loan repayment ratio

Loan recovery ratio (including default and administrative costs)

Kenya Venezuela Chile Honduras Indonesia Brazil Ecuador Jamaica Bolivia Denmark Mexico Japan Peru Australia United States (GSL) Finland Guatemala Norway Colombia Hong Kong Dominican Republic Province of Quebec, Canada Barbados Sweden

30 7 52 49 43 38 54 44 68 48 56 50 62 52

⫺3 ⫺8 18 27 29 29 30 30 33 38 39 40 40 43

71 65 82 67 71 57

47 48 50 52 53 53

75

57

69 87 72

63 67 67

a

Source: Carlson, 1992; Ziderman & Albrecht, 1995.

larger is the hidden grant element, and the more likely is the loan to be successful in encouraging educational enrollment of needy youngsters. However, lower hidden grants imply greater public subsidy. From the perspective of the government, the overall recovery rate depends not only on cash repayments (reflected in the size of the hidden grant received by students) but also on administrative costs (which are usually not passed on to the student) and the probability of repayment default. The financial efficacy of any loan program depends centrally on the size of the recovery ratio — the extent to which loans are repaid. It provides an indication of the program’s efficiency in achieving cost recovery or, alternatively, the extent to which a loans program is subsidized. The recovery rate on student loans depends crucially on four main factors: whether repayments are linked to inflation; the amount of interest subsidy on the loans; repayment losses owing to nonrepayment (default); and administrative costs. It is apparent from the table that, in practice, few student loans schemes have proved to be financially viable,

A. Ziderman / Economics of Education Review 21 (2002) 367–380

particularly in developing countries, in the sense that loan recovery typically is less than 50% and seldom exceeds 70%. The table provides a comparative background against which the subsidized Thai loans scheme may be viewed. A central finding of the research reported in this paper, developed in detail in Sections 5 and 6, is that the level of loan recovery in the Thai loans scheme is very low. With an interest rate far below market rates, repayments in nominal rather than real terms, and a long period for repayment, the Thai student loans scheme is highly subsidized. Even under the best possible scenario of minimal repayment default and low administration costs, repayments on the average loan will fall considerably short of loan disbursements. Given the comparative data for SLSs in other countries, as reported in Table 1, this outcome may not be an unacceptable one. As will be emphasized below, the objectives of the Thai scheme are predominantly social (designed to assist disadvantaged students) rather than aimed at enhanced cost recovery. This, in part, is expressed by the inclusion, most exceptional in terms of international experience, of upper secondary education within the scheme. A low average loan–recovery ratio, resulting from the subsidization of the scheme, may be less the result of financial negligence and more that of deliberate design intended to further the social objectives of the scheme. Many publicly provided goods and services are highly subsidized. But what level of subsidy is acceptable for the Thai scheme? Looking on comparative international experience, we may ask: how does the level of loans subsidy in the Thai scheme (or its inverse, loan recovery) compare with that of other countries with student loans schemes? Is Thailand to be located in the upper section of Table 1, together with the poorly performing countries? Or does Thailand share company with the better performing schemes listed in the lower section of the table? The greater is the weight given to purely social objectives in a loan scheme, the lower is likely to be the acceptable level of loan recovery. However, there are limits to this. From a purely financial point of view, very low levels of loan recovery, requiring a heavy, continuing call on the public purse, might be politically unacceptable over the longer term. 3. Loans schemes: alternative objectives and policies While student loans schemes are in place in many countries, the central objective and type of policies pursued differ from case to case. We may identify four different sets of objectives for student loans schemes, which in turn will influence the design and operation of the scheme as a whole, as well as its financial sustainability; in practice, any given scheme may incorporate more than one objective.

369

3.1. Pure cost recovery Universities throughout the world, and particularly in developing countries, are under financed. Governments have responded to the growing social demand for higher education through policies leading to expanded student enrollments; yet, owing to national budgetary constraints, the growth in student numbers is largely unmatched by similar additional government funding.5 Many university systems have turned to greater cost recovery in an effort to tap additional sources of funding. The main thrust of these policies is to be seen in the introduction, or increase, of student payments for services received; those may take the form of higher, more realistic tuition fees or increased payments for subsidized lodgings and meals.6 However, policies aimed at raising students’ fees, whether for tuition or living expenses, may clash with vested interests. This will generate opposition to the imposition of student fees, which may represent a multiple of current salary levels. Student loans, obtained through the banking system, could ease these payment burdens, but they may be unavailable; banks are notoriously loath to lend for education courses. Hence the role for a government-backed students loans scheme, offered at commercial rates, to fill this gap. Students are able to finance their education and living expenses through borrowing; the repayment burden is eased by the expected enhancement of earnings, which the additional education makes possible. While loan schemes primarily concerned with cost recovery are also frequently subsidized and targeted at the poor, these elements are not an integral part of a cost recovery loans scheme. In principle, it should be offered at market interest rates and available to all, not only the poor. 3.2. Facilitating higher education expansion Responding to the pressures of growing social demand for higher education expansion will require sizeable increases in public expenditures; these increases, as we have argued, could be contained by the generation of offsetting additional revenues from increases in student fees, leading to reduced public institutional support. A complementary measure is to encourage the growth of private educational institutions. Students pay full costs,

5 For a discussion of the ongoing financial crisis facing universities in developing countries, see World Bank (1994) and Ziderman and Albrecht (1995). 6 Incidentally, the introduction of a regime of sizeable student fees will have additional effects on the education system. It leads to a changed relationship between institution and students, in the direction of a provider–client relationship and a greater student voice in the affairs of the institution.

370

A. Ziderman / Economics of Education Review 21 (2002) 367–380

with a minimal burden on the public purse. However, full cost fees are likely to be sizeable and beyond the reach of large segments of the population. A student loans scheme may have a central role to play in easing the burden of private fees, particularly so if private education is to be widely available and not to remain the purview of the rich. However, a student loans scheme introduced for this purpose does not, in principle, need to be subsidized. 3.3. Easing student financial burdens Even when tuition and other education-related fees are minimal, students (including the non-poor) may face considerable financial burdens. University-level students are at an age of legal and financial independence; yet, potential earnings are forgone while studying, and living expenses may be sizeable, especially when the student does not attend a local university. Financial pressures, which may have negative effects on student motivations and performance, can be mitigated by the broad availability of student loans. While such burdens may fall relatively heavily on the poor, in principle loans for this purpose could be made available for all students, including the non-poor, but not subsidized.

loans, in terms of grace periods for repayment, belowmarket rates of interest and repayments in current rather than real terms. However, these subsidies will result in only a partial recouping of loans overall. The part of the loan that is not repaid may be seen as a ‘hidden grant’ to the borrower, the size of which is a reflection of the degree of subsidy built into the scheme. In this sense, loans offered at commercial rates on the one hand, and outright grants on the other, might be seen as two extreme points on a continuum, with a subsidized loan lying somewhere in between. The larger the loan subsidy, the greater is the hidden grant element. From this arises a central policy issue: given the higher administrative costs of loans compared with grants and the probabilities of repayment default, at what level of subsidy does a grant become a more cost effective instrument than a subsidized loan? This issue, as we shall see, is very relevant to the situation in Thailand, notably at the upper secondary level of schooling. 3.5. The Thai loans scheme How far are these four objectives reflected in the Thai loans scheme? A note to the act giving formal statutory status to the loans scheme (Educational Loan Fund Act, 1998),7 explains its purpose as follows:

3.4. Increasing access None of the three above reasons for introducing a student loans scheme make out a case for highly subsidized loans. Subsidies are justified when the objectives of a loans scheme is to increase the educational participation of the poor. In many countries the relatively low enrollment of poor and disadvantaged youth in non-compulsory education is a cause of social concern; increasing the access to schooling among these segments of the population has become a major element in educational and social policy. There is a broad consensus that clear financial incentives need to be offered, not only to overcome the burden of fee payments and living expenses but also to offset parental resistance to reductions in family income and the risk that the benefits of the educational process may not be sizeable. The traditional, and most effective, method of enhancing the educational access of the poor has been through the provision of means-tested grants (scholarships) to cover tuition fees (where schooling is not free) and usually for living expenses as well. This still remains the dominant approach in place for secondary education. However, a widespread scholarship scheme is likely to be expensive; the use of loans rather than grants offers a method that both increases the access of the poor and contains public expenditure over the longer term as the loan repayments build up. To increase effectively the education access of the poor, loans have to be offered at attractive rates. Hence the justification for subsidized

…there is a need to develop human resources in order to achieve economic growth and to increase the competitive capacity of the country. Given these needs, educational development needs to be accelerated. Enlarging the educational opportunities of students from low-income families can solve existing problems of educational inequality within society. This will play a major part in improving general living standards. To achieve these goals, it is necessary to establish the SLS. In the act itself, the only objective mentioned is “to lend money to poor students for tuition, educational expenses and other expenses necessary for living during studying” (Clause 5). From these and other citations, it is clear that the objectives of the SLS are primarily social: the availability of student loans would lead to greater educational opportunities for the poor, higher living standards and greater degree of equality in the population. The longer run objectives — of enhancing the nation’s human capital, competitiveness and development — are economic; but the human capital development will be secured by

7 The decision to establish the student loans scheme was made by the Cabinet in March 1995 and came into effect in 1996. It operated under this Cabinet resolution until the promulgation of the Act in March 1998.

A. Ziderman / Economics of Education Review 21 (2002) 367–380

targeting the needy. No other objectives are recorded. The exclusive concentration on social objectives is surprising. We have noted above that most student loans schemes are multi-objective, with due emphasis placed on economic outcomes — particularly, enhanced cost recovery through higher tuition fees and lower public expenditures. Thus the central aim of the Thai loans scheme is to increase the access of the poor to upper secondary and tertiary education. For this reason, the scheme attempts to target loans to needy students, under extremely favorable repayment conditions. However, the way the scheme has worked in practice has resulted, de facto, in other objectives for loans schemes, discussed above, also being covered by the SLS. The SLS has served as a means of easing the financial burdens of the non-poor. An upper eligibility limit on parental income of 300,000 Baht has been set by the national loans committee. This may be compared with the official poverty measure of 53,280 Baht per year for a family of five. In practice, the Ministry of Education has reduced this upper eligibility limit to 150,000 Baht for loans to non-university students, administered by that ministry; but this lower ceiling still exceeds the official poverty line by a factor of three. Thus many non-poor students, particularly those enrolled at universities, are in receipt of loans. While the SLS is not seen as a vehicle for extensive cost recovery, the introduction of the scheme in 1996 coincided with increases in tuition fees at some public educational institutions. General upper secondary school fees, stable for many years, were raised by over 70%, placing them at par with fees at public vocational schools. Tuition fees at Rajabhat institutions (teacher training colleges) were also raised sharply; fees at RIT institutions (technical institutes) and public universities have shown a more steady, upward trend over recent years. While tuition fees at public universities are now three times higher than some seven years ago, they still represent only a small proportion of unit costs It appears to be government policy to encourage the growth of private education institutions (particularly universities) as a means of increasing the number of student places, at low public cost. The average size of loan given to students at private institutions exceeds by a considerable margin those for students enrolled at public institutions, because of large differences in tuition fee levels. As all student loans are highly subsidized, the availability of loans to students enrolled at private education institutions results in the subsidization of the costs of private education. In 1998, 20% of all private university students were in receipt of loans and for private upper secondary schools the proportion approached a third.

371

4. National loans scheme budget

In this section, we present data on the sums committed to student loans in these early years of the scheme; in the following sections we discuss how much of this loan funding is likely to be repaid into the loans fund. Table 2 shows the main developments since the scheme came into operation in 1996, in terms of the level and growth of government spending on the loans scheme. While the main pattern of spending seems to be quite clear — an increasing level of expenditure overtime as the scheme builds up — we shall argue that in reality the details contained in Table 2 display, on the contrary, a decreasing level of government commitment to the loans scheme. Confusion concerning the level of financial commitment to the loans scheme has arisen because the information reported often uses alternative definitions. The figures may relate to the fiscal year (October– September) or the academic year (May–April); they may be presented on the basis of a semester or for the whole academic year; they may comprise all loan recipients or only new borrowers. Table 2 presents details of expenditure under the loans scheme, employing alternative definitions. Formally, the level of government commitment to the SLS is defined by that part of the national budget that is approved by parliament for allocation to the revolving fund concerned with the SLS — the Student Loan Fund. This annual allocation, which covers the fiscal year, is shown in the top block of Table 2. These fiscal year allocations may be redistributed according to the academic year, the actual calendar period covered by individual loans. The third block shows actual loan disbursements, by academic year. In both cases, a sharp upward trend in annual allocations to the loan scheme is indicated. However, an examination of block two in the table shows that this increase is more apparent than real. To explain this, a brief comment on how the annual budgetary allotments are made to the student loan fund is necessary. Current student borrowers who apply for a loan for an additional year of study (‘old recipients’) are regarded as a fixed charge on the fund; loans tend to be renewed at the same monetary level as of the previous year. The Budget Bureau estimates the budget allocation required for old recipients on the basis of last year’s allocations. These allocations have risen each year, with the increasing number of students entering the scheme. In contrast, annual budgetary allocations for new borrowers in the scheme (‘new recipients’) have remained stable in current terms (except for a temporary increase in 1997) at 6000 million Baht. However, in real terms, allocations for new recipients have fallen sharply each year owing to the fall in the purchasing power of the Baht resulting from ongoing inflation. In actual terms, the 6000 million Baht allocation that was planned for 1999 is equivalent

372

A. Ziderman / Economics of Education Review 21 (2002) 367–380

Table 2 Student loan scheme: planned and actual budget, number of loan recipients, 1996–2000a Fiscal year

1996

1997

1998

1999

2000

I Budget allocation (million baht)

3000

8450b

18,000b

20,000b

[26,800]c

Academic year

1996

1997

1998

1999

2000

II Budget allocation (million baht) Old recipient New recipient Total

– 6000 6000

4836 9900 14,736

12,250 6000d 18,250

18,300 6000 24,300

– 6000 6000

III Actual loan disbursement (million baht) 3764

12,151

17,605e

Loan recipients (number) Ministry of Education Ministry of Univ. Affairs Total recipients

371,130 64,288 435,426

510,187f 103,615f 610,802f (632,639)e

a b c d e f

127,740 20,704 148,444

Source: Bureau of Budget (04.12.98) and Ministry of Finance (24.09.98). Revised downwards from 10,950 million Baht. Estimate. Actual allocation reduced to 5,500 (500 carried forward to 1999): minutes SLC././1998. As of 24.11.98. As of 24.09.98.

to only some 4800 million Baht, an erosion of nearly a third of its nominal value. The fixed allocation of 6000 million Baht in current terms for new recipients is arrived at arbitrarily, and is not based on any realistic examination of the expected student demand or need for new loans. The budgetary justification, prepared by the Ministry of Finance (in the future, this is to be the responsibility of a new loans scheme office, with expanded administrative functions), is made out in terms of supplying loans to 100,000 new recipients at an average loan value of 60,000 Baht. These figures, which seem to be drawn from early planning documents, have no relationship with actuality; e.g. in 1997, there were already well over 300,000 new recipients, with an average size loan of only some 30,000 Baht. Clearly, the decision on the size of the budget allocation precedes its justification, rather than the opposite. With the growing numbers of new loan recipients annually, a sharp decline in the average value of individual loans has resulted, both in real and in nominal terms; the figures are presented in Table 3. This might seem to imply that the loans scheme is under financed; however, in a real sense, the case is opposite. A dramatic expansion of the number of student loans, largely uncontrolled from the center and accruing to non-poor students, has resulted from insufficiently tight criteria for loan eligibility and weak targeting. It is clear that the scheme has now grown considerably beyond the original plans. Table 4 presents some of the major outcomes of the

scheme, in relation to plans. We note both the revisions of planned number of borrowers and the tendency for outcomes to outstrip plans. Original plans (Column 1) are drawn from the Ministry of Finance planning document (Ministry of Finance, 1995) on which the scheme was based. The plan envisaged a build-up of loan recipients in the early years, reaching a steady state of 300,000 borrowers in the fourth year of the scheme. These plans have been revised upwards and in practice, have been overtaken by events. The target of 300,000 borrowers was exceeded in the second year of the scheme; currently the target number of borrowers is outstripped by a factor of 2.3. The table shows planned loans expenditures (with an assumed 5% increase in the rate of individual student loans, in line with inflation) and actual expenditures which are not raised in line with inflation. While planned and actual loans expenditures were broadly in line for the first two years of the scheme, actual expenditures have now forged ahead. In the fourth year of the scheme, actual expenditures will be 45% in access of planned levels. The scheme was planned originally to cover some 9% of enrollments at each educational level; the coverage of the scheme is far in excess of this. By 1998, the overall coverage exceeded 20% of all enrollments, nearly 25% for secondary schooling and 15% at the tertiary level. Clearly, large amounts of public funding are being put into the SLS in its early years. But will these allocations be returned to the loans fund over the longer term? The next two sections are concerned with this question.

A. Ziderman / Economics of Education Review 21 (2002) 367–380

373

Table 3 Average loan size, old and new borrowers, student loan scheme, 1996–1998 (Baht)

Upper secondary Public general Public vocational Private Tertiary RIT Rajabhat Public university Private university

1996

1997

1998

New borrowers Old borrowers

New borrowers Old borrowers

New borrowers Old borrowers

13,725 27,813 22,122

16,894 36,927 28,692

13,874 23,544 18,891

14,622 26,851 20,962

9,809 20,061 17,900

13,753 36,652 24,678

33,950 29,626 43,558 66,199

42,343 39,010 55,120 85,231

35,048 29,560 55,822 86,237

37,367 32,740 55,574 85,937

26,386 30,459 44,320 62,146

38,609 35,443 55,093 85,146

Table 4 Number of borrowers and total loans expenditure, student loan scheme: planned and actual Year

Number of borrowers

1 2 3 4 5 a b c

Loans expenditure (billion Baht)

(1) Planned a

(2) Actual

(3) (2/1)

(1) Planned a

(2) Actual

(3) (2/1)

132,000 214,000 278,000 300,000 300,000

148,444 435,246 632,639 [700,000]c –

1.12 2.03 2.28 2.3 –

4.000 10.726 14.592 16.750 17.588

3.700 12.151 18.250b 24.300b –

0.93 1.13 1.25 1.45 –

As set out in MOF Loans Scheme Planning Document (Ministry of Finance, 1995). Budget allocation. Estimate.

5. Financial analysis: the individual loan account

5.2. Computing the repayment ratio

5.1. Repayment conditions

Based on the formal repayment conditions of the loan, we compute the repayment ratio on typical loan programs (total loan repayments as a ratio of the initial loan). Equal annual loan disbursements are assumed; repayments of the total nominal loan are made according to the schedule shown above, starting from 1.5% of the loan size and rising to 13%, plus 1% interest on outstanding debt. Details of the methodology used for the financial appraisals that follow are provided by Ziderman and Albrecht (1995). Discounted cash flow techniques are used to compare the total of annual loans disbursements with the stream of loan repayments, requiring the discounting of payments received in the future, to reflect their lower value in terms of the present. We make four alternative assumptions about the appropriate discount rate, and test the sensitivity of the results to alternative discount rates.

Loan repayments commence after a two-year grace period, following completion of schooling. Repayment, spread over 15 years, is in nominal terms and covers the principal and interest charged at 1% annually. Annual repayments of principal, as a percent of nominal loan size, are graduated, as follows: Year:

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Total

Repayment (% of loan)

1.5

2

3

3.5

4

4.5

5

6

7

8

9

10

11

12

13

100

We estimate the repayment ratio (and hidden grant) on typical loan programs from the borrower’s perspective in this section; in the following section we add in the effects of default and administration costs to measure the overall level of recovery of the loan: the recovery ratio.

374

A. Ziderman / Economics of Education Review 21 (2002) 367–380

First, the cash flow is discounted, using estimates of future opportunity costs of capital (the loss from the failure to use these funds in alternative uses), which may vary from year to year. Opportunity cost estimates are derived as the difference between estimates of future nominal interest rates and future rates of inflation.8 We also provide estimates using three alternative fixed discount rates: of 8, 6 and 4%, respectively. Results for three typical loans programs are provided in Table 5: three-year programs such as upper secondary and RIT courses; four-year undergraduate programs, including those at universities and Rajabhat institutes; and combined programs, such as upper secondary courses followed by study for a first degree. Reading across the table, it is noted as would be expected, that the longer the course of study, the lower are the total repayments received in relation to loan size. Reading down the table, we may see the effect of alternative discount rate assumptions on loan repayment. It is noted that the repayment ratio with varying forecast discount rates is very similar to that of the 6% discount rate assumption. All the repayment ratios shown in the table are relatively low, within a range of only 14–28%. In all the results presented subsequently, we use a 6% discount rate. The very low repayment ratios shown in Table 5 may seem surprising for a loan program, but they reflect four ingredients of the present loans scheme, that inescapably Table 5 Repayment ratio and hidden grant on selected student loans programs: alternative discount rates (%) Discount rate

Program Upper Undergraduate Upper secondary (4 years) secondary (3 years) plus undergraduate (7 years)

4% Repayment ratio Hidden grant 6% Repayment ratio Hidden grant 8% Repayment ratio Hidden grant Opportunity cost of Repayment ratio Hidden grant

28 72

26 74

22 78

23 77

21 79

17 83

19 17 81 83 capital (varying %) 22 20 78 80

14 86 19 81

8 Opportunity costs of capital estimates were provided by Dr Kitti Limskul, of Chulalongkorn University

lead to low levels of repayment. These are: that interest is levied at a token rate of 1% only; that no interest is charged on the outstanding loans during periods of study or grace periods; that repayments are spread over a 17year period after completing the study, initially at very low rates of repayment; and, most important, that repayments are made in nominal terms. 5.3. The repayment burden One justification for soft loan repayment conditions in many national student loans schemes, leading in turn to low levels of repayment, is to avoid imposing unduly harsh repayment burdens on borrowers, especially needy students. To minimize this disincentive to taking up loans, most governments subsidize loans. But over large subsidies (as in the Thai case), may undermine the very purpose of the loan, as opposed to a scholarship, scheme. To probe how far such considerations apply to the Thai scheme, we estimate very roughly the annual repayment burden falling on borrowers (annual repayments in relation to expected income). Again, we do so for three programs: upper secondary schooling; undergraduate courses at the university; and a combined course of study comprising upper secondary followed by university study. As labor market earnings of females are usually lower than those of males, separate repayment burden estimates are provided, by gender. For the repayment calculations, we measure the annual repayments from data on average annual loan disbursements (as shown in Table 3). Expected future annual income estimates are derived from age–earnings data by education level, drawn from the 1996 Socio-Economic Survey (National Economic and Social Development Board, 1996). It is assumed that real incomes grow at a rate of 3% annually. Two sets of calculations are made. The first expresses repayments as a percentage of income and provides an indication of the weight of the burden of repayment falling on the borrower, during the years of repayment. The second calculates the ratio of repayments to incremental income (the additional income received as a result of completing the stage of education in question) and measures how much of the increment in earnings resulting from extra education remains, after loans repayment. Results for the repayment burden related to the total income are shown in Table 6, for loans for upper secondary programs of study, for university and for upper secondary followed by university study, respectively. The table provides summary results (in terms of present values) of the ratio of total repayments to total income over the 17-year repayment time horizon. Overall, the total loans repayment comprise only some 2–4% of the total income over the period of repayment.9 Interesting 9

The annual repayment burden (annual loan repayment in

A. Ziderman / Economics of Education Review 21 (2002) 367–380

Table 6 Loan repayment as a percentage of annual income (in terms of present value at 6% discount rate) Program

Upper secondary University Upper secondary followed by university

Percent of annual income Males

Females

2.0 2.2 2.4

1.9 3.5 3.6

differences in the repayment burden by gender are shown in Table 6. On loans for university study, the annual repayment burden of females exceeds that of males by some 50%, because of higher male wages. However, on loans for upper secondary schooling, there is little difference in the repayment burden by gender; indeed, it is somewhat lower for females, reflecting the slightly higher annual wages for females at that education level. Repayment as a percentage of incremental income is obviously much higher. While the effect is to reduce the private rate of return on university education investment, repayments are not onerous. For university education, the repayment burden (in terms of the present repayment) is 3.5% for male graduates and 6.1% for females. The general conclusion emerging from this analysis is that the Thai student loans scheme is overly generous, providing borrowers with unnecessarily highly subsidized loans (effectively hidden grants), which may be repaid effortlessly out of higher income received on completing courses of education.

6. Loan recovery with default and administration costs (the recovery ratio) Thus far we have looked at the repayments paid by the typical borrower and at the size of the hidden grant received if the loan is repaid in conformity with the established repayment conditions of the loan scheme. However, the repayment ratio (hidden grant) fails to show the overall recovery (loss) to the loans fund, from a national point of view. On the whole, the fund will receive back less than is indicated in the repayment ratio, because some students will not meet their repayment obligations and there are sizeable administration costs, which are not passed on to borrowers. In this section we measure the average recovery ratio, taking account of relation to annual income) is very light, in the range of only about 1.4–4.3% annually. Full results are available from the author.

375

the probability of repayment default and the incidence of administration costs. 6.1. Repayment default While there has been some limited voluntary repayment of loans, obligatory loan repayment under the scheme had not yet commenced at the time of writing. International experience from loans schemes in developing countries shows considerable variations in default levels; however, many schemes display a default ratio in the range of 5–10%. In the case of the Thai scheme, the low repayment ratio on loans and the light repayment burden are likely to lead to minimal default due to repayment difficulties by borrowers. In the absence of any information on the expected levels of loan default for the Thai scheme (including evasion and repayment in arrears), we simulate the effect of default on loan recovery, using alternative assumptions about the extent of default. The probability of annual default is expressed as a percentage of annual repayment obligations (repayment of principal plus interest). Simulations using alternative rates of default ranging from 0 to 10% (not reported here) rendered rather similar results. Given the lack of sensitivity to alternative default rates, we adopt the optimistic assumption of a low 5% default probability. 6.2. Administration costs To establish the recovery ratio, administration costs must also be taken into account. For our present purpose, these would be expressed in terms of the average (per borrower) annual administrative cost of running the loan scheme. No reliable data are available; official budgetary allocations for administrating the loans scheme are only partial, because a significant proportion of costs are absorbed by the institutions running the scheme, including the Krung Thai Bank central office and all the educational institutions and schools which distribute the loans. Moreover, the existing administration cost allocations are thus far concerned only with initial processing of the loans; subsequent overall maintenance costs and collection costs are still unknowns. Some guidance is available from international experience. While there is no detailed comparative study of administrative costs of loans, a study of loans schemes in Latin America found many countries with administrative costs of some 2% of outstanding debt; this represents an overall cost of approximately 10% of the total value of the loan (Carlson, 1992). Similar cost estimates were found in many other schemes in developing countries (Ziderman & Albrecht, 1995). Table 7 presents alternative estimates of the net average recovery ratio on loans programs. The first column shows the repayment ratio (as given in Table 5), i.e. with

376

A. Ziderman / Economics of Education Review 21 (2002) 367–380

Table 7 Recovery ratio on loans (with 5% default, alternative administration costsa

6.3. Revolving fund or open-ended commitment?

Administration costsb No default

Default rate: 5%

0%

0%

2%

1%

0.5% 0.25%

22 20 16

10 8 5

16 14 11

19 17 13

Upper secondary 23 Undergraduate 21 Upper secondary 17 plus undergraduate a b

ery rate increases to about double than that shown for the 2% administration cost case.

20 18 15

6% discount rates. Percent of outstanding debt.

no default and zero administration costs. Compared with this, subsequent columns assume a default rate of 5% and alternative levels of administration costs. With an assumed 2% rate for administration costs (as a percentage of outstanding debt) the recovery ratio falls sharply to single digit estimates. For the 6% discount case shown, the recovery ratio ranges from 5 to 10%, indicating that loans would be only marginally preferable to scholarships on purely financial grounds.10 However, it is possible that administrative costs of the Thai scheme will fall significantly below 2%. The small average size of loan, particularly at the upper secondary level, does make loans relatively costly. Against this, the Thai scheme is very large in terms of annual number of borrowers; loans schemes in developing countries typically number in the tens of thousands only; this should yield sizeable economies of scale in managing the loans scheme, as the scheme expands. This view is buttressed by the findings of a special exercise we conducted, to estimate total current administration costs of the loans scheme (including hidden costs absorbed by the participating institutions and organizations).11 Administrative costs are estimated to lie at a level of some 600 million Baht in 1998 (or some 3% of the annual loan disbursements). Rough calculations, including augmented costs estimates to include future debt maintenance and collection costs and estimates of probable levels of total loan debt, suggest that administration costs (as a percentage of outstanding debt) could fall to 0.5– 1%. At these administration cost levels, the loan recov-

10 The recovery ratio is somewhat lower for an 8% discount rate, ranging from 4 to 7% (again, with 2% administration costs and 5% default). 11 Walairat Asaves of the Office of the National Education Commission, Bangkok, carried out the survey. Results are available from the author.

A self-financing student loans scheme requires initially large and growing disbursements, as the scheme builds up. In principle, total disbursements should then level off as the first cohorts of borrowers complete their studies and exit the education system; subsequently the loans fund becomes self-sustaining as loan repayments increasingly finance loans to new borrowers. The expectation that the Thai SLS will operate in this way seems to be widespread both in central government circles and at the educational institution level. However, such a positive outcome represents little more than wishful thinking. Using a set of realistic assumptions concerning administration costs, default and the appropriate method of investment appraisal, our calculations have shown that it is unlikely that more than 10% of the average loan, in real terms, will be returned to the fund. More optimistic assumptions may partially raise the recovery ratio, but still to less than 20%. Thus, as operated at present, the loans scheme will not be self-financing, even over the longer term; indeed, present arrangements imply a large and continuing call on the public purse. Given the clear social objectives of the SLS, some continuing level of commitment of public funds to the scheme is both to be expected and broadly acceptable. However, it seems clear that a 90% level of subsidy is excessive, particularly given that the scheme is not well targeted, as intended, to reach poor students.12 Moreover, if student numbers continue to grow and the demand for loans continues apace, this commitment will be increasingly onerous, particularly if the current extent of support is maintained (in terms of the ratio of the number of borrowers to total student enrollments). Indeed, the rhetoric aside, signs of retrenchment on the part of the government are already visible, as we have observed in previous sections. It is true that the value of total annual loan disbursements continues to increase, but this is mainly to meet the existing commitments to established borrowers. The loans budget for new borrowers is planned to remain fixed at 6 billion Baht in nominal terms, at least until the year 2001; this implies a declining level of total support in real terms. The fixed loans budget, together with a growing demand for student loans, has resulted in a drastic reduction in the average loan size for new recipients in 1998. The average loan size in 1997 was some 30% higher in nominal terms than in 1998; measured in real terms, the gap would be even greater. This unplanned outcome further

12 The efficacy of loans targeting in the Thai SLS is discussed in a complementary paper by Ziderman (2001).

A. Ziderman / Economics of Education Review 21 (2002) 367–380

militates against the objectives of aiding needy students and the principle of horizontal equity. It is government policy to encourage the growth of private universities as a means of increasing the number of university places, at low public cost. Loan levels to students at private institutions (particularly universities) exceed those for students enrolled at public institutions, because of considerable differences in the tuition fee levels. The low level of expected repayments in real terms (in relation to the size of a loan) implies something like an 85% subsidy of the costs of private education. Thus the expansion of private education, with a growing commitment to provide highly subsidized loans, may not represent a cost-effective method of increasing university enrollments. Clearly, the loans scheme, as it now stands, with its excessive level of subsidy, is not financially viable. Change can go in one of two directions. With such a high hidden grant element, a strong case could be made for replacing the present loans scheme by a regime of student grants (or scholarships). In order to maintain the present level of net costs, the scholarship scheme would operate at a somewhat lower level of total budgetary support. There would be considerable savings in administration costs and, of course; the issue of repayment and default does not arise. Alternatively, steps could be taken to enhance the financial viability of the scheme by reducing the hidden grant element, by replacing the over-generous repayment conditions with more realistic ones that are somewhat closer to those of the market. This would imply, as we argue in Section 8, a combination of higher interest rates, a shorter repayment horizon and the revision upwards of annual loan repayments, to correct for inflation. Our central recommendation, elaborated in the following sections, is a dual one. At the upper secondary level, the loans scheme would be replaced by a system of targeted scholarships. In parallel, the loans scheme would be retained for students at tertiary level institutions, but it would be radically reformed. Reform, in addition to taking the direction of improved targeting and the unification of loans conditions across all tertiary level institutions, would improve the financial viability of the scheme through a hardening of repayment conditions. 7. Replacing loans by grants at the upper secondary level The extremely low level of loan recovery for upper secondary student loans indicates the need for reforming the scheme. While one possibility of reform is the financial strengthening of the scheme, we argue strongly for the dismantling of the system of student loans at the upper secondary level and its replacement by scholarships (grants) to needy students. A number of considerations underlie this radical recommendation.

377

First, it would be more effective in achieving the objectives of the SLS at the upper secondary level. At present levels of subsidy, the loans scheme is not sustainable. The hidden grant element is so substantial that in practice the loans scheme almost amounts to a grants regime at present; yet the effect of an overt grants scheme in influencing poor families to continue with their children’s schooling is far sharper than for loans. Moreover, administration costs of a grants scheme are considerably lower than those of running a loans scheme. The alternative, of increasing the scheme’s financial viability through stricter repayment conditions, would considerably weaken its role of encouraging the poor to continue study through the upper secondary level. Very few countries have loans for students of school age: over 50 countries do have loans programs for college level students. Technically, there is little difference whether the borrower is a school or a university student; however, in the former case, there may be an ethical issue involved with parents entering underage youngsters into debt repayment obligations that extend up to a 20year time horizon. Ongoing plans to make upper secondary education free for all youngsters, as outlined in the education reform act in 1999, considerably weaken the case for loans at the secondary level. Indeed, it must call into question the continuation of loans for secondary school students. At the present time, a student loan may be divided into two elements: a loan for educational expenses which is paid directly to the institution and a loan for living expenses which is received by the student. Educational expenditures represent investments in human capital for which loans are clearly appropriate; living expenses are consumption items for which subsidized loans are less clearly justified. Currently, the loan is provided as a package covering both elements. Once living expenses alone remain, the justification of a loans program is less clear, though it is true that very many countries do have loans programs that cover living costs, but at the tertiary educational level. The move to a straight grant for living expenses, targeted to the poor, would be more appropriate, given the smaller average loan size involved, following the implementation of government policy to cancel upper secondary school tuition fees. Indeed, the introduction of grants would bring the upper secondary level segment into line with lower levels of education, where the practice of providing grants to needy students is established. For example, an emergency scholarship scheme for primary and lower secondary school dropouts was introduced in 1998, at the height of the economic crisis. 8. Loans for tertiary level students: improving financial performance We have recommended (though not only for financial reasons) that the upper secondary element of the loans

378

A. Ziderman / Economics of Education Review 21 (2002) 367–380

scheme be converted to one of grants. At the tertiary level, even if administration costs and default are kept optimistically low, the resulting net recovery ratio of 14– 17% for the 6% discount rate case, clearly indicates that the scheme is unsustainable over the longer term. However, it is possible to enhance the financial viability of the loans scheme, bringing it into line with best international practice. While this can be achieved by hardening the conditions of repayment, such changes will imply higher levels of annual repayments which may be excessive in relation to expected incomes. In this final section, we discuss alternative policies that may be adopted to raise the rate of loan recovery. We then probe the efficacy of these alternative measures in improving the recovery ratio and follow this by an examination of the feasibility of such reforms. We confine the results presented to the case of undergraduate study at university (simulations for other programs of study are available on request). There are two reasons for this. First, the intention is to provide general broad areas of magnitude only and the results for different levels and types of courses do not differ greatly. Secondly, and more important, as we have recommend that the loans scheme be abrogated at the upper secondary level, it did not seem appropriate to report detailed simulations for the upper secondary level case here.

8.2. Improving the recovery ratio

Similarly, as the scheme becomes more effectively targeted on the poor in the future, default will increase as lower income students are higher-risk, displaying a higher probability of repayment default (Greene, 1989; Dynarski, 1994). However, we have noted above that recovery ratios are not sensitive to repayment default in the 5–10% range. For this reason, the results presented in Table 8 assume a 5% rate of repayment default; however, as the recovery ratios are sensitive to the assumed level of administrative costs, we simulate recovery ratios for alternative levels of administration costs. The recovery ratio under present policies is shown in Box A-I, for alternative administration costs and 5% default. Moving across the table shows the effect of higher interest rates in raising the recovery ratio; moving down the table we see the effects of other policies (shorter repayment horizon and inflation linkage). With full inflation linkage, the shorter payment horizon and a 5% interest rate, the overall recovery ratio reaches 70%; on the basis of a comparative international experience, this would be an extremely favorable outcome for a student loan program of this type. However, it is unlikely to prove feasible. Not all the policy options presented in the table are viable, because of their deleterious effect in raising the repayment burden falling on borrowers. In the remaining part of this section, we examine the efficacy of introducing stricter repayment conditions by simulating their effect on the repayment burden. Separate simulations were run measuring the repayment burden for each of the options shown in Table 8, in terms of the ratio of payment on the loan to the probable level of annual income (for each year of repayment). As it is cumbersome to present the full detailed simulations, we provide only the summary results in Table 9. The figures in the table show the ratio of the present value of repayments, to the present value of annual incomes, over the total repayment period. We have noted (Table 6) that under present repayment conditions first degree graduates incur a very light repayment burden of only 2.2 and 3.5%, for males and females, respectively, in the present value terms (Box A-I of Table 9). From Table 9, we may note how the repayment burden rises in response to increasingly stringent repayment conditions. As expected, the tightening of loan repayment terms results in a higher burden for females than males; an exception is the last scenario, number IV (linking to inflation plus shortening the repayment period to 8 years).13 The simulations in Table 9 provide a general framework which can aid the decision-maker in developing new repayment conditions. The table suggests that raising the rate of interest payable on loans or shortening the

Recovery ratios for alternative repayment conditions are given in Table 8. The tightening of repayment conditions, resulting in increased repayment burdens on borrowers, is likely to lead to greater repayment default.

13 A referee comments that this is likely to be the result of gender differences in the pattern of lifetime earnings (presumably linked to child care patterns of graduate women).

8.1. Lowering student subsidies: three instruments The recovery ratio may be enhanced by increasing the level of repayments in relation to loans (reducing the hidden grant) or by economies in administration costs. Assuming that the latter are forthcoming, we concentrate on the effects of a policy change to decrease the subsidy by raising the repayment level per loan. Basically, this requires a tightening in the present liberal repayment conditions. We consider three instruments, which may be applied individually or in concert. Rates of interest on the principal can be raised closer to market levels; annual repayments can be made in real rather than nominal terms; the 15-year repayment period can be shortened to provide a more rapid payback. We assume the following repayment schedule spread over eight years (derived by merging repayment percentages for adjacent years), as follows: Year:

1

2

3

4

5

6

7

8

Total

Repayment (% of loan) 1.5 5.5 7.5 9.5 13.0 17.0 21.0 25.0 100.0

A. Ziderman / Economics of Education Review 21 (2002) 367–380

379

Table 8 Recovery ratios on loans: effect of changes in repayment conditions at alternative levels of administration costs undergraduate programsa Loan repayment conditions

I. Present conditions (non-linked to inflation and loan repayment horizon of 15 years) II. Non-linked to inflation and short repayment horizon (8 years) III. Linked to inflation and long repayment horizon (15 years) IV. Linked to inflation and short repayment horizon (8 years) a

A

B

C

Interest rate: 1% (current rate) Interest rate: 3%

Interest rate: 5%

Administration costs

Administration costs

Administration costs

2%

1%

0.5%

0.25% 2%

1%

0.5%

0.25% 2%

1%

0.5%

0.25%

8

14

17

18

13

19

22

24

19

25

28

29

18

23

26

27

22

25

30

31

26

31

34

35

30

40

44

47

41

51

56

58

53

62

67

69

44

52

55

57

52

59

63

64

59

66

70

72

Assumptions: 5% default; 6% discount rate.

Table 9 Present values of repayment burden (6% discount): Loan repayment as a percentage of annual income, alternative conditions, undergraduate programs (%)a Loan repayment conditions

I. Present conditions (Non-linked to inflation and loan repayment horizon of 15 years) II. Non-linked to inflation and short repayment horizon (8 years) III. Linked to inflation and long repayment horizon (15 years) IV. Linked to inflation and short repayment horizon (8 years) a

A

B

C

Interest rate: 1% (current rate)

Interest rate: 3%

Interest rate: 5%

Males

Females

Males

Females

Males

Females

2.2

3.5

2.8

4.4

3.3

5.3

5.3

8.3

6.0

9.4

6.6

10.4

5.7

8.5

7.0

10.3

8.2

12.1

11.7

8.7

13.0

9.0

14.0

10.6

Excluding years of grace.

loans repayment period would leave the borrower with a still relatively light repayment burden; but the recovery ratio would rise to only some 30% (Table 8). Linking repayments to the inflation rate (Boxes A-III in Table 8), will raise the recovery ratio to some 40–47%, depending on the level of administration costs. This outcome is in line with comparative international standards, as indicated in Table 1. If, in addition, the rate of interest were raised to 3% (Boxes B-III), then recovery ratios in excess of 55% would be possible; this would be a very acceptable outcome in comparison with those for loans schemes in other countries.

Reform on these lines seems to be feasible in terms of the repayment burden. While we have no hard information on what is an acceptable annual repayment burden level for Thai graduates, we may follow the practice in other studies of assuming a maximum feasible repayment burden of around 10% (see Carlson, 1992).14 Table 9 shows the repayment burden for these two options to

14 Woodhall (1988) provides a concise discussion of what constitutes a ‘reasonable’ level of debt in relation to annual income.

380

A. Ziderman / Economics of Education Review 21 (2002) 367–380

be consistent with this 10% limit, though the burden is again somewhat heavier for women. At a 3% interest rate, the repayment burden is 7.0 and 10.3%, respectively, for men and women. Moreover, even with these higher loan repayments, the bulk of incremental earnings (additional earnings resulting from the completion of an undergraduate course) would still accrue to the graduates. These results are not shown in Table 9, but for the repayment conditions depicted in Boxes B-III, male graduates would still receive about 90% of incremental earnings; females some 82%. Thus the return from educational investment is unlikely to be eroded unduly if higher levels of loan repayments are introduced. Keeping the repayment burden relatively low decreases the probability of repayment default. However, to the extent that evidence was forthcoming that the repayment burden could be raised further without deleterious effects on student incentives (say to about 12– 13% of income), then repayment conditions could be further hardened, with commensurate improvements in the recovery ratio. References Albrecht, D., & Ziderman, A. (1993). Cost recovery for higher education: are student loans an effective instrument? The World Bank Research Observer 8 (1). Albrecht, D., & Ziderman, A. (1991). Deferred cost recovery for higher education: Student loan programs in developing countries. World Bank Discussion Paper No. 137. Washington, DC: The World Bank. Albrecht, D., & Ziderman, A. (1992). Student loans and their alternatives: improving performance of deferred payments programs. Higher Education 23 (4).

Carlson, S. (1992). Private financing of higher education in Latin America and the Caribbean. Regional Studies Program, Report No. 18. Latin American and the Caribbean Technical Department. Washington, DC: The World Bank. Dynarski, M. (1994). Who defaults on student loans? Findings from the national post-secondary student aid study. Economics of Education Review 13 (1). Educational Loan Fund Act (1998). Rajakijja-nubegsa (Vol. 115, Section 15A, pp. 16–31). Greene, L. L. (1989). An economic analysis of student loan default. Educational Policy and Policy Analysis 11 (1). Johnstone, B. (1986). Sharing the costs of higher education: student financial assistance in the UK, the Federal Republic of Germany, France, Sweden and the USA. New York: The College Board. Ministry of Finance (1995). The student loan scheme. Report Submitted to the Cabinet, 31 March. National Economic and Social Development Board (1996). Socio-Economic Survey. NESDB. Woodhall, M. (1988). Designing a student loan programme for a developing country: the relevance of international experience. Economics of Education Review 7 (1). Woodhall, M. (1994). Student loans. In: T. Husen, & T. N. Postlethwaite, The international encyclopedia of education (Vol. 10, 3rd ed.). New York: Elsevier. World Bank (1994). Higher education: The lessons of experience. Washington, DC: The World Bank. Ziderman, A. (1999). The student loans scheme in Thailand: a review and recommendations for efficient and equitable functioning of the scheme. Prepared for UNESCO-Bangkok as part of the Asian Development Bank Social Sector Program Loan, Bangkok, July. Ziderman, A. (2001). Targeting the poor: Secondary school student loans in Thailand. In preparation. Ziderman, A., & Albrecht, D. (1995). Financing universities in developing countries. The Stanford Series on Education and Public Policy. London: The Falmer Press.