Official PDF , 152 pages - World bank documents

22 downloads 455246 Views 8MB Size Report
eight grades and reduced the number of years of vocational training in secondary .... to primary and high school only differ in one year the length of each level.
Public Disclosure Authorized

DiscussionPaper EDUCATION

AND TRAINING

SERIES

EducationCostsand FinancingPolicies in LatinAmerica

Public Disclosure Authorized

Public Disclosure Authorized

Public Disclosure Authorized

Report No. EDT6O

EmestoSchiefelbein February1987

Education and Training Department

Operations Policy Staff

The views presented here are those of the author(s),and they should not be interpreted as reflectingthose of the World Bank.

ABSTRACT

In this review of research on Latin American educational costs and policies, the author discusses the available use of alternatives for better use of resources.

He shows that a more active role by the private sector

will help meet the increasing demand for educational expansion.

The author

emphasizes that change will result only from policies applied over the long term.

Table of Contents Page No. Su nmmary ........................................................

1

EDUCATION COSTS AND FINANCING POLICIES IN LATIN AMERICA

3

A. 1.

...

.....

INTERNAL EFFICIENCY ........................................ The effect of management on internal efficiency .... ........ Management Strategies ...................................... Management Practices ....................................... Cost Studies ............................................... Information For Better Management. ......................... Summary .................................................... A.2 Effect of Alterable Variances on Internal Efficiency. ...... Class Size ................................................. School Calendar ............................................ Teacher Qualifications ..................................... Other Alterable Variables .................................. Siummary .................................................... A.3 Student Flow Characteristics. .............................. High Levels of Repetition .................................. Factors Related to Wastage ................................. Reported Attempts to Reduce Repetition ..................... A.4 Effects of selection processes on internal efficiency ...... Effects in Primary Education ............................... Effects in Adult Education ................................. Effects in Secondary Education ............................. Effects in Higher Education ................................ Summary ....................................................

4 6 6 10 12 12 14 14 14 15 15 17 17 18 18 20 22 23 24 24 25 25 27

B. PRIVATE CONTRIBUTION TO THE FINANCING OF EDUCATION .... ..... 27 B.1 Cost Recovery Policies by Level of Education. .... .......... 28 Factors related to increased public share .................. 28 How to measure private contributions ....................... 29 Cost Recovery by Levels .................................... 29 Subsidies as a Cost Recovery Strategy ....................... 31 Other Cost Recovery Strategies ............................. 32 Summary .................................................... 33 B.2 Private Schools ............................................ 35 Proportion of Private Enrollments .......................... 35 Types of Private Schools ................................... 35 Financial Impact of Private Education ...................... 37 Subsidies to Private Education ............................. 38 Fee Levels ................................................. 39 Public Control of Private Schools .......................... 41 Comparison of Achievement Levels ........................... 41 Summary .................................................... 43 B.3 Employer-financed Vocational Training ....................... 43 Professional Training Institutions ......................... 43 Alternative Strategies for Training ........................ 44 Higher Education Institutions .............................. 47 Summary . ....................................................48

- ii -

Page No

B.4 Experience with Student Loans and Voucher Systems .... ...... 48 ......... 48 The Impact of Inflation on the Lending Capital .... 50 Collection Costs and Performance ........................... Impact in Selectivity ...................................... 50 Su.rmmary .................................................... 50 B.5 Systems Used to Allocate Scholarships ...................... 52 Allocation to Assessed Needs ............................... 52 Allocation to Academic Achievement ......................... 52 Problems for Allocating Scholarships ....................... 54 Use of a Market Mechanism .................................. 54 Summary .................. .... 54 C. LESSONS OF EXPERIENCE FOR IMPROVING ALLOCATION OF RESOURCES 55 55 The Extent of Excess Demand for Education .................. Excess Demand In the 60's and 70's ......................... 56 Actual Evidence of Excess Demand ........................... 56 Summary .................................................... 59 Responsiveness of Enrollment to Fee Increases .... .......... 59 Price and Income Elasticities of Demand .................... 59 Effects of Changes On Fee Levels ........................... 61 Effects of the Supply of Education ......................... 62 Sulmmary .................................................... 62 Experiments In Cost Recovery and By Price Rationing ... ..... 62 Other Experiences in Cost Recovery ......................... 65 Cost Recovery Through Teachers Salaries .................... 66 Price Rationing ............................................ 67 Suminary ..................................................... 67 68 C.4 Non-Price Rationing Systems ........... The Impact of Public Education in the 60s and 70s .... ...... 68 Discriminatory Allocation of Resources ..................... 71 Public Systems Allocation Processes ........................ 72 Lack of Relevant Signals for Improving Allocation .... ...... 76 Sutmmary .................................................... 76 C.5 Resource Allocation within the Education Sector .... ........ 76 Laws, Decrees, Norms and Instructions ...................... 78 Budget Allocations .......................................... 79 Institution Building ....................................... 80 Leadership of Participation Processes ...................... 82 Suimmary .................................................... 85 D. Policy Implications from the Analysis of Available Research 85 D.1 Reducing Unit Costs of Graduates by Increasing Efficiencies. 85 D.2 Increasing Private Contribution ............................ 87 ANNEXES .................................................... 90 REFERENCES ................................................. 117

Summary Although we still have only a limited knowledge of the specific ways in which internal efficiency and resource allocation to education may be improved, the review of available research results can focus our attention on the right problems; help us to understand some of the apparent contradictions observed in the educational system; narrow the search to a limited number of alternatives; identify the direction in which they should be modified, provide some hints about the areas where policy modification is feasible and suggest how to implement the policies. One set of results from this study deals with identification of the right problems. Even though diagnostic reports prepared by the Ministries of Education in Latin America have usually singled out drop out as the main source of wastage in the educational system, research shows that repetition is truly the main source of wastage in Latin American education. Students are dropping out after five or more years of schooling, yet they are only at second or third grade level. Therefore, from all students attending school, only a fraction is being promoted each year to the next grade and eventually-finishingprimary education in the normal amount of time. The source of the wastage is chronic repetition finally ending in dropping out at an early grade. The analysis of available research results helps us to understand some.of the main apparent contradictions observed in the educational systems. Available research results suggest that Latin American countries have been able to provide schooling for a high proportion of their population in school age, with relatively low costs per enrolled student and high costs per graduate student. Although this apparent contradiction is mainly explained in terms of the low teacher salaries and high rates of repetition, other mechanisms also contribute to reduce costs. For example, the Latin American countries have extended primary education to seven or eight grades and reduced the number of years of vocational training in secondary education, thus cutting costs of the seventh to ninth grade in half. In addition, the number of class hours per week has been reduced, parents contribute to finance current costs, the average size of schools is reaching more efficient levels, salaries of high school teachers tend to be reduced to the level of the primary teachers; and schools are used in double shifts. On the other hand, there are built in factors contributing to inefficiency. The most pervasive problem is the lack of incentives to allocate resources more efficiently, because the Ministry of Finance periodically cuts all non-salary expenses in recession times. These cuts may partially explain four other negative aspects identified in the study: (i) untrained teachers are still being hired for rural areas; (ii) good teachers tend to leave the education sector and move into better paid jobs; (iii) there is lack of on-the-job supervision, and (iv) there are wide differences in quality among schools of each country. The analysis of available research suggests that there are few financial alternatives to continue expanding education. Although the public sector has made a substantial effort in the last three decades to provide more educational opportunities, the public funds are now dwindling and any additional efforts will have to be made by the private sector. Over one third of the educational services are already provided by private institutions and several countries are trying to expand the contribution of

-2-

the private sector through fees. If informal on-the-job-training is considered, the role of the private sector could reach near half of the total amount the countries are allocating to education. Although most private education caters for medium and high socioeconomic families in urban areas there is also some private education even in isolated rural areas. At the university level, publicly financed student loans have provided additional incentives for private institutions to expand enrollments or raise quality. One way to tap other private new resources is to improve the allocation of resources through a combined price and non-price rationing system. Fees could be introduced or those in place increased for higher education while testing could help to select suitable candidates including those that should receive scholarships. The large demand for education would ensure the additional resources through fees without reducing enrollments. There are several pressures increasing the demand for more education unemployment, the high social value of education, salary scales, the raising of parents' education beyond certain thresholds, the population explosion, and the political value of educational activities add up to increased demand for education. Although the demand is mainly related to income levels there is some responsiveness to changes in fees. With a free tuition system there is a high excess demand at all levels. In primary level the excess demand seems to be located mainly in rural areas, while the demand for the other levels is mainly urban. Research also provides some warnings about the speed at which changes can occur and the predictability of results. Most of the reported changes seem to be the result of slow long term processes rather than clear cut decisions. Fifteen years after research detected the effects of textbooks, or lack of them, they are now accepted as one of the required conditions for minimum quality results; nuclearization that was actively supported on theoretical grounds in the sixties is now under heavy criticism because it has not produced the predicted results. All in all, decision making is being based now on a broader basis of information than before. There is also a greater number of experiments being tried out in all countries, but in order to extract suitable lessons from them better processing of the results of those experiments into usable information should be fostered.

-3-

EDUCATION COSTS AND FINANCING POLICIES IN LATIN AMERICA 1 Although Latin American research reports on the financing of education are usually focused on understanding concrete problems of education cost and financing policies rather than on a more general type of knowledge, a systematic analysis of those reports permits drawing conclusions that could be used by decision makers when designing educational policies for any country of that region. Most the large number of available reports on the financing of education deal with country specific situations. Many reports take into account many complex interactions in an attempt to draw conclusions in specific situations. Many resorts present good descriptions of financial mechanisms and discuss the pros and cons of potential alternatives in terms of efficiency and equity. There are too many reports on this topic to simmarize all the relevant information in this report and to draw more general conclusions. Abstracts prepared by the REDUC system (network for exchange of educational information) helped the search for relevant reports. In this review only reports published before May 1985 are considered. The references provide the number of the abstract where additional information may be obtained or even to use that number to obtain the corresponding microfiche with the original document. In spite of the large number of reports in many cases it was not possible to obtain updated comparative figures in available reports. 2 At the outset it must be mentioned that while Latin American countries have many common traits, there are also large differences. Thus all experiences described in this report may be appraised in terms of the specific situation of each country. For example, although admission ages to primary and high school only differ in one year the length of each level may differ up to three years (see Annex 1). All Latin American countries have 6 or 7 as the admission age. Most countries have a 6-year primary level, but since the mid sixties it is frequent to find extended primary with 7 to 9 years' length.

1/ The REDUC (Red de Educacion) exchange of information centers provided invaluable help in gathering-the basic information. Marta Zeballos and Maria Clara Grossi helped in the identification of the relevant references. Himelda Martinez, Hilda Szklo and Osvaldo Kreimer helped in gathering materials from international agencies. Alfredo Sarmiento contributed from Colombia. Susanne Mowat in a previous project on educational financing and George Psacharapoulos provided the opportunity to prepare the present report as background material to a World Bank Policy Paper in this topic. Mary Brady helped in editing part of this version. Liliana Longo, Livia Mitchell, Mary Ann Heraud, and Rosita Estrada, transformed the final manuscript into the present version. 2/

The number of each RAE (Resumenes Analiticos en Educacion) can be used for retrieving the abstract or the microfiche of the original document in any of the 20 centers in 16 countries participating in the REDUC network.

-4-

In spite of a relatively common structure, there are large differences in the size of the educational systems. While in the less developed countries only 10 or 15% of the population attends school, 30 to 35% do so in the more developed countries of the region. These figures are somewhat misleading, because countries with a larger older population (for example Argentina or Uruguay) show smaller percentages, but provide an idea of the differences. There are large differences among countries when comparing the total national effort in education during the economic peak level of the mid seventies (see Annex 2). If total cost of education is compared with the gross national product, figures suggest that Haiti was making the least effort (1.88%) while the countries more interested in education were spending nearly 10%. Present economic recession will probably reduce the share of education in gross national product. In summary, when using research on costs and financing of education as a basis for policy making both the positive and negative aspects must be considered together with an assessment of those aspects where no research is available. Figures suggest common traits and differences among countries which must be taken into account when analyzing actual financing policies in Latin America. Conclusions to be drawn on policies will be limited when the criteria for judging those policies include basic philosophic issues (freedom equity, social cohesion, self-realization,welfare and, of course, efficiency) on which it is difficult to reach full agreement or where little or no research is available. A.

INTERNAL EFFICIENCY

Although Latin American countries have been able to operate with low costs per enrolled student, too many of those students are repeating. Therefore, the analysis of units costs presented in Table 1 must be carefully done in order to avoid misleading conclusions. The analysis presented in this paper takes into account that figures usually published on repetition levels grossly underestimate the real level of repetition. In this context of low unit costs and slow students flows both the management strategies and the internal variables affecting achievement are presented and discussed. The effect of the selection process is finally commented on as an aspect that links the internal educational process with the management of the system. Although in this chapter the overall budget is assumed to be fixed, in order to analyze the effect of management on internal efficiency, Table 1 suggests that Latin American countries may substantially increase the educational share of gross national product (Corvalan, 1982). A few comparisons with other.regions will provide a useful framework for analyzing internal efficiency. Each education level is more developed in Latin America than in Africa or Asia. Gross schooling rates of 100 percent are observed in L.A. primary education while the figure for other LDC is 71%. The high enrollment rates in primary education observed in Latin America are not reflected in the distribution of resources. Only one third of educational expenses are devoted to primary education. Reaching larger schooling rates with a smaller proportion of GNP is explained by lower teacher's salaries in terms of the GNP per capita.

-5-

Table 1:

EDUCATIONAL EXPENDITURES INDICATORS FOR LATIN AMERICA AND OTHER DEVELOPING COUNTRIES (percentages)

Indicators

1.

2.

3.

4.

5.

6.

7.

Notes:

a/ b/ c/ d/

Public Expenditures in education as percentage of GDP

Latin LDC America (70 (22 ctries) ctries)

3.47

4.00

Africa (35 ctries)

4.75

Asia (13 ctries)

4.75

Public expenditures in education as % of total public expenditures

17.4

18.2

Capital expenditures in education as % of public expenditures in education

n.

n.a

Distribution of current expenditures by levels: - primary - secondary - higher - other

34.0 1.0 30.0 15.0

50.0 27.0 15.0 8.0

43.0 26.0 16.0 15.0

50.0 23.0 15.0 12.0

71.0 15.0

58.0 22.0 b/

76.0 11.0

8.4 38.0

6.7 44.0

35.0

Indicators for primary education: schooling rates (gross) 101.0 unit cost as % of the GDP per ca 11.0 teacher salary divided by GDP per ca3. 4.0 6 students-teacher ratio 35.0 Indicators for secondary education: 35.0 schooling rates (gross) 22.0 unit cost as % of GDP per capita students-teacher ratio 18.0 Indicators for higher education: schooling rates (gross) unit cost as% of GDP per capita

9.7 121.0

16.0 52.0 23.0

2.0 362.0

20.9

7.8 a/

11.0 133.0 ct 23.0

1.3 1,166.0

17.3

8.0

31.0 27.0 25.0

6.5 205.0

Figures for Latin America are obtained from Dagnino P. 1978 circa 1975. Figures for other LDC, (Africa and Asia) are obtained from an updated version of the 'EducationSector Working Paper,' IbRD, December, 1974, and correspondto several years around 1972-74. From Artus, Kuzuku, 'Cross-countrydata on government expendituresclassified by functions,' World Bank studies in domestic finance N.3, December 1973. 20.0 for East Africa and 24.0 for West Africa. 124.0 for East Africa and 142.0 for West Africa. 927.0 for East Africa and 1,405 for West Africa.

a

-6-

Low teacher salaries in Latin America may be explained in terms of lack of expatriate teachers and low social prestige of native teachers. There are no examples of highly paid teachers. Although some countries during the sixties have made some efforts for substantial raises in their salaries (the cases of Chile and Peru will be explored in C.3), the general trend is toward a reduction in teachers' salaries expressed in real terms. Several other explanations for lower unit costs will be explored below. Low unit costs may also be explained by the existence of one third or more of untrained teachers, use of buildings in double shifts, lack of funding for maintenance and operational expenses (partially provided by the parents, but not included in the cost estimations), large class sizes, reduction of boarding facilities (in part due to the increasing urbanization rates of Latin America, with some countries reaching three quarters of urban population) and large average size of schools. In the following four sections, four relevant factors influencing internal efficiency in Latin America will be described, and their interaction with some specific allocation mechanisms will be further commented in the last chapter. 1.

The effect of management

on internal efficiency

Although educational management in countries following revolutionary processes seem to differ from those following evolutionary. political processes, the analysis will only focus on the latter. Educational management in countries affected by revolutionary processes seems to reduce attention to internal efficiency as observed in the case of Cuba and Nicaragua (Arrien, 1980). In those two countries most efforts are devoted to improve the external efficiency of the system according to economic, social, political of psychological objectives of the new forces in power. Management Strategies In Latin American countries following evolutionary political processes three different management strategies may be observed. First, those strategies used in countries attempting to expand the coverage of the school system and the quality of the education through the improvement of the centralized administration. Second, those strategies used in countries trying to improve specific aspects of the educational system through ad-hoc groups (proyectos or programas), and third, those strategies used in countries trying to develop decentralized mechanisms for decision making and administration. Most countries used the first model during the sixties and early seventies. National plans were the main tool for reaching higher levels of Several countries still rely in this internal (and exterral) efficiency. approach (Venezuela, 1980; Ecuador, 1983; Dominican Republic, 1979). The large increases in the schooling rates and in the percentages of GNP devoted to education (see Annexes 3 and 4) seem to be related to this approach. However, those plans have sometimes little relation with the annual budgets. One study of the budget process in Mexico suggests that the budget process is shaped by internal forces (inside the Ministry of Education) in isolation from the national political mainstream (McGinn et al, 1980).

-7-

It is possible to assume that the macro-planning approach have helped to reduce the large differences observed between unit costs in primary and secondary education. Table 2 shows that the difference is now smaller (a reduction from 2.04 to 1.60 in the 1965-1975 period) probably due to lower salaries for the secondary teachers. Something similar is observed for the unit expenditures in tertiary education. The extension of primary education to seven, eight or nine grades (see Annex 1) has represented considerable savings given that the unit cost of primary education is around half of the unit cost of secondary education. In addition some countries have further reduced the number of years spent in vocational education (see Table 3). The decisions have been taken on intuitive basis given that only for Brazil and Colombia there are comparative studies on rates of return of regular .vocational and the training institutions and in both cases small differences were estimated (Castro, 1971, 1972, 1973, Psacharopoulos, 1985). In all countries the best graduates from vocational schools go into higher education and in many countries the graduates from vocational schools cannot compete with those trained in a short time in training institutes. This is an area where more studies are required. Follow up studies on graduates from agricultural and industrial schools are specially relevant for future decision making. Table 2:

MEAN UNIT EXPENDITURE COMPARISONS BY LEVEL OF EDUCATION (a)

Secondary Expenditure Tertiary Expenditure per Student per Student Primary Expenditure Primary Expenditure Per Student Per Student Other Middle Other Middle Latin Income Latin Income America Countries America Countries

Tertiary Expenditure per Student Secondary Expend. Per Student Other Middle Latin Income America Countries

1965

2.04 (0.40)

1.80 (0.34)

16.60 (1.20)

9.10 (0.56)

7.90 (0.75)

5.20 (0.42)

1970

2.40 (0.39)

1.70 (0.40)

13.60 (0.72)

13.10 (0.76)

7.10 (0.97)

8.60 (0.67)

1975

1.80 (0-35)

1.50 (0.27)

11.50 9.92)

9.30 (0.56)

5.10 (0.65)

6.00 (0.38)

1978

1.60 (0.34)

1.70 (0.28)

7.80 (0.60)

3.90 (0.35)

4.80 (0.58)

3.10 (0.22)

(a) Number in bracketsequalsthe ratioof the mean to the standarddeviation. Source: Hellerand Cheasty,1983, p. 11. Elaboratedby Jimenez,1984. Source: AlyaroReyesPosadaand PatriciaRestrepo,wCostode la Educaci6n Piblicaen Colombia,'ECIEL,Rio de Janeiro,1976 p. 82. G. Psacharopoulos, 'To Vocationalize or not to Vocationalize", discussionpaper,IBRD,June 1986. E. Schiefelbein, *Diagn6stico del SistemaEducacional Chilenoen 1970,"Departamento de Economfa,Universidad de Economra,Universidad de Chile,Mayo,1976,p. 177 (a)The best technicalschoolhad a 40% higherunit cost and an agricultural schoolhad a unit cost twiceas largethe averace.

-8-

Reduction in the unit costs of secondary education is specially important in the present stage of the Latin American educational systems. In most countries access to primary education is becoming universal. Although too high repetition rates reduce the number of students graduating from primary education, this number is increasing and there is an excess demand for secondary education. Lower Unit costs in secondary education may ease up the access to that level (Chinchilla, 1981; ODEPLAN, 1982). The second model has been supported by international agencies. Since the early seventies Project Offices have been created to meet specific targets. In most cases those offices have been effective in setting up the required inputs, but little may be said in terms of the outputs. Even though several attempts to measure efficiency have been reported little is known in terms of the efficiency of projects such as MOBRAL (Castro, 1981) or ACPO (Brumberg, 1974). The third model aims to decentralize decision making. Although most countries frequently mention their interest in this model few attempts have been actually carried out (Le Boterf, 1982). Colombia, has created the Fondos Regionales, Brazil and Chile have turned the schools to the Municipalidades (Sobrihno, 1978, CPEIP, 1982), and Peru has relied on the "nuclear system" (Peru, 1982). Since 1982 Peru is trying to return to a *more centralized system, given the negative effects detected in internal evaluations. Brazil created a national institution (PROMUNICIPIO) to help the Municipalidades to improve their low efficiency to improve their low efficiency. Improvements in the number of school days attended by children have been detected in Chile, but also low teachers morale seem to be another outcome of the decentralization process. In Colombia the decentralization process has generated large pressures on the contributions of the central government to each region, given the little information is available in the central level to allocate resources to each of the regions or to estimate the efficiency of each of the regions (additional information is presented in C.4). The effect of the different models is probably limited by the high proportion of costs made up of salaries for teaching (Table 4) and lack of efficient management. The administrative wages probably account for another 5%. In the remaining expenditures there are many fixed charges for maintenance, books, stationery, power, or services. Elasticity of substitution between different factors of the educational process seems to be very low. Buildings cannot be substituted for teachers (and viceversa) even in the medium term.

-10-

Management Practices Although many reports illustrate the poor level of educational management, attempts for improving management practices are also frequent. A report on 10 Latin American countries suggests that buildings are inefficiently used and resources are allocated with poor criteria (Castro et al, 1980) and in section A.3 the high level of repetition rates will be commented. In fact most educational managers do not seem to be aware of the possible impact of the key elements affecting the educational process. An important exception are textbooks. Managers have been gradually accepting textbooks as one of the factors that should be provided to all students in order to improve their achievement. The lack of training in management of educational officials may also be reflected in lower levels of execution of the budget. Data presented in Table 5 shows that the Ministry of Education used a lower proportion of its budget than the central government as a whole. This data reflect the situation of one country, but it is also representative of the region. However, the lower level may be caused by the directives issued by the Ministry of Finance establishing low priorities for education in case of financial stress.

Table 5:

Year

PARAGUAY. COMPARISON AMONG THE ALLOCATED BUDGET AND THE ACTUAL EXPENSES OF THE CENTRAL GOVERNMENT AND THE MINISTRY OF EDUCATION. (millions of Guaranies)

Central Government Allocated Expended

1970

10,697

9,401

1975

18,935

1980

%

Ministry of Education Allocated Expended

Ratio Expenses % MOE to CG

88

1,389

1,232

89

13%

19,562

103

2,977

2,737

92

14%

54,795

56,028

102

7,688

7,513

98

13%

1981

84,072*

80,539

96

9,816

8,855

90

11%

1982

96,167

78,209

82

13,304

10,658

80

14%

Source: IBRD Mission. *

The originalfigurewas 83,978,but repaymentof governmentloanswas reprogramed from 13,047to 32,136.

Notes: The allocatedbudget includesall additionsapprovedduringthe fiscalyear.Actualexpenses includepaymentsin the monthof Januaryof the next fiscalyear. The figuresincludeboth currentand capitalexpenses.

-1l-

In general, the Ministries of Finance do not stimulate the interest of the Ministries of Education to introduce more efficient use of resources. The reduction of expenses through a more efficient use of resources is seldom reallocated in the same sector and any attempt to internally transfer funds to supervision or educational materials finally goes into general funds when the next financial crisis arrives. The previous comment may explain, in part, failures in trying PPBS methods in many Ministries of education. Those methods have usually been formally followed, but no attempt to really try to find better ways to meet the targets has been observed. However, countries have tried to achieve the traditional goals using less resources or to improve education using better techniques. For example, during the sixties and seventies several countries reduced the total number of class hours per week (in some cases from 35 to 30 hours per week). In secondary education a second foreign language has been eliminated (unsuccessfully in the case of Chile due to pressures from France). Several countries (Paraguay, Peru, Ecuador, Guatemala, Mexico) are supporting experiments to determine efficient methods for teaching students. In the long run such experiments (and the exchange of results among Latin American countries) should have a powerful impact in efficiency. For example, there are impressive gains using bilingual education for students whose mother language is different from the national language (Spanish or Portuguese). Repetition may be reduced (50% of students in first grade are repeating in areas where bilingual education is required) and achievement may be substantially raised. (Some further comments are presented in C.3). There are attempts to solve the problems created by rural teachers living in distant cities and commuting to the school or living in the village while their families stay in the city. Commuting teaches tend to reduce the number of hours worked with the students and as a result "time on task" of students is also reduced. Mexico has implemented a successful system of "Instructores Comunitarios" whose salaries are paid by the government through a patronato or group of elders in each community. Thus the teacher becomes accountable to the community and must work the whole week schedule (CNFE, 1978). Several countries (Colombia, Costa Rica, Chile, Paraguay, Peru) have attempted to improve the supervision of the system, but monitoring traditions are not easy to generate. Those countries have trained the "best" teachers and appointed them as "supervisors". However, in most cases the Ministries of Finances have eliminated the budget allocation for travel and per diem of the supervisors. Therefore, the salary is paid, but the supervision work cannot be done and good teaches are sifted into In Costa Rica an amount of money is provided to bureaucratic channels. supervisors together with their salary whether they visit schools or not and the results do not seem to be positive either. Provision of motorcycles to supervisors in Paraguay and Costa Rica has been reported as a positive action, but further information is required to assess its real impact on supervision. Supervisors maintain their motorcycles and the reduced cost for gas makes it easier for them to use their own money to operate the motorcycles. Still, even if the supervisors may visit schools the information systems are not yet providing the list of schools with special problems for them to pay special visits.

-12-

Cost Studies Many countries have carried out studies on "unit costs" as a Table 6 shows device for allocating resources in a more effective way. that in Venezuela, Colombia and Brazil there is room for a better allocation given that the differences between the extremes of the range of costs are substantially large. This is noteworthy given the little variety in the way in which schools are run. Most of the difference in unit costs may be explained in terms of the variance in wage levels of teachers, class sizes and non-wage costs per class. The Venezuelan situation seems to be representative of most Latin American countries. Even if differences in salaries are due to legal increments related to experience, there are still important differences in class size and non-wage costs per class that could be due to inefficiencies or at least suggest large inequalities in the quality of the education provided to students of different areas. In any case there are not well established systems to compute cost regularly (CEPADES, 1982). The studies on unit costs may have also helped Latin American countries to keep costs in line with the overall economic situation or supported cost recovery strategies. Table 1 showed that education represents a relatively low burden on the public finances of Latin American countries (at least in comparison with other regions) and although unit costs in universities could be overestimated these unit costs have been used in countries looking for ways to get the students to pay back part of the costs. Studies of unit cost have also helped to detect inefficiency in graduate studies. The administrative costs in graduate studies are usually too high. Most of the master and doctoral programs should be offered pooling together the resources of several universities. However, that strategy has not been followed (Oteiza, 1982). Cost studies have also had an indirect impact on resource allocation as a component of studies on the rate of return to education. In the last decade most countries have made a substantial effort to provide primary education to all children (apart from Haiti, Guatemala is the lowest and it is now enrolling over 70% of each new cohort although their attrition rate is quite high). This effort may suggest that the managers have listened to research results suggesting that rates of return to primary education are much higher than the rates for other levels. Now, estimations for the rate of return to pre-primary education and the evaluation of such projects suggest that investment in the pre-primary is also promising (ODEPLAN, 1977 and 1980). Information For Better Management. The Ministries of Education are becoming aware of the need to learn from past experiences in a more systematic way. Mexico, Costa Rica, Colombia, Chile, Brazil, Peru, Panama and Guatemala have been developing data banks in the eighties and information for improving decision making is gradually being used. Even though the retrieval of information is still slow and little information flows to the desks of the decision makers, those attempts show the interest of authorities for improving management

TADlK 6 : Primary behooln, average oharmoteriatlonand raotoreafrecting expenditure (current currency)

CIIABACTERISTICSOF SCHOOLS

fluinriRg expenditure:per Tcaclierq

pupil

VENEZUELA,1916/7-78/9 AVERAGE RANGE HINIMUH HAXIH 80.5

57.8

BRASIL, SAO PAULO, 1980 COLOMIA', 1950 AVERAGE RANGk AVERAGE RANGE HIHIHII HAXIHIH HTNI?II BAXI

109.6

22.857

0.808

599Q70

5.235

3 i00

77790

1 9;0.

6.200

wages as % of running

expendllttire5

Teachero per clans Average teachier'o

--

74%

wage per

month

Otber wage-costs per class

1.1

1.1

1969

1769

2053

743

467

1017

Non-wageconts per class

25

Non-wage coats as expenditure

1%

% or

_

--

-

1.2

_

3.393 J

207

---

---

---

--

---

---

---

25.A

-- -

---

---

---

running

uhmmberof pupils per class Contrlbutions of parents

36.6

-41.8

_ ---

-

---

1.410

Soimce: Claude Tibi, 'Preliminaryreaults of case studies carried out In Brazil, Indoneola, Nigeria and 1 Valeuzuela", IIEP/S631 1, Paris, 17 November 1980, pp. 46-57. Paro, Victor 11."0 ouoto do enaino publico no estado de Sao Paulo, Estudo de custo/aluno ua rede totadual de primnero e oegundo graus". Cadernoo do Peaquisa NO 43, 1982 pW. 3-29. Paulsen, Alba, "Costo escolar y mejoramiento cualitativo do In.Educacldn", Rlevist Colombiana de Educaoi6n NO 8, 1981 (RAE 2507). --- Not available

1.220

-14-

systems. The success of the Latin American Exchange of Educational Network (REDUC) now operating in 16 countries also confirms the interest in learning from past experience. The development of a management information system (MIS) may be a realistic way to create an institutional memory able to reduce the effects of the high turnover rates in the civil service and the tradition of "taking home" the files accumulated during the time the officials were working in office. No Ministry is yet receiving lists of those schools with the lowest achievement indexes (in terms of repetition, attrition during the school year, overage students or approval rates at the end of the year) for the supervisors to visit and to eventually improve achievement. It may be forecasted that objective information will soon be calling the attention of the authorities about anomalous situations and information will probably be a way to permanently improve the management educational systems. Summary Although management practices in Latin America are poor, there are The expansion also some signals that educational management is improving. of enrollment in primary and secondary schools has consisted of children from increasingly poorer segments of the population (Selowsky, 1980 DEALC, 1981). However, in spite of this fact repetition and dropout rates are somewhat constant or improving overtime. Therefore, some improvement in management should be explaining that performance. Cost studies and information systems are positive factors, but there are no built in incentives for improving the efficient use of resources. A.2

Effect of Alterable Variances on Internal Efficiency.

Although the results of the educational process are multidimensional, the effects of alterable variables will be analyzed only in terms of achievement. Effects on achievement, motivation and motor skills cannot be expressed in a single scale. Furthermore, there are These problems for measuring each of those dimensions with precision. limitations explain that most of the available research deals only with academic achievement. Class Size Available research suggests that class size (in the range found in the four countries studied in the available research) does not affect achievement (Schiefelbein, 1981), but the policy implications should be carefully examined. Only in one of seven multiple regression analysis including "class size" a negative effect on achievement was statistically significative. This conclusion seems to be valid for most of the teachers, but it may not be so for good teachers able to provide individualized teaching with a small number of students per class. It might also not be true for teachers working with several grades in the same class-room (multigrade teaching). Practitioners suggest that there is a lower limit for multigrade courses (may be 25 students per group, provided that all students may work with their own textbook).

-15-

School Calendar Several interesting experiences with the school calendar have been carried out. In rural areas of Uruguay the students start at 10:00 a.m. (instead of 7:00 a.m.) in order to let the children to help in agricultural work at home. Substantial increments in attendance has been reported. In the Southern part of Argentina winter vacations are longer to reduce the effects of floods. In Cuba the long vacations coincide with the "zafra" and in certain areas of Costa Rica and Guatemala with the coffee crop. In Bolivia there are positive experiences with teachers that follow the stational migration related with local crops. In any case the total number of annual school days (length of the school year) has shown significant results on achievement (Schiefelbein and Clavel, 1974). Teacher Qualifications Teacher qualifications do not seem to produce a clear impact on achievement. (See Table 7). Eight out of 19 studies show positive effects on achievement (Schiefelbein, 1981). Detailed studies have shown that the main effect on achievement is observed in the threshold between having some training in teaching and not having any training at all. In any case increments in the length of training seem to have decreasing marginal returns in terms of achievement of students. The analysis of available research results should be carefully used, because some studies of teachers effectiveness have not controlled by key variables such as teachers turnover (according to a report by L. Wolff turnover is an important variable), availability of textbooks-, amount of homework demanded to students, and multigrade teaching, therefore some available results could be biased. Effects of these variables are also summarized in Table 7. Experience of the teachers has been detected as statistically significative in four out of 12 studies. But some of the results suggest that experience may be related in a non-linear pattern with achievement (positive relationship in the beginning and negative in the final period). Therefore, salary increments beyond a certain level of experience are no longer productive. Teachers up-grading has been detected as non significant in terms of students achievement. But it must be recalled that up-grading usually refers to traditional classes that may increase knowledge of certain topics, but may have little impacts in teachers techniques actually used in the classroom. Up-grading through stages in good schools or in discussion groups may have a positive impact on achievement and should be experimented in controlled environments.

-16-

Table 7:

LATIN AMERICA: FINDINGS OF STUDIES ON PRODUCTION FUNCTION IN EDUCATION (Number of Studies)

Variables Analyzed

Expected sign of the relationship

Teacher variables 1. Teacher qualifications (education) 2. Teacher experience 3. Turnover 4. Up-grading S. Maintains discipline 8. Innovative in teaching methods 7. Education of teacher's father School Variables 1. Average class size (student/teacherratio) 2. Equipment availability 3. Duration of school year 4. Two shifts (one group in morning and another in afternoon) 5. Coeducational 8. Out-of-schoolactivities 7. Homework .8Textbooks available 9. Expenditures per student 10. Use of library Student factors 1. Reading matter available in nome 2. Kindergarten 3. TV 4. Socio-economic level S. Family size 8. Repetition 7. Educational aspirations 8. Malnutrition 9. Self-concept 10. Intelligence (Raven)

Source: E. Schiefelbein, 1981.

Statistically significant and with sign expected

Not statistically significant or with opposite sign

+

8

11

+

+

4 1 0 2

8 0 2 3

+

0

1

+

0

1

2

8

2 1

2 0

0 0 1 2 3 8 0

3 2 0 2 0 4 2

2 1 5 5 3 2 2 4 1 1

0 0 0 8 5 0 1 4 0 0

+

+ e

+ +

+ + +

+

+ +

+ +

+ +

+

-17-

Other Alterable

Variables

Malnutrition, lack of sanitation, low levels of of psychological stimulation and other environmental deficits surrounding children in poverty areas are some of the factors affecting performance that may be influenced by public policies. Both the preschool abilities and achievement levels in primary education seem to be positively affected by TV specially through programs such as Plaza Sesamo or Villa Sesamo. (Schiefelbein, 1981). Therefore, this is another aspect that may be explored in the design of future public policies. Of special interest are the effects of two additional school variables: preschool training and school size. The marginal product of additional schooling seems to depend on the level of preschool abilities of the child (Selowsky, 1980). As mentioned above (see A.1) many efforts are being carried out in preschool education. Textbooks for parents and teachers have been developed in several Latin American countries. (Lira, 1980). Participants in those programs have shown substantial development of cognitive skills. (Llanos and Winkler, 1982). Malnutrition has also been reduced and there is a positive relationship between the cost of the nutrition programs and the improvement in the nutritional standards. (Latorre, 1980) Little evidence is available on scale effects and results are not conclusive. Table 8 presents data on scale effects for Bolivia and Paraguay. Changes in average costs are related to the student-teacher ratio that increases with the size of the school (the definition of the school size is not available). Effects of size in academic achievement measured by two different test are positive in Bolivia and are negative in Paraguay. Fragmentary evidence suggests that primary schools with less than 200 students tend to have higher costs than average and schools with more than 1000 students tend to be difficult to manage (it is possible to assume that through the years the upper extreme of the range could be increased due to improved management abilities in the region). On the other hand, one-room schools can provide high quality education when the teacher has training in multigrade teaching and the students have their own textbooks. Summary The evidence suggests that payoffs associated with two costly policies (lengthening teacher training and reducing class size) are small, while cheaper factors may be more relevant. It is true that the number of untrained teachers should be reduced in order to increase internal efficiency, but more research should be available in order to provide an empirical basis for finding effective upgrading methods (lengthening their training or requiring longer years of previous schooling is not cost effective). There is still a long way to reach the target of 100 percent of trained teachers in all Latin American countries (see Annex 3). Furthermore, the level of teachers' training modify the effect of reducing class size. Only good teachers will benefit from smaller classes. On the other hand, substantial gains in efficiency (mainly for small rural schools) may be obtained through multigrade teaching (when students have textbooks), that allows the operation of complete schools (offering all grades in primary) that reduce early attrition and allow larger school sizes (Annex 4 illustrates the proportion of incomplete schools in the system). Finally, adjustments to the school calendar that allows students to attend a larger school year period have been reported as positive. (Further comments on the imnlementat4on of chanzes are -A 4 r ? . _resnri-

-18-

A.3.

Student Flow Characteristics.

Statistical data based on forms filled by school principals have generated dangerous misconceptions about students' access to the education system and survival of students in school systems in Latin America. Therefore, decisions based on such information are adding to the inefficiency of the educational system. For example, figures showed a paradox: according to published data, although Uruguay has one of the most developed educational systems in the region it shows one of the highest rates of repetition. However, while the excellent supervision system in Uruguay is one of the reasons why accurate reporting of repetition is provided, in most of the other countries teachers do not look into their files in order to fill the statistical forms, (in many cases there are no files) and they simply ask their students to raise their hands if they are repeaters, then they count them and report the figure in the form. Therefore, the paradox is explained by repetition rates that are grossly under-reported in many countries of the region. High Levels of Repetition Table 9 presents a comparison of reported (official) rates and estimations of the true rates through more elaborated methods. (Unesco, 1981). In many countries the estimations may double the reported figures. Therefore, the dropout rates are usually over estimated (they are the complement of the promotion and repetition rates) and newcomers are Table 8: AVERAGE COST PER STUDENT-YEAR--CROSS-SECTION EVIDENCE (1975)

Country Bolivia

Paraguay

School Size Small Medium Large Overall Small Medium Large Overall

AVC 711 414 315 493

AC 800 497 444 592

3,620 2,959 3,078 3,225

4,563 3,946 4,173 4,230

SCHOOL-LEVEL

ST 17 23 39 25

RS 9.9 12.3 14.9 12.2

SS 9.6 10.7 12.0 12.7

16 25 21 20

14.7 13.2 10.8 13.0

10.9 10.5 8.5 10.0

Source: ECIELdata tape presentedby Jimenez,1984 (TableA.4.7). AVC = Averagevariablecost AC = Averagecost ST = Student-teacher ratio RS = Mean readingscore SS = Mean sciencescore

-19-

Table 9:

Country A Year

COMPARISON OF THE RATES OF GRADE REPETITION, BY COUNTRIES REPORTED AND ESTIMATED DATA (a)

First Grade Second Grade Third Grade Fourth Grade Reported Estimated Reported Estimated Reported Estimated Reported Estimated

Ecuador 1972-3 1974-5

14.1 12.8

48.7 44.0

12.1 11.5

35.9 35.5

7.8 8.3

Brazil 1970 1971 1972 1973

27.8 23.0 23.3 23.4

61.2 57.0 64.8 84.8

15.7 12.5 17.7 17.6

43.6 39.4 48.8 42.7

11.6 10.6 11.0 11.5

Colombia 1968-7 1967-8

41.3 40.8

56.8 67.8

31.1 38.1

44.2 36.9 47.8 34.7

8.4 7.3

31.2 38.1

8.2 8.2 10.4 10.7

48.2 57.4 39.0 28.9

28.0 27.4

12.3 12.9

Costa Rica 1980

27.7

Chile 1974

35.0

El Salvador 19s8

23.8

60.4

Mexico 1967

21.0

44.0

Panama 19s8 1968

30.8 30.5

37.2 38.0

Paraguay 1972 1973 1974 1975 1976

25.0 28.0 25.0 24.0 22.0

48.4 48.s 48.2 44.8 44.1

21.0 21.0 20.0 19.0 18.0

38.8 38.1 41.7 34.7 34.8

18.0 18.0 16.0 15.0 14.0

33.0 29.1 32.1 30.9 30.4

12.0 11.0 11.0 10.0 9.0

24.3 24.9 25.0 25.5 24.9

Venezuela 1961-2 1964-5 19se-8 1968-7 1967-8 1968-9 1969-70

31.8 28.4 24.2 20.3 18.2 18.3 17.8

58.1 45.7 47.3 45.5 56.0 45.3 32.9

14.0 14.2 12.4 9.4 9.4 10.1 10.2

32.3 18.0 24.1 26.3 25.7 25.9 21.8

13.3 15.7 14.3 10.8 10.0 11.2 12.0

34.1 29.8 28.3 28.2 27.3 27.0 18.1

13.9 15.1 13.8 10.7 10.0 10.6 11.2

29.2 25.2 22.7 31.2 27.3 27.2 21.1

Guatemala 1979

15.9(b) 51.6

15.9(b)

25.7

38.8

33.8

30.8

15.9(b) 29.4

27.8

15.9(b) 30.2

Source: E. Schiefelbeinand McGrossi, 1980. (a) Date used were for consecutive years and for Pupiis

by grades and aces.

-20-

over estimated (Alfonso et al, 1985). This data have made the decision makers to put more attention to exogenous factors (economic and social) rather than to pay more attention to the internal efficiency to the system (teachers methods, educational materials, and attendance). The difference between repetition levels reported by principals (that are the base for published statistics) and the true repetition levels is awesome. The reported repetition levels tend to overestimate attrition and to distort the analysis of "apparent cohorts" used to define efficiency (graduates from the nth grade compared with the enrollment in first grade "n" years before). In fact while the enrollment in first grade is much larger than the newcomers, enrollment in the nth grade tend to be much closer to the number of graduates (there is a selection process that makes repetition rates to decrease with higher grades). Thus cohort analysis like those presented in Table 10 are biased and distort the analysis of the true efficiency of the system. Global analysis of flows hides the differences between rural and urban areas or the differences by social strata. Students from rural areas or from low socioeconomic background usually join the educational system at older ages than the rest of the student; they attend less days during the school year, have higher rates of repetition and finally drop out when they reach the working age (frequently near the 12 to 14 years of age). Although students from low socioeconomic levels are enrolled during 5 to 7 years, they do not approve more than one to three grades. An important number of urban students from low socioeconomic levels (mainly primary students) switch from day shifts to night shifts when they join the labor market. In fact most of the "adult education" student body in Latin American countries is less than 20 years of age. This also might explain that according to the figures in the Census of Population the lowest literacy rates correspond to the 15-19 years old groups. The lowest of the literacy rates by single ages provide a good estimate of the access of each cohort to schooling and of the future literacy rates of that cohort. Such lowest rate is used for estimating more realistic repetition rates. Repetition rates are much lower in secondary (high) schools than in primary schools. However, in one tracer study carried out in Chile students repeating five or more times during high school were found. The certificate (diploma) received at the end of high school opens new areas in the labor market and there is some social reward to finishing that level. Latin American societies assign considerable value to certificates and diplomas and the students and their families are well aware of this fact. (Bruner, 1982). However, the expansion of the educational system has generated a quick devaluation of the certificates received at the end of primary and secondary levels. (Munoz, 1974) Factors Related to Wastaze The high repetition in first grade seems to be associated with the criteria for pass: students must know how to read or write a few words and to do simple additions. These are difficult barriers especially when students do not have textbooks or when they learn to read in a language different from the mother tongue. The lack of textbooks is even more crucial when the teacher does not have training as educator and has no methods for teaching those basic skills.

-21-

Table 10:

INDICATORS OF EFFICIENCY IN PRIMARY EDUCATION

0S

~t

Country

_

|

L.

a'

students cohort) 1

_~~Q

Ai"n*ft

Ui

commulis Cum R oxw 0Ch.

1mANOW 85 5 ih.' GUmIna GWuW

_d)

b .

7 a 5 a

173 117111.07 lS37 U2 1i70 7.2

4 3'

117 17"

5

137 1171 1570

S 6 6

vu~a

Sourco:

2

*

0

_

17 t.m 1.52 1.D

rw 218

ms 3 3 7=

7." 12.72

1.3 1.fl

13n5 453

547

8S. 1734

143 1.t3

M3 729

1s50

12.15 7.3

2=. 1.22

77 127

1974

8.2

1.3 2.0 1= 2. 1.42 1.3

34

104

S

pani

Student, I144n the Ieat # * _according # o*grade to thernumber of repetitIton. (from a ISSo

.

U

1570 2.3 £__~6 1971 7.73 - 8 1970 12.18 6 1372 8.32 a 1974 11.14 a

197

7.70

1.

6

1970

7.

1=

UNESC3, 1980:48

=

34 2 21 12

275

73

23 1564 181M 14

Ut 11 171 7 17 104 3 543 615 24 171 26 1US 2 ff7 271 =

734 2?0 50 254 OO

72 I¶0 1o4 s S= 11s 47 511 t 7 52

-

= 401 IO S44 =33 SW4 121

SU

21

A2

34

4 IS

i37.

Is

so

18

27 9 = 41 134 67 40 21. 11 35 162 t 127

S

2t6

617

97

IS 15

104 2

-22-

The criteria for promotion in the other grades is less precise and each teacher has his own judgment. One study shows that in different schools of the same country the requirements (in number of pages the student has mastered) are quite different (Villarroel et al 1984). There are no standardized tests applied to students in different grades of primary and secondary education, but in only one country. However, there are standard university entrance examinations in five countries (Brazil, Colombia, Chile, Costa Rica and Mexico). Results in those university entrance exams suggest relatively low achievement levels of student and there are continuous demands from university professors to raise the standards in secondary education (even though the claims of university professors are probably biased). New figures from the International Education Achievement Study (IEA) may shed light in this aspect in the near future. Factors affecting repetition are also the factors affecting achievement. The effects of the school calender and teacher qualifications on student flows have already been explored above (see A.2). Incomplete schools, late entrance, mother language, transport and lack of textbooks are additional factors detected in other reports. Several strategies to modify some of these factors in order to improve achievement have implemented, but few have been evaluated. Some of the experiments are very simple. For example, schools with few students in certain classes have been encouraged to ask students with learning problems to attend the second (Schiefelbein, (other) shift in order to provide some remedial attention. 1982). Other experiments are sophisticated and have been evaluated. For example, the use of radio helped Nicaraguan students to raise their achievement in mathematics and to reduce their repetition rates (see C.5 below). The project implies an increment in unit cost per year, but reduces the cost of each graduate student from primary education. (Jamison, 1978). Reported Attempts to Reduce Repetition Although Costa Rica, Ecuador, Venezuela and Chile have enacted "automatic promotion" laws or norms for eliminating the number of repeaters in primary education, repetition has been reduced gradually by a change in teachers' attitude toward repetition. In Ecuador teachers are assigned to the same group during the first to second grade to reduce claims against the "teacher of the previous grade," but there still remain many students repeating first grade in spite of being illegal to do so. In Chile students are only required to attend over 120 school days in order to be promoted. In several countries remedial teaching time is offered to slow learners. Although the laws do not drastically change the patterns of promotion, the laws make teachers aware of their responsibility and, in the long run, repetition rates have been reduced in countries where automatic promotion has been enacted, but the net effect of this strategy is not yet conclusive. Teacher training is now placing more emphasis on the responsibility of the teacher in the success of normal children. (MuEoz and Schmelkes, 1983). This emphasis may have an important impact in future student flows.

-23-

Teachers working in one-teacher schools have been trained to work with several grades at the same time (provided that textbooks are available) and to offer even all primary education to a group of 45 students or less in a single classroom (for example first and second grades in the morning shift and third to sixth grades in the afternoon shift). The "escuela nueva" en Colombia, and the experiences in many other countries shows that when teachers receive a minimum training and students have textbooks to work by themselves, the promotion rates increase and dropout tend to be reduced (at least forced drop out is reduced, because no longer students must leave the school because there are no further grades). Preschool training both in schools or at home have proved to have positive effects in promotion rates in first grade. (Llanos and Winkler, 1982; Filp et al, 1982). Preschool training may also have an impact in following grades (as suggested in results from research carried out in developed countries). In a few countries a special effort has been made to use mass media for raising educational aspirations. (Mayo et al, 1974). Given that aspirations seem to be related with achievement they may have reduced repetition and had some positive impact on student flows. In most countries children and adults have special opportunities for certification of knowledge. At least in one country an adult may be Given the certified even in only one subject (for example, mathematics). high value of certificates these mechanisms have reduced the pressure on enrollments on the regular system. The training centers have discussed for many years the need for a vocational certification system, but little has been implemented. (Agudelo, 2983) Enrollment as special students on university courses where there are available unused places (with no prior requirements) has been discussed as an alternative way to certify knowledge; those special students meeting the promotion requirement procedures would receive specific certification for the course approved. However, although this option was enthusiastically received in a seminar organized by CINDA in May 1986, still no experiment of this type has been carried out. Su=mary All in all there are substantial increments in the promotion and retention of students in the last two decades. However, still in many countries there are no policies for reducing.repetition. In most countries the private schools do not allow their students to repeat and the public system must accept those less able students. Thus comparisons of achievement in both types of schools should be carefully done (Uzategui and Muelle, 1974). A.4.

Effects of selection processes on internal efficiency

Although the selection process tend to increase the achievement levels of those selected, it also tends to reduce the achievement of those screened out, to raise unit costs in the secondary level and to reduce equal opportunities for studying.

-24-

The educational selection process starts at the cradle, even before entrance into the school system, and the selectivity impact is reduced only when students are accepted into a specific university career. Children from affluent families send their children to kindergartens while families from lower socioeconomic levels seldom obtain that service free of charge, mainly in some industries of the modern sector. The differences in initial attention have an impact on all school life. (Selovsky and Taylor, 1973; Schiefelbein and Farrell, 1982). In the most prestigious schools sometimes the children can only join the school at that level or even the family must own a highly valued share of the educational institution for the children to became a student. Effects in Primary Education In private schools most children enrolled in a grade have the same age and use the textbooks designed for their age. In public schools there is a more heterogenous distribution of ages; necessarily some students must use textbooks that do not motivate them, because they are older, and the teacher must make an extra effort to cope with the differences in ages. These differential factors add to other factors commented above (for example socioeconomic factors) and generate higher repetition levels in public schools. The selection process is also reinforced by the different attitudes of teachers depending in the type of preschool or school they work. (Filp et al, 1982) Teachers in schools for students from low socioeconomic levels seem to be more authoritarian, and rote learning and some physical or verbal abuse prevails; sometimes the environment and traditions of those students are critically commented by the teachers thus reducing self assurance of students and increasing anxiety levels, (Guida et al, 1984; Dahse, 1982; Putzey et al, 1983). Students from rural primary schools are selected to continue their education when the students are able to enroll in other schools offering continuation beyond the last grade offered in the local school. In general there are other informal ways to select students through requirements of uniforms, shoes, textbooks and copybooks or special contributions requested by the school. All these mechanisms reduce both the class size (increasing the unit costs) and the graduates/newcomers ratio (graduation rates). Effects in Adult Education The lack of alternative strategies that suits each group of adults have been suggested as one of the causes for poor results in adult education. (Castro et al, 1982) More diversified types of education seem to be required for adults to become interested in using the supply of learning opportunities. The analysis of population cohorts suggests that beyond the age of 24 it is very difficult to become literate and to keep the literacy skills (when the environment does not stimulate the use of those skills) unless there is a special drive. However, most literacy campaigns seem to be the result of sudden political interest rather than the result of careful analysis of problems and experience and little professional advice is used in their design and implementation.

-25-

Effects in Secondary Education At the secondary level the selection processes are expressed in terms of multiple pressures for offering more opportunities of costly vocational education to students from lower socioeconomic levels. Students are streamed into the academic and vocational education, with the underachiever students going into the second track. However, most students reaching secondary education expect to go to the university (given the high payoff associated with success in university and the fact that graduates from high school, on average, belong to the upper 25% of the socioeconomic distribution of each cohort). Therefore, the best students graduating from vocational schools usually get their way into the university. (Mayo et al, 1974). The paradox is that the better the quality of the vocational education, the higher the probability of reaching the university level and the lower the probability of using their vocational skills. Given the cost of vocational education usually doubles the cost of academic education (see Table 3), such a selection process generates inefficiency in secondary education. Effects in Higher Education At the university level the selection is expressed in terms of a hierarchy of universities and careers as well as in terms of the cost of the studies. University Entrance Examinations in countries with good free public university education have stimulated many families to invest in good fee-paid private high schools in order to get access to those tuition free public universities, while students attending poor tuition free public high schools end up in poor fee paid private universities (mainly in Brazil and Colombia). This is a topic of present interest in Latin America, (Soria, 1980) and some studies have suggested the need to ask the students to pay at least the same amount they have been paying in high school (Paraguay or Chile). Those payments may represent 10 to 20% of the educational budget (see details in point C.3). Selection processes represented by University Entrance Examinations have forced the teachers to train (coach) the students for those tests rather than emphasizing creativity or vocational aspects (Schiefelbein, 1982). This is specially true in those countries where demand for places in the university exceeds the available places (DEALC, 1981). Private paid schools and universities usually have lower repetition rates than the public institutions, but the rates are affected by the higher SES of private students and by their better entrance achievement levels as well as higher aspirations and motivation. (De Andraca et al, 1982) Charging fees in Chile has been associated with a reduction of students from lower socioeconomic levels. (Briones, 1981). Increments in fees in the Universidad Rafael Landivar (from $600 in 1980 to $800 in $1984) did reduce the newcomers but did not affect the drop out rates (see Table 11). As in the lower educational levels, comparison of achievement between private and public institutions must control the analysis for intervening factors.

-26-

Lack of selection in the entrance process has generated other selection mechanisms, mainly "remedial courses" to up grade high school graduates to the accepted entrance levels required by the university or a drastic selection at the end of the first school year. In the Instituto Tecnologico of Ecuador most students are accepted in the preparatory course Table 11:

Year 1962 1967 1972 1977 1979 1980 1981 1982 1983 1984

GUATEMALA.

Annual Fee 400 400 400 400 400 600 600 600 600 800

UNIVERSIDAD RAFAEL LANDIVAR

Number of New Students 138 506 586 1,662 2,465 2,257 2,420 2,028 1,778 1,503

(CENTRAL CAMPUS)

Drop out Rates n.a. n.a. n.a. n.a. n.a. 17% 27% 24% n.a. 25%

n.a.= non available Source: Annex 5

and only the best are accepted as full fledged university students the first year. In the National University of Ecuador all postulants must be accepted by law, but then the selection is carried out at the end of the first year. (Pozo et al, 1981). The selection at the end of the first year generates a large premature drop out, but some students that would have been screened out through an entrance exam-may attend classes during the first year and eventually succeed in their studies. The more selective the access to a career the lower is the wastage during the career. In those careers where students must approve the more demanding exams (for example, Medicine or Engineering in many countries), the lower attrition rates are observed. But again, the students accepted to those careers are those that have benefited from high family socioeconomic background, well equipped elementary and secondary schools, good teachers and stimulating environments. Selection to graduate studies is open to all graduates from higher education and the selection is done among an already selected group. Therefore, at that level social background seems to play a minor role. Latin America is now expanding graduate studies. There is still an opportunity for most people interested in graduate studies (and willing to pay the opportunity costs for reducing their paid work) to have access to available programs. Attrition rates are still relatively low. However, this is an area where scholarship programs must be offered in order to get the best possible candidates (Pozo et al, 1981).

-27-

Summary Socioeconomic, geographic and ethnic factors affecting present educational selectivity allow increments in achievement levels of a selected group, but probably reduce the overall achievement level of the region. Selectivity allows that all children from upper income levels (including the children of decision makers) be attending good schools, therefore no automatic incentives are built into the society for improving the quality of schooling for those negatively selected. Most countries have made efforts to open vocational opportunities that reduce the number of candidates for university admission, but no one has yet succeeded, mainly due to a salary structure and traditions that associate a high payoff to university graduates. Free tuition does not reduce an educational selection process that starts at the cradle and continues in the admission to specific university careers and in the graduation. In fact in some cases, as in the admission into the university, the free tuition system exacerbates inequities. B.

PRIVATE CONTRIBUTION

TO THE FINANCING OF EDUCATION

The previous section has suggested that reducing the internal inefficiency of the educational system involves an increment in the unit costs per student and a reduction in the unit cost per graduate. At the same time, the total amount of resources must increase to make those improvements and a source of funding has to be found. Although the public sector has played a key role in funding the expansion of primary and secondary education during the last three decades, the private sector has substantially increased its share in higher education in the last decade. Given the dwindling public funds resulting from the economic recession and the service of the foreign debt, the private sector is now the obvious source of funding. This section presents available evidence on past efforts made by the private sector in order to assess whether additional contributions could be expected. The magnitude of the present crisis is outlined and four experiences of tapping new resources from the private sector are explored. Attempts to introduce or raise user charges in government schools as well as expanding private schools and earmarked taxes for training institutions are commented on and compensatory mechanisms such as the provision of loans and scholarships are reviewed. Although during the sixties and seventies expenditures increased at a fast pace even when economic growth was sluggish (Annex 11), information about the recent recessive years suggests a large negative impact in education financing. It has been suggested that the increment in educational expenses during the sixties and seventies reflected an effort to compensate previous underinvestment in human resources vis-a-vis other factors. (Chenery and Syrquin, 1976). If that was the case, now it may be no longer true. Public funds for education-have sharply declined in the eighties in spite of the increased number of candidates. This decline has been caused by a reduction in GNP per capita. In turn there has been a general reduction on the percentages of GNP devoted to the state budget and a specific reduction in public funds for education. Almost all countries reduced their GNP per capita in the 1980-1985 period (a negative annual rate of 1.64% for the region), with some countries reducing their GNP per capita levels to the levels registered in the late sixties. This decline

-28-

led half of the countries to reduce the central government total expenditures as percentage of GDP during the 1980-1985 period. Finally, there was a decreasing public spending on education as a share of the state budget (from 18.9 in 1970 to 15.3% in 1980 for the region). In other words, education is receiving a decreasing share of a decreasing amount of social resources. Cuts in educational budgets are undermining attempts to raise the quality of education. Budget officials in several countries reported that educational expenses not related to salaries are the first to be eliminated to solve financial troubles. Private contributions to formal education are also growing at a slower pace and the attempts to accelerate that pace will be reviewed below. However, the fact that at a given moment a government made a special effort to increase public schooling rates should be kept in mind for analyzing the future development of private education. Private contributions to in-service training seem to be growing at a faster pace. The percentage of GNP spent in education by the public sector in Latin America is still low in comparison with developed countries (between 1% and 6.5% and between 4.1 and 8.% respectively in the early eighties) and the percentage of GNP spent in private education is even lower (around 1% in Guatemala and Paraguay where private education is well developed). In other words, there is ample room both for an expansion of public funds for education and for privatization of provision of educational services. With respect to private financing, the problem is to find some mechanisms for encouraging the private sector. This is the object of the analysis in each of the next five points of this section. B.1

Cost Recovery Policies by Level of Education.

Although there is a historic tradition of centralized taxes systems in Latin America, there are still important and increasing contributions from private sources that could be even further increased. The faster growing sources of revenue fall in the fiscal domain and have reduced the importance of local sources of revenues. Rapidly growing activities like education have forced an increasing state activity or increased transferences of funds to local authorities. In general, most reports mention the government as the only source for further expansion, (Franco, 1980; Ecuador, 1983; Donoso and Hawes, 1983) but those that have studied the actual contribution of fees have a different view (Boenninger, 1980; Schiefelbein, 1984). Table 12 shows a wide range of proportions in which different sources share the funding (tax contributions to vocational training institutions commented in B.3 are included in this table). For example, Brazil has managed to mobilize large private contributions to run the educational system, even though its educational system is less developed than the other four countries included in the comparison. Factors related to increased public share Several sources of pressure may be identified for increasing public funding of education. In many countries education is a communication channel more efficient than those provided by the political parties, army, mass media or the church. For example, construction of new schools may be offered to local groups in political campaigns or teaching jobs may also be used to pay political favors and to secure the loyalty of

-29-

officials. At the same time, governments prefer to finance students rather than to have open youth unemployment and the expansion of teaching staffs may provide a job to the annually increasing crop of graduates. Political ideals for academic freedom and university autonomy it is also a powerful force. For example, in Guatemala, the Constitution assigns 5% of the fiscal budget to the Universidad San Carlos (assigned 3% until 1983). In this case there is no way to modify the fiscal contribution short of changing the basic law. Sometimes the pressure comes from members of the congress that usually belong to a reduced number of traditional universities. Whatever the ideological differences, the alumnae rally together to approve laws that benefit their Alma Mater. This fact may explain the advantages obtained by universities (specially the traditional universities) in many Latin American countries. (Castro 1980; Frigoto, s/f). In sulmmiary,the dynamics of public and compulsory education may reflect, in some cases, strong influences of vested interests that forced increased public expenditures. How to measure private contributions The contribution of the private sector is sometimes underestimated. In the analysis of who pays for university expenses it could be assumed that about 50 percent of the total operating expenses for some higher education institutions are allocated to education while the other 50 percent correspond to other university functions such as research, or diffusion (in many private universities the proportion for research and diffusion may be substantially lower). The percent of university income financed by fees is only a portion of the percent of expenses for higher education. Table 13 shows that tuition fees in several public universities reach near 10 percent of total income, but the amount financed by fees represents a larger proportion of expenses in education. On the other hand, there are also important proportions of students receiving financial help from the government (see further comments in B.4). Therefore the net contribution may be also different from the figures presented in Table 13. Although the cost recovery figures suggest that there is ample margin for raising user charges in higher education, the figures also suggest that half a dozen Latin American countries have done a serious effort to raise as much as possible the university fees. In fact the differences in fees among Mexican public universities suggest that it has been politically unfeasible to raise the fees in the UNAM. (Levy, 1979). While the Chilean experience (see details in C.3) shows that tuition fees as proportion of university income may be as high as 25 percent, the attempt carried out in Guatemala in 1984 to raise the fees from US$5 per year to US$10 per year ended with a machine gun pointed to the heart of the president of the university. Cost Recovery bv Levels There are limited attempts for introducing user charges in primary and secondary education. Table 13 suggest that registration and tuition fees in formal education are cuite small (range 0 to 10% for all cases, but two) and frequently when parents have financial problems even those small fees could be easily waived. However, additional user charges are

TABLE 12;

RATIOS AMONG SOURCES OF EDUCATIONALEXPEINDITURES

RATIOOF EXPENDITUIIES OF TIJE MINISTRYOF EDUCATIONTO TOTAL I)DUCATIONAL EXPENDITURES

RATIOOF EDUCATIONAL EXPENDITURES RATIOOF PRIVATEEDUCATIONAL OF OTIERMINISTRIESTO EXPENDITURES EXPENDITURES TO TilEEXPENDIOF TIIEMINISTRYOF EDUCATION TURESOF TIHEMINISTRY.OF [LxCATIOIO

Averaqe

.50

Averaqe

.1

Average

.42.

Chile

.67

Brazil

.24

Brazil

1.11

Venezueld

.66

ColoMbia

.23

Argentina

.58

Colon)ia

.41

Brazil

.19

SOURCE: Zynce1iitan, 1973:

Chile

.14

Venezuela

.13

Argentina

.005

Coloobia

.51

Chile

.32

Venezuela

.13

Argentina

.04

0

-31-

introduced using a variety of different mechanism operating at the school level. The main mechanism is the so called "parent's voluntary (or somewhat forced) contributions." Two variations of this mechanism are operating in Latin America. One operates in subsidized private education and the other in the public schools (usually administered by the principal). These two mechanisms are discussed in the following sections. Subsidies as a Cost Recovery Strategy Subsidized schools may cut government expenses by half. Subsidized schools are private schools receiving some funding from the state and are allowed to obtain a "small" voluntary contribution from students' parents. In certain countries the subsidy is an amount per student enrolled, while in other there is subsidy per day the student attends school (the case of Chile), in other countries, lump sums (the case of Paraguay) or payment of a few teachers (the case of new schools in Paraguay or Guatemala). There is even a system where the government pays the teaching staff including the school principal, but the principal is a member of a religious group (the case of the "Fe y Alegria" schools in the Andean countries) that provides some support for the non-wages costs of the school. In all these cases the state pays only a fraction of the total-cost of education and the balance is paid through "voluntary" contributions from parents, teachers or other community groups. Table 14 presents a summaary of the direct contribution in tuitions and registration fees in the case of Paraguay. When all the contributions are taken into account they represent 21% of the total educational expenditures (not including family expenditures related to education). Further comments on these subsidized school are included in the next section (B. 2.). Table13: COST RECOVERYBY EDUCATIONAL LEVEL IN LATINAMERICA,AROUND1980

Country

User Fees as % of Unit PublicCost Primary Secondary Higher (a)

Bolivia Brazil Chile Colombia CostaRica DominicanRepublic Ecuador Guatemala

.8 1.6 .3 0 0

.4 .9 .5 0 0

Haiti

6.8

3.4

Honduras Mexico Paraguay Uruguay

0

9.6

4.1 .5

2.0 .4

Average

0.9

1.7

1.0 5.0 25.0 3.4 8.0 1.0 2.0 10.0 10.0 3.-70 .7 6.0 6.8

(a) Giventhat unit cost in highereducationmay includeresourcesfor researchservices(for examplehospitals)or extramuralactivities, the figuresonly correspondto a 'minimum, percentage.In the case of Chilethe realpercentage of the unit cost per studentis estimated in 30 to 50e. = not available

Source: Calculated by Psacharopoulos, Tan and Jimenez (1988) from Schiefelbein (1985), except Colombia: Gomez,H., (1984); and Bolivia and Haiti, Ainsworth(1984). See Annexes6 and 7.

-32-

Table 14:

PARAGUAY.

ESTIMATED FAMILY CONTRIBUTIONS IN 1983 (in Guaranies) No. of Students

Annual Tuition

Total

TO EDUCATION

PRIMARY Public Private Subsidized Private

473.8 40.9 35.0

100 40,000 1,000

47.4 1,636.0 35.0

BASIC SECONDARY Public Private Subsidized Private

67.6 18.1 5.8

1,300 47,900 2,000

87.8 862.0 11.6

BACHILLERATO Public Private

IN HUMANITIES 29.2 13.9

2,000 52,640

58.4 731.7

BACHILLERATO Public Private

IN COMMERCE 5.8 2.1

2,000 20,000

11.6 42.4

1.5 0.1

2,900 20,000

4.1 2.0

18.0 7.0

3,000 80,000

54.0 560.0

TECHNICAL BACHILLERATO Public Private UNIVERSITY Public Private

Contribution 1,718.4

961.4

790.1

54.0

6.1

614.0

TOTAL

4,144.0(b)

(a) The estimatesincludetuitionfees but do not includeclothing,food or transportation. (b) This total represents 21% of the totaleducational expenditure. Source: Numberof studentswere providedby SEC - planningoffice. Annualtuition- estimated by IBRD mission,1984

Other Cost Recovery Strategies Local funding of public schools is an important source of funding primary education and is contributing about one third of the total student cost. Both contributions in food for teachers and manpower for construction and maintenance of school building have been provided in several countries (Paulsen, 1982; Lopez, 1980). The contributions in species are mainly provided in rural areas specially in the poorest areas (Schmelkes et al, 1983). Private contributions (including contributions from the teachers' pocket) may average 30% of the cost or rural primary schools in Colombia (See Annex 8). Comments from teachers in six countries suggest that similar percentages may be exacted in rural areas of other countries. Table 15 suggests that the private contributions are mainly

-33-

used for individual materials, furniture, and construction or maintenance. In certain cases parents feel that there is too much pressure for them to contribute; their children sometimes are sent back home when the parents have not sent their contribution. What is clear is that in L.A. most public schools could not operate unless the parents contribute. Teachers' salaries may be another mechanism through which the government may be recovering part of the cost of training the teachers in tuition free public institutions. Teachers' salaries mainly depend on decisions of the government and are relatively low with respect to other professions that require a similar training. If the government reimburses teachers for their service at rates below the value of those services there would be a case of "bonding" or social service that would be reducing the real cost of education. It should be noted that the recent economic crisis has not reduced the magnitude of the enrollments, but only the salaries of the teachers (See C.3). The high rates of teachers moving out of their profession (Census of Population show a large proportion of teachers working in other sectors while untrained teachers are working in rural schools) may also suggest lower salaries than the levels required to keep teachers in their profession (Munoz and Schmelkes, 1983). Private education provided to preschool students or to special students (blind, deaf or retarded) only caters to high income families, with only a few exceptions. In several L.A. countries there are now operating once a year national TV "Telethons" which are collecting money for special education for people with little resources (in Chile each year around US$2 millions are collected and similar efforts are being carried out in Colombia, Peru and other countries), but still most poor families have little hope for their children when they need special treatment. Non formal education is also an additional mechanism for channeling private resources into educational activities. Non formal education is able to mobilize resources from the private sector (including massive amounts of time provided by the monitors and instructors in those courses) or the participants themselves into educational activities well fitted to the needs of special groups in each country, especially the poorest groups. Contributions are provided in varying shares by the leaders of the non-formal activities and the participants in those activities only one third of the funds are provided from public sources (Swett, 1983). Specific remarks on employer-financed vocational training are presented later on in section B.3. Summary Latin America has been able to reach an important (30 to 40%) level of privatization of provision of all educational services in the region. Subsidies to non-fee private education have reduced the total amount of public resources devoted to education up to 50% in the extreme cases. Although subsidized systems are available only in a few countries,

-34-

Table 15: COLOMBIA COST STRUCTURE IN SIX SCHOOL - 1980 Item Total annual cost (Col$000) Distribution Current Expenditures Personnel teaching staff principal materials groups individuals Services Transport Water and energy Boarding Recreation First Aid Public relations

Case 1 Case 2 1,584.0

Case 3

Case 4 Case 5

254.1

591.6

295.7

100.0% 93.2

100.0% 80.1

100.0% 89.6

100.0% 100.0% 90.7 82.7

62.5 .6

40.4 .5

55.1 1.7

.3

.7

1.5

43.7 1.9

.3

25.3

31.2

24.2

39.4

1.1

1.2 .8 .3

1.4

2.0

2.8

348.7

65.8

.5 13.0

Case 6 994.0 100.0 % 90.4 51.0 1.1

.9 29.8 1.8 .1 1.5

.3

2.2 .1

.6

.7

3.4

3.4

3.4

2.0

Investment Land Buildings Classrooms Health services. Recreative services Teachers' house Circulations/deposit Equipment Furniture Teaching aids Library Other

6.8

19.9 .6

10.4 .2

9.3 .3

17.3 .1

9.6 .6

3.0

10.2 1.4 1.3 3.8

3.6 .8 .5 .5 2.6

5.7

5.1 .7

4.1 .5

2.4 5.3

.5 1.6

1.4

2.2

1.3 .3 .7 1.5

.8 .8

Maintenance

1.3

Maintenance

Other Programs

2.2

1.0 .8 .4 .2 .1

.6 .9 .9 .3

.1

.1 .7 1.0

.4 .3

(31.6)

(Not included in the

total annual cost)

Note: Indirectcosts (forexampleadministrative costsof the regionor centralgovernment) are not included. Source: Project'Costosde la EducacionPrimariaRuralen Colombia- Estudiode Casos", OFISEL,1981.

-35-

those systems could be extended to the whole region. Several countries have been trying to raise user charges in university education, but there is a need for providing more objective evidence to the public opinion of the rationale for doing so and for diffusing government reliable strategies for introducing change in the present tuition free policies. Therefore, the future contribution of private funds may be still expanded in several countries of the region. B.2.

Private Schools

Although private schools are now the most important mechanism for training students and for channeling resources from the private sector into the educational activities of each country, governments do not have clear policies for further expansion or improvement of private schools. The main role of Latin American private schools is to provide the upper social groups a better quality of education or a different type of education, than the education provided by the public system. Therefore, different types of private education may be identified. All types of private education tend to be concentrated in urban areas where the population is more dense and family income is higher. Only some non fee Church schools or schools provided by landowners are sometimes found in rural or isolated areas. Proportion

of Private Enrollments

Table 16 presents the proportion of enrollment in private institutions in each of the three levels. In primary education between 5 and 18% of the students were studying in private schools (although Haiti had 42%) in 1975. Percentages in secondary private education were much higher (14 to 45% with the exception of Costa Rica with 6%), but they were slightly lower in higher education (between 10 to 30%) with the exception of Brazil and Colombia where private higher education exceed 50% of total enrollment in that level. It is difficult to detect a trend in the proportion of private education over time. The figures in Table 16 may include some mistakes and it is questionable whether trends may be identified. However, small changes are detected in most countries, with a few exceptions (such as Haiti or Bolivia). Some countries even increase the private contribution in primary education and decrease it in secondary education (See for example Colombia). In countries that have finally reached 100% access to primary education (for example Chile and Mexico) in the 1965-1975 period, the proportion of the private contribution has been reduced. The high rates of growth in secondary education (mainly explained by the governmental effort) have also reduced the proportion of private secondary/ education in total enrollments. More data are required for higher education in order to identify significant trends. Types of Private Schools Behind the common label of "private education" there are several types with special characteristics. Differences are observed between schools sponsored by religious groups (probably three quarters of private schools are sponsored by the Catholic Church in Latin America) and schools sponsored by laymen. Church schools attend both high and low socioeconomic

-36-

groups and some redistributive role is played with the income from their operation. Schools sponsored by laymen mainly cater to high and medium socioeconomic levels of the population, but those rural schools sponsored by small philanthropic groups and schools that the law demands landowners or firms to operate. On the other hand all the private schools usually follow quite closely the curricula instructions of each Ministry of Education (Brahm and Cariola, 1971; Balderrama and Del Castillo, 1976; Mun5oz and Hernandez, 1978), with the exception of the internationally oriented schools. Fee paying schools are academic oriented (most of thefr graduates go to universities) and tend to provide instruction in a foreign language (usually English) since the first grade, while free and public schools tend Fee paying schools also to offer a second language only in high school. provide better sport programs and provide better life "connections" through the classmates or their parents. Although there are private schools for students requiring special treatments their number is reduced and are usually sponsored by Extreme climatic conditions, risky zones, philanthropic institutions. isolated points, or students requiring high unit cost tend also to limit private activities. Table 16:

PERCENTAGE OF PRIVATE SCHOOLS IN TOTAL ENROLLMENT, LATIN AMERICA, 1965-79

Country

First Level 1965 1975

Argentina Bolivia Brazil Chile Colombia Costa Rica Dom.Rep. Ecuador El Salvador Guatemala Haiti Honduras Mexico Nicaragua Panama Paraguay Peru Uruguay Venezuela

14 26 11 27 14 4 7 18 4 19 26 7 9 16 5 10 14 10 13

17 9 13 18 15 4 12 17 6 14 42 5 6 15 5 13 13 17 22

Second Level 1965 1975 41 26 49 38 58 24 38 47 54 43 53 29 44 17 24 24 17 23

45 24 25 23 38 6 30 47 43 76 51 25 14 37 17 18

Third Level 1965 1975 50(a)

57

-

21(a) 13(a)

17 15(a)

13 20 (d)

13

28 (d) (b) 10

1970/71 1973 Only in September 1984 the private Catholic University activitieswere given official approval. 1983 1_ Meansnot available Sources: E. Jimenez, 1984: Appendix 4.3 (Data from the Statistical Division of Unesco Education Sector Review, 1980, the World Bank). S. Lourie in Brodersohn and Sanjurjo, 1978: Tibi, C. Financement et Development de l'education: Le Cas de la Colombiew, IIPE, Paris, 1976, and Hay, George Andrews, 'EducationalFinance and Educational Reform in (a) (b) (c) (d)

-37-

Financial

Impact of Private Education

Table 17 provides an idea of the financial impact of private education. In certain countries private education shares one half of the total resources. The share of private education increases in the higher levels of the system. The relative size of private education in each country seem to be negatively affected by the rapid growth public education and positively affected by incentives (for example grants) given by the government (DEALC, 1981). In fact in some countries the large size of the grants to universities (for example in Ecuador and Chile) make it difficult to determine whether in fact it is possible to establish a difference among public and private institutions. Although there were no clear trends in the specific effects of changes in GNP on changes in private expenditures in education during the seventies, the recent economic crisis probably reduced the share of private educational expenses in many countries. During the seventies, some countries tend to increment the proportion of private expenditures when higher levels of GNP are reached, while the opposite happens in other countries (see Table 18). Therefore little can be said from the data on five Latin American countries during that period. The trends depend on the specific policies designed by each country. However, present trends may have better defined patterns. Fragmentary evidence suggest that in recent years governments have been reducing public expansion of higher education and fee paid private primary and secondary schools are facing reduced enrollments. Public discussion of fees in private schools (mainly for mid class is also becoming more frequent and public regulation mechanisms have been able to affect the level of fees (see C.2 and Annex 18). Table 17.

PERCENT PRIVATE IN TOTAL NATIONAL EXPENDITURE ON EDUCATION a/

Country Honduras b/ Panama b/ Venezuela b/

1970-74 a! 57.1 55.0 52.3

1975-78 a/

1984

51.6 52.8 51.2

Paraguay

21.0

Selected OECD Countries Australia Belgium Greece Japan b/ Spain United Kingdom United States

11.6 2.2 37.4 61.2 52.7 25.0 21.6

5.6 1.9 35.3 55.3 51.4 22.2 20.5

Source: E. Jimenez,1984:Appendix4.2. IBRD,ParaguaySectorMemorandum,1985

a/ Tan, J.P. 'PrivateEducation:Some MacroTrendson Enrollmentand Expenditure, mimeo, EDC, The WorldBank,Dec. (1982). The individual countryfigure in each periodwas calculatedas the averageof the data for as many years as are availablein that period. b/ Figuresfor educationincludeexpenditure on recreation, entertainment, educationand culturalservices.

- 38 -

Subsidies to Private Education Subsidies tend to be allocated in terms of specific cases rather than considering the overall impact in the educational system. In general, government subsidies are only given to non-fee paying private schools. However, there are a few countries (for example Paraguay) where some feepaying private schools get special transfers from the education budget. There are also special arrangements for appointing as civil service principals of public schools some educators belonging to religious groups; those principals are allowed to appoint as civil service teachers the candidates they select (some or all of those teachers are paid by the government). In general, those schools operate as private schools (for example "Fe y Alegria" in Colombia, Peru or Bolivia, as described by Munoz and Hernandez, 1978), and receive some extra support from religious groups. Their efficiency is usually considered much larger than the average public school.

Table 18: ELASTICITY OF PRIVATE EXPENDITURE ON EDUCATION, WITH RESPECT TO GDP, SELECTED COUNTRIES AND YEARS

Region/Country

El Salvador Honduras a! Panama a/ Puerto Rico Venezuela a/

Year

1970-77 1970-75 1970-72 1970-78 1970-78

Elasticity

Private Expenditure as % of Consumption (1975)

.62 .57 .82 1.37 1.11

2.1 5.1 7.8 1.0 8.9

.38 1.00 .99 1.06 .96 .95 1.07 1.04 1.00

.5 .2 2.8 .3 1.8 9.0 2.2 2.2 2.0

Selected OECD Countries Australia Belgium Canada France Greece Japan a/ Spain United Kingdom United States

1970-77 1970-78 1970-78 1970-78 1970-78 1970-78 1970-77 1970-78 1970-78

Source: Tan (1982)quotedby E. Jimenez,1984. a/ Data for educationincludesprivatefinalconsumption expenditure on recreation, entertainment, educationand culturalservices.

In a few cases (for example Paraguay and Guatemala), the Government gets slowly involved in the financing or private schools under the pressure of community groups. For example, in rural areas some parents build the school and pay the first grade teacher. Later on they ask the Ministry of Education to pay for that first grade teacher and the community

- 39 -

hires on its own a second grade teacher. The process is repeated until they reach the normal level of operation of the school. While this procedures stimulates early community involvement it may also generate inefficient allocation of new schools or inability to catch enough students in the future to reach the range of optimal school sizes (or even to offer complete schooling). Subsidies to private schools seldom account for more than 10% of public expenditures in the corresponding level (not including grants or loans to students for them to pay school fees). In most cases subsidized schools are not legally allowed to use their own sources of funds to raise the level of the inputs for increasing educational quality. However (as shown in Table 14 and 15) for Colombia and Paraguay) many of those schools get "voluntary" contributions. All private schools receive indirect subsidies because they are not taxed in their services or in their income. Contribution from firms are tax deductible as expenses and there is no transfer (donation) tax on those contributions. In a few cases the state has helped capital formation. In the last decade several private universities have borrowed international development banks in favorable conditions with the state guarantee. Campuses of the University of Ladivar in Guatemala, University of Costa Rica in Costa Rica and Catholic and Concepcion Universities in Chile have been financed by the IDB, with the guarantee of the state. This is a mechanism that could also be explored for application to private secondary education. Fee Levels Although fees are usually established by each school, in some countries (for example, Colombia and Chile) fees have been under the control of the Ministries of Education or Economics (see section C.2). In some cases, Parent Associations fight for keeping low fees, but the most exclusive schools sometimes get the approval of parents for increasing fees in order to select the students only among families that can afford those higher fees. Table 19 presents a distribution of fees in Latin America. There are wide differences in each city, but the higher fees are detected in the capital city of each country and in the English speaking schools (Munoz and Hernandez, 1976; Brodersohn and Sanjurjo, 1978). Private institutions have experimented with success differential pricing higher fees for families from higher socioeconomic levels since the seventies (Boeninger, 1980). In general, differential pricing is a positive experience, but requires a constant supervision to detect people not willing to pay according to their real economic income. While the income of those household heads receiving a salary is easily established, there are serious problems to do so with respect to the income of people working as self-employed professional or high level managers. The selective impact of fees is probably related to income. income, Household expenditures in education rapidly increased with higher 1976). The with elasticities in the range of 1.3 to 2.3 (Musgrove, increment in these expenditures is related, among other factors to the a greater propensity to buy of children to fee-paying schools, sending other types of non-formal education (dance, piano, foreign languages, crafts and other complementary skills) and the increase probability that

Table 19; DISTRIBUTION OF ANNUAL FEES IN PRIVATE EDUCATION - 1984 (in US$)

Country Ecuador a/ Guatemala Dominican Republic Haiti b/ Costa Rica Colombia c/ Uruguay d/ Chile e/ a/ b/ c/ d/ e/ f/

Primary Education Average High

Low

100 100 100 100 150 30 100

-

800 -

-

-

-

500

600 300 400 400 300 1,000

Secondary Education Low Average High -

-

-

-

200

100 300 300 150

600 500

-

600 300

1,000 -

800 1,200

Thereis an initialbonusof USS250. Plus an annualbonusof USS20. Transportation and lunchmay represent someUSS600to USS1,200to compensatelowfees. Thereis an initialbonusof US$30. Lunchrepresents USS120and transportation USS160. with no publicsubsidies. Takingintoaccountonly the new threeuniversities

Source: Estimated by officials workingin the statistical unitsof the corresponding MOEs.

Higher Education Average tH

Low

400 -

400 1,200 f/

-

500 -

1,400 -

1,900 2,000 f/ 0

- 41 -

the children will reach more expensive (higher) levels of education. In any case, higher fees should have a selective effect (see Table 11). The study of effects on tax deductions for educational expenses has been explored, but the direct taxes are still quite low in Latin America and its impact does not seem to be substantial yet (Tanzi, 1978). Public Control of Private Schools Although private schools are relatively free to design their own programs, their curricula is usually similar to the official one. In most countries in order to open a new private schools need a permit from some authority in the Ministry of Education and sometimes from the Ministry of Health must be obtained. Private schools are sometimes loosely supervised and must provide statistical data to the Ministry of Education. In general, the academic and administrative controls are very flexible, if any, but those mentioned above in terms of fixing fee levels. However, from time to time, there is always a small number of private schools (and in some countries there are universities) that are accused of "selling" diplomas without regard to the academic work of the student. Therefore in many countries students had to take exams with special commissions designated by the Ministry of Education; however, this is no longer in operation in most of the countries. Comparison of Achievement

Levels

In general, students from paid private schools:show the highest levels of achievement, but the analysis of such data must be careful. Table 20 shows the results of testing in all Chilean schools. The analysis is done by types of schools (P= paid; private; F = public/Ministry; M- public municipal; S = subsidized private); by regions (AM = metropolitan areas; GC = large cities; RP rest of country) and by socio-economic levels (A = high; B = medium; C = low). Chilean primary students from paid private schools are followed by those from the-subsidized private schools and by the students from public and municipal school, even when the socio-economic levels and regions are taken into account (however, similar studies that have used the students as the unit of analysis have shown that in Chile there are no significant differences in the quality of schools). High school Colombian students in private education got higher scores, even when the regime and sex was controlled (see Annex 26). However, in each country private education is "polluting" the rest of the educational system by Sometime, excluding those students with learning or behavior problems. repeaters are not allowed to continue in private education, while the public schools are forced to accept them. There are no regulations with respect to the failures created by private schools. Therefore, scores from public schools are negatively affected and the scores from private schools are upward biased with respect to the "true" scores corresponding to each type of school. Therefore, a simple comparison of scores by types of schools may be misleading. Table 21 also shows that achievement levels in elementary private schools in Paraguay and Bolivia are better than the achievement levels in public schools. These good results reached by students in private schools are obtained in spite of larger student-teacher ratios in private educaLion (in general they do not affect results as seen in Table 7). Given that urban schools use similar technology as reflected by inputs and curricula the larger student-teacher ratio in private schools may explain that the ratios of unit costs in urban areas are lower than the unity (better teachers or better management would be possible explanations for better achievement levels with lower costs).

_ 42

CHILE - STUDENT ACHIEVEMENT BY TYPE OF SCHOOL AND SOCIO-ECONOMIC BACKGROUND - 1982 (Achievement Scores in Mathematics)

Table 20:

Region (Level of Urbanization)

-

Background (SES)

Public Schools

Municipal Subsidized

Private Subsidized

Private Fee Paying

Metropolitan (Santiago)

High Middle Low

75.9 53.4 50.0

64.7 53.5 49.5

64.3 57.0 51.8

77.3 71.2 a/

Other Large Cities

High Middle Low

58.7 50.6 63.5

62.4 57.1 51.2

68.5 60.2 49.1

73.6 68.8 a!

Rest of the Country

High Middle Low

63.5 52.6 47.9

61.8 53.7 45.6

69.7 57.9 44.8

75.5 71.0 a/

Source: Serie de Estudios, No. 81, Centro de Perfeccionamiento,Experimentaci6ne Investigaciones Pedagogicas,Lo Barnechea, July 1983. Note: Data by regions is available for Mathematics (Serie de Estudios No. 117, June 1984) and in Spanish (Serie de Estudios, No. 90, November 1983). More details are presented in Annex 31 a/ Families from low.socio-economicstrata cannot afford to send their children to private fee paying schools. Therefore, there are no cases for this cell.

Table 21:

Country

INDICATORS OF PRIVATE VERSUS PUBLIC EDUCATION SECTOR EFFICIENCY. 1976

Variables

Overall

Bolivia

Average Variable Unit Costs Average Total Unit Cost Student-Teacher Ratio Reading Score Science Score

0.716 0.886 1.76 1.21 1.30

Paraguay

Average Variable Unit Cost Average Total Unit Cost Student-Teacher Ratio Reading Score Science Score

0.827 0.742 1.21 1.21 1.06

Ratio of Private to Public Rural Urban

Sources: E. Jimenez, 1984. Table A.5.8 with the ECIEL Data Tape/ Note: Blanks indicate that data are not available.

1.90 1.90 0.889 1.61 1.51

0.674 0.877 1.63 1.36 1.25 0.827 0.742 1.21 1.21 1.06

-

43 -

It is interesting to note that many leading private schools have larger class sizes than the national averages. In good private schools there are constant pressures for receiving more students, therefore classrooms with 40 or 45 students are common practice. It should also be noted that in spite of those large class sizes students obtain much better achievement levels than students in public schools, but the differences tend to disappear when controlled by the family socioeconomic level. Summary Countries have no clear policies for the role that private education should play in the whole system. Private education cater mainly for both the upper socioeconomic suburban class, but it also enrolls some students from the population living in urban-marginal or rural areas. The magnitude of the private educational activity in the lower socioeconomic groups depends on the amount and conditions of public subsidies. A fair number of experiments have been mentioned above. (See also Munoz and Hernandez, 1978) It is necessary to analyze in detail their pros and cons with respect to well defined criteria. Capital loans from international organizational (with state guarantee) could also be considered as one type of subsidies to be included in such study. There seem to be a contraction in the educational investment of the Catholic Church schools during the seventies and eighties (Catholic Church is now less interested in education in comparison with pastoral activities) that should also be studied in more detail, while there are new activities of industrial groups that have funded or supported regional universities that should also be analyzed, may be through a series of relevant case studies. B. 3

Employer-financed

Vocational Training

Although Latin America developed since the 40's a strong system of training institutions patterned in the ILO hands-on methods, employers have also tried out other experiments in the region that deserve careful study. Professional Training Institutions Since the forties Latin America has developed an excellent handson vocational training system financed with taxes from the payrolls. The Brazilian and Colombian models helped other countries to define autonomous institutions with directive Boards made of representatives of the industrial organizations. Later on other institutions have been created with goverrment support and training is now also provided for peasants. Most of the institutions are linked to Ministries of Labor or Industry and have weak ties with the Ministry of Education of the country (see Annex 9). Most of the training is provided to adults. Only 15 to 30% of the teaching time is allocated to youngsters (apprentices) in half of the institutes, while the other half only trains adults (for a detailed analysis of the apprenticeship system see Corvalan, 1981). Table 22 presents a description of the main financing sources used by the institutes. The payroll tax ranges between 0.5 and 2% but income and export taxes are also earmarked for training in a few countries. The contribution of the private sector exceeds the contribution of the public

- 44 -

sector. However, the total amount spent in this type of vocational education represents less than 1% of the total expenses of the formal educational systems in Latin America (Castro, 1979). Only in five countries the enrollments in these institutions exceed 4% of the enrollments in primary and secondary education of the respective country. Costs per student and per hour of class are presented in Table 23. The figures for administrative expenses are quite high, but there seem to be some inconsistencies in the data, and the figures should be revised in the future (for example the development prograzmned instruction has been included in the administrative costs in stationery, rather than considering the costs as an investment in preparing the programmed educational materials). Alternative

Strategies

for Training

Although in average the professional training institutions train around 2% of the labor force each year, this activity seems to be much lower than the training efforts carried out inside the firms. A survey carried out in Bolivia suggest that 9% of the urban labor force would have each year some training activity financed by the firms themselves [PREALC, 1975). However, little attention has been paid in most countries to help the internal training process. The fragmentary evidence on impressive amounts of internal training has been the basis for experimenting in Chile a new training schema. Each firm may deduct from its taxes the amount of resources devoted to training (with a maximum of 1% of the payroll) if the training programs have been previously approved by a public institution (the Servicio Nacional de Capacitacion y Empleo SENCE). This procedure is equivalent to provide public funds for all "publicly approved" private training and reduces the negative impact of the payroll tax on the use of the labor force in comparison with capital. This schema has been complemented with some limited public funds for financing training courses for prospective workers, but the funds must be returned to SENCE if less than 65% of the trainees are employed three months after the end of the course. This requirement is very difficult to meet in countries with high unemployment levels and few funds have been actually demanded for training (Oteiza, 1983). Table 24 shows the reduction of enrollment in training courses in 1981 when the requirement of 65% of the trainees becoming employed in three months was established. During 1984 the training courses lasted (in average) 117,8 hours and US$10 per hour were paid to the training institution (US$0.50 per hour/student), but each proposal could bargain the price per hour. There was a bidding system to allocate the total amount of subsidies available for the year. The subsidy was paid when a proof of the required percentage of working graduates was provided. In the last four years the system has been used by firms that are hiring new workers and get them trained with public money or by institutions willing to falsify the proof of "working graduate". SENCE has not paid those institutions where suspecting "proofs" have been identified. Finally, instead of reducing the minimum percentage of employment after training (that seems to be too high in a country with 20% unemployment at that time) to 30 or 40%, SENCE decided to cancel the program.

_ 45 -

Table 22:

Country

FINANCIAL SOURCES FOR THE PROFESSIONAL TRAINING INSTITUTIONS IN LATIN AMERICA - 1974

Institution

Sources (% of Total Amount) Private Other State Sector Foreign Sources

Total Amount Of Resources (in US$'000

2.8

-

1.9

-

6.8

100.0 100.0

100.0 100.0 -

-

-

-

70,434.62 30,702.62 4,886.18 5,574.64 25,163.52

Argentina

COMET

85.2

12.1

Bolivia

FOMO

91.3

Brazil

SENAI SENAC PIPMO DMMO

135,823.40 278.42

Colombia

SENA

1.1

98.9

-

-

Costa Rica

INA

-

97.8

-

2.3

2,438.72

Chile

INACAP

89.9

2.8

0.6

6.7

1,995.36

Ecuador

SECAP

11.5

47.3

41.1

1,718.73

Guatemala

INTECAP

23.2

59.6

Honduras

INFOP

10.6

88.8

Mexico

ARMO

88.0

Nicaragua

INA

94.2

Paraguay

SNPP

-

Peru

SENATI

-

Uruguay

UTU

98.5

Venezuela

INCE

13.8

-

-

1,964.00

-

0.6

1,769.00

5.6

-

6.4

2,210.80

-

5.8

-

1,204.13

100.0

-

-

562.52

100.0

-

-

20,378.73

-

-

1.5

5,892.05

-

4.1

54,408.84

82.0

17.3

Sources: Claudio Moura Castro, 1979, pp. 20-21, with data available in CINTERFOR.

-

46 -

Table 23: UNIT COSTS PER YEAR AND PER CLASS-HOUR IN LATIN AMERICAN PROFESSIONAL TRAINING INSTITUTIONS. 1974 (in US dollars)

Country

Institution

Argentina Bolivia Brazil Brazil Colombia Costa Rica Chile Ecuador Guatemala Honduras Mexico Nicaragua Paraguay Peru Uruguay Venezuela

CONET FOMO SENAI SENAC SENA INA INACAP SECAP INTECAP INFOP ARMO INA SNPP SENATI UTU INCE

Total Expenses

278,430 88,873,000 23,895,750 22,832,680 2,281,050 1,911,310 1,238,760 1,833,000 1,087,000 1,951,200 1,203,430 488,310 20,378,730

5,713,490 54,411,180

Enrollment

192,731 682 415,836 396,382 389,583 9,142 31,934 10,351 9,687 4,148 18,871 1,S33 2,353 21,453 37,945 155,2S1

Unit Costs Per year Per Class

420 180 80 80 250 80 120 190 280 120 780 210 950

150 350

% of Expenses Allocated to Administration

1.25 0.94 0.24 0.55 0.32 1.79 1.84 4.47 4.81 2.07 1.71 5.14 1.80

89% 47% 51% 87% 42% 18% 49% 53% 83% 40% 88% -

Source: Claudio de Moura Castro, 1979, p. 28-30 with data from CINTERFOR. "" Means not available.

Table 24: CHILE SERVICIO NACIONAL DE CAPACITACIONY EMPLEO VOCATIONAL SCHOLARSHIPSPROGRAM.

Year 1979 1980 1981 1982 1983 1984

Number of Scholarships 50,401 50,748 19,493 18,600 15,201 18,780

Note: In 1979 and 1980 the program was administeredby Guillermo del Campo. Since then it has been administeredby Maximo Silva. Source: SENCE, unpublished data, 1984.

-

47 -

The decision to ask the market forces to do the job performed by the traditional training mechanisms developed in latin America has faced heavy criticism (Salame and Barrera, 1982). It has been commented that market forces focuses all training in the short term and that training heavily relies on the criteria of the person or group that approves the proposals (allowing in certain cases to consider as training the European vacations of top executives). Whichever the specific results obtained in the initial trail the experience deserves some further future analysis and additional experimenting with alternative conditions. Colombia has developed another interesting employer-financed training schema. The Fundaci6n Grupo Social is now the fourth more powerful industrial group in Colombia (US$500 million in sales per year). This holding controls some 20 different firms ranging from insurance to computation and construction. Profits from the commercial firms are used to expand business activities, but also to broadcast Educational TV, support cooperative development or formal and non-formal education. The are now providing vocational "Centros de Capacitacion" of the Grupo Social training to -some 15,000 workers per year and mass media advertising is reaching a larger audience. However, the main effect of the Grupo Social may be the on-the-job training or socializing of a new breed of Latin American executives that blend efficiency with care for the socially deprived. The mixture of personal of efficiency, trustworthy (the Grupo Social pays twice as much in taxes as the other groups), and common social development goals seem to explain the rapid growth of the group and its commercial success (Londono, 1984). -A similar attempt is being now launched in Venezuela and there is a group studying the idea in Chile. Higher Education Institutions In several opportunities when employers felt that better highly trained personnel was required (or when their offspring required a better education) they provide the resources and the management for launching suitable educational institutions for higher education. There are several universities funded by industrial groups such as The Tecnologico de Monterrey; Universidades del Valle, Santander y Los Andes en Colombia, Ingenieria in Peru or Santa Maria in Chile (Munoz y Hernindez, 1978; Vivas, 1972). University laboratories and research projects have also been funded by industries, mining companies and commercial firms. As suggested above is should be necessary to produce case studies on the relevant experiences in order to identify the elements that may help those groups interested in developing similar projects in the future. Training of teachers may also be considered an employerfinanced vocational training program given than most of the teachers are employed by the government. The possible "bonding" procedure has already been commented above (See B.1). The frequent agreement between the public health services and the Schools of Medicine may also be mentioned as mechanisms to be explored under the topic of this section. Finally the Civil Service Training Schools (The Peruvian ESAP and the Brasilian CENDEC are the best known) have been created in the seventies in several countries for improving management of public institutions, but their initial success has somewhat vanished later on.

-

48

-

Summary Employer-financed vocational training represents only 2% of the total number of students in the formal education, but the number of worker trained in their (usually short) courses is at least similar to the university enrollment and there are also many in-service training activities in private firms. Studies on the rates of return of their graduates show excellent results (Castro, 1978). Although the training system works efficiently and the system may be used as a model for countries that have not yet implemented this type of institution, additional resources to be obtained from payroll taxes may be relatively modest. Some interesting new experiments on employer-financed training and research are now being developed and future evaluations may provide valuable information for designing new employer-financed training mechanisms in countries that have not yet experimented them. B. 4

Experience with Student Loans and Voucher Systems

Although most of the Latin American countries have been able to operate student loans institutions, real recovery rates have been low due to negative real rates of interest( the higher rate of inflation transformed small fixed nominal rates into negative real rates of interest) and lending has limited equality of opportunities when a collateral is required as security for payment of the debt. The Impact of Inflation on the Lending Capital In 1952 ICETEX launched its initially successful lending programs. In 1983 there were 78,337 outstanding student loans representing US$38 millions (ICETEX, 1984). In 1983 there were 5,199 new grants approved and 17,347 were renewed with a total of US$6 millions. Although an 8.45% of debt had already reached repayment state and was in arrears, only 1% was considered absolutely lost. However, the capital of the institution had reduced dramatically its purchasing power due to the recovery in highly devaluated pesos. Low fixed rates of interest (only Argentina and Chile have interest rates linked to a cost of living index) in countries where high inflation rates have prevailed means that "loans" are too subsidized and, in fact partially correspond to scholarships. The effect of inflation mainly explains that although now some 20 countries have some modalities of Student Loans Systems (SLS), the present level of lending is much lower than a decade ago. Table 25 shows that some 500,000 loans were outstanding in 1978 and over 600,000 in 1981. 80% of those loans were made in Brazil. High rates of inflation in Brazil have virtually converted those loans in full scholarships (Velloso, 1984).

-49-

Table 25:

TOTAL NUMBER OF OUTSTANDING STUDENT LOANS, LATIN AMERICA 1978

Country

Argentina (INCE) Bolivia (CIDEP) Brazil: (APLUB) (Caixa Economica Federal) Colombia (ICETEX) Costa Rica (CONAPE Chile (Catholic University) Ecuador (IECE) El Salvador (Educredito) Honduras (Educredito) Jamaica (Students' Loan Bureau) Nicaragua (Educredito) Panama (IFARHU) Peru (INABEC) Dominican Republic (FCE) Venezuela (Educredito) (SACUEDO)

Total Loan Awarded Excluding Loans Already Repaid 1978 1981 1,400 476 3,084 354,588 53,865 1,286 1,982 15,803 2,350 1,740 6,875 630 5,800 274 10,097 2,866 2,770.

450,000 78,337

336

4,386

Source: M. Woodhall, 1983, with data from APICE, VIII Congreso Pan Americano de Credito Educativo, 1979 (RAE 2638).

- 50 -

Collection Costs and Performance Although a substantial amount of the resources are expanded in administration and subsidies have been generated by inflation, cost recovery rates are poor. In 1983 administrative expenses of ICETEX absorbed 24.32% of the resources for the SLS. The percentage is higher in other less efficient institutions. The subsidies to the students (borrowers) generated by rampant inflation some countries have reduced the real debt and made easier to collect. However, the managers are having some problems for collecting payment. While well managed institutions like ICETEX still are able to have rates of loss below 10%, programs in Ecuador, Chile, Peru and Venezuela have high rates of loss (around 20 to 30%). Colombia and Brazil have looked for alternative mechanisms and authorized commercial banks to provide loans to students (rediscounting the loans in the Central Bank with low rates of interest that provide utilities to the banks), but commercial banks have also problems for recovering the debts (in Colombia commercial banks have worse collecting records than ICETEX). Impact in Selectivity The limit to the total amount that a student can borrow has been estimated in 7.5% of the probable individual income (Dominguez, 1978). If this estimate is true the possible contribution of this source of financing is in fact limited, given the low income levels of Latin American countries. High attrition rates observed in higher education have forced ICETEX and other institutions to finance only students that have approved the first year of higher education (thus in fact denying the entrance to higher education to those students that most need a loan for studying). But still, students that have not finished their course work for a degree have problems to repay the loan. On the other hand when the laws of the country requires a collateral guaranty it is difficult for poor students to find someone (willing to act as a guarantor) that can be accepted as a collateral guaranty for the payment of the debt. Although Table 26 shows that a substantial coverage of higher education enrollment was reached by 1978, the coverage may be much lower now. In spite of the large coverage the impact on the demand for higher education might be limited. Given that the loan system has mainly benefited students from medium-high socioeconomic backgrounds that would demand higher education even without loans, the loan system should not have stimulated a large increment in the demand for education. Summarv The experience with student loans in Latin America shows that it is possible to have a reasonable efficient administration of a student loan system in a variety of conditions, but better systems must be designed for taking into account the situation of those not finishing their careers, the adjustment in payments due to unemployment, the effects of high inflation rates, the demand of collateral guarantees, the length of the initial grace period, the lack of suitable insurance systems and the effects of the criteria used for selecting the recipients of the loans (Woodhall, 1983; Velloso, 1984; ICETEX, 1984). Educational voucher systems have not been tried out in Latin America. In Chile a voucher system has been experimented in health. Particinants in the Social Security System may pay a percentage of the medical bills with vouchers issued by the SSS.

-51-

Table 26:

PROPORTION OF OUTSTANDING STUDENT LOANS DUE FOR REPAYMENT LATIN AMERICA 1978 AND NUMBER OF STUDENT LOANS AWARDED IN BRAZIL, COLOMBIA, ECUADOR, PANAMA & VENEZUELA 1976-78

Country Argentina (INCE) Bolivia (CIDEP) Brazil (APLUB) (Caixa Economica Federal) Colombia (ICETEX) Costa rica (CONAPE) Chile (Catholic University) Ecuador (IECE El Salvador (Educredito) Honduras (Educredito) Jamaica (Students' Loan Bureau Mexico (Bank of Mexico) Nicaragua (Educredito) Panama (IFARHU) Peru (INABEC) Dominican Republic (FCE) Venezuela (Educredito) (SACUEDO)

% Loan Due for Repayment 14.2 46.4 37.7 3.0 44.3 3.4 46.4 24.4 43.0 52.8 33.3 71.5 47.1 36.7 52.5 82.7 36.8 7.2

1978 Total No. of Students Higher Education

No. of Students. Receiving Loans 1976-78

1,251,116 211,302

388,415 56,422

235,274

14,271

-9,596

2,466

34,302

4,502

282,074

2,202

Source: U. Woddhall, 1983 with data from APICE (As Table 3.4); UNESCOStatisticai 1981 and UNESCO 1979. A. Franco, 1978: 482 Honduras, 1980

Yeark Book

- 52

-

In general the administration of such a system is quite complex and administrative costs are high. Differential pricing according to the family socioeconomic level has been successfully used by private schools and universities (See B. 2) and there are also small scale experiments in selected public institutions. The use of differential pricing could be further diffused.

B. 5

Systems Used to Allocate Scholarships

The allocation criteria for scholarships assessed needs and the academic achievement, however performance in sports has also been considered in a way for a few candidates to gain access to secondary Allocation

mainly consider the in recent years good few institutions as a and higher education.

to Assessed Needs

Scholarships for assessed needs are mainly provided in kind or in services. For example shoes, books and food are provided to students in There are different delivery mechanisms (for example primary education. food may be provided in a breakfast or lunch or in packages of food that may be taken home) in use in the primary level (Ecuador, 1983). Banking, books and some small amounts of money are given as scholarships in In several countries private secondary schools secondary education. provide free education to 5 to 10% of their students, especially for those students whose parents are in financial trouble or have died while the students are enrolled in the school (seldom are provided scholarships to newcomers not previously linked in some way with the schools through parents or close relatives). Similar arrangements are operating at the university level (Latorre, 1982). In Colombia the Ministry has asked private schools to provide scholarships as a condition for letting the schools to raise their fees. Access to lunch and scholarships is usually defined in terms of some criteria, but sometimes those criteria are not operational or are not easy to apply. In fact in most cases the teacher or the principal distributes a certain number of scholarships assigned to the school (seldom the criteria allows a centralized application that would reduce the number of scholarships to be assigned to the wealthier schools). Pressures and vested interests are influencing the number obtained by each school. Allocation to Academic Achievement Scholarships received for outstanding academic achievement are usually a small number. Table 27 shows that the total amount for scholarships is reduced (4 or 8% of unit costs) and most of that amount is allocated to assessed needs. The government (for example the Presidency or the Ministry of Education), private firms, philanthropic groups, or individuals are the main source of funds for those scholarships (See Annex 10). Although sports ability is considered for admission of students into the university in spite of their weak academic credentials and to waive or reduce the size of the fees, there are no scholarships for athletes. The distribution of scholarships for graduate studies at the university is made based on academic criteria (Oteiza, 1982). However, most graduate programs

-

Table

27:

53

-

SHAREOF DIRECT PAYMENTSTO STUDENTSIN THE EDUCATIONBUDGETBY LEVEL, CIRCA 1980 (pertent)

Country

Primary

Argentina Bolivia Chile Costa Rica Dominican Republic Guatemala Haiti Eonduras

1/

S-condary

1.9 10.5

6.5 2.2 0.4

-

8.7

-

0.8 1.3 1.3

2.3 4.2 4.8

0.9 2.3

4.1

1.8

1.8 0.1 -

-

Jamaica Mexico Nicaragua Trinidad & Tobago Uruguay Venezuela

5.2 0.5 3.7 5.3 5.8

2.0 0.5 6.4 1.1 0.3 20.6

Average

4.0

4.4

Figures are aggregate for secondary

Source: Unesco,

Eigher

-

.

/

1.5 7.6 2.5

29.9 20.6 1/ 8.2

and higher education.

Statistical Yearbook, 1984 prepared

by Psacharopoulos,et al 1986

Note: Figures refer to all forms of financial aid given directly to students, such as boarding, meals, transport, and medical services. They do not include iimplicit'subsidies as a result of free tuition. -"

means not

available.

-

54 -

rely on private financing. Only graduate courses in medicine and other professions mainly working for the government may get some government support (partial scholarships). Problems for Allocating Scholarships It is difficult to assess how good are the allocation criteria because the initially measured problems of the students should be reduced by the scholarships and it is expected that their achievement improve over time. Therefore pre and post test must be applied in order to detect how good is the application of the criteria. But available reports only provide cross section results in one point in time. The data shows that primary students receiving scholarships (related to assessed needs) have lower achievement levels than the average student even after benefiting from the scholarship (Donoso et al, 1980). There is need for a more systematic discussion of the allocation criteria for scholarships and the feasibility of applying those criteria. Public reports show that there is no acknowledgement of how different are the students that benefit from scholarships in terms of academic achievement. Even the descentralizationof decisions for the allocation process is mainly produced by the lack of management skills rather than by the belief in the efficiency of local decisions. Therefore no mechanisms for participation of community groups are usually provided (Mc Ginn et al, 1980). Use of a Market Mechanism An interestihg mechanism has been operating in Chile:since 1981. As described above (See the SENCE experiment in B. 3), costs of training unemployed people are paid by the government to the training institution if 65% of the graduates from the program are employed in the three months period after the course is completed (Oteiza, 1983). Of course this strategy places a heavy burden on the training institution, but it might be worthwhile for further explore the possibilities of this idea. In most cases the training institution already has a contract to provide trained workers for a commercial or industrial enterprise, thus distorting the expected effect. However, experimentationwith lower percentages of required after training employment and larger time periods for the graduates to become employed may lead to define an efficient mechanism (See Table 24). Summary In general the experience shows that the distribution of scholarships not directly related to assessed needs will tend to increase inequality. Precise criteria for selecting postulants and use of local groups to allocate available scholarships may contribute to an efficient allocation process. Lack of serious sanctions for perjury (given the Latin structure of the law) preclude the use of declaration under each as the only source for allocation. The weak tax control system also makes unreliable the use of tax declarations as the basis for allocating scholarships.

- 55

C.

-

LESSONS OF EXPERIENCE FOR IMPROVING ALLOCATION OF RESOURCES

Although the previous section suggests that tuition fees and employers' contributions seem to be the most probable source of additional contributions to education and training from the private sector, the amount of the contributions depends both on the excess demand and the responsiveness of enrollment to fee increases, as well as to other variables that are interacting when changing policies into practice. Tuition fees in public and private schools are a sure way to increase educational financing if such policies can be implemented in an efficient way. Employers' contributions to training are already substantial and the issue is how to improve the training process and support the economic development of the country. In this section the emphasis is in the analysis of attempts to improve the allocation of resources. The initial topic is the extent of excess demand at the current prices. The second topic is the evidence on the reductions in excess demand when fees increase. The third and fourth topics deal with experiments to ration student places through price and non-price systems. The final topic included in this review of research is the analysis of strategies used to alter the allocation within the education sector. It is difficult to avoid overlapping in the analysis of experiments in the financing of education with the topics analyzed above. For example, greater coverage of the population may be accomplished by lowering the required inputs per capita (moving toward the technological frontiers) or getting more funds by raising fees among those that can afford them. However, in this case the topics are discussed in the more dynamic context of policy implementation.

The Extent of Excess Demand for Education Although several factors seem to reinforce each other to generate a sustained social demand for more and better education, these factors are common to most Latin American countries. The basic interest in more education is enhanced by factors such as: social prestige of education; population explosion; the concentration in urban areas of a sizeable part of the demographic increments; exposure to radio and TV; access of household heads to minimum levels of education (four or five years of school) that make them able to profit from written information; social acceptance of universal primary education (and older school leaving age) as a goal to be achieved; and provision by the schools of food (breakfast and lunch) and health services. (Arriaga, 1972; Mayo, 1974; Zymelman, 1973; Anderson, 1971). Those factors are present in different magnitudes in each country thus creating particular patterns, to generate demand for education. Although the effects of each of the factors cannot be analyzed in this paper, their overall effect on part and present demand must be assessed.

-

56

-

Excess Demand In the 60's and 70's There was a clear evidence of a large social demand for education during the sixties and seventies when a large expansion of the supply of education was occurring and pressure groups were organized to demand even more education.. Both the political sensitivity, the rational provided by the Human Capital theory and the international forums (for example, the UNESCO/ECLA meeting in 1962, the Alliance for Progress and the Meeting of the Presidents in 1966) prompted the countries to substantially expand the supply of educational opportunities. The large annual growth rates in enrollments at all levels show that there was large excess demand for education in the sixties and seventies (Annex 11). The analysis of the net enrollment rates by single ages suggests that at present most of the demand for primary education has been net, but past increments in the primary level continues to press for further increments in the following two levels (UNESCO, 1983). Excess demand in higher education has been evident in the past both in terms of growth in enrollment and street demonstrations and The regional pressure groups demanding expansion of higher education. demand for higher education may be illustrated by a comparison of the actual enrollment rates at the university with the rates observed 30 years ago (see annex 15). The highest enrollment rate among all countries in 1950 is larger than only two figures in 1980. During the 1970-1976 period the third level enrollments raised its proportion to the 18-23 years old populatian from 5.2 to 1Q.4%. Women accounted for the largest increment, but they still trails the male percentage (Table 28). Although some factors like population growth have reduced their impact on the demand for education, unemployment is now one of the most powerful sources of demand for more education, the higher levels of education of the new generations and the lagging job opportunities (Annex B) explain that the minimal levels of education required for entry to a In spite of the higher traditional job have been rising steadily. educational requirements minimal salaries or even industrial or construction salaries do not show related increments in their real purchasing power (annex 14). Therefore, people feel the pressure of unemployment and want to have a relative educational edge with respect to other people competing with them for a reduced number of job opportunities. This competition has led to further expansion of demand for education (Uzategui and Muelle, 1974.)

Actual Evidence of Excess Demand Excess demand may only be measured now at the university level and available measures show a large excess demand. Table 29 shows that the ratio of applicants to entrants varies from 1:1 to almost 1:4. Countries like Guatemala or Ecuador have places for all applicants, including those willinc to nay $40 to $160 per month in private Guatemalan Universities and everybody in Ecuador given that national universities may accept all applicants since 1969-70 as shown in Table 29. The number of applicants in Chile include students graduating from high school the year before, but also high school graduates from previous years that are taking for second In Colombia the or third time the University Entrance Examination Test. number of applicants include each of the students times the number o-

universities the students applied to. In the case of Chile and Colombia, the number of students in the last year of high school is also included to help the reader to estimate the backlog of applicants or the multiple applications, respectively. Table 28:

LATIN AMERICA - FEMALE ENROLLMENT AS A PERCENTAGE OF TOTAL ENROLLMENT BY EDUCATION LEVELS 1960-1977

Educational Level

Primary Secondary Higher Total:

Source:

1960

1970

1977

48.1 46.4 29.1

48.8 47.2 34.7

49.0 49.2 42.1

478.6

48.2

48.6

UNESCO,1980:33.

The level of excess demand depends, of course, on the fee levels or costs associated to become a student. Figures from Table 29 suggest a quick answer to changes in the fee level. In 1976 a $5,fee for taking the university entrance examination test was established in Chile; later on has been raised to US$18. In 1976 there was an important reduction in the number of Chilean applicants and the demand has been drastically reduced since that year to levels lower than the level observed in 1974. The reduction could be related to the fee but the reduced number of vacancies in first year (places) may also explain the reduction in terms of awareness that students with poor academic background have no real chance to be accepted into the university. 'while the average applicants-entrants ratio for the region is 2.3, the ratios observed in specific regions, universities or schools, "Facultades" or countries may be quite different. The ratio of applicants to vacants in provincial or regional universities is usually much smaller than in the capital (Hawes & Donoso, 1983). Applicants to the Schools of Medicine or Engineering are sometimes ten times larger than the number of available places. However, Argentina moved back to oven admissions in 1983. The University of Buenos Aires increased its enrollments from 60,000 in 1983 to 240,000 in 1985 (report from Rector Delich in 3rasilia, October 1985). In a few countries it is possible to detect lack of excess demand in secondary education or in certain regions. In Guatemala primary education is still screening effectively the access to second and third levels. Secondary schools in Guatemala still have not enough demand to enroll as many students as places are available in the first year of secondary education. In Mexico there was a demand in rural areas that was met when the "Instructores comunitarios" were hired as teachers (Consejo Nacional de Fomento Educativo, s/f). In 1964 Chilean parents were asked through the mass media to enroll all their seven year old children in first grade (85% were already enrolled) and once the list of excess postulants was comDleted some 200 prefabricate classrooms were built "on demand". Now Colombia is planning to implement a similar exercise. On the other hand

-58-

Table 29: DEMAND FOR UNIVERSITY EDUCATION IN LDCs

Country

Year of Data

1981 1980-1981 1984 1980 1977 1973 1974 if 1975 1978 if 1977 1978 1979 1980 1981 Colombia 1988 1973 1982 1983 : Post Grado 1983 Ecuador 1985-88 1968-87 1987-68 1988-89 1989-70 1970-71 1971-72 1972-73 1973-74

Last High School Grade Enrollment

El Salvador Costa Rica Guatemala Paraguay Bolivia Chile

Average (Non-Weighted)

59,282 68,330 70,738

124,181

Total University Enrollment

Number Of First Year University Places Applicants

48,437 48,800 60,700 30,000 41,408 148,451 144,881 147,049 134,149 130,678 130,208 126,434 118,978 95,925 118,400 280,540 292,573 6,783 12,900 14,200 16,600 19,500 27,800 31,000 40,800 50,200 67,182

(19,885) 5,922 (Intake) 22,495 11,800 v 11,800 8,000 3 11,200 6,000 ' 20,400 47,214 2 118,193 42,S555 128,870 41,044 w 142,272 34,542 * 99,598 33,320 0 91,734 34,277 W 107,104 32,509 115,670 32,954 120,239 34,223 127,015 21,045 33,959 43,380 95,890 88,389 287,239 98,315 285,515 3,700 6,215 4,700 7,100 5,700 7,800 6,200 8,400 7,300 10,400 12,500 12,300 13,100 13,800 17,300 18,100 20,200 18,800 28,299 25,940

Ratio Applic./FY

n.a. 3.8 1.0 1.9 3.4 2.5 3.0 3.5 2.9 2.7 3.1 3.6 3.7 3.8 1.8 2.2 3.0 3.0 1.7 1.4 1.4 1.4 1.4 1.0 1.1 0.9 0.9 1.0 2.3

Note: Applicants graduated from high school in the previous year are shown in the column 'Last High School Grade Enrollment.'

the central provinces of Paraguay have reduced their total population now there is excess supply in those provinces.

and

There is an increasing demand for many alternative delivery systems of adult education (lifelong learning) reflected in the ever increasing number of non formal educational institutions (there are detailed descriptions of large numbers of those institutions for Colombia, Chile and the Dominican Republic). However, there is little effective demand for literacy courses (Castro, 1981) and the analysis of the census data shows that literacy campaigns have no effect at all in the literacy patterns of each age cohort. There are many applicants to special schools that are rejected each year because of lack of space to attend them. Special education is a very expensive type of education and it may be postponed for some time before the Latin American countries can afford it (till most of the normal children are enrolled).

Summary There is still excess social demand for education in all levels in almost all countries with the present fee level structure. In primary education such excess demand is mainly located in rural isolated areas. In secondary there is excess demand for academic education and eventually for university training. Those countries offering some primary education to all members of cohorts reaching the school age seem to be reducing their fertility rates and the reduction in fertility rates as a built-in control mechanism for constraining further social demand. On the other hand, unemployment seem to be a built-in mechanism for increasing social demand in secondary and higher education. (El Salvador, 1981.) Responsiveness

of Enrollment to Fee Increases

Although estimated price elasticities of demand for education suggest that increment in fees will generate small reductions in enrollment, relatively large changes in enrollments were observed in three cases of fee increments. The apparent conflict may be explained in terms of the different ways in which estimated price elasticities may be measured. Given that households and individuals have to make many different choices regarding education, demand for education can be measured in many different ways. For example, demand may be measured from "household spending in education", to "number of years of completed schooling" or "probability of having one child or all children in school".

Price and Income Elasticities

of Demand

Although only two reports on estimated demands for education have been identified both suggest that changes in fee levels will not affect enrollment. The price measures for both studies are indirect and "incorporate various components of private costs since it is difficult to obtain enough variation in observed fees for any given sample" (Jimenez, 1984). Table 30 presents the "price elasticity of-demand for education", which measures the percentage change in demand with respect of one percent change in the cost to user of the service.

Table 30.

Country

PRICE AND INCOME ELASTICITIESOF DEMAND FOR EDUCATION

Latest Year Dependent Variable of Data

Colombia

?

Income Measure

a. Amount of Spending Husband's Income on education

b. Share of HH Budget on Education c. Actual Expense/ Predicted/Expense d. HH Educ. Achievement Index El Salvador (UrbanPop.)

1980

Price Measure

Income

HH Characteristics 1.045 (No. of Children of Different Age Groups. 0.334

Inelas

1.035

Inelas

1.343

Inelas

Amount of Spending in Household Permanent Income (Providedby Education HH Characteristics Exp.)

a. Santa Ana b. Sonsonata

Source: Jimenez, 1984: 91 Data was obtained from householdsurveys. Note: HH = Household

Elasticities Source Price

Inelas

---

0.967 0.023

Birdsall 1980

Inelas Inelas

Jimenez 1983.

-

61

-

The price elasticity of demand summarizes the responsiveness of the propensity to utilize an educational facility with respect to a tuition fee. All price elasticities of demand in Table 30 are low. In general it may be assumed that the responsiveness tends to be less, the smaller is the share of educational expenditures that tuition payments make up and the smaller is the share of educational expenditures in total expenditures. Although, the analysis of the distribution of household expenditures in education show that enrollment fees (including "voluntary" contributions) are over 50% of total expenditures in education in most cases (annex 16) the proportion of total expenses devoted to education ranges between 1 and 7% (annex 17). Although higher price elasticities than those'reported in table 30 should be expected according to this high importance of fees in educational expenses, it prevails the small incidence of educational expenditures on total expenditures. In summary, the evidence suggests than an "a%" increase in tuition fees is likely to imply a decrease in utilization much smaller than "a%". The income elasticities, on the other hand, suggest that the educational consumption patterns are mainly determined by the income level. In other words more is expended in education when the income is higher. Data from table 30 and the analysis from household expenditures in nine countries support the last sentence. Table 30 shows that in three cases out of six an increase of "1%" in income will be associated with an increase in educational expenses larger than "1%". The household survey also suggest that the percent of household expenditures devoted to education increases with income and that the highest quarter has expenditures in education that are 10 to 30 times larger than the lower quarter (annex 17). The rapid increase of expenditures in education with higher income is related, among other factors, to providing textbooks and other educational material, sending children to private fee-paying schools, a greater propensity to buy other types of non-formal education and the increased probability that the children will reach more expensive (higher) level of education. (Musgrove, 1976.) All in all the information suggest that the education has a high priority in Latin America and that further increments in demand should be expected in the near future.

Effects of Changes On Fee Levels Although above commented data suggest excess demand and low price elasticity and high income elasticity of demand, three cases iniwhich data about reactions to fee changes is available suggest a higher price elasticity of demand. The raise in fees of Chilean universities that was enacted in 1981 led to a reduction of the proportion of students from low socio-economic levels. (Briones, 1981.) In the case of a country like Guatemala with no excess demand, at the university level, fee-paying students reduced their demand when fees were raised from $400 to $600 in 1980 and to $800 in 1984 (see Table 11). It is interesting to note that attrition during the school year was relatively constant in 1984 (25%) with respect to 1981 and 1982 (27 and 24%), thus suggesting that a rational decision was made before enrolling. The US$5 fee established in the 1976 Chilean University entrance examination test mentioned above (see C.1.) substantially reduced the number of applicants for the examination (see table 29).

-

62

-

The magnitude in official fees of public schools and universities is very low (see Table 13), and poor parents are usually dispensed of paying both the official fee and the "voluntary" contributions set up at the school lever. Thus, the analysis has been restricted in this section to the fee-paying public and private universities.

Effects of the Supply of Education Date from Colombia allows to detect a substantial response of private schools to a "freezing" of the fees in spite of sizeable inflation rates (annex 18). Since 1978 increments in fees had to be approved by Colombian government officials and increases in fees were effectively constrained. Although *a 10% increment in primary enrollments and a 25% increment in secondary enrollments were observed in the 1973-1977 period, a 10% decrease in private primary enrollments (from 621,360 to 559,533) was observed, in the 1977-1981 period, and a slight reduction in secondary in the 1977 - 1980 period was also observed followed by slow increments in the 1981-1983 period. Therefore, data show that the rapid increments observed in the 1973 - 1977 period were drastically changed after the intervention in the prices of private schooling. On the other hand a substantial increase in private school supply was observed in Chile when restrictions in fees were eliminated in the earlier eighties. These two cases suggest increments in supply may be elicited when controls on fees are eliminated. However the pressures of less wealthy groups that must switch their offspring to public schools due to fee increases have been substantial in both countries. Summary The evidence suggest some responsiveness of enrollments to fee increases through changes both in supply and-demand levels. The existence of flexible mechanisms to analyze the case of families with financial problems subsidize that allow mixtures of public and private support and options among alternative schools for the children to move from one type of fee to a different level may reduce the impact of taking the students off the system. Experiments

In Cost Recovery and By Price Rationing

The analysis of many "plans" for improving education does not reveal specific proposals of "how" to expand educational opportunities and to increase internal eff.iciency, beyond some too vague statements on budget allocation. There is little systematic analysis of the political options open to decision makers in education. In this section and in the next section price and non-price experiments to improve allocation of resources In the experiments with prices some proposals are commented respectively. have succeeded and some proposals have failed. For example, no program aimed to public university self financing has gotten very far with the exception of Mexico and Chile (see Table 13). In Mexico, Presidents DiazOrdaz and Echeverria tried to establish tuitions fees in the Universidad

- 63

-

Nacional Autonoma de Mexico (UNAM) in order to reduce its share of 40% of the total public expenditures in higher education, but students are still paying US$16 per year (3% of the unit cost). However, the new Universidad Autonoma Metropolitana (UAM) is a public university now charging US$400 per year. (Levy, 1979). The Chilean Experience. In Chile, a long public discussion was held in the 1968-1970 period on the need to raise tuition fees at the university. Carefully documented reports prepared by the educational planning office in 1968 led to drafting a proposal on cost recovery presented to the Congress, but political support was not strong enough and the proposal was not approved. The military regime that took power in 1973 tried to apply a neo-liberal economic model and decreed in December 1981 that secondary and higher education should be paid (now or later). Although the fees were not applied in the secondary level, the government succeeded in applying fees in the higher level and launched a loan system. Fees and loans were part of a more general reform aimed to introduce competency among universities. Each university would continue receiving 50% of the fiscal subsidy already received in 1980 and the other 50% would be pooled and distributed among all universities as a lump sum for each enrolled student of the best 20.000 student scores in the University Entrance Examination Test (UEET). The UEET were taken by some 100,000 applicants and some 30,000 are admitted as students (see Table 29). There was a four year transition period for implementing the new norms gradually adjusting from the old situation. (Castaneda, 1986). Given that no additional funds would be available all expansion of university activities should be funded from increments in tuitions. (Briones, 1981). New universities could be open (albeit should be approved by the Ministerio del Interior) in order to ease up a market equilibrium. The government hoped that competency among universities (and eventually with new universities) would raise the quality of the academic alternatives given that students would select only the best universities and would also raise the motivation of students to increase higher achievement levels. Although there was intense competition among universities, it was based on the expansion of those careers chosen by students with high scores in the UEET rather than by raising quality in every career. Funds were channelled to advertise in newspapers and magazines and to provide fringe benefits for the students included in the "best 20,000" group rather than In sunmmary, places in medicine into improving laboratories and libraries. and engineering were doubled (Universities without those careers started their creation), each university hired a marketing agency, scouts were sent to each of the best high schools and university "scholarships" were offered to the best students (Schiefelbein, 1982.). The Chilean government realized during the initial year of the change that the free market adjustment would be too expensive in terms of poor quality of some of the physicians and engineers and in terms of too many graduates competing for a small number of jobs. The government concluded at the end of 1981 that the highly expensive university training would end up with unemployment (for the least trained), with underemployment for most of the graduates in terms. of not working in their main training area and with migration for the balance to other countries.

- 64

-

Therefore, in 1982 the government fixed the number of entrants in each career of the subsidized universities and did not allowed creation of new universities (Donoso and Hawes, 1983.). The allocation of half the university subsidy through the "best 20,000" system was eliminated two years later. In 1983 a basic bonus on US$100 was established for students with scores of 572.5 or less; a bonus of US$300 was paid for students with scores of 573 and 622.5 points; US$600 for the group with 623 to 662.5; US$900 for those with 663 to 703.5 and US$1200 for scores over 704 points. However, this bonus system keep the total subsidy constant in Chilean pesos of 1983 in spite of an annual inflation of 20%, thus creating serious financial problems to the universities in 1984 that were solved through numerous transferences from the public budget worked out on a case by case basis. Although the subsidy and loan systems was originally designed to be applied to all operating universities (including those universities that would open in the future) in fact only was only applied to the eight public and private universities that were operating in 1981. The bonus system included the private universities existing in 1981, but not those created after that date (three new private universities were operating in 1986). The elite small best private university (UCCH) was benefited the most from this system given that in 1981 most of its students were among the "best 20,000" used as allocation criteria in the initial year of operation of the system and that the university has an overshare of students with high scores in the present modified system. Therefore the UCCH has substantially increased its public subsidy. The former National University is the university that has lost the highest share of the total amount of subsidies. As mentioned' above the reform established also in 1981 a loan system (credito fiscal). Each student able to prove that he or his family could not pay the tuition (some cases were studied by social workers) was allowed to receive a loan (the average loan was US$750 in 1984) to be paid with a 2 year grace period after graduation, but those dropping our should start payments the following year (this group has been unable to pay in 1984). Table 31 shows the distribution of students of a regional university receiving the total loan (one third of the student body) or fractions of the maximum loan. Students from the three new private universities created after 1981 have not been able to receive student loans.

- 65

Table 31:

CHILE -ALLOCATION PROCESS OF SUBSIDIES UNIVERSIDAD DE LA FRONTERA, 1984

Percentage of Tuition Fees Provided as Loan 100% 90 to 80 to 70 to 60 to 50 to 40 to 30 to 5 to

-

Number Students Obtaining the Benefit

98 88 78 65 55 45 35 25

Students provided with Loan (Subsidies) Rejected Applications for Loan (Subsidies) Students Not Demanding Loan (Subsidies) Total Number of Students

Percentage With Loans Student Body

1685 632 475 301 192 133 77 85 105

45.7 17.2 12.9 8.2 5.2 3.6 2.1 2.3 2.9

36.3 13.6 10.2 6.5 4.1 2.9 1.7 1.8 2.3

3,685

100.0

79.4

115

2.5

837

18.1

4,637

100.0

Source: Victor Nazar, Unpublisheddata, Temuco,1984.

Public universities are now getting 20 to 30% of their income from tuition (and around half of such tuition comes from the "Credito Fiscal") and another 1 to 20% is obtained from selling services or from odd incomes. Given that "teaching expenses" may be estimated in 60% of total expenses, present tuition levels represent, in fact, between one third and one half of the teaching costs. Other Experiences

in Cost Recovery

A study for Paraguay showed that in 1984 some G$125 millions (10%) may be added each year to the public university income by demanding new university students to pay fees of the same value as those paid in high school (World Bank, 1985). Assuming that careers are five years long, there would be a constant additional income flow of G$725 million (some 15% of total university budget) starting the fifth year of the operation of the tuition systems (Annex 19). In 1985 Council of the National University studied increments in the fee levels and although for the first time in more than three decades university students demonstrated in the streets of Asuncion the attempt allowed an increment in fees. In 1984 the University Council of the Universidad Nacional Autonoma de San Carlos in Guatemala proposed an increment in fees from US$5

- 66

-

to US$10 per year. The proposal was sent to a Committee for "further" study after armed students stormed the office of the President of the University and promised to kill all Council members if the proposal was to be approved. The history of university student demonstrations in Latin America against increases in fees is long and frequent. The number of death resulting from much violence is high. Decision makers must take into account both the economic, political and moral elements involved in the decision on fee increments. Cost Recovery Through Teachers Salaries Although it is not quite clear whether governments reimburse teachers services at rates below the market levels, teachers' salaries in Latin America are much lower than in other regions (see Table 1). A substantial amount of the adjustment generated by the recent economic crisis has been solved in terms of lower teachers' salaries rather than in terms of reduced enrollments (Annex 11). Poor motivation of teachers seems to be a key factor on poor student achievement that to low teachers' salaries with respect to other professionals and to decreasing salaries both in constant terms and in (Figures for Paraguay, Peru and Chile are presented percentage of the GNP. in Annexes 20, 21 and 22.) It is interesting to note that in Peru differences among primary and secondary teachers were reduced and as a result the top salaries were slightly increased in primary and vocational education while in Chile the ratio between the maximum and minimum salaries was reduced from 2.4 in 1970 to 1.7 in 1979. As a result of these processes, both in Peru and Chile salaries of high school teachers were reduced to near the level of the primary teachers. The low level of teachers' salaries may also be detected in terms of the high proportion of untrained teachers. One third of teachers were untrained in the early seventies and around one, fourth are still without training nowadays (Annex 3) in spite of the high unemployment rates. Rates of return for teachers were negative in Chile in the mid seventies (Morales et al, 1977) and although substantial differences were observed in Colombia for those with a teaching degree (between 50 and 100%), the low absolute levels of salaries did not attract enough candidates (with a teacher degree) for the rural areas.

The effect of the regionalization on salaries will be commented in the next section (C.4). Table 32 shows differences in salaries are for not having a teaching diploma in the case of Colombia. As in many countries, increments in salaries related to years of experience in spite of the poor relation with students' achievement seen in A.2) may compensate lack of training; an untrained Colombian teacher at the end of his/her career earns This increment the same amount as a well trained teacher (2nd Category). in salary in spite of not having a title (as well as the reluctance of teachers trained in urban zones to move into isolated areas) and especially the low salary level may explain the slow improvement in the percentage of trained teachers observed in the last 15 years.

-

Table 32:

67

-

COLOMBIA - DIFFERENCES IN SALARIES OF TEACHERS WITH AND WITHOUT DIPLOMAS IN EDUCATION (in pesos)

Primary Teachers

No Training (No Diploma) 4th Category 3rd Category 2nd Category 1st Category

Minimum Salary Per Year

Maximum Salary Per Year

12,250 15,850 17,840 19,450 22,640

19,490 24,850 25,500 26,160 28,130

Source: Claude Tibi, uFinancementet Developpmentde i'Education: Le Cas de la Colombie: Rapport de Recherche, No. 22, IIEP, Paris, 1978

Price Rationing At the preschool level the rationing is mainly determined in terms of price. Children from well off families usually have access to fee paid kindergartens. The small number of public non-fee kindergartens mainly cater to children from families living in some isolated modern sector% factories or in public primary schools. Entrance to the new public kindergartens attached to primary schools is usually obtained through different forms of kinship. Therefore, public kindergartens tend to cater also to children from families from socio-economic level above average. Very few non-fee kindergarten are operating in the rural area. Lack of enough places in secondary public education has forced many families to make an extra effort to enroll their children in feepaying private schools. This is especially evident in the large proportion of private education in countries like Colombia (49.8%) and Mexico (32.8) where only a fraction of the graduates from primary public education could find places in public secondary education. Table 33 shows the differences of primary and secondary private education for those countries as well as the relative growth rates of the public and private education (see also Table 16). In secondary education the students with the best achievement levels (or with better kinship connections ) that have access to the free public secondary are usually from higher socio-economic levels than those that are denied access. Therefore, those that have to pay (in the cheaper private schools) for getting secondary education belong to families that are in worse economic conditions. Summary Although the number of successful experiments in cost recovery is small, they suggest that more attempts can succeed in the future if the public opinion is well informed with relevant data from the country and Cost recovery policies should be with suitable regional ccmparisons. developed in a medium-term range, unless extremely favorable public support

- 68

-

is obtained beforehand. Initial allocation of cost recovery funds to scholarships for needy students may ease up the acceptance of cost recovery policies. Further cost recovery through teachers' salaries should be avoided and eventually salaries should be raised. Although reduction of teachers' salaries has been one of the long term strategies used by government to expand education with the available amount of resources allocated to education, such a reduction is preventing the attraction of good candidates into the teaching force, damaging teacher motivation and eventually reducing students' time on task. C. 4

Non-Price Rationing Systems

Although during the sixties and seventies almost all Latin American countries increased the amount of educational resources and tried to improve the allocation of those resources, there are important questions about the rationality of such attempts and about the outcomes. The Impact of Public Education in the 60s and 70s Although there is an obvious agreement on the explosive increment in enrollments at all educational levels in L.A. on the last 30 years, there are many questions on the final outcomes both in terms of efficiency and equity. Impressive figures on the increments have been commented in the introduction to the initial Chapter (see Table 1 and Annexes 11, 12 and 15) and that type of evidence can be easily expanded. Comments on the poor internal efficiency were discussed next in the rest of the initial chapter (see Sections A.1 and A.3). Now some of *the improvements-that happened in the selectivity of the system and remaining differences will be outlined. Differences in selectivity have been reduced dramatically in the last 40 years. Although Table 34 presents the changes observed in Chile, the figures are representative of the changes in most Latin American countries. Still, in 1970-1974, the higher the educational level the higher the selectivity (the ratios between extreme selectivity indexes are 5.5 versus 9.8 when comparing 8th and l2th grade), but the ratios have been drastically reduced from the 1929 levels (30.1 and 545.8, respectively). The reduction in the selectivity indexes reflects the fact that most Latin American countries have now enough school capacity to enroll all newcomers reaching the school entrance age. However, countries like Guatemala and Haiti, and in less extent some other countries, where governments played a lesser role in the last decades, are not providing enough student places in the rural areas. However, even in those cases the positive impact of the government policies can be traced. For example, there is an impressive increment in student enrollments in Guatemala in the last 15 years. It is important to remark that the mere operation of the price allocation system in the countries where the government did not provide more education was not able to generate similar results to the public rationing system. However, the key question is which rationing system should be used now to improve the education of marginal groups in tetms of reducing early attrition and increasing application of knowledge to their "praxis" (Carvalho, 1982), and how to develop a system able to follow the right signals.

Table 33: LATIN AMERICA - LEVELS AND TRENDS OF PRIVATE EDUCATION 1970-1974

Countries

Primary Level Percentageof Rate of Growth of Private Education Private Education In Total Enrollment Divided by Rate of (1974) Public Education

Secondary Level Percentageof Rate of Growth of Private Education Private Education In Total Enrollment Divided by Rate of (1974) Public Education

Higher Education Percentageof Rate of Growth of Private Education Private Education In Total Enrollment In Total Enrollment (1974) (1974)

Bolivia

13.6

1.88

25.9

0.54

Colombia

14.6

1.07

49.8

0.73

51.6

1.73

7.8

0.98

32.8 b/

1.28

12.5

0.98

leru

12.9

0.77

17.3

0.46

27.1

2.03

Venezuela

11.4

0.48

18.3

0.81

9.6

0.68

MAexico

-

a/ The ratio between the private rate of growth and the public rate of growth has been computed for the following period: Bolivia (1970-74); Colombia (primary 1984-1974;secondary 1986-1974 and higher 1970-1974);Mexico (1970-74),Peru (1964-1974),and Venezuela (1969-1973). b/ Junior High School. Source: Carlos Munoz I and Alberto Hernandez M. Financiamientode Ia educaci6n privada en Am6rica Latina, 1978.

- 70

Table 34:

Occupational

Groups

Managers and Professionals Clerks, Salesmen and Small Business Craftsmen, Skilled Workers and Drivers Peasants, Unskilled and Service Workers Difference Between Extreme of Range Ratio Between Extremes of Range

-

EVOLUTION OF SELECTIVITY INDEXES IN THE 1929-1974 PERIOD

Selectivity Indexes a/ 1970-1974 1929 7th Grade 12th Grade 8th Grade 1979 12th Grade 1974

8.4

27.3

3.3

3.9

9.6

14.9

1.7

2.5

3.6

3.0

1.1

1.0

0.3

0.1

0.6

0.4

8.1

27.2

2.5

3.5

30.1

545.8

5.5

9.8

Source: Schiefelbein and Farell,'Desigualdad escolar: Analisisy Polfticas,*Mensaje,No. 277, March-April 1979, pp. 148-151. 3/

The indexfor each occupational groupw; is computedas the ratiobetweenthe *percentage of males in groups 'j'with respectto the total activemale population" and the 'percentage of parentsof studentsworkingin groups' with respectto the totalnumber of students.'

For Several important changes occurred in the last decades. example, secondary and higher education substantially expanded and primary education reduced its share of public current expenditures (Annex 23 presents data for the 1965-1975 period). Also, not only more resources were allocated to secondary education to expand enrollments for the still small group benefiting from this education but also the unit cost in secondary education is almost three times the unit cost of primary education (see Annex 24). Finally, university also improved its share, but the units costs have been kept under control (only 12 times the unit cost of primary education), and fragments evidence show a continuous effort of government for avoiding too large public contributions to the small university elite. Discriminatory

Allocation of Resources

Although urban areas are clearly favored by public educational expense and metropolitan areas are still more benefited, the government is providing almost all rural education with only a few exceptions. Table 35 illustrates the heavy concentration of higher education in the two main cities of Ecuador. Both the concentration of population and the visibility of the education urban problems in a mass media mainly aimed to urban consumption, generate pressures for coping with educational problems in

- 71 -

that area rather than in the isolated rural area (Uzategui and Muelle, 1974). In addition, scattered rural population tends to raise unit costs and the bonuses for teaching in isolated places or under more difficult living conditions also tends to raise unit costs. Several factors explain the high proportion of public schooling in rural areas. The main factor is lack of effective demand. Research suggest that interest in education for the offspring is only expressed in demand for enrollments when the parents have at least some five years of education. Given that rural areas have traditionally lower levels of education (and the average schooling is much lower than five years) there are less pressures for demanding education in rural areas, but in those areas where there are strong community organizations. As a result of the low demand, the lack of pressure on government officials, the higher unit cost, and the low income levels (and the fact that schools in isolated areas have a natural monopoly situation) most of the supply of education in rural areas is provided by the Government. All in all, the government has corrected the lack of educational services generated by the inoperation of the market system in the rural areas. Table 35.

Regions

ECUADOR - REGIONAL DISTRIBUTION OF HIGHER EDUCATION ENROLLMENTS, 1976

Population (000)

University Enrollments (000)

Schooling Rates (per 000)

Distribution of Enrollments (%)

Capital

1,828.9

78.8

43.1

42

Second City

1,736.6

66.3

38.2

35

Rest of the Country Total:

3,740.5 7,306.0

45.4 190.5

12.1 26.1

23 100

Source:

Annex 25.

On the other hand, Governments' non-price rationing system have led to misallocation of resources in certain activities. For example, it has been questioned the efficiency of MOBRAL (the national institution for literacy in Brazil) and it has been suggested that its resources should be devoted to the regular primary education (Castro and Franco, 1981). In general, literacy campaigns frequently sponsored by new governments after taking power, have not been successful in most Latin American countries. In summary, the distribution of school resources by socio-economic characteristic is in part a consequence of the availability of free (or subsidized) public educational services and in part a consequence of the ability of well-off parents to keep their children at school during a longer period and to provide them with special help for their learning and health problems. The higher the level of the education system the more selected the socio-economic level of the students (see Table 36). In certain countries (for example, Brazil or Mexico) it was argued that the

-

72

-

provision of education is becoming more and more unequal (Frigoto, s/f; Carvalho, 1982, Barkim, 1972). The question, however, is whether the increasing inequality is a result of too much government intervention or the result of a lack of intervention. Unfortunately no definitive research is yet available to elucidate this question. Public Systems Allocation Processes Although some scholars assert that changes in the educational system are the result of decisions made by the groups in power, there are also cases where changes were not generated by deliberate decisions, but were the resulted of slow change processes where no specifi'- "decisions" could be traced. There are many references on the so called decisions made by the groups in power to meet their goals (Carnory, 1975; Cox, 1984). On a similar vein, other reports suggest that some social groups have been able to influence the Government in order to obtain special privileges (DEALC, 1981); several studies have identified the "medium" socio-economic levels as the group benefiting from public subsidies, and other reports criticize the concentration of resources (public and private) towards the richest groups (Frigoto, s/f.). On the other hand, the lack of clear cut decision processes has also been examined by several authors. In Mexico the decision process seems to develop inside the government, but independent from the central government directives (McGinn et al, 1980). In Brazil it is possible to identify specific decision processes, but the context in which those decisions are made evolves in a non-rational way where no deliberate decision could be traced (Velloso, 1980). Some studies may lead to specific action, (as in the case of graduate training after the study prepared by T. della Senta), but other studies, for example the Brazilian studies financed by the Ministry of Education) have not clear impact on decision making (Castro, 1980). In other cases, it is assumed that it is Possible to make decisions and, therefore, it is only necessary to create mechanisms (for example participation channels) for making the required decisions (Franco, 1980). What is clear is that the educational diagnostics prepared during the sixties showed that certain areas (usually the poorest areas in each country) were lacking educational infrastructure. Public opinion, political parties from center and left, and international institutions supported the need to reduce the differences. Governments responded to the pressures by building the infrastructure while private provision lagged, given that the families of the new students were willing to get education for their offspring but were not able to pay for the educational services (see for example Venezuela, 1980). Although in most cases, only educational "plans" with global targets were prepared, more information on inputs, quality, imbalances and management was made available, allowing problems to be better defined and priorities (formally) established. In sulmmary, the increasing state activity in the education sector was the result of combined pressures exerted by beneficiaries, bureaucrats and politicians. As commented above, the allocation of public funds to higher education has been affected by authorities and congressmen more loyal to the universities where they have studied than to their ideologies) (see B.1). There was a tendency to provide an overshare of resources to universities, thus benefiting the higher income groups to which the

- 73 university students belong, but in recent years more control on the allocation by levels has been exerted by wider groups of each country. All in all, government-providededucation was welcomed for apparently equalizing access to education, and for taking into account benefits generated by education that cannot enter into a private firm's calculations because they cannot be appropriated by each firm (external economies). However, no serious public discussion about financing of and provision for education was carried out in Latin America during the sixties and seventies. Only one regional seminar on the financing of education was sponsored by UNESCO and OAS in the sixties (UNESCO, 1966; 1967; Rojas, 1966; Ordoinez,1966; and OEA, 1970). The state-university relations are influenced by many independent factors such as population explosion, economic situation, unemployment, political arena and academic or social tradition (Levy, 1979). It is in such complex web where de facto governments act. In the case of Chile, it has been mentioned above that the attempt to make use of market forces (forgetting the long term effects of educational decisions that may lead to cyclical changes well described in the so called "spider web" theorem) led to a reduction of students from poor families and then to stopping the operation of most of the market mechanisms (see C.3). Bureaucracies, congressmen and military regimes have adopted and dismissed many models. The interest in planning, was then replaced by PPBS (Planning, Programming, Budgeting system), projects, zero budgeting and market mechanisms. Still, there are some pervasive ideas that have gained public support during the last decade. The most important of those ideas is the need for "decentralization." Although little effective decentralization has been implemented, there are some interesting experiments. Large countries like Colombia, Argentina, Brazil or Mexico where regions or states have some independence in their decisions, the geographical distribution is affected by the power structure defining the amounts transferred to each state or region. In centralized countries some geopolitical national strategies sponsored by the Army Forces may give some advantages to schools located in future poles of development or the frontier (for example Guyana in Venezuela as described by Davis et al, 1970 or the city near the Itaipu Dam in Paraguay). Centralized governments may decentralize decisions, for example taxes to certain export products also have been earmarked to certain regions, or even to private institutions as the case of FEDECAFE in Colombia. There may be some inconsistenciesin all these decentralized allocation processes-. For example, FEDECAFE is willing to build and equip new schools and is very efficient in doing so, but it was reluctant for some time to complement the salaries of the teachers in order to have excellent schools with students effectively working with good educational materials. The frequent "revolutionary"processes observed in Latin America are also a potential power for altering public resource allocation. The temporary reduction of university activities in Cuba or the emphasis in literacy both in Cuba and Nicaragua are examples of such changes (Arrien, 1980). The military regimes in Argentina, Brazil, Chile and Uruguay, identified in the late seventies the universities as centers of opposition to their power and those regimes restricted expenditures in the third level and established control mechanisms of their operations (Levy, 1982; Donoso and Hawes, 1983; DEALC, 1981). On the other hand the explosive expansion of the Universidad de Buenos Aires reported above is also an example of perturbatory factors generated by returning to a democratic regime (see C.1 above).

-74-

Table 38: CHILE. SURVIVAL RATES IN PRIMARY AND SECONDARY EDUCATION. 1962-1976 FATNtrs OCCUPATION

puas

- 0-

77wn M2

-

31i

!4 -

.

X

'a~

~

~

~~a

a

I

32(im

ate a3 ct *h onguau pno

I osn.

Source: Schi~efelbein

rtuan

mupwwmIbea

az pne

and Farrell, June 1978: 338

Sma

12 i-d. are gnTa 12 fl a

-

/ j

-

Lack of Relevant Signals for Improving Allocation Although teachers unions, political parties, associations of private schools, officials, statistical services, scholars, professional groups, church, mass media and community groups convey messages about detected problems and proposed solutions, there are no sure ways to provide solutions. There are no systems providing the right signals for improving allocation of resources. All those pressure groups (with the exception of scholars) do not have the time for careful assessment of the distance of the proposed or endorsed technologies with respect to the optimal technological frontier (for the accepted educational goals to be reached by the society). None of the above mentioned groups will probably detect the difference between the number of students enrolled in 1980 in several Paraguayan programs and the potential enrollment described in Table 37. This inability to detect opportunities for improvement suggests the lack of relevant signals for allocating more resources to valuable activities. However, given the low socioeconomic levels of those potential customers only a voucher system would allow those potential costumers to express their demand in real terms. Management Information Systems continuously exploring unit costs, distributions of inputs, achievement indicators and timely access of students to the system can complement the present price and non-price allocation system and help the, authorities to timely identify prospective problems. The information is already available in the ministries and processing is becoming more and more cheap. Eventually, massive or selective testing may help to identify achievement problems and to help in solving the exceptional cases. Summary Although public education extended educational opportunities to the rural areas and to the marginal urban groups, thus reducing previous inequalities, there are also cases where discriminatory effects of public non-price allocation processes are observed. There is very little competition from private education in rural or poor areas. With only a few exceptions most of the changes generated by public allocation processes in Latin America are the result of slow processes that gradually alter the "status quo" rather than the result of quick changes triggered by a single decision. As a result of those slow evolutionary processes secondary education has expanded, higher education has been forced to expand through better use of their resources (and partially becoming a refined secondary education, rather than a research oriented university type of education) and now there is a new interest in preschool education and in improving the quality of the primary education (especially in terms of reducing repetition and improving some basic skills in reading and writing). C.5

Resource Allocation within the Education Sector

Although the price and non-price rationing policies determine who is accepted and the amount of resources that may be used in his education process, there are still many options that should be faced both at central, regional, and school levels and even at the classroom level. In previous sections the selectivity and discrimination effects of applying price and non-price rationing policies have been commented. Now the role of law,

Table 37: PARAGUAY - SPECIAL DELIVERY SYSTEMS IN PRIMARY EDUCATION

Name of Program

1. Plurigrados (Multigrade Teaching)

2.

Accelerated Primary

3. Grados Libres (FlexibleExams)

Starting Year

Number of

Number of

Students in 1980

Teachers in 1980

1987

52,000

1978

15,726

68

1948

1,268

n.a.

4.

Special Education

n.a.

6.

Bilingual Education

1978

2,160

160

809

2,780

86

Potential Enrollment

100,000

10,000

2,000

14,000

200,000

Special Features

The same teacher provides education to children of different grades in one classroom - a sound strategy for small rural communities. Average student/teacherratios are 26 to 1. These, ratios could be increased. The program lacks adequate instructionalmaterials and teaching learningmechanisms. If strengthenedcould reach about 100,000 children. Overage studentsare allowed to do two school grades on one calendar year. Admission tests are administeredto select children for the program. OAS provided support in 1978-1979. Children who have been absent from school are allowed to take a test and if they pass it they are promoted to the following grade. A sound strategyfor children who must assist their families at work, who are sick and who find a tutor at home. Education for the Blind and Deaf. To be expanded in a more advanced level of developmentof the country. Teachers are encouragedto use Guarani In the classroom. Special workbooks In Spanish were provided to the Guarani speaking children. An evaluationof the program is being conducted. Further work is needed to determine whether children should read first in Guarani or in Spanish as they are now doing.

Source: IBRD, Projects Department,LAC, "Paraguay:Basic Education SubsectorMemorandum,'Washington,September 20, 1984.

3

V

t

0'

budget, institutions, traditions and chance or the shaping of the school system will be commented upon and examples of their effects and constraints will be discussed. Laws, Decrees, Norms and Instructions Even though there is a big difference between enacting laws and solving specific problems, Latin American governments have heavily relied on laws and instructions as tools for shaping their educational system and the related processes. The Hispanic legal traditions and their lack of enforcement in the Colonies (la ley se acepta, pero no se cumple) have still a mark in the limited power of law. For example, most countries have defined in laws or constitutions approved half a century ago, the number of years of obligatory primary education, but no attempt to enforce those commitments have yet been done. The law or the instructions seem to advertise the direction in which the Parliament or the dictator expect the education system to move and eventually to generate some advance toward the expected goal. For example, several countries (Costa Rica, Chile, Ecuador, Venezuela and Colombia) have enacted continuous or automatic promotion systems for the first grade and in some cases for the first four grades in primary education now the teachers know that their duty is to help normal students to reach minimum achievement goals. Even though in some of these countries the statistical forms show no repetition after the law has been enacted, other data show that still there are repeaters, but repetition is being finally reduced. In Ecuador the Ministry has instructed that the same teacher should work with a group of students during their first two years and the group is promoted automatically from first to second grade. In .Venezuela all students are supposed to be promoted from first, second and third grades, but in fact many finally repeat as it also happens in Ecuador in first grade. In Chile students attending classes more than 75% of the school days obtaining passing grades in Mathematics and Spanish are promoted in first and second grades, but those with bad grades must attend classes in an additional remedial period. In Honduras teachers force students to re-enroll as newcomers, but there is an awareness about the need to reduce high repetition levels (Spector, 1984; Cuadra, 1986). In Argentina "sunmmer schools" have been able to recover 81% of first grade repeaters and similar figures have been recovered for students in all grades in primary attending remedial schools in Chile that support the students' self esteem (Vaccaro, 1978; Castillo, 1975). Although, some institutions have been forced by law to implement certain actions, the institutions frequently defend themselves from unwanted actions. For example, free access to public higher education to all graduates from high school is guaranteed by law in countries like Uruguay, Ecuador, Dominican Republic and Guatemala. However, in those countries some schools have managed to be allowed to provide only a fixed number of vacancies. For example in Uruguay there are a fixed number of vacancies in law, medicine and engineering, while other schools must accept all applicants. In the National University of Ecuador there is a required "preuniversitario" lasting three months that is used as a screening device. In the Tecnologico of Quito the applicants much spend a whole year in a "preuniversitario" and only those promoted are accepted for joining the institution as regular students. (Pozo et al, 1981). Therefore, several strategies are implemented by the universities to reduce the large number

-

78

-

of applicants into those that can perform at the required levels. succeed in one goal and fail in other competing goal.

Laws may

For example, the private Catholic University of Ecuador may accept only those students that pass an entrance examination. Thus their students can save one quarter or one whole year with respect to those in public institutions that must pass the "preuniversitario." Therefore, although the law pretended further equality of opportunities, and succeeded in providing a place in the university for all candidates, it also created other types of discriminations in terms of alternative lengths of the careers. Countries that have fixed by law the number of vacancies (numerous clauses) may also find that discrimination persists. In those countries well off families have been sending their children to other countries where universities have free access, when the children fail to be accepted in their own countries. No statistics of students by country or origin have been obtained to quantify the magnitude of the migration flow, but it is estimated to be over 2,000 students in the region. Budget Allocations Although changes in budget have each year a direct effect in resource allocation, the final expected effects may be different from reality. For example, appointment of the principal with an extended day schedule (or an assistant to the principal) plus the required teachers may be enough for using buildings in double shifts. However, if a second principal is appointed in the, same building frictions will be generated and students achievement may be affected. The general practice in most countries is to have only one principal in charge of both shifts in order to avoid conflicts and research in Chile and Venezuela has shown that Budget double shifts have no detrimental effects on student's achievement. allocations are also an efficient mechanism for launching nutrition programs, early stimulation, and free transportation that seem to have possible effects on achievement, but allocation of the benefits may not reach the really needy (Schiefelbein, 1978). Budget allocations may be enough to launch a preschool program and it has already been mentioned the promising results detected in experiments with preschool as a prevention training for repetition in first grade (A.3). Given that half of the students are repeating first grade, the introduction of a kindergaten (as a readiness work period) does not represent additional costs (students are repeating first grade anyway) and reduces the trauma of repeating first grade. Therefore, through expansion of kindergarten, budget may change the school from a process of repeated failure into being the starting point of a history of successes. Preschool has a positive impact in reducing first grade repetition (Filp et al, 1982). Even though there is still no clear picture of which are the most efficient programs (Llanos and Winkler, 1982), there are materials available (Lira, 1980), and experiences to share. There is also evidence of the impact of early education in the educational achievement levels of students receiving many years after, (Schiefelbein and Farrell, 1982). Although, participation of mothers has proved to reduce an important part of the financial cost of such programs, increments in attendance have been relatively slow (Latorre, 1980). Another example of activities that can be supported by simply allocation of budget is Distance Education. Distance education has been widely experimented in Latin America at all levels. The knowledge is available and the expansion of Distance Education is a matter of hiring the

right people. Unfortunately there are only descriptions of such programs,. but no evaluations with the exception of the radio mathematics project in Nicaragua (Jamison, 1978). This project represented an "additional cost" of US$3 per student (5 to 10% of the cost per student) but was able to reduce the "total cost" of graduating students from primary education. On the other hand, the evaluation of the TV project in El Salvador suggested that there were no direct effects of TV on achievement, but indirect advantages generated through a better management (Speagle 1972; McAnany, 1970). The use of pre-school programs like"Plaza Sesamo" and the like have proven to produce positive impacts in achievement (Schiefelbein, 1978) and their diffusion only depends on buying broadcasting time and some complementary advertising of the advantages for children to watch the programs. Sometimes budget allocations are made without enough knowledge about the impact of budget items in education, especially when making budget cuts in financial crisis. The result of those random cuts have been the elimination of funds for maintenance, educational materials, power, services and supervision that have been commented above (see B.1 and Table 15) Although parents (through monetary or in kind contributions) have provided the required resources for minimal operation (Paulsen, 1981) funds for district supervision require central allocation. Although several attempts of relating earmarked taxes to education have not finally showed a large impact, the Colombian participation experiment with cafe taxes administered by the "Federacion de Cafeteros" have successfully financed many schools. There is still the contention that the central government has allocated less resources for school construction in those areas benefited from the cafe taxes than in the rest of the country. Anyhow, the zones benefiting from such tax seem to be substantially better shape than the rest of the country. The problem with earmarking taxes is that they constrain the use of public resources. The difficulties for evaluating results in this example can be replicated in many of the examples commented above. Institution Building There are cases where the desired educational activities will take place only if (in addition to legal and budget inputs) management inputs such as terms of reference, PERT, technical assistance or guiding principles that allows to launch a new institution are provided. For example, Testing Institutions may be created by law and provided with resources (money and technical assistance) to screen students' access to higher education through tests (with minimum entrance scores) or to generate information to be used as an indicative device for decision makers. Countries like Brazil, Colombia, Chile and Costa Rica have designed University Entrance Examination systems. Now in those countries the students getting the best scores have the opportunity to chose the university and the career they want. The screening process, however, involves many other steps. For example, 4500 students ask permission to take the exam in the Instituto Tecnologico de Costa Rica; 4000 take the exam; 1500 are approved, but only 750 finally sign for enrolling at the institute, the rest have also been accepted in other universitites and have signed with them. At the end of the process, however, the final goal, "objective screening of the best candidates," is far from being reached. In general there are correlations of only .5 to .6 between scores in the Entrance Examinations and further performance in the higher education courses.

- 80

-

Some discrimination is generated by the use of these Entrance Exams because these exams give some advantages to students from higher socioeconomic backgrounds. In countries where fee paid private higher education is relatively large (for example Brazil and Colombia) students from higher socioeconomic backgrounds have access to good tuition free public universities, while the students from low socioeconomic backgrounds can only be accepted in fee-paying private universities (of lower quality in the case of Brazil) (Castro, 1981). In addition there are still some special admission systems for good athletes (or good kinships) to be accepted in the university (Zunino and Rodriguez, 1981). However, their number is small and do not influence the main trends above commented. Another example of a testing program requiring the creation of an institution is the National Evaluation Service created while in Chile in 1968. In 1969 and 1970 the Evaluation Service administered in Chile a national test to all students graduating from the 8th grade in primary. The real purpose of the test was to show the teachers which were the real objectives of the primary education, i.e. application of knowledge or inductive reasoning instead of mere rote learning. Therefore, the test had only 10% weight and the course marks had a 90% weight for computing the score that was-used to select candidates in secondary schools when there was an excess demand. There is no empirical data to assess its impact but there are many reports stating that the teachers became aware of what was expected from them to achieve with their students. Although experiments with "nuclearizat-ion" would also require an "institutional" allocation process, it may be good to restrain from further experiments because of past mixed up results. To those countries trying to launch new nuclear experiences it should be good to remind them that in spite of the efforts carried out in Peru to establish the nuclear system in the mid seventies, the government is now dismantling the nuclear structure after a detailed report on their operation (Peru-1982; Diaz, 1980). That nuclear structure had become only an additional administrative step, but little supervision was actually provided due to lack of funds for transportation and per diem. In the case of Colombia the best teachers were appointed as directors of nucleos. Their role was not clear and they stay in their locations without supervising nor teaching as they had been doing before. Conflicts arose between those "directores de nucleos" and the "supervisors". In summary, the building up of a nuclear structure seems to be a more complex task than originally envisaged. Although promising, the nuclear system is still in the process of developing viable ways to successfully implement it. The prestige of European vocational training methods has led distinguished educators to experiment with "apprenticeship" as one of the main alternatives to prepare future skilled workers. CINTERFOR has supported the idea and several of the Training Institutions have offered apprenticeship programs (see Table 38). The evaluation of such programs has been positive but the magnitude of those programs is still relatively small (Corvalan, 1981). Many other examples that would require institutional development to be implemented may be identified. For example, textbooks design and production supply of primary education through multigrade teaching, and

-

61

-

provision of free transportation. In all these examples there are some past or on going experiments that are promising, but have not been enough evaluated to draw solid conclusions (Schiefelbein, 1978). There have been many efforts for designing cheap educational materials. There are good designs and some diffusion has been attempted, (Jairo, 1981), but there is still little knowledge of what is available for wider use in the whole region. The analysis of production function studies also provide some further suggestions that should be experimented in the future, (Schiefelbein, 1981) such as: reduction of teachers' rotation, use of homework (may be experimental with an alternate attendance system), stimulation of self-esteem and use of commercial TV spots in the educational process (see Table 7). Institutional building has sometimes generated improvements achieved by chance. For example, schools-are now averaging larger sizes (around 200 students per school) that are associated with higher achievement levels (see Annex 28). Once detected these positive effects from larger-sized schools are detected, the size of the schools may be further expanded in a systematic way. Leadership of Participation Processes Last but not least the subjective dimension must be brought into consideration in this case by mentioning two factors not easily amenable to harnessing: leadership and participation processes. Change of behavior and tradition goes beyond mechanic allocation processes, but in many cases behavior and tradition must change if instruction processes are to be improved. Changes in teacher attitudes or.community participation cannot be imposed from outside, but through central or local leadership and feasible challenges. While leadership is difficult to be replicated (as in the non formal education), participation may be fostered. Although participation has lagged, the municipalization experiments seem to be positive. In future experiments more implementation problems should be Table 38: TRENDS IN APPRENTICESHIP OF YOUNGSTERS - 1973-1979 (Percentage in the total amount of activities carried out in each institution)

Country

Brazil Colombia Costa Rica Chile Peru Venezuela

Source:

Institution

SENAI SENAC SENA INA INACAP SENATI INCE

Claudio de Moura Castro, 1979.

Percentage Training Course Hours 1973 1979 1973 1979

15.0 8.3 8.6 10.6 1.3 4.7 1.9

15.0 6.0 7.1 5.0 3.2

20.0 46.3 55.6 14.5 67.0 14.7

20.0 30.1 20.0 15.4

-

82

-

taken into account in order to trigger real community involvement, avoid teachers' anger, reduce both differences in quality of education, and poor schooling opportunities. However, it must be noted that other countries (i.e. Ecuador) are eliminating municipal schools. Municipalization and nuclearization experiments are the two main attempts for implementing "decentralization". There is a widespread belief that educational decentralization should be implemented in the region. The attempt of Colombia in 1970, when the "Fondos Educativos Regionales (FER)" were created, has been relatively successful in the more advanced states, (departments), but little advances are observed in the rest of the country. The "regionalization", among other things, may explain that Colombia remained in the seventies as one of the countries with lowest educational schooling rates and higher illiteracy. There are wide disparities among the unit costs between departments. Table 39 also shows the differences in the minimum and maximum teachers' salary levels by states (regions). The regional differences were quite impressive in the early seventies although now they are much reduced (see annex 39). Mexico has incorporated the local "patronatos" for paying local teachers working in small villages (mainly bilingual education). Costa Rica has gone midway with regional groups able to hire teachers and to reallocate resources (Rojas, 1982). Brazil and Chile have launched the "municipalization", but it seems to be yet too premature for a careful evaluation. Brazil had to create PROMUNICIPIO to help the countries to develop their educational administration (Sobrihno, 1978). Chile eliminated municipal schools in the 19 century and started once again to transform public into municipal schools since 1981 and completed the process in 1986. To reallocate the schools to the municipalities the government had to pay a high subsidy per student in the beginning to get the counties interested in running the school systems (Valdivieso, 1981). Later on subsidies per student were frozen and now the counties are in financial problems. (CastaReda, 1986). Teachers associations have demanded back their old status, but students seem to be attending more school days per year given that the subsidies are paid per "student day attended" (Schiefelbein and Apablaza, 1984). There are no successful cooperative attempts of public programs operating together with non formal education. The non formal educational experiences have shown the potential of using monitors from the community (Garcia Huidobro, 1982), but little has been done in terms of using monitors by the regular schools.

- 83 -

Table 39: COLOMBIA: MINIMUMAND MAXIMUMANNUALSALARY OFPRIMARY TEACHERS IN DIFFERENT REGIONS BY CATEGORIES - 1988-1971

Categories

1988 Minimum Maximum Diff. Salary Salary in X

1969 Minimum Maximum Diff. Salary Salary in X

Minimum Salary

1971 Maximum Diff. Salary in X

No diploma

8,120

10,880 +74.5

7,400

14,400

93.6

13,752

18,800 +22.2

4th Category

7,320

12,000 +83.9

9,800

15,984

88.5

18,752

20,160 +20.3

3rd Category

8,040

13,800 +71.8

10,580

16,992

80.9

18,276

21,660 +18.5

2nd Category

8,780

15,000 +71.2

12,000

18,240

52.0

20,100

23,280 +15.8

1st Category

9,720

18,200 .88.7

14,400

19,440

35.0

23,500

24,840

t

8.2

Source: Unpublished documentsof the Ministryof Educationin Tibi, Claude,'Financement et developpement de l'education: le cas de la Colombie,*, Rapportde RechercheNo. 22, IEEP,Paris,1978 Allocation of subsidies through the "best 20,000 students" taking each year the UEET may also be considered another attempt to stimulate a competition process among universities and to decentralize decisions. Although the results have not been positive in the Chilean experience (see C.3 above), it represents a potential mechanism for taking into account student preferences that should be studied in more detail. Central authorities have seen special problems for developing nonformal education programs. An impressive amount of non-formal educational experiences have been carried out with private funds. Most of the operations remain in small scale. Experiences like ACPO in Colombia were massified only with the help of government and international funds. Many styles can be identified in the non-formal education, but little can be said in terms of their future massive impact (Garcia Huidobro, 1982). The most interesting examples of generation of participation processes are probably the bilingual projects. Although conclusions are still to be refined, learning to read in the mother language (relating signs and sounds) and then learning to read Spanish as a second language seems to be most efficient way to learn to read and write (Corvalan, 1984). The solution of bilingual education problems may represent a drastic reduction in repetition rates in primary education. If the bilingual education cut in half the 10 million repeaters in first grade in 1985, the saving would equal some US$300 million per year.

- 84 -

Summary Resource allocation within the education sector depends on the specific process that must be affected and the agents involved. Legal and budgetary macro decisions may have an effective impact on access, and even evaluation services and new types of education may also depend on suitable laws and budget. However, changing more complex institutions, management and classroom processes may require additional efforts in terms of intellectual personal leadership, specific models or technology (including information). In any case, educational change is extremely slow and special care should be taken to avoid changes that teachers can feel they are not prepared to implement. Proposed reallocation of resources should not be threatening to teachers in order to be successful. In brief, elasticity of substitution between different factors of the educational production function is very low. Buildings cannot be substituted for teachers. and vice versa, and both factors account for a very large percentage of total costs (see tables 4 and 5). Therefore, costs studies and Management Information Systems may help decision makers to improve efficiency and to suggest the best way toallocate future resources, but there is a strong warning on quick changes because the high proportion of costs made up of salaries is intractable. On the other hand, actions aimed to reallocate resources such as bilingual educational, multigrade teaching, use of textbooks, discussion circles (teachers defining problems and together looking for possible solutions), use of mass media spots to show how to use social equipment (used newspapers, the memory of the olders, the countryside, commercial TV or museums), or how to provide early stimulation may have a large impact in speeding up the flow of students through the system and in allowing the student to reach higher achievement and self-esteem levels. Only if half of the repetition is cut down $300 million would be saved per year. D. Policy Implications from the Analysis of Available Research The analysis of available research (published before May 1985) provides some guidelines that could be used by decision makers. First, results from available research suggest that unit costs per graduate may be reduced by increasing the internal efficiency. Second, research results suggest that the private sector is now the obvious source of additional funding and several mechanisms for encouraging contributions from the private sector may be assessed. Third, research results show some of the elements to be considered in assessing specific policies aimed at improving resource allocation. D.1 Reducing Unit Costs of Graduates by Increasing Efficiencies. There are no built in incentives for improving the efficient use of resources in the education sector, but efficiency may be increased by improving management, finding the right combination of input factors, changing promotion problems and reducing side effects of selectivity. Although management practices in Latin America are poor, there are also some signals that educational management is improving. The expansion of enrollment in primary and secondary schools has consisted of children from increasingly poorer segments of the population (Selowsky, 1980 DEALC, 1981). However, in spite of this fact repetition and dropout rates are

- 85

-

somewhat constant or improving overtime. Therefore, some improvement in management should be explaining that performance. Cost studies and information systems are positive factors, but there is a long way to go before information becomes a powerful sources of continuous improvement. The evidence suggests that payoffs associated with some costly policies (for example lengthening teacher training and reducing class size) are smaller than cheaper factors like multigrade teaching textbooks or a good school calendar. It is true that the number of untrained teachers should be reduced in order to increase internal efficiency, but more research should be available in order to provide an empirical basis for finding effective upgrading methods (lengthening their training or requiring longer years of previous schooling is not cost effective). There is still a long way to reach the target of 100 percent of trained teachers in all Latin American countries (see Annex 3). Furthermore, the level of teachers' training modify the effect of reducing class size. Only good teachers will benefit from smaller classes. On the other hand, substantial gains in efficiency (mainly for small rural schools) may be obtained through multigrade teaching (when students have textbooks), that allows the operation of complete schools (offering all grades in primary) that reduce early attrition and allow larger school sizes (Annex 4 illustrates the proportion of incomplete schools in the system). Finally, adjustments to the school calendar that allows students to attend a larger school year period have been reported as positive. (Further comments on the implementation of changes are presented in D.3). Repet-ition is still high in Latin America, especially in the first grades of primary education, and is the main source of inefficiency. Both unrealistic promotion standards and lack of training in multigrade teaching are main causes of the high repetition levels. Although there are substantial increments in the promotion and retention of students in the last two decades, still in many countries there are no policies for reducing repetition. In most countries the private schools do not allow their students to repeat and the public system must accept those less able students. Thus comparisons of achievement in both types of schools should be carefully done (Uzategui and Muelle, 1974). Socioeconomic, geographic and ethnic factors affecting present educational selectivity allow increments in achievement levels of a selected group, but probably reduce the overall achievement level of the region. Selectivity allows that all children from upper income levels (including the children of decision makers) be attending good schools, therefore no automatic incentives are built into the society for improving the quality of schooling for those negatively selected. Most countries have made efforts to open vocational opportunities that reduce the number of candidates for university admission, but no one has yet succeeded, mainly due to a salary structure and traditions that associate a high payoff to university graduates. Free tuition does not reduce an educational selection process that starts at the cradle and continues in the admission to specific university careers and in the graduation. In fact in some cases, as in the admission into the university, the free tuition system exacerbates inequities.

-

D.2

86

-

Increasing Private Contribution

Latin America has been able to reach an important (30 to 40%) level of privatization of provision of all educational services in the region but there is ample room for further increase in the share of the private sector. Although subsidies to non-fee private education have increased private educational services reduced the total amount of public resources devoted to education most of the private activity has no public support. Subsidized systems reducing the use of public resources are available only in a few countries and those systems could be extended to the whole region. Several countries have been trying successfully to raise user charges in university education. There is a need for providing more objective evidence to the public opinion about the rationale for raising fees and for diffusing government reliable strategies for introducing change in the present tuition free policies. Therefore, the future contribution of private funds may be still expanded in several countries of the region. Countries have no clear policies for the role that private education should play in the whole system. Private education caters mainly to both the upper socioeconomic suburban class, but it also enrolls some students from the population living in urban-marginal or rural areas. The magnitude of the private educational activity in the lower socioeconomic groups depends on the amount and conditions of public subsidies. Although there is a large number of experiments some have been mentioned above and more are described in Mun±oz and Hernandez, 1978, it is necessary to analyze in detail their pros and cons with respect to well defined criteria. Capital loans from international organizational (with state guarantee) could also be considered as one type of subsidies to be considered in the analysis of incent-ives for private activity study. There seem to be a contraction in the educational investment of the Catholic Church schools during the seventies and eighties (Catholic Church is now less interested in education in comparison with pastoral activities) that should also be studied in more detail if the trend could stress demand for public schooling. There are new activities of industrial groups that have funded or supported regional universities that should also be analyzed, through a series of relevant case studies, if government want such initiatives to be replicated. Employer-financed vocational training represents only 2% of the total number of students in the formal education, but the number of worker trained in their (usually short) courses is at least similar to the university enrollment and there are also many in-service training activities in private firms. Studies on the rates of return of their graduates show excellent results (Castro, 1978). Although the training system works efficiently and the system may be used as a model for countries that have not yet implemented this type of institution, additional resources to be obtained from payroll taxes may be relatively modest. Some interesting new experiments on employer-financed training and research are now being developed and future evaluations may provide valuable information for designing new employer-financed training mechanisms in countries that have not yet experimented with them. Although the experience with student loans in Latin America shows that it is possible to have a reasonable efficient administration of a

- 87 -

student loan system in a variety of conditions, the capital of all institutions has been reduced dramatically due to recovery in highly devaluated money. Better systems must be designed for taking into account the situation of those students not finishing their careers, the adjustment in payments due to unemployment; the demand of collateral guarantees, the length of the initial grace period, the lack of suitable insurance systems; and the effects of the criteria used for selecting the recipients of the loans (Woodhall, 1983; Velloso, 1984; ICETEX, 1984). Although educational voucher systems have not been tried out in Latin America, in Chile a voucher system has been experimented with success in Health (participants in the Social Security System may pay a percentage of the medical bills with vouchers issued by the SSS). In general the administration of loans or voucher systems is quite complex and administrative costs are high. Differential pricing has been successfully used by private schools and universities and there are also small scale experiments in selected public institutions. The use of differential pricing could be further diffused in the region. There is a general agreement that loans systems must include scholarships programs to reduce a negative impact in equality. In general the experience shows that the distribution of scholarships not directly related to assessed needs will tend to increase inequality. Precise criteria for selecting postulants and use of local groups to allocate available scholarships may contribute to an efficient allocation process. Lack of serious legal sanctions for perjury (given the Latin structure of the law) preclude the use of declaration under oath as the only source for allocation. The weak tax control system also makes unreliable the use of tax declarations as the basis for allocating scholarships. There is still excess social demand for education in all levels in almost all countries with the present fee level structure. In primary education such excess demand is mainly located in rural isolated areas. In secondary there is excess demand for academic education and eventually for university training. Those countries offering some primary education to all members of cohorts reaching the school age seem to be reducing their fertility rates and the reduction in fertility rates as a built-in control mechanism for constraining further social demand. On the other hand, unemployment seem to be a built-in mechanism for increasing social demand in secondary and higher education. (El Salvador, 1981.) The evidence suggest some responsiveness of enrollments to fee increases through changes both in supply and demand levels. The existence of flexible mechanisms to analyze the case of families with financial problems subsidize that allow mixtures of public and private support and options among alternative schools for the children to move from one type of fee to a different level may reduce the impact of taking the students off the system. Although the number of successful experiments in cost recovery is small, they suggest that more attempts can succeed in the future if the pub2ic opinion is well informed with relevant data from the country and with suitable regional comparisons. Cost recovery policies should be developed in a medium-term range, unless extremely favorably public support is obtained beforehand. Initial allocation of cost recovery funds to

-

88 -

scholarships for needed students may ease up the acceptance of cost recovery policies. Further cost recovery through teachers' salaries should be avoided and eventually salaries should be raised. Although reduction of teachers' salaries has been one of the long term strategies used by government to expand education with the available amount of resources allocated to education, such a reduction is preventing the attraction of good candidates into the teaching force, damaging teacher motivation and eventually reducing students' time on task. Although public education extended educational opportunities to the rural areas and to the marginal urban groups, thus reducing previous inequalities, there are also cases where discriminatory effects of public non-price allocation processes are observed. There is very little competition from private education in rural or poor areas. With only a few exceptions most of the changes generated by public allocation processes in Latin America are the result of slow processes that gradually alter the "status quo" rather than the result of quick changes triggered by a single decision. As a result of those slow evolutionary processes secondary education has expanded, higher education has been forced to expand through better use of their resources (and partially becoming a refined secondary education, rather than a research oriented university type of education) and now there is a new interest in preschool education and in improving the quality of the primary education (especially in terms of reducing repetition and improving some basic skills in reading and writing). Resource allocation within the education sector depends on the specific process that must be affected and the agents involved. Legal and budgetary macro decisions may have an effective impact on access, and even evaluation services and new types of education may also depend on suitable laws and budget. However, changing more complex institutions, management and classroom processes may require additional efforts in terms of intellectual personal leadership, specific models or technology (including information). In any case, educational change is extremely slow and special care should be taken to avoid changes that teachers can feel they are not prepared to implement. Proposed reallocation of resources should not be threatening to teachers in order to be successful. In brief, elasticity of substitution between different factors of the educational production function is very low. Buildings cannot be substituted for teachers. and vice versa, and both factors account for a very large percentage of total costs (see tables 4 and 5). Therefore, costs studies and Management Information Systems may help decision makers to improve efficiency and to suggest the best way to allocate future resources, but there is a strong warning on quick changes because the high proportion of costs made up of salaries is intractable. On the other hand, actions aimed to reallocate resources such as bilingual educational, multigrade teaching, use of textbooks, discussion circles (teachers defining problems and together looking for possible solutions), use of mass media spots to show how to use social equipment (used newspapers, the memory of the elders, the countryside, commercial TV or museums), or how to provide early stimulation may have a large impact in speeding up the flow of students through the system and in allowing the student to reach higher achievement and self-esteem levels.

-

89

ANNEX 1

STRUCTURE OF THE LATIN AMERICAN SCHOOL SYSTEM, 1976

Country

AntiguaArgentina

Barbados Bolivia Brazil Colombia

Primary education Ag_e Length

General Age Length

5 6 5 6

11 13 11 14

3 + 2 3 + 2 3+ 3 4

15 11

3 4+ 2

6 7 6

SecondarvEducation Teacning Training Other Age Length Age Length ... 14

2 4

16 18 11 14

2 1-6 1-4 4-6

15 15

3 2

15 15

.3,4 3

-

7 6

8 8 5

Costa Rica

6

6

12

3 + 2

-

--

Cuba

6

6

12

4+ 3

12

5

14

3,4

Chile

6

8

14

4

Dominica Ecuador El Salvador

5 6 7

7 6 9

12 12 16

5+ 2 3+ 3 3

Granada

5

7

12

5 + 2

7 5 .6 6 6 6

6

13

6 6

Hontserrat

Nicaragua Panama Paraguay Peru

-

3

-

-

14

4,5

2 2 3

15

3+ 3

15 18 16 15 16

11 12

5 + 2 3+ 4

6 6 6

12 .12 12

3+2,3 3+ 4 3+ 3

5

7

12

5 + 2

-

7

6

13

3+ 3

16

3

13

6 7 6

6 6 6

12 13 12

3 + 3 3 + 3 3 + 2

15 -

3 -

15 15,18

Rep. Dominicana 7

6

13

2+ 4

14

4

12

3-5

tliaves-Anguila San Cristobal Santa Lucia San Vicents Suriname Trinidad and

5

7

12

4 + 2

-

-

17

2-4

S 6

7 7 6

12 12 12

3 + 2 5 + 2 3 + 3

18 18 14

2 2 3-5

16 12

1,2 3,!

5 6 7

7 6 6

12

3 + 2 3 + 3 3 + 3

-

-

17 18

4 5

15 12 13

2,3 3-7

Guatamala Guyana Haiti

Honduras Jamaica ".exico

obago Uruguay Venezuela

Source:

5.

UZNSCO, 1980

,

.

12 13

16

2,3 1-3 .3

1

12

1-5

3

13

18

3

14 - 13

3 1-5

15 17 15

3 3

12 15 12

-

.

3,4

2,3 1,2 1,4 2 5-6

11 .

1-3 2-3

2,6

.

-

90

-

ANNEX 2

EDUCATIONAL INDICATORS IN LATIN AMERICA: UNIT COSTS PER STtJDENT, AND GDP IN LATIN AMERICA

Percentage of GDP Allocated to Education 1965 1970 1975

Country

Argentina Barbados Bolivia Brasil Colombia Costa Rica Chile Ecuador El Salvador Guatemala Haiti Honduras Jamaica Mexico

Nicargua Panama Paraguay Per-u Republica Dominicana Trinidad y Tobago Uruguay Venezuela TOTAL

1/

1971

fl 3/

1977 1975

Source:

Unit Cost Per Inhabitant 1970 1976

6.19 4.11 7.53 7.02 3.33 3.53 3.80 4.01 6.76 4.80 5.40 5.42 4.59 -5.57 3.82 5.79 3.18 3.44 2.45 2.55 1.75 1.86 3.01 3.21 3.24 3.75 2.66 3.00 2.16 2.66 4.87 6.02 1.93 2.84 5.70 4.39 3.06 3.28 3.39 4.18 4.35 4.05 4.22 5.19

5.34 9.95 4.02 3.66 7.01 5.33 9.81 3.77 4.40. 2.65 1.88 4.00 5.46 3.03 2.86 7.12 2.83 5.17 2.66 3.99 4.52 6.48

41 1/ 44 8 13 11 1/ 29 46 12 8 7 1 9 26 17 10 39 5 17 10 30 31 51

4.26

4.51

19

J.M. Dagnino Pastore, UrSCX, 1980:85

4.04

1978:187

45 2/ 116 16 29 76 28 3/ 36 2/ 19 12 2 16 100 53 22 66 10 31 16 73 -

134 41

Unit Cost Per Student 1970 1976

200 I/ 156 51 70 53 1/ 117 129 56 45 62 14 51 97 78 51 161 26 71 47 111 160 220 92

212 2/ 444 82 129 -

297 99 3/ 134 2/ 89 91 15 80 341 207 104 222 48 113 61 313 526 172

STUDENT-TEACHER RAI'IOS AND TEACIIER TRAINING

P R I M A R Y Trained Teachers Student-Teacher

(M)

Ratio 1973 1/

CouLItr-y

1965

1970

Aigenitina Barbados BolIvia Brazil Colombia Costa Rica Chitle Ecuador El Salvador Guatemiiala . Ilaiti hlonduras Janicalca Mexico Nicaragua Panamisa Paraguiay Pern Republica Dominicana 'Irinidady Tobago Uruguay Venezuela

20.1 35.7 30.1 28.3 36.1 34.4 42.7 38.1 32.2 34.3 44.1 28.7 57.7 46.7 35.5 30.7 26.9 35.1 53.4 34.2 30.5 34.4

18.8 18.6 s.d. 37.2 s.d. 25.8 28.0 26.2 40.4 38.6 29.8 29.6 s.d. 41.7 39.2 39.4 36.4 .42.0 36.0 36.3 s.d. 43.9 s.d. 36.9 51.9 43.4 50.1 47.1 27.5 38.7 29.3 29.4 29.6 28.5 38.8 31.3 58.5 55.5 36.0 35.9 29.5 25.9 34.5 33.8

35.9

36.8

Total 3 1/ 2/ 3/

S E CO N D A R Y Trained Teachers Student-Teacher

35.3

1967 2/

(%)

Ratio 1971 2/

100.0 s.d. 44.6 60.3 59.2 93.6

100.0 s.d. 52.7 65.0 65.1 96.3

88.3 76.1 83.2 s.d. 42.7 s.d. 74,2 52.8 95.8 79.9 s.d. 19.8 76.9 s.d. 88.2

93.8 98.5 89.4 s.d. 55.8 s.d. 79.5 68.8 70.2 85.4 s.d. 20.4 74.4 s.d. 89.4

71.0

75.,0

In a few cases data correspond to 1971 or 1972. In a few cases data correspond to other years. Non-welghted averages.

SoItic'e:J.1. Dagnifo Pastore, 1978:158 in IIrodersohn and Sanjurjo, 1978.

1969

1971 2/

1965

1970

1973 2/

7.1 s.d. 17.3 14.9 13.3 21.2 s.d. 12.6 23.5 7.7 16.0 s.d. 23.0 12.7 15.5 20.8 8.2

7.1 s.d. 19.3 13.2 17.2 24.1 s.d. 14.3 28.1 13.7 16.9 s.d. 25.2 13.7 26.0 20.7 9.4

7.3 s.d. s.d. 13.6 19.7 29.4 s.d. 16.1 35.7 14.5 s.d. s.d. 26.3 14.0 27.3 20.8 10.1

100.0 s.d. 53.3 s.d. 61.7 82.2 98.5 12.2 92.2 88.8 s.d. s.d. s.d. 50.6 s.d. 71.0 98.5

100.0 s.d. s.d. 's.d. 65.6 85.5 98.9 19.0 93.1 90.9 s.d. s.d. s.d. 49.4 s.d. 61.8 s.d.

23.1 24.1 s.d. 19.1

23.3 24.7 s.d. 21.4

25.9 s.d. s.d. 22.7

30.7 47.9 s.d. 56.4

28.9 48.7 s.d. 52.6

16.4

19.0

18.0

66.0

66.2

-

92

-

ANNEX 4

PERCENTAGE OF SCHOOLS ONLY OFFERING PARTIAL PRIMARY EDUCATION (NON-COMPLETE PRIMARY EDUCATION)

Countries

Year

Percentage o-f Incomvlete Schools All Urban Rural

Argentina Barbados Colombia Cuba Dominica Ecuador Granada Guatemala Guyana Haiti Montserrat Panama Paraguay Peru Republica Dominicana San Cristobal Nieves - Anguila Santa Lucia San Vicente Venezuela

1970 1970 1968 1971 1970 1971 1970 1969 1970 1970 1971 1970 1968 1968 1970

14.7 21.2 71.3 0.0 5.2 46.3 3.4 75.2 0.0 0.9 20.0 31.2 65.4 50.8 78.9

5.9 29.6 31.7. 0.0 28.6 11.6 11.4 21.6 0.0 1.4 0.0 1.0 12.7 29.4 21.8

20.3 14.1 92.5 0.0 2.0 56.5 0.0 90.8 0.0 0.0 27.3 33.1 80.4 60.8 87.4

1971 1970 1971 1970

13.8 64.1 9.8 63.2

11.1 94.1 0.0 18.9

15.0 30.0 11.8 87.2

Source:

UNESCO, Statistical Yearbook,

1972.

GUATEMALA

UNIVERSIDAD RAFAEL LANDIVAR (CENTRAL CAMPUS) NEWCOUENS BY ACADEMIC UNITS

-

1982-1984

1971

-

174

191

-

108

-

141

108

-

-

-

-

-

290

260

170

121

118

12

21

-

1983

1984

1966

1988

1987

1988

Economia

44

103

177

221

209

217

192

Ingenieria

-

-

104

106

-

-

-

204

228

377

-

-

-

-

47

77

91

89

81

;118 300

401

639

695

Arquitectura Politicas llumianidados

Ambientales

88

34

163k

-

1465

288"

-

-

Derecho

Maostria Ing.

-

466 -

-

103

103

118

1 98 9 b

-

267*

619

281

309

-

81

-

-

84 -

1973

1974

1976

1978

1977

1978

1979

1980

1981

1982

1983 1984

206

188 72

197

158

199 160

287 278

632 404

743 423

879 376

810 460

791 397

830 269

686 209

18

13

23

17

29

19

22

10

22

29

24

82 18 3

106

108 -

d 4

203

-

-

-

-

-

-

Integ. Cultural

-

88

80

Juridicas Doct.

-

116

1 9 72 b

1970

1982

Acadoinc Units

-

-

302

127

135

124

208 - 138

117

87

70

83

80

147

184

206

136

133

-

-

-

-

-

-

-

-

135 20

140 16

142 2

-

-

-

84

-

-

-

12

-

80

37

24

37

42

82

92

S3

-

88 -

-

_

_

_

-

-

-

-

_

-

-

-

_

-

87 -

80c

_

_

_

_

_

21

24

28

-

-

-

-

-

-

-

96C

64

-

Total

138

269

a42

488

483

608

541

606

846

637

688

618

687

882

892

1882

2080

2486

2267

2420

2028

1778

1603

Annual Foe

400

400

400

400

400

400

400

400

400

400

400

400

400

400

400

400

400

400

800

800

800

800

800

17%

27X

24X

Estudioa Gen.

-

-

Attrition s/ b/

Including students In the *Bolgao Not available by acadomic units.

C/

Estimated

d/

Includas students beyond first year.

Sources:

Detailed by poriods:

Oficina do Planificaci6n, Boletin Informativo, Nos. 1-4, 1988. 1970-71 and 73: Socrotaria General, hoy Diroccion do Registro. 1974-78: Of icina de Planificaci6n 6 1977-80; Dr. Santos Peroz. Matodolo ia para elaborar proyocciones de poblaci n estudiantil 1982-B8:

1981-82; 1983-84:

Departamento de Estadistica. Informe estadistico del Centro de Cilculo, de 8-111-83 y 3-IV-84, resp*ctivamente.

en

al URL, 27 Oct. 1980.

26X

FEE POTICIES

FOR PUBLIC SCIIOOLS, SELECTED COUNTRIES

Prlmary

Schooling

Secondary

Schooling

Coui!t-ty

Year

Bolivia

1977

Registration fee of US$0.50.

Registration

Ilaiti

1983

Tuition fee of US$2.50, of which at least $1.00 is placed in a school 'dossier" account maintained by the Ministry of Finance. Most dossler funds were used for construction of secondary schools in 1982. Destination and use of funds is unclear. Fees unchanged for at least a decade. Parents purchase uniforms, supplies and textbooks (latter costs $6-8 per year).

Tuition fee of US$3.00

Paraguiay

1983

Tultion fee of US$4

Tuition fee of US$6. Vocational schools demand purchase of materials.

Clille

1984

Registration fee of US$4. contribution to.the PTA.

Ecuador

1984

Only voluntary contributions fixed by the Comite de Padres de FAmilia (up to US$4)

Similar to primary education. schools also ask for school materials.

1984

Tuition free, but Sociedad de,Padres y Amigos'de la Escuela asks for c,ntributions.

-Similar to primary educatton.

1984

Junta de Padres asks somneUS$0.50 of voluntary

US$2.00 of voluntary contributions

Doniican

Republic

Costa Rica

There is also a variable

fee

of

US$0.50.

Registration fee of US$5. There are variable contributions to the PTA and to each class.

contribut iOns

Uruguay

1984

IIondlurati

1984

S(MLC'Z(i.;IWO), A f.i:SW01L11, It1.12111a

inll,

.O

lt

II

Comision Fomento asks around US$0.10 to US$2.00 per month. Funds are used by the principal. T'eachercontributes. T'uition free. Sociedad for small contributions.

1984 for s..

llaiti.

Sector

Reports

de Padres

de Familia

ont Paraguay,

1984.

Similar to primary education '

asks

Reports

'Registration Examination fromii teachers

fee US$12. ' fee US$5.50. in

each

of

the

But

ANNEX 7

COMPARISON OF PARTICIPATION AND FINANCING OF HIGHER EDUCATION IN SELECTED COUNTRIES, AROUND 1980 (Percentages)

Relevant Age Group Entering Higher Education (a)

Country

Bolivia Costa Rica Colimbia Chile Guatemala Uruguay Honduras Dominican Rep. Brasil Ecuador Mexico

Sources:

a/

Tuition Fees As Proportion of University Income Public Private

8 7 13 6

1 8 25 10 5 10 1 5 2 3-70f/

E. Boenninger (1978) pg. 337

30a/ lOOa/ 75a/ lOOb| lOOc/ 100 80 80 100 50e/

Proportion of Students Receiving Financial Aid from Government

8 75

10 2

8So/

In 1986 fees for a 12 credit

semester were US$40.

b/ C/ d/

el

Unpublished data, Consejo de Rectores (only for the new private universities created in 1982-1983). Mission estimates, October 1984 R. I. de Acuna et al., 1983 G. Pozo de Ruiz et al., 1977; cuadro 17

f/

D. Levy, 1979

gA

Munoz and Hernandez, 1978: 266

COLOMBIA; UNIT COSTS PER STUDENT BY SOLURCE OFiFINANCING 1980

(as. 1

_ Case 2

Casw 3

-

Case- 4

8

XS

SUBTOTAL PUBLIC

3,881

89.8

A.

3,749

87.1

Public

Education

1. MLn-ICCE-E.N. 4.6 0.8 3.6 2. FER 3. ICEIEX

2b0 50 40

Case

Case 6

Sources

20

47.1

8_660

72.7

1,9J0

62.4

8,000

84.9

4,330

84.0

47.1

3560

39.4

_1_

29.4

8,000

84.9

2 880

39.3

1,480

16.6

2

1,920

46.3

38.8

1,100

29.4

6,400

87.9

2,340

34.8

3 000

33.3 24.7

0.8

3,490

d2.8

3,600

1.2

a80

-

1,90

4. DAINCO

140

--

1.6

--

6. SCRIA ED. DEP.

560

--

B. 141 80

Non Public 2.6 23.08 1.

70

1.2

Ed. 8

ECOPETROL

1.2

2. CNT 20

0.4

3. SENA --

__

00

6.8 4. ICEF

-

0.1

--

_

--

--

2,120

210

l

DRI

b. UUNICIPIO 60 0.8 --1,680 23.3 8. DFPARIAUEHI() --

6

--

--

2.3

170

1.9

220

6.8

23.6

840

--

17.2

--

StOBIOIAL CAUASIPUB1.IC __

--

670

16.7

1. COMITE CAFEIERO 870 16.7 SUJTOTAL PRIVATE 6.0

210 80 10 20 40

00

1. 2. 3.8 1.8 0.1 3. 0.4 3.7 2.1 4.

FAMIIY TEACIIERS

30.4

1,680

37.2

2,480

27.3

1,780

47.6

1,420

15.12,4303

1,480

28.2

1,310

30.9

1,880

18.8

1,670

41.9

1,220

12.9

9,420

100.0

1.4 2.0 2.8

1

70 190 120 140 190

2.8 3.7 1.3

a 2

e0

2,110

29.8

COUUUNITY

NON PROFIT 280

--

1,890

-3.1

--

6. CIIURCIJ --

--

10

__

8. PRIVATE --

30

10

0.1

--

a.3

0.8 --

7. FIRMS --

--

__

TOTAL

60

1.2

40

0.7

6,681

100.0

4,240

100.0

9,010

100.8

3,140

1

0

8,780

Souru*:

104.0

P.4l**

.............

, Alb&, I'8l:su.d,

Latin Ameica:

Percentage Oistribution

of Public Current Ezpenditutr S.cna,t

Py,mx

A,gernU,a

43.8

204

270 1

8jg0arcoI

'i6.4

34?

27 3 644

Bo,,a

4*J6 23.7,1

349

.19.5

06.5 52s 2

44 27' 37:' J2 dI

50.4 44 4103 '656

zt '4 1

± 0 11 4 71 1 49.4

±8.0 46 5 54 2 44 7 sa0?

*taraqV4

60.9

57 i

Plramt

526 Ss.65

a3 9 848

41 3 54 2

41 4c 52 5

44J9

45 1

432 52.5

*

EJS*vw

0Gy4fa

*59

PM,@ Tr.4= am

roaac

~;~U4yN Vemi.7eLt UiitdA4fIa

Source:

faveags)

IDB, 1978,

(a) 1988;

(b) 1967;

PSi. (c)

70

14gc

32

i 3 'I 9

_-

"3

_5

-

l't t 56.

6 '° -

A 4 3 :7 2

;7

175 t_0

1.

-

'4 6

253

-

95

283

10

22

'3 2

46.8

412

I

(d) 1971;

244 x

007

22.fft

t4 31 I

3 i s '

'22

-

'

'76 -

'IC

10a 4

255

22.7 I 1424

2 6 !24A '72

;2? 16 1 .3;

53

21

-

'21

33 '2 53

't 3t

I'3

:

02

03 t1 5

14 7

4

19 8 '221 6 -

12.6 17 9

160 .44

1?

'S'

-

'14

121

-

:5 6 5

183 9a

-

I 15.l

:oa6

t.0

84

_-

00

ll/

:5

55

070

's9

'55

22.4

20.

20.3

15.9

11.7

19.5

13.5

13.1

18.3

(f) 1974;

3j^z;zes

;7

it S14

pp. 559-581

(9) 1978

:4-.3 '.,,4cor

;,,Jtt^4tiF

te45\C

;

.

;

,

! . j *-