9547 -and c~~~~ :-Economic Devnelopm ' FILE COPY FILE C.:WPY edited by Stephen P Heyneman Daphne Siev White LC 65 .Q38 1986 - AAATxA4n1^l Os4ifl JOINT BANK-FUND LIBRARY LC65 .Q38 1986 The Quality of education and economic development / JLC039928 The Quality of Education and Economic Development A World Bank Symposium .03f I~ ~~92 The Quality of Education ans Economic Development edited by Stephen P. Heyneman Daphne Siev White The World Bank Washington, D.C., U.S.A. Copyright (D 1986 by the International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing February 1986 The World Bank does not accept responsibility for the views expressed herein, which are those of the authors and should not be attributed to the World Bank or to its affiliated organizations. The findings, interpretations, and conclusions are the results of research supported by the Bank; they do not necessarily represent official policy of the Bank. The designations employed and the presentation of material are solely for the convenience of the reader and do not imply the expression of any opinion whatsoever on the part of the World Bank or its affiliates concerning the legal status of any country, territory, city, area, or of its authorities, or concerning the delimitation of its boundaries or national affiliation. Library of Congress Cataloging-in-Publication Data Main entry under title: The Quality of education and economic development. "The papers were originally prepared for a conference on school quality sponsored by the World Bank's Research Committee and held in Harper's Ferry, West Virginia, in May 1983"- Foreword. 1. Education-Economic aspects-Congresses. 2. Education-Aims and objectives-Congresses. 1. Heyneman, Stephen P. 11. White, Daphne Siev, 1955- III. International Bank for Reconstruction and Development. LC65.Q34 1986 370.19'341 85-31530 ISBN 0-8213-0702-9 Foreword The issue of quality in the educational systems of vocational skills depended on the quality of prior basic T developing countries is a relatively recent one, and education, that the quality of facilities was as important not much has been written on the subject. For many as quantitative targets, and that the quality of education years most of the emphasis was on growth and on the required the same capital investment and managerial need to provide more schools and more classrooms for attention as did the quantity of education. students in these countries. Since 1963 the Bank has supported projects for As these school systems have grown, however, atten- teacher training, curriculum development, mass media tion has been shifting toward the quality and value of the education, research and evaluation institutes, and ex- education being received. The papers in this volume are amination systems-areas that were hardly considered some of the first to address this topic. Two types of potential investments when the sector was inaugurated. studies are presented here: the first relates to the eco- Early projects supported the construction of schools, nomic growth that might result from expenditures on but more recent projects also support the equipping of the quality of education: the second concerns the best schools. For example, 40 percent of the Bank's educa- way to implement changes intended to improve the tion projects now contain textbook components. quality of education. The papers were originally pre- The seminar described in this volume was designed to pared for a conference on school quality sponsored by take stock of two things: the recent research on the the World Bank's Research Committee and held in Har- economic implications of investments in the quality of pers Ferry, West Virginia, in May 1983. education and the operational experience acquired as a The World Bank has been assisting developing coun- concomitant of this investment. The Economic Develop- tries in the field of education for more than twenty years. ment Institute (EDI) is helping to disseminate these re- During that time 284 projects costing more than suits to create a more informed and informal dialogue. US$12.1 billion have been approved for support with This is one of a series of symposium publications de- Bank loans and IDA credits in more than ninety coun- signed to help officials in developing countries and Bank tries. This assistance has helped to create more than 2.6 staff keep up to date on each other's current thinking. million new student trainee places in approximately 21,000 educational institutions: 185 universities, 651 teacher training colleges, 2,903 secondary schools, and Christopher R. Willoughby 18,000 primary schools. Director The effect of this experience has not been confined to Economic Development Institute the developing countries, however. The Bank's oper- Operations Policy Staff ational concept of the education sector has grown and, perhaps more significantly, has diversified. In the begin- Aklilu Habte ning the Bank emphasized quantitative targets in sec- Director ondary and technical education. But gradually it recog- Education and Training Department nized that the efficiency of transmitting technical and Operations Policy Staff Contents Foreword v Contributors ix Other Participants x 1. Overview 3 Stephen P. Heyneman Educational Quality and Economic Growth 4 Implementing Investments in Quality 4 PART I. EDUCATIONAL QUALITY AND ECONOMIC GROWTH 5 2. The Quality of Education and Economic Growth: A Review of the Literature 7 Lewis C. Solmon Measuring Quality 8 Quality Inputs 9 A Review of the Literature 9 School Learning and Job Performance 10 Decisions about Methodology 11 Microstudies in the United States 11 The Conflict between Quality and Equity 13 References and Bibliography 14 3. School Quality and Economic Growth in Mexico 17 Bruce Fuller, Kathleen Gorman, and John Edwards Investment in Schools and National Growth 18 Historical Background 20 Linking School Investment and Quality to Economic Growth 24 Methodology and Procedures of the Study 25 Findings of the Study 27 Discussion and Future Work 29 Notes 30 References 31 PART II. IMPLEMENTING INVESTMENTS IN QUALITY 35 4. The Stages of Growth in Educational Systems 37 C. E. Beeby Success Is Illusive 37 vii viii Contents Stages of Growth 38 Qualitative Change in the Classroom 39 External Constraints on Change 43 Conclusions 44 References 44 5. Lessons from Bank Experience 45 Administration of School Systems 45 Examination and Selection Policies 46 Promotion and Repetition Policies 47 Teacher Training 49 Time on Task 50 Prevocational Subjects 50 The Production and Distribution of Textbooks 52 The Use of Electronic Media 54 Bibliography of World Bank Research on the Quality of Education in Developing Countries " 57 General Estimates of School Quality and Gains in Achievement 57 Analyses of Specific School Qualities and Specific Countries 57 The Economic Effects of School Quality 58 School Quality as Policy 59 Contributors C. E. Beeby helped to found the United Nations Educational Stephen P. Heyneman is chief of the Education and Training Scientific and Cultural Organization (UNESCO). He is cur- Design Division in the Economic Development Institute of rently at the New Zealand Council for Educational Research. the World Bank. John Edwards is an assistant professor of economics at Tulane Lewis C. Solmon, an economist, is a professor and associate University in New Orleans. dean in the Graduate School of Education at the University Bruce Fuller is a member of the Graduate Faculty of the of California at Los Angeles. University of Maryland and a researcher at the World Bank. Daphne Siev White is an educational writer and a consultant to Kathleen Gorman is a Ph.D. candidate in education at the the World Bank. University of Maryland. ix Other Participants The World Bank Douglas Keare, chief, Education Division, Latin America and the Caribbean Regional Office. Albert Aime, general educator/planner, Education Division, Sherry Keith, educator, Education Division, Latin America and Eastern and Southern Africa Regional Office. the Caribbean Regional Office. Gunnar Berlin, educator, resident representative, Nairobi, William Loxley, formerly consultant, Education and Training Kenya. Department, Operations Policy Staff. Victor Billeh, general educator. Education Division, East Asia Himelda Martinez, general educator, Education Division, and Pacific Regional Office. Latin America and the Caribbean Regional Office. Aurelio Cespedes, general educator, Education Division, Latin America and the Caribbean Regional Office. James McCabe, general educator, Education Division, Eastern and Southern Africa Regional Office. Hon-Chan Chai, general educator, Education Division, East- ern Asia and Pacific Regional Office. Robert McMeekin, economist, Education and Training Design Division, Economic Development Institute. Mae Chu Chang, formerly consultant, Education and Training Department, Operations Policy Staff. Abdun Noor, educator, Education and Training Department, Richard Durstine, general educator, Education Division, East- Operations Policy Staff. ern and Southern Africa Regional Office. Delia Pitts. formerly economist, Education Division, Western Frank Farner, general educator, Education Division, East Asia Africa Regional Office. and Pacific Regional Office. George Psacharopoulos, manager. Research Program Unit, Education and Training Department, Operations Policy Birger Fredriksen, educational planner, Education Division, Staff. Western Africa Regional Office. Shigenari Futagami, mass media specialist, Education and Andre Salmon, general educator, Education Division, East Training Department. Asia and Pacific Regional Office. Ralph Harbison. operations advisor, Education and Training Barbara Searle, general educator, Education Division, East Department, Operations Policy Staff. Asia and Pacific Regional Office. Mats Hultin, formerly senior advisor, Education Division, East Lennart Swahn, training adviser, Education and Training De- Asia and Pacific Regional Office. partment, Operations Policy Staff. Dean Jamison, chief, Education Policy Division, Education and Eric Swanson, consultant, Development Research Depart- Training Division, Operations Policy Staff. ment, Economics and Research Staff. Richard Johanson, senior adviser, Education and Training Jee-Peng Tan, economist, Education and Training Depart- Department, Operations Policy Staff. ment, Operations Policy Staff. x Participants xi Hans Thias, economist, Education Division, Western Africa Consultants Regional Office. Jacob Van Lutsenburg Maas, sociologist/educator, Education Philip Altbach, professor of Education, State University of New Division, East Asia and Pacific Regional Office. York at Buffalo. Mulugeta Wodajo, general educator, Education Division, Henry Levin, professor of Education and Economics, Stanford Western Africa Regional Office University (California). Laurence Wolff, general educator/planner, Education Divi- Benjamin Makau, fellow at the Institute of Development Stud- sion, Eastern and Southern Africa Regional Office. ies, University of Nairobi. Abdelwahed Zhiri, general education specialist, Education and Anthony Somerset, lecturer, Institute of Development Stud- Manpower Development Division, Europe, Middle East, and ies, University of Sussex. North Africa Regional Office. 4 The Quality of Education and Economic Development Overview Stephen P. Heyneman 4 T wenty years ago the issue of quality in education cal review; and students with special needs must be was regarded as delicate and politically sensitive. identified and assisted. Curricula were thought to be culture-specific, and And so during the past twenty years it has become national authorities were concerned first with localizing clear that much (although by no means all) of the educa- educational objectives and purposes. But the experi- tional process is universal rather than culture-specific. ences gained in two decades of reform and intervention An educational system, for example, is frequently the experiments have led both national and international nation's largest employer: it typically accounts for an authorities toward the view that, although cultural dif- average of 15 percent of a nation's resources, 25 percent ferences remain, the quality of education can-and in- to 30 percent of public sector employees, and on any deed should-be subjected to technical scrutiny. given day it directly involves a substantial portion of a There are several reasons for this. Many educational nation's population. Thus the efficient delivery of educa- objectives are the same from school district to school tional services is much more than a local problem: it is a district and from country to country. For example national and even an international issue. elementary mathematics taught in the fourth grade in Indeed the demand for international capital to im- Nicaragua overlap considerably with the mathematics prove the quality of education has risen markedly in the skills taught at the same grade in Thailand although the past decade. Previously most educational loans from the colonial history, culture, and govemments of the two World Bank were directed at expanding educational sys- countries are entirely different. Similarly technical skills tems by building more schools, hiring more teachers, in agricultural engineering taught in Bangladesh are and providing access for more students. Yet experience similar to those taught in other countries in Asia, as well with this growth showed that merely building schools as in Latin America, and in the Middle East, particularly and adding teachers without providing them with text- where agricultural cropping is similar. This being the books, classroom materials, or proper teacher training case, it is possible for education officials to ask whether does not bring about the desired changes in educational students in their school system are able to master skills achievement. as well as students in a neighboring system. This realization is reflected in recent Bank lending. It is also recognized that there are some common Although none of the thirty-one World Bank education problems the world over, particularly management projects appraised between 1963 and 1969 contained a problems concerning "inputs" to school quality. specific component for classroom materials, since 1976 Although classroom pedagogical style may be locally one in five projects has included significant funding for determined, the ingredients required to make class- electronic print classroom materials. One out of every rooms function properly are not. Teachers need to be two projects, moreover, has included funding for pre- trained, paid, and supervised; reading materials, news- service training. This does not imply that developing letters, and reference books need to be provided for each countries are less concerned with expanding education; subject; curriculum objectives require constant techni- what it does imply is that, in addition, they are now 3 4 Overview giving more attention to the problems of financing edu- school quality has substantially greater effects on cational quality. changes in students' cognitive and behavioral outcomes The Bank has been sponsoring research on the quality in developing than in advanced countries. He also finds of education in developing countries through three proj- that school quality seems to affect the lifetime earnings ects. One concerns the economics of educational radio, of all students, regardless of the level of development of one is about textbook availability and educational qual- their countries. Solmon concludes that in the long run, ity, and the last one deals with school quality and educa- the quality of schooling provided may be more impor- tional outcomes. The first two projects, both of which tant for the future economic prospects of developing were started in 1977, have attempted to measure the nations than expanded access to poor quality education. effect of textbooks and radio, respectively, on student It is left to Bruce Fuller, however, to test the proposi- learning, attendance, and internal efficiency of student tion. He chose Mexico as his example and has taken data progression. The third project is an attempt to measure from several periods ranging from 1880 to 1945. Using the effect of improving the full range of educational states with differing levels of investment in school qual- inputs: textbooks, teacher training, supervision and ity and controlling for other intervening influences, management, salaries, and classroom physical condi- Fuller demonstrates that states and communities that tions. Bank staff have also summarized findings from invested more heavily in the quality of education for the published literature and investigated trends across their children tended to experience higher rates of studies. Such reviews have been conducted on teacher growth or higher levels of economic development in training, class size, promotion and repetition policies, most sectors a decade later. and the contribution of educational attainment to eco- nomic growth and agricultural output. But two types of issues have thus far remained rel- Implementing Investments in Quality atively unexamined. One is the economic yield to be anticipated from an investment in educational quality Knowing that educational quality has an economic (as opposed to expansion), and the second is the degree effect does not determine whether investments in educa- to which such investments can be effectively imple- tional quality will be successfully implemented. This is mented. These two issues were the main topics of dis- the subject of the rest of this volume. C. E. Beeby argues cussion at a seminar, "The Quality of Education in that educational systems proceed through a series of Developing Countries," held at Harpers Ferry, West Vir- stages during which they are able to implement in- ginia, in May 1983. This seminar was sponsored by the creasingly more complex educational investments. His Research Committee through RPO 673-10. The papers conclusions and recommendations reflect almost sixty in this volume stem for the most part from that forum. years of experience in advising educational systems in The seminar was attended by thirty-five Bank staff mem- each stage of development. bers with expertise in the different countries and regions The last chapter summarizes the experiences of World involved in Bank lending, as well as staff from the policy Bank staff who attended the conference. Sessions were and research departments. Education experts from out- dedicated to very specific issues, such as teacher train- side the Bank also participated in the sessions. ing,- examinations, prevocational subjects, and elec- tronic media. It is fair to say that some educational programs fi- Educational Quality and Economic Growth nanced through World Bank loans and credits have been disappointing, that there have been errors in judgment, The time lag between school attendance and eco- and that there have been instances of unrealistically nomic activity, as well as variables such as home back- high expectations. But it is also fair to say that there have ground, general ability, and shifts in the economy, make been successes, with tangible and lasting improvements. it difficult to isolate the effect of education on economic This is true, for example, in the case of textbook produc- production. Accounting for quality factors within educa- tion, teaching by radio, and teacher training. Moreover tion makes the matter even more complex. Nevertheless both Bank research and experience suggest that coun- it is important to try to measure the effects of invest- tries that have placed a higher level of emphasis on ments in educational quality on economic production. educational quality have experienced a higher level of Lewis Solmon's paper reviews the relevant research economic growth and social prosperity. on this question. His review of the literature reveals that Part I 14 Educational Quality and Economic Growth The Quality of Education and Economic Growth: A Review of the Literature 4 Lewis C. Solmon This chapter reviews evidence on ways to develop tives, particularly in population control, health, nutri- Teducational policies that would stimulate economic tion, literacy, and communication. Colclough claims growth in developing countries. More specifically, the that these other goals also affect national development question is: to what extent can improved educational and growth: "The evidence suggests that the economic quality advance economic growth? and social returns to investment in primary schooling in Educational quality has many dimensions. In the very most developing countries are higher, at the present poorest countries qualitative changes may simply mean time, than other forms of educational investment" more resources. A certain minimum is required to get (p. 19). any result at all. As more resources become available, In some of the poorest countries, where real rates of the types of goods and services available in the school return on industrial and infrastructure projects are must be considered: the basic abilities, education, pro- often small or even negative, the returns to investment fessional training, and in-service programs for teachers; in primary schooling appear to be very high indeed and availability of textbooks; availability of other support more attractive.than-many alternatives. "In countries materials; ph ysical plant, and so on. Quality is also a where a large proportion of the working population is function of how students and others spend their time, dependent upon farming, and where rates of illiteracy both in and out of school. In addition, the importance of are very high, primary schooling thus provides an in- school quality depends on the availability of other learn- vestment opportunity which ought to have high priority ing modes at home, in the media, and in the community. on economic grounds" (p. 19). All of these factors will contribute to student learning, The literature overwhelmingly concludes that more career development, and earnings. They affect human years of schooling provide a plethora of private and capital, which in turn contributes to national economic social or national benefits. It -is- also both logical and growth. The research to date tempers the standard pol- empirically established that there is a tradeoff between icy of expanding primary schooling in developing coun- the quantity (years) and quality (expenditures per stu- tries without much regard for improving the quality of dent) of schooling. Thus most benefits derived from the education provided. low-quality schooling can be achieved faster if school To assess the role of educational quality in national quality is increased. To carry this argument further, if economic growth, one must ask whether schooling in quality is as low as it is in some developing countries, general affects growth. Colclough (1980) has provided improved quality could result in more benefits in fewer an excellent review of the evidence. "The case for invest- years. For example rather than needing six years to teach ment in primary schooling is overwhelmingly that it a labor force to read at what would be a fourth grade level makes people more productive at work and in the in the United States, increased availability of school home . . ." (p. 19). For example provision of a primary materials could result in the achievement of a fifth grade school education to a substantial percentage of the reading level in five years. In this case the nation would citizens facilitates the attainment of social policy objec- not have to give up the earnings of the students during 7 8 Educational Quality and Economic Growth their sixth year of schooling and would have new work- * The recent value of classroom materials and other ers entering the labor force a year earlier and with more nonsalary recurrent investment per student en- skills. Where educated labor is scarce, the earlier entry rolled in primary schools ranges from US$0.80 in of more accomplished workers into the labor force could Bolivia, to US$2.24 in Indonesia, to US$75 in Italy, contribute greatly to national income, to US$220 in the Netherlands and the United States, Moreover the dropout rate increases the longer a to more than US$300 in Sweden (Heyneman 1983). cohort of students remains in school. Thus if a certain * In the United States each elementary school pupil ,educational goal can be achieved more quickly, a larger has access to an average of fourteen library books in ,hare of the participating students will complete it. In addition to textbooks, reference books, and visual addition, a more amenable school atmosphere will en- aids. In the Philippines ten pupils must share each courage students to remain in school. primary school textbook. Thus typical pupils in the "The evidence shows that the benefits of primary United States have more than 140 times the amount schooling arise from the cognitive and non-cognitive of reading material at their disposal than their Fili- behavioral changes which the schooling experience pino counterparts (Heyneman 1983). brings," Colclough says (p. 20). In fact it appears that attitudes and behavior are now changing even in school To equate a similar number of years attained by stu- systems of very low quality. dents in different regions and schools in the United "The implication is that even if resources are scarce, States, to say nothing of comparing attainments of stu- and if the affordable quality of schooling is low, a further dents in dissimilar nations, is ridiculous. According to extension of the coverage of primary education can still unpublished estimates cited by Psacharopoulos (1983), be expected to bring benefits. In this sense the evidence in 1977 the average Western country invested fifty times from fertility studies, and from studies of farmer produc- more per pupil than did any of the thirty-six countries tivity, suggest that the individual behavioral changes with per capita income below US$256. that result from schooling are stronger when literacy is Even if educational outcome or school effectiveness widely spread than when it is more concentrated. There vary depending on resources available and on various seems to be then an interactive effect between individual school characteristics, there is a serious question about and community attitudes and values which significantly whether measures of expenditure per pupil adequately strengthens the economic and social case for univer- reflect such differences in schools. In general one would salizing access to primary schooling," Colclough says expect that where there was more money per student, (p. 20). each student would have access to "more" resources. It At the present level of expenditure per pupil in de- is more accurate to say, however, that more money veloping countries, it is quite apparent that if more years spent per student would on average enable students to of schooling have positive effects, higher-quality school- have more access to more expensive resources, not ing for any level of attainment will make an even greater necessarily that all students are getting a better educa- contribution. This can be deduced from the low absolute tion. levels of material resources that exist in most developing High average expenditures per student in a nation or countries, from the studies of the effects of school ex- in a school district may mean that a few students are penditures and school quality in such countries, and obtaining a large number of resources while many are from the findings of studies in the United States that all getting very little. Thus the variance in the distribution imply that increased resources have the greatest effect in of resources around a given mean must be considered. schools that have the fewest available resources. Even within a classroom, different students may have access to different amounts of resources, such as teacher Measuring Quality time. In addition, inputs to schooling that do not have It is clear that every year of attainment, even at the market prices may vary greatly. Highly educated parents sam nomina , level of education, doest rult in te may contribute services in schools located in rich neigh- same nmominalo levelinof eduCaton,idoes oth reultoing: t borhoods, whereas the parents in poor neighborhoods same amount oflearning.Considerthfolmay be able to contribute only much less skilled labor or * In 1980 in the United States current expenditures no labor at all if they must work at other jobs. Yet per pupil ranged from US$3,462 in New York State technically, expenditures per student may be equal for to US$1,574 in Arkansas (World Almanac 1983). the two schools. In addition, certain families, particu- - In 1980 expenditures per student in four-year col- larly those with educated parents, can provide more leges in the United States ranged from less than complements to school training at home. US$2,000 to more than US$10,000. Hanushek's study of a few elementary school systems ltze Q'uality ot EAucation and Economic Growth 9 concluded that "factors which are purchased by the A Review of the Literature school systems are not for the most part the characteris- tics of schools and teachers which are important in In low-income countries the power of educational determining achievement levels. The bulk of instruc- attainment and of school achievement to determine tional expenditures goes toward the purchase of three occupational success may be substantially higher than classes of inputs: class size, teacher experience, and the power of socioeconomic status or sex (Currie 1977; teacher graduate education.... The characteristics of Fry 1980; Heyneman 1980; Schiefelbein and others teachers which appear important in the estimated mod- 1983). Even when school and teacher quality are quan- els include teacher verbal ability . . . , recentness of tified and entered into regression models similar to the teacher educational experiences, and proportion of non- production-function paradigms of the 1960s, they ap- white teachers (which may be interpreted as a measure pear to be the predominant influence on student learn- of the quality of educational experiences of nonwhite ing around the world. The poorer the economic sit- teachers)" (Hanushek 1972, pp. 108-10). uation in a country, the more powerful this "school Consideration must be given to efficient combina- effect" appears to be. tions of resources. For example, if highly complex equip- " Children who attend primary school in countries with ment that no one can operate is purchased or if highly low per capita incomes learn substantially less after skilled and highly paid teachers do not have access to similar amounts of time in school than do pupils in textbooks, blackboards, or decent classrooms, both the high-income countries. The lower the income of the effectivensss of the expenditures and the quality of the country, however, the lower the correlation between education will be low. pupils' social status and their achievement. Conversely The evidence seems to show that there are substantial in low-income countries the effect of school and teacher diminishing returns to per-pupil expenditures. At some quality upon students' academic achievement in pri- point a certain percentage increase in expenditures will mary school is greater. From the data in numerous lead to a smaller percentage increase in quality (output). studies it is possible to conclude that the predominant Thus where expenditures are five times higher, quality influence on student learning is the quality of the will not necessarily be five times as high. schools and teachers to which they are exposed. It seems clear that in a world of less than perfectly Schiefelbein and othersl1983) studied the influence competitive markets (that is, where dollars spent may of school resources in Chile and concluded that invest- not reflect productivity perfectly), differences in expen- ments in school quality are likely to have a substantial ditures may not accurately reflect real differences in and positive effect upon students' success both in school quality. When policy decisions are made, however, it and in the labor market. If they are to have significant seems that certain increases in spending per pupil could effect on students from lower socioeconomic strata, be directed so as to increase quality. improvements in school quality must start early in the educational process and be accompanied by attention to out-of-school problems, such as health and nutrition. Quality Inputs For both sexes educational variables are more power- ful predictors of occupational attainment than is family In developing countries a wide variety of quality com- social status. For females years of schooling is more ponents have been shown to affect student achievement. important than educational achievement, and the latter Quality matters more the poorer the setting. Class size has a negligible effect on occupational destination. does not seem to be an important factor within the Among males, however, the reverse is true; educational relevant ranges in either developing or developed coun- achievement suppresses the relatively small effect of tries. years of schooling, but it operates negatively in relation Teacher quality does seem to matter in developing to level of first job. For both sexes, educational quality countries, however. This is contrary to the apparent variables are much more powerful predictors of occupa- situation in the United States. When teacher salaries are tional attainment than are either years of schooling set according to seniority or degree level, however, or educational achievement. This relation is stronger salaries are an inappropriate proxy for teacher quality; among males than among females. Changes in levels of yet usually this is the only measure available. achievement are explained primarily by differences in Thus the available data suggest that the poorer the the quality of the educational experience received rather country, the greater the effect of school and teacher than by differences in family background: this holds for quality on achievement. This reflects the fact that, as a students of both higher and lower socioeconomic status. commodity, education is both scarce and in high de- Loxley (1983) studied primary school quality in Egypt mand. and found that school quality affects achievement more 10 Educational Quality and Economic Growth than does the influence of the home. Apparently Egyp- Psacharopoulos (1983, p. 5) has summarized the tian primary schools do provide a learning environment findings on the effects of school quality in developing that is independent of home resources and that affects countries: "The dimension of educational quality is typi- pupil performance on tests of basic literacy skills. The cally missing in quantitative measures of education." incremental effects of school quality on the poor are Yet the few existing studies of educational quality in this greater than those found for children of advantaged respect have shown quality to be economically and backgrounds. socially productive. Regardless of whether educational The net effect of the school is greater than the effect of quality has been measured in terms of school buildings, the home: 14 percent of the variation in math achieve- laboratories, textbooks, teacher qualifications, nature of ment can be attributed to schooling and 6 percent can be the curriculum, class size, composition of the student accounted for by preschool influences entered into the body, or expenditure per pupil, evidence more recent regression equation first. For reading the percentages than the Coleman study of 1966 shows that such mea- are 13 and 12, respectively. Clearly the school predicts sures have an effect on enrollment (Birdsall 1982), on the largest portion of variation in achievement in both student achievement, and on adult earnings (Solmon math and reading performance in Egypt (Loxley 1%83). 1975; Wachtel 1975; Rizzuto and Wachtel 1980). Also, Although low-quality, poorly equipped schools may the lower the per capita gross national product (GNP) of a inhibit learning for any student, children of poor social country, the higher the effect that school quality appears background appear to perform appreciably better in to have on student learning and the lower the effect that high-quality schools than they do in the least productive factors of socioeconomic status appear to have on learn- schools. Students coming from illiterate households are ing (Heyneman and Loxley 1983). The effect of enroll- more sensitive to school quality than students from ment within at least one country, Brazil, was greater advantaged households. among poor households (Birdsall 1982). Such different In an extensive review of teacher training on student findings between rich and poor households or countries achievement in developing countries, Husen and others suggest decreasing returns to inputs designed to im- (1978) conclude that teacher characteristics positively prove school quality; that is, such inputs are most effec- affect student achievement. Contrary to arguments pre- tive when administered to a small existing base. sented in studies of schooling in the United States, the "Consideration of school quality might contribute to evidence in this study suggests that trained teachers do explaining the widening gap in economic performance make a difference. between developing and advanced countries or the Numerous other studies of students in developing alleged failure of some economies to grow in spite of the countries further confirm these general findings. Such rising educational attainment of their populations. studies include the work of Birdsall in Brazil (forthcom- Jamison and others (1981) report that whereas in 1960, ing) and that of Sabot (1983) on Tanzania and Kenya. the industrialized countries (members of the Organisa- The only negative findings regarding quality measures tion of Economic Co-operation and Development, OECD) and educational effects pertain to class size. According on average invested sixteen times more per pupil than to Haddad (1978, p. 12), did any of the thirty-six countries with per capita in- comes below US$265, by 1970 the difference had grown It cannot be concluded than an increase in class to 22:1 and by 1975 to 31:1. According to unpublished size will necessarily lead to a decrease in the level estimates this ratio stood at 50:1 in 1977" (Psachar- of academic achievements of pupils. Likewise, a opoulos 1983, p. 5). decrease in class size does not guarantee an im- provement in the social environment of leaming. What seems to be more important is what the School Learning and Job Performance teacher does with the opportunities provided by the size of the class ... . It may be more cost- Since the amount students learn (achievement) is one effective to increase class size moderately, and in- (although not the only) factor in determining job per- vest the annual savings in areas that prove to be formance and earnings, the fact that improvements in effective, such as textbooks and qualified and moti- quality increase achievement in developing countries is vated teachers. Farrell and Schiefelbein calculated very important. The links between quality and achieve- that a 15-percent increase in the average class size ment are stronger in these countries than in the more in Chile would produce an annual saving equiva- developed world. Students in developing countries are lent to 5 percent of the 1970 annual budget. more likely than, say, those in the United States, to which together with savings from other measures improve their chances for success in careers (and so can finance the components of a major educa- contribute to national economic development) when tional reform. they do better on achievement tests. ilze Quality ot laucanton ana tconomic urowt 11 In developing countries the amount learned at one search can be conducted. In addition, measures of both level of education more directly relates, through en- inputs and outputs have been somewhat standardized trance examinations, to one's probability of advancing to from survey to survey. the next level. In the United States many factors besides The most obvious problem of applying findings from test scores determine progression through the ranks of the United States to the developing countries is one of education. At the early stages, merely completing one relevance. Can policies inferred from observing different grade gets a child into the next. At higher levels, grades effects of expenditures of US$1,000 and US$3,000 per and personal attributes are considered along with test student be recommended to nations that spend US$2 per scores. In some developing countries, however, student student? In a simplistic sense the answer is clearly no. progress is based exclusively on test scores. But that choice is too trivial. The real question is what In many developing countries the civil service is a can we leam about principles, relations, and trends by much larger segment of the formal labor force than is looking at particular aspects of the U.S. studies? And of the case in more advanced countries. Since exams fre- course the transference from developed to developing quently are required to obtain civil service jobs, again countries must utilize the comparative educators' ability to perform on such examinations is important for 4knowledge of differences in institutions, customs, and a high probability of career progress. wealth, among other things, across nations. If the Finally, in many developing countries there are choice is to base policy for developing nations upon no avenues to learning other than schools. Whereas scanty data from those countries or upon much better youngsters in advanced countries can avail themselves research from nations less like them, it seems clear that of television, libraries, newspapers, neighbors, and edu- the latter source of evidence must at least be considered. cated family members, those in developing countries A second argument for the usefulness of U.S. research must learn in school or not acquire any human capital lies in the great diversity of the United States both over at all. time and across regions. This has enabled groups of For all these reasons one would expect school quality people to be compared according to how they resemble to more strongly affect leaming and, through attain- the people in developing countries. In particular, blacks ment, career success in the developing world. The stud- in southem U.S. states have typically had fewer educa- ies to date confirm this expectation. Unfortunately, the tional resources than have whites in northern and west- bulk of the evidence deals with leaming. Information on ern states. Therefore a comparison of the effects of the links between school quality and earnings is avail- different quantities and qualities of education for blacks able primarily in studies on the more advanced nations. and whites or northerners and southerners will produce insights into the effects in other nations that have rel- atively few educational resources. The experience of Decisions about Methodology U.S. blacks is quite useful in predicting how increased educational resources may affect students in developing The evidence on the relation between the quality of countries. education and economic growth is quite varied. Some is clearly more convincing than others. It usually is obtained from studies of the most advanced nations. Microstudies in the United States To assist policymaking for developing countries, an important methodological decision must be made at the In the early 1970s various economists began to ana- outset. "Comparative" educators must deal constantly lyze several data bases of individuals to determine the with the question of what evidence is relevant or helpful effects of quality education on eamings. In general, the for developing nations. For the problem at hand much of proxy for school quality was expenditure per student, the research focuses on the most developed countries usually in the state where the student was educated, but and particularly on the United States. A small literature at times data were available for individual school dis- is now emerging, primarily from World Bank research- tricts. A significant problem in most of the studies was ers, on the effects of educational quality in the devel- that resource data for the specific schools that students oping world. attended were unavailable. Nevertheless the results from The advantages of the work on the United States are the different studies were consistent and strong. the greater completeness (types of variables), compre- Johnson and Stafford (1973) looked at people in the hensiveness (nationally representative), and accuracy of 1965 University of Michigan Survey Research Center the data used. Thus, fuller models can be tested, and Panel who had attended school in the 1930s, 1940s, and studies using alternative data bases can replicate and 1950s. The sample was restricted to white males who confirm the findings. In some cases data have been had earnings in 1964; the dependent variable was aver- collected over a number of years, so longitudinal re- age hourly earnings. The study found that there were 12 Educational Quality and Economic Growth positive but diminishing returns to per-pupil expendi- Johnson and Stafford's findings of a tradeoff between tures on schooling; the estimated coefficient on the log quantity and quality of education. That is, a student of expenditure per pupil variables was 0.198, significant might benefit as much from fewer years of better school- at the 0.1 percent level. When students in the North ing as from more years of worse schooling. were considered separately, the coefficient fell to 0.159. For blacks they found that years of schooling had a This implies that returns to quality in the South, in negligible direct effect on earnings. In contrast in- which expenditures were lower, were higher than re- creased expenditure per student had a larger and statisti- turns in the North. cally significant direct effect on earnings; the estimated An extra US$100 of expenditure per pupil reduced the elasticity of earnings with respect to expenditure was amount of schooling necessary to reach a certain stock about 0.57, or more than three times the estimate for of human capital by about nine-tenths of a year. There whites. Since blacks start with much lower expenditures are, in general, much higher returns to quality of per student than do whites, their circumstances are schooling than to years of schooling. This indicates that closer to students in developing countries than to whites too few educational resources have been invest%d in in the United States. That is, these findings imply that school quality. The rate of return to quality, however, increased expenditures per student in developing coun- appears to diminish rather sharply as expenditures are tries would have huge payoffs in increased earnings and increased. At the highest levels of expenditure, returns productivity, as seems to be the case for blacks in the to quantity and quality are about equal. United States. Johnson and Stafford also conclude that the quality of Link and Rutledge also confirmed Welch's (1975) the public school attended should have a positive effect finding that improvements in the quality of education on the decision to stay in school longer. That is, quality have played a significant role in narrowing the differ- lowers the psychic costs of remaining in school longer, ences in earnings between blacks and whites. The calcu- which results in greater attainment and therefore higher lated negligible returns to quantity of education indicate income. that public programs that increase years of schooling for The findings of this paper and others are not neces- blacks without simultaneously increasing the quality of sarily inconsistent with those of Jencks (1972) and Cole- schooling will most likely fail. man and others (1966), who argue that the quality of Also in 1975 Ribich and Murphy used National Longi- schooling has little effect on measures of output, such as tudinal data on workers five years after high school tests of cognitive skills. Their findings are based on three graduation to determine how educational quality, abil- arguments: there are no precise links between test ity, and socioeconomic status affect educational attain- scores and earnings; even if there was a strong link, most ment, ability, and subsequently income. They concluded tests of relations between input and output rely on that just because expenditures do not affect test scores measures of statistical association, not causation; and does not mean they do not affect income. Better quality Jencks relies upon goodness of fit (R-squared), whereas education develops personal attributes that are useful in Johnson and Stafford's interpretations are based on the job market or conducive to greater interest in con- point estimates from the regressions. tinuing schooling. Despite small direct effects between Johnson and Stafford were not able to include mea- expenditures and income in the North, the indirect sures of individual ability or of parental background. effects through increased years of schooling are impor- Nevertheless subsequent studies, which did include tant. Moreover in the South expenditures positively such factors, appear to confirm their findings. In 1975 affect both test scores and educational attainment. Link and Rutledge used the 1968 National Longitudinal Finally high socioeconomic status has a role in directing Sample ("Parnes") data to corroborate the Johnson and children toward schools with a high level of expenditure Stafford study. Their dependent variable was the log of per student. annual, rather than hourly, earnings; they had district- Akin and Garfinkel (1977) used a later version (1968- by-district, rather than state data on expenditure per 72) of the Michigan "SRC Panel" of the longitudinal student; they included measures of ability and experi- data to look at black and white males between the ence in the labor market; and they included blacks in the ages of 30 and 55 in 1972. They concluded that high- analysis. Of course, differences in expenditures within quality schools provide socialization, increasing self- districts were still unknown. confidence, and academic and vocational skills, which Link and Rutledge concluded that the return to qual- are unlikely to be taught in schools with low levels of ity (expenditures) for whites was 0.176 and that returns expenditure per student. Expenditures had a small effect to quality in general were high but diminishing. The IQ on ability scores, and such scores had a similarly small measure reduces the coefficient for years by 15 percent effect on earnings. But expenditures have a greater effect and for quality by 10 percent, but Johnson and Stafford's on years completed and, directly, on income. Overall, conclusions do not change. In fact they further confirm rates of return on expenditures per student for whites The Quality of Education and Economic Crowth 13 range from 9 to 15 percent and for blacks from 13 to 20 where resources (quality) are more abundant. Second, percent. They also stress nonfinancial returns. Like increased school quality affects earnings in more ways other researchers Akin and Garfinkel found the rate of than by merely increasing achievement. Finally, the return to be uniformly higher for blacks than forwhites, translation of human capital into earnings depends on despite the fact that many of the school expenditure institutional arrangements, the relative scarcity of pro- coefficients for blacks are smaller than those for whites. ductive workers, and complementary factors in the pro- Two reasons for this are suggested. First blacks attend duction process. school for fewer years and so begin to accrue benefits earlier. Second the mean expenditure per black student is lower than for each white one, so any increment will The Conflict between Quality and Equity have a greater effect. Finally Rizzuto and Wachtel (1980) used data from Since the "Great Society" days of the mid-1960s there the 1960 and 1970 censuses of population and concluded have been serious efforts in the United States and other that rates of return to quality are higher for blacks (0.33 countries to expand educational opportunities at all to 0.35) than for whites (0.13 to 0.17) and are higher levels for the poor and for minorities. "Equal educa- than returns to quantity (0.07 to 0.08 for whites and tional opportunity" has been viewed by many to mean slightly less for blacks). They argued that there has been "equal access," however, and equal access was thought underinvestment in quality of schooling in the 1970s, to have been achieved if everyone who so desired could particularly for blacks. There may be benefits in allocat- get into some educational institution. Thus, access to ing a greater proportion of educational budgets to pro- education of high quality was of less concern. grams that improve the quality of elementary and sec- If indeed it matters not only how long one is in school ondary schooling and a smaller proportion for programs but also what quality of education one gets, then to that require or encourage higher attainment. ignore quality is a serious problem. At the elementary Rizzuto and Wachtel also tested other possible mea- and secondary level the problem is compounded if sures of school quality. They found that the ratio of policymakers are unable to define what quality is. Since pupils to teachers and measures of teacher salaries were universal access has generally been achieved at the pre- less important than were expenditures per student or college level in the United States, an equal distribution longer school years. of quality was, at least implicitly, sought. But since no Some might argue that these strong and consistent one knew how to identify quality, it was assumed that if findings for the United States are not directly relevant to resources were equally distributed-perhaps by pro- developing countries. Evidence on diminishing returns hibiting wealthy districts from spending more than to expenditures per student, however, along with the others for education or by distributing the students greater observed returns to blacks than to whites (blacks according to various busing plans-at least everyone have traditionally been in lower quality schools) suggest would get the same quality education, even if not a high that in developing countries with very low expenditures quality education. The resources used to achieve equal per student and commensurately few available real re- opportunity-programs like compensatory education, sources, the payoff to quality (increased expenditures) busing, and greater administrative overhead-have would be immense. If the experience in the United States alternative uses. can serve as a guide, this would be the case particularly if If these incremental costs could be paid for by addi- additional resources were used to achieve things other tional funds, then there would be no reason to sacrifice than increase teacher salaries or decrease class sizes. existing programs. But if resources are diverted from A review of the research has shown that: traditional uses, the move toward equity may lower the * School quality has a much greater effect on student quality of education, particularly in the public schools. test scores in developing countries than in the Resources for public schools have diminished under the United States. double burden of tax limitation initiatives and inflation. * School quality appears to have a substantial effect And in many areas "white flight" has lowered the effect on subsequent earnings of students in the United of peer groups on children with low socioeconomic States; however there is little evidence either for or status. Loss of resources, changes in student composi- against this relation in developing countriesr tion, and crime, among other factors, have led many good teachers to leave the public schools as well. These two results can be reconciled in several ways. On balance the situation for students who formerly First, U.S. evidence on diminishing returns to school were receiving the lowest quality elementary and sec- quality would imply that where the level of quality is very ondary education may have improved. It remains to be low, as in developing countries, the effect on either shown how the quality of education has changed for achievement or earnings will be greater than it would be others, however. If the gains in productive capacity by 14 Educational Quality and Economic Growth those whose quality of education has improved are tage in the labor force of college-educated compared greater (or worth more) than the losses in productivity with noncollege-educated persons has declined (Free- of those whose educational quality has declined, then man 1976). It has not yet been determined how much of society has become better off. But we will not know the this is the result of an increase in supply of, relative to net effects of the tradeoff between quality and equity demand for, college gxa4uates and how much is a result until such determinations are made. of the reduced productivity of graduates from poorer- At the postsecondary level in the United States the quality colleges. But it has been widely stated that qual- problems are similar, but there are also additional ones. ity will once again become the primary goal of educators The first stage in seeking equal access was to assure that in the United States in the 1980s. Such a refocusing may everyone desiring to enter a postsecondary school could result in part from the realization that tradeoffs between do so. Beginning in 1972 the federal government estab- quality and equity may have to be made. lished massive student aid programs to help the dis- In the developing world the problem of the tradeoff advantaged. Even though most very poor students can between equity and quality may be even more severe. In receive sufficient financial aid to cover the direct costs of many instances attendance at even the lowest levels of expensive colleges, the costs covered exclude fotgone schooling has been reserved for those with high earnings. And many poor students still do not attend the socioeconomic status. Thus any attempt to slow ex- most expensive private colleges and universities, some of panded access to improve quality of the education may which are generally viewed as providing the highest be viewed as an attempt to divert even more resources to quality education. In the past several decades two-year the rich. colleges have proliferated and expanded, and these are Politically it will be difficult for many governments to the schools to which most minority and poor students achieve higher-quality education at the expense of ex- have had access. panded access. This might be an important reason for an Moreover some formerly high-quality colleges, such outside agency such as the World Bank to promote as those in the system of the City University of New York, quality. This would enable the domestic government to experimented with open admissions programs, which continue to use internal resources to widen access. But were costly and disruptive and which diverted some of if the quality of education is very low, merely expanding the financial and human resources from their former access will serve the national interest poorly. purpose of educating the better prepared students. Even Improvements in the quality of schooling provided in universities that stopped short of open admissions but developing countries may be more important for the that tried to increase the diversity of their student bodies future economic prospects of these nations in the long became involved in remedial programs; special courses, run than will expanded access to poor quality education. such as women's and minority studies; and new counsel- Thus the quandary of equity versus efficiency will have ing activities. Even though many of these activities were to be addressed in these countries as it has been in the funded from new money rather than from ordinary United States and other developed nations. funds, certainly some funds were reallocated, and so Ultimately, additional questions not related to educa- fewer resources could be devoted to traditional efforts of tion will also have to be addressed to facilitate develop- educating only the best prepared students. ment. These may range from issues of debt service and Thus at the college level many new students have been trade balances to agricultural policy and procedures for educated. But the real question is whether, as some utilizing natural resources. people have argued, the average quality of education All of these problems will require money to solve. But being received by all students has declined. Certainly unless sufficient resources are committed to improving some students are receiving the high-quality education each nation's stock of human capital, and until these that had previously been available. But it is unclear resources are effectively allocated within the educational whether the increase in number of college students has systems, the developing countries will be at a serious compensated for the decline in the quality of education disadvantage in the continuing quest for greater wealth received in terms of the productive capacity of the col- and prosperity. lege-educated labor force. On the one hand there may be a higher economic retum, both to the individuals edu- cated and to society, to providing high-quality education References and Bibliography to the most able than to providing lower-quality educa- tion to more students. On the other hand some have Akin, J. S., and 1. Carfinkel. 1977. "School Expenditures and argued that the overall social benefits would be greater if the Economic Returns to Schooling." Journal of Human even more resources were diverted from the more-or- Resources, vol. 12, no. 4, pp. 460-81. less able (Astin 1973). This is an empirical question that Astin, A. W. 1968. "Undergraduate Achievement and Institu- must await further analysis rather than more rhetoric. tional 'Excellence."' Science, vol. 161, pp. 661-68. During the past decade the relative economic advan- . 1973. "Measurement and Determinants of the Out- Il t: UULLLy ul £UULU1U11 Unu rcOnomlC Urowrn 1i puts of Higher Education." In Does College Matter? edited . 1983. "The Distribution of Primary School Quality by Lewis C. Solman and Paul T. Taubman. New York: within High- and Low-Income Countries." Comparative Academic Press. Education Review, vol. 27 (February), pp. 108-18. .1977.FourCriticalYears. San Francisco: Jossey-Bass. Heyneman, Stephen P., and others. 1984. "Textbooks in the Becker, G. S. 1964.Human Capital: A TheoreticalandEmpir- Philippines: Evaluation of the Pedagogical Impact of a ical Analysis, with Special Reference to Education. New Nationwide Investment." Educational Evaluation and Pol- York: National Bureau of Economic Research. icy Analysis, vol. 6, no. 2, pp. 139-50. Behrman, Jere, and Nancy Birdsall. 1983. "The Quality of Husen, Torsten, and others. 1978. Teacher Training and Stu- Schooling in Developing Countries: Quantity Alone Is Mis- dentAchievement in Less Developed Countries. World Bank leading." American Economic Review, vol. 73, no. 5 (De- Staff Working Paper no. 310. Washington, D.C. cember), pp. 928-46. Jamison, Dean T., and others. 1981. "Improving Elementary Birdsall, Nancy. Forthcoming. "Child Schooling and Public Mathematics Education in Nicaragua: An Experimental Inputs in Brazil." Journal of Development Economics. Study of the Impact of Textbooks and Radio on Achieve- Bowen, H. R. 1980. The Costs of Higher Education. San Fran- ment." Journal of Educational Psychology, vol. 73, no. 4, cisco: Jossey-Bass Publishers. pp. 556-67. Iencks, Christopher S., and M. D. Brown. 1975. "Effects of Chiswick, B. R. 1973. "Schooling, Screening, and Income." In High Schools on Their Students." Harvard Educational Does College Matter? edited by Lewis C. Solmon and Paul T. Review, vol. 45, no. 3, pp. 273-324. Taubman. New York: Academic Press. Jencks, Christopher S. and others. 1972. Inequality: A Reas- Colclough, Christopher. 1980. Primary Schooling and Eco- sessment of the Effect of Family and Schooling in America. nomic Development: A Review of the Evidence. World Bank New York: Basic Books. Staff Working Paper no. 399. Washington, D.C. Johnson, C., and F. P. Stafford. 1973. "Social Returns to Coleman, J. S., and others. 1966. Equality of Educational Quality and Quantity of Schooling." Journal of Human Opportunity. Washington, D.C.: U.S. Government Printing Resources, vol. 8, no. 2 (Spring), pp. 139-55. Office. Layard, R., and George Psacharopoulos. 1974. "The Screening Currie, J. 1977. "Family Background, Academic Achievement, Hypothesis and the Returns to Education." Journal of Po- and Occupational Status in Uganda." Comparative Educa- litical Economy, vol. 82, no. 5, pp. 985-98. tion Review, vol. 21, pp. 14-28. Link, C. R., and E. C. Rutledge. 1975. "Social Returns to Denison, E. F. 1962. The Sources of Economic Growth in the Quantity and Education: A Further Statement." Journal of United States. New York: Committee for Economic De- Human Resources, vol. 10, no. 1, pp. 78-89. velopment. vreemamen,. B196TeOeduaemrcneYok Loxley, William A. 1983. "The Impact of Primarv School Qual- Freeman, R. B. 1976. The Over-EducatedAmerican. New York: . Academic Press. ity on Learning in Egypt." International Journal of Educa- tion Development. vol. 3 (March). Fry, G. W. 1980. "Education and Success: A Case Study of Thai Psacharopoulos, George. 1972. "Rates of Return to Investment Public Service." Comparative Education Review, vol. 24 in Education around the World." Comparative Education (February), pp. 21-34. Review, vol. 16, no. 1, pp. 54-67. Haddad, Wadi D. 1978. Educational Effects of Class Size. - 1982. "The Contribution of Education to Economic World Bank Staff Working Paper no. 280. Washington, D.C. Growth: International Comparisons." Paper presented at Hanushek, Eric A. 1972. Education and Race-An Analysis of the conference on International Comparisons of Productiv- the Educational Production Process. Lexington, Mass.: ity and Sources of the Slowdown. American Enterprise Insti- Heath Lexington. tute for Public Policy Research, Washington, D.C., Septem- - 1979. "Conceptual and Empirical Issues in the ber 30 and October 1, 1982. Processed. Estimation of Educational Production Functions." Journal . 1983. "Educational Research at the World Bank." of Human Resources, vol. 14, no. 3, pp. 351-85. Research News, vol. 4, no. 1, pp. 3-33. Heyneman, Stephen P. 1980. The Evaluation of Human Capi- Rizzuto, R., and P. Wachtel. 1980. "Further Evidence on the tal in Malawi. World Bank Staff Working Paper no. 420. Returns to School Quality." Journal of Human Resources, Washington, D.C. vol. 15, no. 2, pp. 240-54. - 1983. "Improving the Quality of Education in De- veloping Countries." Finance and Development (March), Ribich, T. I., and J. L. Murphy. 1975. "The Economic Returns pp. 18-20. to Increased Educational Spending." Journal of Human . 1984. "Research on Education in the Developing Resources, vol. 10, no. 1, pp. 56-77. Countries." International Joumal of Educational Develop- Schiefelbein, E., and others. 1983. "The Influence of School ment, vol. 4, pp. 293-304. Resources in Chile: Their Effect on Educational Achieve- Heyneman, Stephen P., and William A. Loxley. 1983. "The ment and Occupational Attainment." World Bank Staff Effect of Primary-School Quality on Academic Achievement Working Paper no. 530. Washington. D.C. across Twenty-Nine Low-Income Countries." American Schultz, Theodore W. 1961. "Investment in Human Capital." Joumal of Sociology (May), vol. 88, pp. 19-23. American Economic Review. vol. 51 (March). pp. 1-17. 16 Educational Quality and Economic Growth Schultz, Theodore W. 1961. "Education and Economic the Effects of College. Current Issues in Higher Education Growth." In SocialForces InfluencingAmerican Education. vol. 8. Washington, D.C.: American Association for Higher Chicago: National Society for the Study of Education. Education. Solmon, Lewis C. 1973. "The Definition and Impact of College Trow, M. 1975. "The Public and Private Lives of Higher Educa- Quality." In Does College Matter? Some Evidence on the tion." Daedalus, vol. 2 (Winter), pp. 113-27. Impacts of Higher Education, edited by Lewis C. Solmon Wachtel, P. 1975. "The Effect of School Quality on Achieve- and Paul T. Taubman. New York: Academic Press. ment, Attainment Levels, and Lifetime Earnings." Explora- . 1975. "Capital Formation by Expenditures on Formal tions in Economics Research, vol. 2, no. 4, pp. 502-36. Education, 1880 and 1890." In Dissertations in American Welch, F. 1978. "Black-White Differences in Returns to Economic History. New York: Arno Press. Schooling." American Economic Review, vol. 68, no. 5, . 1981. "New Findings on the Links between College pp. 893-907. Education and Work." Higher Education (November), The WorldAlmanac and Book ofFacts 1983. 1983. New York: vol. 10, pp. 61548. Newspaper Enterprise Association, Inc. Solmon, Lewis C., and N. L. Ochsner. 1978. New Findinqs on -~~~~~~~ School Quality and Economic Growth in Mexico Bruce Fuller, Kathleen Gorman, and John Edwards 14 F aith in the belief that education helps drive eco- schools may also be more potent in boosting the indi- nomic growth has dwindled considerably during the vidual's success in the labor market within the develop- past decade. Research has seriously challenged the ing world compared with more industrialized nations. In assumption that higher levels of school investment and Chile, for example, school quality-including textbook quality will boost the individual's academic achievement availability and class size-were strong determinants of or eventual economic success in the labor market (Cole- occupational success (Schiefelbein and Farrell 1984). man and others 1966; Jencks and others 1979). Prior This chapter asks whether investment in schools con- estimates of the contribution of educational attainment tributes to economic output at the national level. The to national economic growth (Denison 1967; Schultz economic effects of schooling quantity and quality are 1971) now also appear to have been overly optimistic usually examined by following individuals from school (Barnhouse-Walters and Rubinson 1983). into the labor market. But the structure of the econ- Yet findings from research on achievement and eco- omy-that through which the individual moves-is nomic effects of school investments within industrial assumed to be constant (Bowman 1976). Work has just nations should not be generalized to developing coun- begun to measure directly the influence of investment in tries, nor even applied to all historical periods within any schools on the productivity and growth of a nation country. Recent work, for instance, reveals that in- and to move beyond growth-accounting methods or cremental improvements in school quality have yielded inferring social returns of education from aggregated sizable gains in student achievement in several develop- individual-level experience (for review, Rubinson and ing nations (for review, see Psacharopoulos 1983). Ralph 1984). This chapter directly assesses the effect of The relative influence of school attainment and quality school investment and quality on the national output of in relation to family background appears to be much a developing country. stronger in developing countries than within industrial The specific topic of this chapter is the contribution of nations (Heyneman and Loxley 1983). Investments in education to early agricultural and industrial growth in Mexico between 1880 and 1945. Mexico offers an in- Note: This work was largely supported by the World Bank. We triguing case. Before the revolution of 1910-17 Mexico especially want to thank Stephen Heyneman, who has provided moral strongly encouraged urban-centered industrial develop- and financial support. Early comments by staff at the Bank aided our ment spurred by infusions of foreign capital. After the analysis. Conversation and correspondence with Susan Holloway, Wil- liam Johnson. Aaron Benavot. Pamela Barnhouse-Walters, John revolution, however, Mexico shifted toward a rural- Meyer, and Jerald Hage have helped to refine our ideas and spark new focused strategy of agricultural development and dis- areas of inquiry. The Institute for the Study of Exceptional Children couraged investment in mining and manufacturing and Youth at the University of Maryland and the Education Depart- (Reynolds 1970a). Patterns of investment in schools also ment of the Baltimore Countv campus at Marvland assisted with computing and clerical support. Initial findings were reported at the changed dramatically between these two periods as a meeting of the American Educational Research Association, New result of the sharp switch in social priorities. Schooling Orleans 1984. moved from locally controlled, largely urban institu- 17 18 Educational Quality and Economic Growth tions to a federally dominated movement, which was integrate individuals and communities into a more designed to help organize rural communities around tightly organized economy and social structure. School- shared economic and social goals. Mexican history thus ing establishes universal knowledge (reading and writ- provides an opportunity to examine the effect of school ing ability), a common language, and a more widely investments on growth under two distinct economic based political order (Bendix 1964; Althusser 1971). structures and contrasting education policies. Research on individuals may actually underestimate the economic effects of school investment by missing the institutional and regional events, such as the accumula- Investment in Schools and National Growth tion of knowledge. In addition regional or national deci- sions about educational investment may shape growth This chapter builds on, yet departs somewhat from, differently in various economic sectors. One question in the way economic effects of school investment have been this chapter is whether post-revolutionary Mexico was studied within the framework of human capital. Current able to boost agricultural production through invest- debate over school investments and whether they boost I ments in schooling. This is one example of how school academic achievement or economic benefits is txposing investments may be an exogenous force, which comple- limitations both in how the issue has been conceptual- ments capital investment strategies that focus on ized and in the research methods that are often used. growth in a particular economic sector (at the commun- This look at the contribution of school investment and ity level, see Jamison and Lau 1982). The central point is quality to economic growth in Mexico responds to sev- that school investments at national or regional levels eral points of criticism directed at past literature. may better demonstrate the economic benefits and costs Much of the research on economic returns to educa- than the level of the individual. tional investment has utilized data on individuals and The growth accounting method of estimating the typically examined how youth with variable levels (and contribution of education has also been used to examine quality) of schooling achieve in the labor market (for the economic effects of education at the national level example, Hansen 1970; Psacharopoulos 1973; Sewell (Denison 1967, 1974). Simply assuming that the eco- and Hauser 1975; Bowman and others 1981). The nomic growth not explained by capital, land, and labor assumption about human capital in this work is that inputs was shaped by education and technological in- schooling imparts skills and attitudes that lead to novation, however, may have overestimated the effect greater productivity, which is measured in wage gains of investments in schooling at the national level. and aggregated social returns. However, the theory of Barnhouse-Walters and Rubinson (1983) directly mea- class conflict suggests that schooling may affect wages sured and modeled the effect of educational attainment because some youths hold higher credentials, not neces- and found that primary schooling significantly affected sarily because they gain higher technical proficiencies national output. The influence was considerably less from their education. Level of schooling could consist- than the growth accounting methods had estimated, and ently covary with the relative positions of individuals in the effect appeared only for the early industrial period in the labor and wage structure; yet aggregate productivity the United States.' The production-function method was and national growth are not necessarily affected (Berg used to model longitudinal effects of school investment 1970; Thurow 1974; Collins 1979). The traditional by assessing the economic effect of alternative measures framework for human capital also assumes constant of investment in school quantity and quality after con- patterns of labor demand: the economic structure pre- trolling the effects of capital, labor, and cultivated land sumably discriminates between individuals (holding (Hanushek 1979).- - variable skills and educational attainment constant) in similar wavs over time (Bowman 1976). At times, how- Dimensions of Quantity and Quality ever, the labor market favors subgroups with low or high in School Investment status. One example of this is the higher rates of return from education for blacks in the United States during Early research defined school investment, for both the 1970s (Hanushek 1979; Rizzuto and Wachtel 1980). individuals and nations, as the quantity of schooling The process of school investment is viewed as largely attained for an individual or a nation's labor force. But occurring at an institutional, as well as at an individual school investments can have several qualitative compo- or family, level. Particularly in early stages of national nents as well. School quality has been measured in economic development investment in school quantity terms of expenditure per student, character of school and quality is part of institution-building processes de- facilities, strength of academic programs, class size, signed to boost educational levels, technological inno- ksocial composition of students enrolled, and teachers' vation, regional trade, and a national consciousness i salary levels or qualifications (for review, see Solmon, (Meyer 1977). Government investment in schools at chapter 2; Welch 1976). These quality ingredients of local, state. and national levels expresses a desire to schooling have been found to influence both individual oLtivuwuualy uru rconomric urowrn in Mexico 19 academic achievement and success in the labor market. in which textbook availability, teaching quality, and Expenditure per pupil, the ratio of students to teachers class size contributed to labor market success more within the school, and teacher salaries appear to in- efficaciously than length of schooling (quantity) or fam- fluence school and economic experience significantly, ily background (Schiefelbein and Farrell 1984). Jamison when family background within the United States is and Lau (1982) also found that persistence through four controlled for (Wachtel 1975; Akin and Garfinkel 1977; years of elementary schooling raised the farm productiv- Rizzuto and Wachtel 1980). Initial evidence suggests ity of individuals by 9 percent. In that study the depen- that school quality (in terms of expenditure per pupil) dent economic effect was measured in concrete output also shapes rates of growth in pupil persistence through (rather than by using wage rates as a proxy for presumed school and state-level income, when earlier efforts at productivity gains), and the effects on growth pertain to school expansion are controlled for (Solmon 1975; Ful- one particular economic sector. This study on Mexico ler 1983). Although a tradeoff appears to exist between extends this line of work and relates levels of school the quantity of investment and school quality, little is investment and qualitative aspects of public investments known about which qualitative factors and what mix to economic output over time. yield the greatest academic achievement and economic 4 effects. The distinction between school expansion and school quality is particularly important in assessing the effects Historcal Specificity of the Effects of educational investment in the developing world. For instance in the thirty-six poorest nations, with per capita School investments may produce particular effects GNP of US$265 or less, average enrollment rose from 48 only within certain stages of development or partic- to 70 percent of school-age children between 1960 and ular economic structures. Research on human capital 1977 (Heyneman and others 1984). Yet within this re- at the individual level has not examined the variation markable spurt of institutional growth and access to in the structure of occupations or the dynamics of labor education lie vast variations in school quality. For demand patterns as change occurs during historical teacher salaries, for example, Malawi spends only one- periods. Analyzing worldwide data on educational at- sixth of the resources allocated by Bolivia (Heyneman tainment across nations, Benavot (1982) found that dur- 1982). Contrasts are even sharper between industrial ing 1950-70 investment in primary schooling boosted and developing nations. In 1975 members of the OECD economic development for poorer countries but not spent thirty-five times more resources per student than for more industrialized nations. Similarly Barnhouse- the world's poorest countries (World Bank 1980). In WaltersandRubinson(1983)foundthatprimaryschool- 1977 ten students are enrolled for every one textbook in ing significantly helped predict early industrial growth the Philippines (Heyneman and others 1980). in the United States (1890-1929), although secondary Variation in school quality also touches levels of schooling nudged development after 1933. Technologi- academic achievement. In twenty-nine low-income na- cal innovation and assumed shifts in job skills may also tions Heyneman and Loxley (1983) found that school have boosted the growth of primary schools in the for- quality shapes sizable portions of variance in students' mer period, then reinforced secondary school expansion achievement. In the poorest nations, such as India and in the late industrial period (Rubinson and Ralph 1984). Colombia, school quality played a stronger role than This work does not emphasize how differing dimensions family background. Incremental improvement in the of the quantity and quality of investment may variably availability of textbooks, instructional materials, and influence economic growth under specific historical even radio instruction have revealed similar gains in conditions; yet a focus on the efficacy of different levels achievement among students in Nicaragua (Jamison and of instruction and types of curricula (for example, pri- others 1981), Malawi (Heyneman 1980), and Malaysia mary versus secondary or academic versus vocational) (Haron 1978). Effects are at times quite significant. A offers a strong beginning point for future research. control-group evaluation of increasing textbook availa- bility in the Philippines, for instance, found an increase in one-third of a standard deviation in first and second grade math achievement across 8 million students. This means that the level achieved by 50 percent of all stu- Past approaches to the study of the effects of invest- dents in the first year was attained by 63 percent of all ment in human capital have been criticized for modeling pupils the next year after additional textbooks were in- the process in the wrong direction. Rather than arguing troduced (Heyneman and others 1984). that school investments boost economic growth, the However little evidence exists that links school quality class-conflict perspective proposes that the structures of to economic gains within the developing world. The the economy and of the state determine the growth and previously cited study in Chile is an important exception character of school institutions (Bowles and Gintis 1976; 20 Educational Quality and Economic Growth Collins 1979; Hogan 1982). Little doubt exists that labor administration of the schools. Seconc is the question of market demand in part shapes school enrollment rates how the content of schooling matched the economic and and school investments by governments. Supportive social priorities of the government. empirical evidence exists both for industrializing set- Education in colonial Mexico was largely controlled tings (Katz and others 1982; Fuller 1983) and within by the church Secular elements of the government developing countries, such as Mexico (Goldblatt 1972). successfully advocated public support of schools only This chapter, however, focuses on a model that views after Mexico achieved independence from Spain in 1821. school investments and quality as antecedents to eco- Liberal reforms of the independence era and during La nomic growth. The rise of school institutions preceded Reforma in the 1850s would eventually constrain the commercial expansion and industrialization in many church's ability even to operate primary schools. Yet an settings, including nineteenth century France (Baker increasingly secular federal government would not rep- and Harrigan 1980), the United States (Fishlow 1966; licate the church's level of centralized administration Meyer and others 1979), and Mexico (Diaz Covarrubias until after the revolution of 1910-17. Throughout the 1875; Wilson 1941). To some extent social values and latter half of the nineteenth century primary education commitment to school institution building, which are occurred within private schools or small institutions often rooted in religious convictions and faith in literacy supported by municipal governments (Larroyo 1948; training, operate somewhat independently of economic Wilson 1941). In 1875, 87 percent of Mexico's 8,103 factors (Lipset 1972; Tyack and Hansot 1982). In ad- primary schools were operated by municipal authorities, dition schooling investments in time 1 historically 603 were supported by state governments, and the re- precede growth in certain emergent industrial or com- mainder by churches or private associations (Hamilton mercial sectors in time 2. This reality requires that 1884; Diaz Covarrubias 1875). economic structure be considered a dependent variable. The federal government did urge that public.free This chapter does not assume that school investment schools be established, both during the reformist ad- patterns are not affected by previous social class struc- ministration of Benito Juarez and in the early years of tures, even those operating within early agrarian and the Porfirio Diaz regime. The organic laws of public local commercial elements of the economy (Soltow and instruction (1867) required municipalities to establish a Stevens 1981). School investments may also allocate primary school for every 500 inhabitants and urged economic gains within unequal patterns (Carnoy 1972; hacienda (estate) owners to create schools for the chil- Levin 1984), as happened in prerevolutionary Mexico. dren of farmworkers and miners (Universidad Nacional Yet the focus here is on the extent to which school Autonoma 1967; Vaughan 1982). In 18~Ihe Diaz ad- investments and quality may spur growth in new or ministration mandated school attendaffc f6r all school- young economic sectors of a developing nation. age children: Ironically this mandate was exercised In summary this chapter looks at how measures of when not more than one-fifth of all children entered school quality among Mexico's thirty-two states have primary school and only 15 percent of all Mexicans were affected economic output from 1880 to 1945. This literate. The federal government did not begin to subsi- approach allows an assessment of the role of educational dize schools run by the states and municipalities until investment and quality on direct measures of economic 1888. output. Before the methodology and findings are dis- Local control of education greatly influencedAhe-in- cussed, however, the policies on school investment and vestment in, and quality of, the schools. In the 1870s patterns of economic growth in Mexico are outlined. annual expenditure per student enrolled in public This sets the theoretical issues within Mexico's own schools equaled only 5.4 pesos, equivalent to two weeks' historical context. wages of a farm laborer. But in private schools 24 pesos were expended per child (Diaz Covarrubias 1875). Urban families and children benefited most from private Historical Background schools. Within Mexico City 40 percent of all primary schools were private. Municipalities that prospered from The theoretical issue of how school expansion and coastal trade or mining were better able to raise public school quality influence economic growth must be situ- revenues for schools. Urban philanthropists and private ated within Mexico's own historical experience and societies also helped to support schools (Banda 1873).2 cultural goals. In addition economic firms, particularly in mining, cre- ated their own schools to ease labor shortages (Maillefert The Aims and Quality of Mexico's Schools 1865; Wilson 1941). Initially municipal and private primary schools simply Two elements of Mexico's educational history are borrowed curricula and materials from the Catholic especially pertinent to the discussion of school quality. institutions, which emphasized classical training in First is the issue of who controlled the financing and morality, religious doctrine, and the humanities. After the mid-1800s, however, the secular state would recur- clothes and meals and by threatening parents with fines rently encourage more Western educational philosophy. or even jail for failing to send their children to school Reformers argued that schooling should not teach just (Callcott 1965). conformity through strict instruction and recitation. Beyond low levels of educational investment per cap- Instead education should nurture the individual de- ita and low enrollment rates, concrete signs of variable velopment necessary for political independence and school quality were apparent. Teachers supported by economic success within a postcolonial, free market municipal governments earned about 25 pesos a month economic structure (Sierra 1969; Knowlton 1976). Cur- in the 1870s, roughly the same wage of a carpenter and ricula and policy emanating from the federal govern- less than that of many mine workers (Wells 1887; ment-along with financial subsidies-increasingly Vaughan 1982). Rural school teachers earned as little as argued that schools should impart job skills and social one-fourth the wages paid in commercial centers. Dur- norms that were closely linked to modern sector jobs ing this period only five teacher training schools pro- and urban lives. Instruction in urban life, the French vided formal teacher training. Quality standards set by language, and moral behavior were integrated with older state or local governments were symbolic at best. In curricula in mathematics, science, and history (Sierra Chihuahua in 1909 all schools were to have labs, a sink, a 1910; Benjamin and Ocasio-Melendez 1984).3 This cou- bathroom, and a playground; only two schools met these pling of public primary schools and the modern urban standards (Sierra 1910). Teachers were at times evalu- economy occurred even though two-thirds of all workers ated on the basis of their students' performance on in the formal labor market in 1900 were engaged in exams, with an 80 percent passage rate deemed satisfac- agriculture, not urban commerce or manufacturing. tory by school inspectors. Yet as late as 1927 more than The growth in school investments before the revolu- 40 percent of all first grade students did not pass the tion was remarkable. The number of primary schools in national exam.' Mexico rose from 5,240 in 1843 to 10,592 in 1902 (Ortiz The quality of rural schools-where they managed to 1930) (Table 3-1). Yet by 1907 less than one-third of all operate-was very low. In 1910 the state of Zacatecas children were enrolled in primary school, and a third of closed many of its 251 rural schools because of inade- those enrolled attended school infrequently. In poor quate buildings and an inability to find teachers states such as Chiapas and Oaxaca less than 20 percent of (Vaughan 1982). Double standards of quality were also all children ever enrolled in a primary school (Vaughan applied to rural schools. In the early 1900s the federal 1982). Twenty-five years after compulsory attendance government approved a national school calendar but laws were passed, states and municipalities were still allowed rural areas to ignore it during planting and inducing higher attendance by giving students free harvest seasons. Although urban schools were inspected Table 3-1. Growth of Schools in Mexico, 1843-1940 Item 1843' 1875 1888-93b 1925-28 1936-40d Number of primary schools 5,240 8,103 9,039 16,692 22,205 Municipal 9,751 4,755 8,831 (state) State or federal 4,284 3,943 13,374 (federal) Private or church 2,017 2,998 School expenditures (millions of pesos) 2.8 3.3 45.0 106.3 (1907 = 11.3) Municipal 1.0 3.6 2.7 State or federal 0.4 41.4 103.6 Private 1.2 Annual school expenditures per student (pesos) Public 5.4 9.6 32.1 53.0 Private 24.0 Literacy rate (percent) 14-18 33 42 Note: All expenditure data are in current dollars. Between 1877 and 1900 the peso's value declined by 86 percent against the U.S. dollar (Sollano Ramos 1961). From 1900 to 1920 absolute inflation in Mexico (not pegged to the U.S. dollar) rose by 21.6 percent. Between 1920 and 1940 the peso's value actually gained 2.5 percent (Wilkie 1970). Blanks indicate unavailable data. Sources: a. Data for 1843 and 1875 are from Dias Covarrubias (1875) and Ortiz (1939). b. Callcott (1931). c. Vaughan (1982) and Stanley (1948). d. Stanley (1948). 22 Educational Quality and Economic Growth for the quality of instruction in math, science, and movement was in part a manifestation of the govern- modern languages, rural schools were expected to focus ment's desire to encourage a national consciousness on manual skills, "intuitive knowledge," and the Span- to emphasize a Mexican identity, was independent of ish language (Sierra 1910). foreign business, and assimilated various regions and Indian groups into a cohesive social framework (Vasconcelos 1923; Ramos 1941). The secular state com- municated themes of cooperation, solidarity, and na- Successive federal administrations redirected school tional self-reliance through the rural schools (Ebaugh, investments substantially following the revolution of 1931).6. 1910-17. Implications for school quality can be ex- Despite these gains in building school institutions amined through the same two elements summarized for quality was mixed. In 1927, for instance, forty-three Ithe eras before 1917: who controlled school investment students were enrolled for every teacher. Ninety percent ipolicy, and how were the quality and character of of all rural schools consisted of one large room with only |schools shaped to help push economic development? tables and chairs and very few instructional materials. In The federal government assumed strong central this same year enrollment in the third grade was just 10 directionofschoolinvestments. Between 1910 an 1923 percent the level of first grade enrollment (Stanley annual federal support of schools grew from 9 million to 1948). As late as 1950, 60 percent of all children enrolled 29 million (constant 1923) pesos. The number of feder- attended the first grade (Myers 1965). In 1920 the state ally built primary schools increased fivefold during this of Sonora was spending only 6 pesos per capita on public time. By 1925 the federal government provided just over education (about two days' wages of a railroad worker; one-half of all public funding for schools. The once Bell 1922), even though the appropriation for education dominant share of municipal government support had was fifteen times higher than before the revolution. fallen to less than 10 percent of all apportionments Trained teachers were in very short supply, particularly (Stanley 1948).5 Spurred by the rising federal invest- given the rapid rate of school expansion. Wage rates did ment, 46 percent of all school-age children were at least not indicate, however, a shortage of people willing to registered in school by 1927. join the movement. Rural teachers entering the federal This burgeoning level of federal investment in schools school system earned 2 pesos a day, which was compara- was focused on rural development-both to boost agri- ble with workers in manufacturing (Wasserman 1984). cultural productivity and to improve the quality of life in I The federal government addressed issues of quality by the countryside. Revolutionary ideals of this era were centralizing standards and encouraging local commit- embodied by the vision of increasing literacy, integrat- ment. Villages were asked to donate land and labor to ing Indians into the modern state, and equitably redis- build and support the rural schools. The national admin- tributing land and wealth to rural peasants (Sanchez istration standardized curricula, enforced national ex- 1936; Ruiz 1963). More than 80 percent of primary aminations, and expanded (both pre- and in-service) schools built by the federal government between 1907 teacher training (Sanchez 1936). Although improve- and 1928 were located in rural areas. Between direct ments in quality took a long time to implement, Mex- federal support and strengthened mandates on ranchers 'ico's expansion of school investment had tangible and factory owners to open classrooms for theirworkers, effects. For example the nation's literacy rate increased the number of rural schools rose from just over 1,000 in from 23 percent in 1910 to 42 percent in 1940 (Wilkie 1921 to 13,700 in 1947 (Secretaria de Educaci6n Publica 1970). 1926; Wilkie 1970). This brief historical outline of Mexico's investments in The content of primary schooling also shifted to education and the resulting variable quality of schools emphasize rural development. The revolution had to (the antecedent factors of interest) should be matched some extent repudiated free market, individualistic eco- against a brief sketch of economic development patterns nomic values and instead urged cooperative agrarian (the dependent variables) from1880 to 1945. The his- and industrial action to boost domestic productivity torical background in these two areas will help to illu- (Tannenbaum 1933). Rural school "missionsbecame inate the empirical models and to interpret the the tool for organizing v-I1ages--in-the-co6inryside to findings. address problems of illiteracy, primitive farming tech- niques, and poor health care in a comprehensive fash- Patterns of Economic Development ion. The rural teacher was part instructor, part commu- nity leader (Sanchez 1936). Many rural schools operated Into the early 1900s Mexico maintained a colonial collective farms, ran small industries such as furni- economic structure based on the export of minerals, ture-making, raised chickens, and operated cooperative tobacco, tropical fruits, and livestock. As early as 1810 banks for their village (Callcott 1931). The rural school rising domestic demand had also spurred modest pro- School Quality and Economic Growth in Mexico 23 duction in manufacturing, particularly within a modest Table 3-3. Annual Growth Rates textile industry. In that year New Spain's output com- of the Economic Sectors, 1900-40 prised mining (15 percent of total value), manufacturing (percent) (29 percent), and agriculture (56 percent); (Meyer and 1900-10 1910-25 1925-40 Sherman 1979). In 1884, US$37 million in gold and _______1910_25_192S-40 silver were exported, which constituted 75 percent of all Gross domestic product (GDP) 3.3 2.5 1.6 exports (Hamilton 1884). Population 1.1 0.1 1.6 This economic structure was reinforced between 1877 Real per capita GDP 2.2 2.4 0.0 and 1910 under the administration of Porfirio Diaz. Diaz Agricultural production 1.0 0.1 2.7 encouraged foreign investment and growth in domestic Manufacturing 3.6 1.7 4.3 manufacturing. Duriln-t is-period fo6relgf tiradein- Mining and petroleum 7.2 5.6 -1.9 crease`d-urfo1d6 By 1916, US$647 if1ioin in capital was Total exports 4.5 3.9 -1.4 invested in the mining industry, 95 percent of it from Total imports 1.3 3.1 -3.5 the United States, Britain, and France. By 1899, 118 4 cotton factories were operating and were centered pri- Sources: Reynolds (1970a, 1970b). marily in the Federal District, Puebla, and Jalisco. Mex- ico's GNP increased 3.7 percent a year from 1876 to eventual state control of rail and petroleum companies 1910, compared with a 1.6 percent a year after the greatly reduced available capital (Mosk 1950; Reynolds revolution from 1920 to 1940 (Derossi 1971). 1970a). Federal policy was, however, efficacious in Agricultural exports grew by 47 percent in the last boosting agricultural productivity. The annual growth decade of the nineteenth century.7 Yet the aggregate rate moved from 1.0 percent between 1900 and 1910 to value of Mexico's agricultural production in 1882 was 2.7 percent between 1925 and 1940 (Table 3-3). low relative to international standards and equaled only The regional geography of Mexico is also an important the value of oat production in the United States. Diaz did factor ineconormic development. Historically centers of little to urge greater rural productivit , which was production were located in rural areas around mines and already constrained by a feudal agrarian order domi- haciendas, while the nation's more than fifty coastal nated by haciendas and ranchos. In 1910, 97 percent of ports grew because of trade, commerce, and government rural family heads owned no land at all. Mexico's 840 activity. At times, both hacienda owners and mine oper- hacendados controlled most cultivated land and relied ators were faced with wage competition from the mod- on inexpensive farm workers rather than investing in emizing sectors of more urban economies. Similarly more productive technology (Hager 1916; Callcott 1931; the textile industry moved from hiring mostly men to Glade and Anderson 1963; Cockcroft 1968). employing more women in the latter third of the Mexico's shift toward rural development and nation- nineteenth century (Wells 1887; Chavez Orozco 1937). alization ot_eavy ifdin try was clearly felt following In addition migration from the city back to the hinter- 1917. Tables 3-2 and 3-3 show the structure of Mexico's lands increased recurrently because factories were be- gross domestic product (GDP) and economic growth coming increasingly mechanized, price competition was rates before and after the period of reform from 1910 working against small-scale household industries, and to 1917. Discouragement of foreign investment and land reformwas expected (Reynolds 1970a). This issue of Table 3-2. Size of Gross Domestic Product (GDP) Components and Labor Market Sectors, 1900.-40 (percent) 1900 1930 1940 CDP Employment GDP Employment cDP Employment Agriculture 29.9 65.9 23.7 70.2 23.0 65.4 Crop production 14.3 n.a. 13.1 n.a. 12.6 n.a. Livestock 15.6 n.a. 10.6 n.a. 10.4 n.a. Manufacturing 13.2 14.4 16.7 13.4 18.0 10.9 Mining 6.4 2.2 9.8 1.0 5.6 1.8 Transport 3.1 n.a. 2.8 n.a. 5.3 n.a. Other 47.4 17.4 47.0 15.4 48.1 21.9 n.a. Not available. a. Wilkie (1970) reports that agricultural employment declined from 68.3 to 63.4 percent between 1910 and 1940. Yet the size of rural labor force participation as a proportion of all active workers grew from 61.9 to 71.4 percent between 1900 and 1930, but then dropped closer to Wilkie's reported level by 1940 (Reynolds 1970a). 24 Educational Quality and Economic Growth migration between sectors is important in conceptualiz- Several indicators of investment and quality varied ing how investments in schools might differentially in- between the two groups of states. For instance the states fluence growth or decline within different economic with the highest levels of investment spent 51 centavos spheres. Textile factories in rural areas paid higher per capita on schools in 1888; the states with lowest wages than working farms (Nash 1958). In turn urban levels of investment expended only 16 centavos annu- manufacturers and commercial firms paid more when ally. High investment states had a literacy rate of 29 they had to compete for labor with government and percent versus only 6 percent in low investment juris- trade enterprises. Since education presumably increased dictions. The data illustrate the complexity of the mea- job alternatives, wage competition and costs increased sures of school quality. The number of students enrolled within certain sectors. In sum the effects of school per teacher was considerably higher in states with the investment on economic growth may vary depending on highest investment, counter to expectations. Higher en- the economic sector (Hage 1980). rollment rates per capita did not necessarily mean that institutional resources could keep pace with demand. t Table 3-4 also shows how the economies of the two Linking School Investment groups of states differed twelve years later in 1900. and Quality to Economic Growth Although causality cannot be inferred based on these simple tabulations, the differences between the groups For this study several variables representing school are interesting. States with the lowest levels of invest- quality and economic development were included for ment were more rural and had lower mine and cigarette Mexico's thirty-one states (or territories) and the Federal production. But these states displayed comparable agri- District (thirty jurisdictions before 1910). The study cultural and textile output relative to states with the examines how variation in school investment patterns highest levels of investment. Manufacturing related to among these units helps to explain growth or decline tobacco was based in cities, while textile factories were within economic sectors between 1880 and 1945. To distributed throughout both urban and rural areas. Thus illustrate covariation between school quality and eco- the link between the wealth of a state and its level of nomic development, Table 3-4 compares states with the investment in schools also appears to be related to both highest levels of school investment with states that were rural agricultural production and urban-centered in- the least committed to educational institutions. dustry. Table 3-4. Comparison of Mexican States with High and Low Levels of School Investment States with States with highest levels of lowest levels of school investment school investment (means) (means) Education features, 1880-88 School expenditures for instruction per capita (peso) 0.51 0.16 Literacy rate (percent) 29 6 Percent of all residents enrolled in school - - 5.6 3.3 Percent of students who presented year-end exams 44 66 Number of residents per municipally supported school 1,420 1,969 Number of residents per state or federally supported school 24,482 18,204 Number of students enrolled per teacher 113 47 Number of residents per published newspaper 29,906 96,441 Economic features. 1900-10 Percent population living in towns with at least 2,500 residents 57 79 Value of agricultural output per capita (pesos) 10.3 9.8 Value of mining output per capita (pesos) 9.3 4.0 Value of textile production per capita (pesos) 3.0 4.3 State government revenues per capita (pesos) 1.7 1.2 Volume of cigarette production per capita (cigarettes) 51 12 Note: States with high levels of investment in schools include Chihuahua, the Federal District. Jalisco, Morelos, Nuevo Leon, Sonora, and Veracruz. States with low levels of investment include Chiapas, Durango, Guerrero. Michoacan, Oaxaca, Queretero. and Tlaxcala. Sources: Wilkie (1970), Wilson (1941), Vaughan (1982), and the authors' data analysis. School Quality and Economic Growth in Mexico 25 Methodology and Procedures of the Study sides of each equation are reported, which allows use of least-squares regression while retaining the model's We turn next to an empirical test of the basic ques- multiplicative character (Walters 1970; Barnhouse- tion: did school expansion and school quality help spark Walters and Rubinson 1983). early growth in Mexico? The level of variables are observed at two times (for both the 1880-1900 and the 1925-45 eras), and the Data Collection variation among Mexico's states are studied. Capital investment, land, labor, and literacy are assumed to Several federal agencies published statistics on affect output in the same year; school investment and schools and economic activity in each of the two periods quality measures are lagged by twelve to twenty years, studied: 1880-1910 and 1920-45. The data come from depending on data availability. This builds from the federal surveys and budget documents related to school traditional production-function approach. Yet rather support, tax reports on agricultural activity and various than using annual time-series data, observations across industries, and commerce reports. In addition the de- th4e Mexican states are used for each panel of data. cennial census, begun in the 1890s, provides figures on The second approach uses a conventional panel- occupations, school attainment, and literacy. (See Note analysis model, in which the influence of school quality 4 for a complete list of the data sources.) Very few of the in times 1 and 3 on economic output in times 2 and 4 is indicators necessary for conventional time-series analy- examined, holding economic activity in times 1 and 3 ses were published annually. Instead both modified pro- constant. This approach assumes that prior levels of duction-function and panel-analysis models have been economic output will largely determine output ten or used. These models use data falling into four panels: the twenty years later, but a portion of the unexplained 1880s (time 1), 1900-10 (time 2), 1923-28 (time 3), and variance may be shaped by patterns of school invest- 1940-45 (time 4). The models examine the influence of ment. Rather than controlling for capital, land, and school investment and quality in times 1 and 3 on labor in time 2, the level of economic activity in time 1 measures of economic output in times 2 and 4, respec- (for each economic sector) is controlled. School invest- tively. ments are then entered to examine the influence on output in time 2. For instance The Models Used ln(EAt) = a + b* ln(EAt- I) Two types of models were constructed to predict levels + c* ln(T, - l) + d* In(E, n 1) of economic output. They differ on when education where EA stands for economic activity, T for literacy, effects are expected to occur: either in the same year or and E for measures of school investment. Again, literacy lagged twenty years. The tvo models also disaggregate and school investment measures are candidates for entry economic inputs, such as capital and land under cultiva- based on the criterion of minimum significance. This tion, differently. The production-function model, which panel-analysis approach allows economic indicators to is constructed for the periods before 1910 and after be combined within sectors in which reliable composite 1920, adapts the Cobb-Douglas approach. The influence measures can be constructed, rather than differentiating of school investment and quality indicators on economic between capital, land, and labor inputs. This may help to output (Q) are examined after controlling for production stabilize models, particularly when a moderate number inputs: capital investment (K), the labor participation of states or cases is used. The natural log of output, rate (L), and amount of land under cultivation (N). For economic activity, and school investment measures are example used to ensure comparability with the production- Qt = aKb e function model. Q, = aK, t 2 t t t- l Exact measures of some economic variables did not Production inputs are assumed to influence output always exist for the two time periods; yet in most cases simultaneously. Literacy (T) is a candidate variable for composites could be constructed that showed strong entry, and school investment and school quality mea- correlation across the two periods, which provided sures (E), lagged from time 1, are allowed to enter the efficacious control measures. This is an important point equation (using a p = 0.10 level of statistical signifi- related to the validity of inferring causality from these cance). Given the limited number of cases (thirty over longitudinal analyses for both production-function and times 1 and 2 and thirty-two cases over times 3 and 4), panel analyses. For example where production inputs no more than three school quality measures were and controls do not show expected association with the allowed to enter any regression equation. The Cobb- output measures, any relation between levels of school Douglas model assumes interaction between all inde- investment and output may be simply related to varia- pendent variables. Logged values for variables on both tion in state wealth. But when the controls and inputs 26 Educational Quality and Economic Growth are highly related to output, the effects of overall wealth measured by combining figures on the number, horse- will be largely partialed out, which strengthens the power, and value of mining machines operating within likelihood that school investment was indeed affecting each state (alpha = 0.91). The best measure of capital the growth or decline in economic output. The panel- investment in textile manufacturing was the number of analysis approach has been used to examine longitudinal factories operating per capita. A measure of urban-based effects of school growth on economic development and capital accumulation was used to examine the contribu- the reciprocal relation (Meyer and Hannan 1979; Be- tion of capital to the output of tobacco manufacturing. navot 1982; Fuller 1983). This includes the level of capital investment in tobacco factories and the concentration of landlords (alpha = Reduction of the Data 0.92). The composite was used as a measure of urban capital, which justified the mix of capital from two Production-function models assume that the different sectors.8 No measure of capital investment was available elements of economic activity (capital, labor, and out- for agricultural production. By most accounts farm put) are collected in a reliable fashion. Panel-arnlyses workwas highly labor intensive from 1880 to 1900. Even also benefit from the use of identical measures at two large haciendas reportedly used very little machinery points when controlling for the prior level of economic (Wells 1887; Redfield 1930). activity. But such pure data are not always available for A direct measure of the value of textile output was developing nations, especially detailed reports on early available for time 2 (1900). Two measures of mining periods of growth. Composite measures were at times output-one for haciendas and one for all other mines- used as proxies for certain constructs. For example the were combined into a composite (alpha = 0.84). A direct value of mining and textile manufacturing was reported measure of the value of agricultural output was also for Mexican states in 1900; yet output for other indus- reported. Finally the measure of urban capital was trial sectors, especially firms located in urban areas, was matched with the indicator of tobacco manufacturing not reported. The volume of cigarette manufacturing output described above. and federal tax on this industrial sector were highly For the panel analyses two control variables for eco- associated measures, however (Cronbach's alpha = nomic activity were available in time 1 (1880) for each of 0.70). This composite was used as a proxy for output in three sectors. In agriculture the value of land converted tobacco manufacturing. In addition various measures of to cooperative farms was highly related to agricultural school investment and quality were reported. Factor output in time 2 (1900). Agricultural output in 1878 was analyses for principal components were performed to also used as a control variable. Measures of the value of determine patterns of clustering among various indica- mineral exports and trade in nonmetal goods (two in- tors and thus to provide reliable composites. All indica- dicators, alpha = 0.86) were highly related to both tors were adjusted to represent per capital levels of textile production and tobacco-related manufacturing activity, except for ratios such as the rate of labor force and offered valid control variables. participation or ratio of students per teacher. Several measures of school investment and quality were narrowed down to two composites. A composite of local educational resources was built from six indicators Measures for Time 1 (1880) and Time 2 (1900) that measured school expenditures, the ratio of student Economic variables for time 2 first consisted of the attendance to enrollment, and the availability of news- overall rate of participation in the formal labor market, papers and libraries Xalpha = 0.79). This composite given that precise workloads were not available by eco- basically tapped the level of municipal commitment to nomic sector. A rate measure was used to match the per schools and institutions linked to literacy. A second capita level of activity used for the other variables. The composite was related to the availability of state and amount of land under cultivation was estimated for each federal support of schools and was independent of the state based on figures for agricultural volume reported first measure (alpha = 0.96). A direct measure of the by crop, from 1900 and contemporary crop yield data expenditures allocated per pupil in 1874 was also used. (Secretarfa de Programaci6n y Presupuesto 1981). In addition state literacy rates were used as candidate Various indicators of capital investment were avail- variables, subject to the minimum standard for entering able; three composites were constructed based on relia- the regressions (p = 0.10). The literacy figure for 1900 bility levels and the expected relation of the measures to was used in production-function models and the 1880 related output variables. Production-functions were not estimated rate for the panel analyses. built when an output measure could not be matched with a significantly associated measure of capital, given that the examination of the effects of school quality is useful only to the extent that other inputs can be ade- More complete figures on economic activity and quately controlled. Capital investment in mining was school investments were published after 1917. For the production-function equations the overall rate of formal * Institution building-relating to overall expendi- labor force participation was also used in time 4 (1940). tures, numbers of schools constructed, and enroll- Direct measures of land cultivated with maiz and frijol ment rates per capita were summed and used as an estimate of land. Invest- * Quality within the school-including measures of ment in irrigation was used to indicate capital in the the student-teacher ratio, pupils' passage rate on agricultural sector. In manufacturing separate mea- annual exams, and teacher salary levels sures of total capital investment and the value of factory * Sector-specific investment-using separate mea- machinery per capita were combined into a composite sures of private, technical or vocational, and rural (alpha = 0.88). A direct measure of total capital invested school investment. in mines was available for time 4. Agricultural output was measured from the value of regionally exported maiz and frijol (alpha = 0.99). Then, since the value of livestock produced and agri- Findings of the Study cultural and commercial tax revenues shared consider- able communality, they were combined into a composite Results for the production-function models for 1880- measure of agricultural output (alpha = 0.86). A direct 1300 are reported in Table 3-5. For the agricultural measure of the value of mining was used. Manufacturing sector the land variable is strongly significant in predict- output was measured from the value of output, the ing output. Literacy (1900) enters significantly and aggregate of industrial wages earned, and the levels of holds a positive effect on output. Higher levels of state state tax revenue (alpha = 0.92). and federal support of schools is associated with lower For the panel analyses a composite economic control agricultural output during this period. Capital invest- for time 3 (1925) for agriculture consisted of the re- ment strongly determined value of output in mining. ported value of rural and all farms (alpha = 0.88). In School investment had no effect. Within the manufac- manufacturing the composite included value-added to turing sector levels of capital investment contributed all manufacturing and capital investment in tobacco substantially to output. Literacy shows a negative effect factories (alpha = 0.97). Finally the value of exports, for textile manufacturing. In the more urban-centered level of municipal taxes, and business tax revenues were tobacco industry, expenditures per student increased combined into a composite measure of urban-based eco- output. nomic activity (alpha = 0.86). This control variable for Table 3-6 reports findings for the panel models during time 3 was matched with a direct measure of value of this same period. In each sector the controls for time 1 export activity for time 4. entered significantly, especially within the textile and Measures of school investment and quality were abun- tobacco elements of manufacturing. In agriculture ex- dant for time 3 (1925). Principal-components factor penditures per pupil boosted output from 1880 to 1900. analysis indicated some degree of independence among The literacy rate in 1880 appears to have had a positive the various measures. Measures of school investment influence on textile output, yet a negative effect on analyzed here fall into three categories: tobacco manufacturing. The negative influence of liter- Table 3-5. Influence of School Investments on Economic Output, 1880-1900: Production Function Model School investments, 1880 Estimate State Expendi- Economic Labor of Local or tures - R- output, partici- cultivated commit- federal per Literacy, square 1900 Capital pation land ment resources student 1900 (percent) Agriculture n.a. 1.10 0.95d -0.51c 0.85h 98 (1.93) (0.03) (0.27) (0.33) Mining 0.22a 0.92 0.01 40 (0.06) (0.83) (0.01) Manufacturing Textiles 1.89k -0.78 0.02 -0.83c 96 (0.08) (2.72) (0.04) (0.44) Tobacco 0.42, 0.15 0.02 1.181 55 manufacturing (0.09) (0.88) (0.01) (0.56) Note: Unstandardized betas and standard errors (in parentheses) are reported. Each model used logged values. Blanks indicate no significant effects. n.a. Not available. a. p = 0.01; b. p = 0.05; c. p = 0.10. 28 Educational Quality and Economic Growth Table 3-6. Influence of Investment in Schools on Economic Growth, 1880-1900: Panel Analysis Model School investments, 1880 State Expendi- R- Economic Local or tures Liter- square output, Time 1 (1880) commit- federal per acy, (per- 1900 Economic control(s) ment resources student 1880 cent) Agricultural Ejidos output (1878) Agriculture -5.06' 0.54 4.11' 38 (2.99) (2.12) (1.26) Manufacturing Nonmetal Metal trade exports Textiles - 19.82' 1.10' 36.02' 54 (3.98) (0.64) (18.38) Tobacco -1.04' 0.13' 1.12' -5.32' 56 manufacturing (0.38) (0.06) (0.22) (2.21) Note: Unstandardized betas and standard errors (in parentheses) are reported. Each model used logged values. a. p = 0.01; b. p = 0.05; c. p = 0.10. Table 3-7. Influence of Investment in Schools on Economic Output, 1925-45: Production Function Model Agriculture Crop Agricultural Manufac- production commerce Mining turing Economic factors, 1940-45 Capital - 0.07b - 0.01 0.011' 0.87' (0.03) (0.03) (0.002) (0.08) Land cultivated 0.14' - 0.02 0.000 - 0.09' (0.04) (0.04) (0.001) (0.02) Labor participation -0.87 9.32b -0.03 8.14' (3.61) (4.59) (0.15) (2.13) School Investment measures, 1925-28 Ratio of school inspectors - 17.50' to primary schools (5.33) Salary expenditures per 0.49' federal school teacher (0.12) Federal expenditures per 0.47b student and per capita (0.20) Proportion of school-age children - 2.31b enrolled in private primaries (1.10) Ratio of rural schools to (-0.12)' rural school-age children (0.06) Proportion of school-age children 31.62' enrolled in federal primaries (5.63) Small class size/student pass rate 1.72' (composite) (0.49) Literacy rate, 1940 7.67' (2.41) R-square (percent) 59 67 52 92 Note: Unstandardized betas and standard errors (in parentheses) are reported. All variables reported are for logged values. a. p = 0.01; b. p = 0.05; c. p = 0.10. School Quality and Economic Growth in Mexico 29 acy, however, occurs in the equation after the positive activity. Here too enrollment in technical schools di- effect of local educational resources has been partialed minished trade activity during this period. out. Findings for 1925-40 are reported in Tables 3-7 and 3-8. The production-function models show that capital Discussion and Future Work investment significantly affects output in each regres- sion, except for the measure of output of agrarian com- These findings suggest that the level and character of merce, in which labor participation significantly enters. investment in schools historically influenced Mexico's Higher levels of school investment and quality appear to economic output, at least within the agricultural and boost output except for the mining sector, in which manufacturing sectors. These patterns of influence dif- commitments for private and rural schools lowered min- fer somewhat between the two eras, and the effects may ing output. Heavier investment in school inspectors also not always be positive. For example the level of school seems to have depressed agricultural growth after the investment from central state and federal governments revolution. before the revolution appears to have been negatively Table 3-8 illustrates findings for the corresponding associated with agricultural growth. This may have re- panel analyses. At least one of the economic controls for sulted from the urban focus of schools during this time. time 3 (1925) enters strongly for each economic sector. State and federal funding may simply have reinforced Federal school expenditures boosted output in agricul- the emphasis of municipal government on improving ture. Sector-specific investments in both private and schools in trading centers. And the local authorities technical schools were negatively related to rural pro- were, after all, the greatest backers of schools at this duction (the private school variable is the ratio of resi- point. After 1917 school investment-now focused on dents to private primary schools). The composite mea- rural development-positively influenced agricultural sure of school quality-lower student-teacher ratios and output. There are two exceptions, however. Investment higher attendance rates-increased export and trade in technical schools and government inspection appear Table 3-8. Influence of Investment in Schools on Economic Growth, 1925-45: Panel Analysis Model Agricultural or crop production Mining Manufacturing Export activity Exports or Value of Exports or Farm value municipal tax textile output municipal tax Time 1 (1925) 0.41' 0.olb -0.3 0.08' Economic control(s) (0.14) (0.005) (43.6) (0.01) Value-added in Value-added in manufacturing manufacturing 0.69b 0.04b (0.32) (0.02) School investment measures, 1925-28 Federal expenditures 0.36' (0.12) Residents per private 0.0001' primary school (0.00004) Enrollment rate in - 6.69c -0.60' federal technical school (3.40) (0.19) Students enrolled per 2.86b federal primary school (0.62) Ratio of students to teachers -0.06' in all primary schools (0.02) R-square (percent) 53 14 16 85 Note: Unstandardized betas and standard errors (in parentheses) are reported. All variables reported are for logged values with one exception. The regression run for agricultural crop output is for raw per capita values, given the weakness of the time I economic control variable when all values were logged. a. p = 0.01; b.p = 0.05; c.p = 0.10. 30 Educational Quality and Economic Growth to be negatively related to output. Yet the number of analysis also should be moved toward an investigation of primary schools per capita, teacher salary levels, and school investment and economic growth within local total federal expenditures for education all appear to economies. School attainment and economic data exist exert a positive influence on agrarian output after con- for Mexican municipios after 1900. Analysis of these data trolling for states' wealth, capital investment, and land may yield a more fine-grained picture of whether, and under cultivation. These findings appear for both the the extent to which, school investments have helped production-function and the panel-analysis models. shape Mexico's economic growth. The contribution of school investment and quality to manufacturing output is less clear. First conflicting positive and negative effects of literacy were observed before 1917 on textile output between the production- Notes function model (which used a control for direct capital investment and the literacy rate for 1900) and the panel- 1. Barnhouse-Walters and Rubinson's findings (1983) are not con- analysis model (in which economic activity controls for clusive, as they point out. Their regression models for the early U.S. time 1 and the literacy rate for 1880 were ent&red). industrialperiodwereattimesnotstablewherethemeasureofcapital These contrasting effects may result from the different took on a negative coefficient. Unstable or weak effects for some control measures or from a change in literacy pattems ndicators of capital investmenht were ptensted. These equations were between 1880 and 1900. Further work is needed to sort subsequently dropped. this out. One possibility, however, is that increased 2. A private association, La Compania Lancasterana, operated literacy drove up wages in textiles as modern sector jobs schools according to Lancasterian ideals advocated in Europe. The rapidly expanded for a fledgling middle-class during the emphasis on strict order and rote instruction blended well with late nineteenth century. Evidence does indicate that teaching of the church catechism. These schools were later absorbed during this period the proportion of women textile by municipal school authorities and provided a transition between during ~~~~~~~~~~~~~~~~religious and secular institutions in both content and method. At the workers rose from about 30 to 50 percent, which indi- end of the nineteenth century Lancasterian schools were outlawed as cates changing patterns of labor demand and perhaps a remnants of a nonmodem, constraining social order of the past (Bar- shift toward higher paying occupations for men (Chavez ranco 1915). Orozco 1937; Gonzalez Navarro 1970). After 1917 school 3. The struggle within schools to appear modern at times took investment and quality (especially low class size and rather awkward turns. One school textbook, urged on municipal high student performance, Table 3-7) appear to have cchools by the federal goveofrThanksgiving (Secretaria de onstruccion raised aggregate manufacturing output. The strength of Publica 1905). this production-function equation is evidenced by the 4. Primary sources of data. fact that all three production inputs affected output fact entered all threentrolsubefore inptsc l afectmed t ma. Population, occupational, social, and literacy data from decen- when entered as controls before school investment mea- nial census documents. sures. This more consistent effect may result from a Departamento de Estadistica, Mexico City. 1925. Talleres Graficos de higher rate of return on education when Mexico's in- la Nacidn. dustrial sector boomed before and during World War II . Cuarto Censo General de Poblaci6n, 1921. Vols. 1-31. (Reynolds 1970a). This positive effect may have been an Direcci6n General de Estadistica, Mexico City. 1900-07. Segundo unintended consequence of investment in rural schools Censo de Poblacion, 1900. Vols. 1-21. interacting with urban migration during this period of . 1918. Tercer Censo de Poblacidn, 1910. Vols. 1-30. economic growth. - . 1932. Quinto Censo de Poblacion, 1930. Vols. 1-32. More thought should be given to the structure of . 1943. Sexto Censo de Poblaci6n, 1940. Vols. 1-30. school quality, particularly after l917when diverse mea- Secretaria de Programacion y Presupuesto. 1982. Decimo Censo sures become available. The approach here has been to General de Poblaci6n y Vivienda, 1980. Resolutados Preliminares move beyond global measures of school attainment or Nivel y por Entidad Federativa. Mexico City. expenditures per pupil, which have been used in past b. School data. research on economic growth and school quality. This Direcci6n General de Estadistica. 1896. Estadistica General de la initial investigation of the factor structure of school Republica Mexicana, 1884-1896. Vol. 1-10. Mexico City. investment and quality yielded a variety of somewhat GobiernodeMexico. 1888.BoletinSemestraldelaEstadistica.Mexico independent dimensions. Future work should reduce City. available measures to clearer indicators of the quantity Penafiel,A. 1901.AnuanboEstadisticadelaRepublicaMexicana, 1900. of investment in schools and the quality of schools. Mexico City: Oficina Tipografico de la Secretaria de Fomento. In addition the pattems discovered here should be SecretariadeEducaci6nPuiblica. 1926.Boletindelnstrucci6nPublica. en addition the patterns ascovered here should be Mexico City. examined for the period after 1945. Rapid growth in ____. 1927. Noticia Estadistica sobre la Educacion Publica de Mexico continued into the 1950s. Some observers feel Mexico, 1925. Mexico City: Talleres Graficos de la Nacion. that investments in education after the revolution came . 1929. Noticia Estadistica sobre la Educacion Publica de to fruition even more clearly during this time. This Mexico, 1927. Mexico City: Talleres Graficos de la Naci6n. .cfzoo! Quality ana tEconomic Growth in Mexico 31 Secretaria de Instrucci6n Publica y Bellas Artes. 1909. Boletin de Vargas in Rio Grande do Sul on the Eve of Brazil's 1930 Instruccibn Pdblica. Mexico City: Tipografia Economica. Revolution." HispanicAmerican Historical Review, vol. 63, - 1911-.Boletin de Instruccion Ptiblica. Mexico City: Imprenta no. 2, pp. 255-75. de Stephan Y Torres. Baker, D., and P. Harrigan. 1980. The Making of Frenchmen: c. Economic data Current Directions in the History of Education in France, Busto, E. 1880. Estadistica de la Reptiblica Mexicana. Mexico City: 1679-1979. Waterloo, Ontario: Historical Reflections Press. Imprenta de Ignacio Compliado. Banda, L. 1873. Estadistica de Jalisco. Guadalajara: Tipografia Departamerto de la Estadistica Nacional. 1926. Anuario Estadistico: de I. Banda. CersoyDemografia, VidaEconomicay VidaSocialyMoral. Vol.2. Barnhouse-Walters, P., and R. Rubinson. 1983. "Educational Mexico City: Talleres Graficos de la Naci6n. Expansion and Economic Output in the United States, - 1930. Anuario Estadistico: Comercio y Navegaci6n. Mexico 1890-1969: A Production Function Analysis." American City: Talleres Graficos de la Naci6n. Journal of Sociology, vol. 48, pp. 480-93. - 1932.Anuario de 1930. No. 16. Mexico City: Talleres Graficos Barranco, M. 1915. Mexico: Its Educational Problems. New de la Secretaria de Agricultura y Fomento. York: Teachers College, Columbia University. Penafiel, A. 1888-89. Estadistica de la RepuiblicaMexicana. Nos. 1 and 2. Mexico City: Oficina Tipografico de la Secretaria de Fomento. Bell, P. 1922. Mexican West Coast and Lower Califoia: A - 1912. Anuario Estadistico de la Republica Mexicana, 1907. Commenriandin OutiasD Mexico City: Oficina Tipografico de la Secretaria de Fomento. einent Printing Office. Secretarfa de Estado y del Despacho de Hacienda, Credito PGblico y Benavot, A. 1982. "Secondary Schooling and the Impact of Comercio. 1913. Anuario de Estadistica Fiscal. Mexico City: Tipo- Vocational Education on Economic Development." Paper grafico de la Oficina Impresora de Estampillas. read at the Midwest Sociological Association, Des Moines, Secretaria de la Economia Nacional. Direccion General de Estadistica. Iowa. 1939. Anuario Estadistico. Mexico City: Talleres Graficos de la Bendix, R. 1964. Nation-Building and Citizenship. New York: Naci6n. Wiley. - 1943. Anuario Estadistico de los Estados Unidos Mexicanos. Benjamin, T., and M. Ocasio-Melendez. 1984. "Organizing the Mexico City: Direccion General de Estadistica. Memory of Modern Mexico: Porfirian Historiography in 5. Rising school appropriations reflected a substantial shift in fed- Perspective, 1880s-1980s." Hispanic American Historical eral social policies. Between 1920 and 1927 the annual budget for Review, vol. 64, no. 2, pp. 323-64. national defense declined by 40 percent, from 124 million to 76 millionr Berg, 1. 1970. Education and Jobs: The Great Training Rob- pesos (Callcott 1931). Between 1920 and 1945 federal budget appro- bery. Boston: Beacon Press. priations increased from 25.7 to 105.6 (1950 constant) pesos per capita bo r. 1974. eaton O r t a ol (Wilkie 1970). Boudon, R. 1974. Education, Opportunity, and Social In- Federal support of Mexican schools has continued to increase. The equality: Changing Prospects in Western Society. New federal share currently is 70 percent of all funding. State governments York: Wiley. provide 22 percent of all funding, and private sources, such as indus- Bowles, S., and H. Gintis. 1976. Schooling in CapitalistAmer- trial and agricultural company schools, cover the remaining 8 percent ica. New York: Basic Books. (Neumann and Cunningham 1982). Bowman, M. 1976. Through Education to Earnings? Stanford, 6. Recurrent efforts by both state and federal Latin American gov- Calif.: National Academy of Education. ernments to encourage cooperative action and attitudes are explored Bowman, M., and others. 198 1. Educational Choice and Labor by Van Young (1984) for the Guadalajara region and Bak (1983) for Markets in Japan. Chicago: University of Chicago Press. Brazil. Callcott, W. 193 1. Liberalism in Mexico, 1857-1929. Stanford, 7. In the early 1880s Mexico's annual agricultural production com- Calif.: Stanford University. prised the following major crops: maiz (US$112 million), wheat (US$17 million>, sugar (US$9 million), frijol or beans (US$8 million), Carnoy, M. 1972. "The Political Economy of Education." In cotton (US$7 million), coffee (USS2.6 million), and tobacco (US$2.0 Education and Development; Latin America and the Carib- million). bean, edited by T. LaBelle. Berkeley: University of California 8. This composite measure of tobacco manufacturing and concen- Press. tration of landlords isavalid indicatorofcapital accumulation in urban Chavez Orozco, L. 1937. Revolucion Industrial, Revolucion centers in 1900. The correlation between this composite and the Politica. Mexico City: DAPP. proportion of residents living in towns of 2.500 or less equals -0.67 Cockcroft, J. 1968. Intellectual Precursors to the Mexican (P =0.001). By 1940, however, the correlation between manufacturing Revolution, 1900-1913. Austin: University of Texas Press. outPut and this rural measure equaled - 0.33 (p = 0.06). Coleman, J., and others. 1966. Equality ofEducational Oppor- tunity. Washington, D.C.: Department of Health, Education and Welfare. References Collins, R. 1979. The Credential Society: An Historical Sociol- ogy of Education and Stratification. New York: Academic Althusser, L., ed. 1971. "Ideology and Ideological State Appar- Press. atuses." In Lenin and Philosophy and Other Essays. Lon- Denison, E. 1967. Why Growth Rates Differ: Postwar Experi- don: New Left Books. ence in Nine Western Countries. Washington, D.C.: Brook- Bak, J. 1983. "Cartels, Cooperatives. and Corporatism: Getulio ings Institution. 32 Educational Quality and Economic Growth Denison, E. 1974. Accounting for United States Economic Jamison, Dean T., and others. 1981. "Improving Elementary Growth, 1929-1969. Washington, D.C.: Brookings Institu- Mathematics Education in Nicaragua: An Experimental tion. Study of the Impact of Textbooks and Radio on Achieve- Derossi, F. 1971. The Mexican Entrepreneur. Paris: brganisa- ment." Journal of Educational Psychology, vol. 73, no. 4, tion for Economic Cooperation and Development. pp. 556-67. Dias Covarrubias, A. 1875. Estadisticas: Las Escuelas de Mex- Jencks, Christopher S., and others. 1979. Who Gets Ahead? ico. Mexico City: Government of Mexico. The Determinants of Economic Success in America. New Ebaugh, C. 1931. The National System of Education in Mex- York: Basic Books. ico. Baltimore, Md.: Johns Hopkins University Press. Katz, M., and others. 1982. The Social Organization of Early Fishlow, A. 1966. "Levels of Nineteenth Century Investment Industrial Capitalism. Cambridge, Mass.: Harvard Univer- in Education: Human Capital Formation or Structural sity Press. Reinforcement?" Journal of Economic History, vol. 26, Knowlton, R. 1976. Church Property and theMexican Reform, pp. 418-36. 1856-1910. DeKalb, Ill.: Northern Illinois Press. Fuller, B. 1983. "Youth Job Structure and School Enrollment, Larroyo, F. 1948. La Historia de la Educacion Comparada en 1890-1920." Sociology of Education, vol. 56, pp. 14S-56. Mexico. Mexico City: Universidad Nacional Autonoma de Glade Jr., W., and C. Anderson. 1963. The Political Economy of Mexico. Mexico. Madison: University of Wisconsin Press. Levin, H. 1984. "Assessing the Equalization Potential of Gonzalez Navarro, M. 1970. Las Huelgas Textiles en el Por- Education." Comparative Education Reviewv, vol. 28, no. 1, firiato. Puebla: Editorial Jose Cajica, Jr. pp. 11-27. Hage, J. 1980. Theories of Organization: Form, Process, and Lipset, S. 1972. "Values, Education, and Entrepreneurship." Transformation. New York: Wiley. In Education and Development: Latin America and the Hager, G. 1916. Plain Facts about Me-xico. New York: Harper Caribbean, edited by T. La Belle. Los Angeles: University of and Brothers. Califomia Press. Hamilton, L. 1884. Hamilton's Mexican Handbook. London: Maillefert, E. 1865. Directorio del Comercio del Imperio Mex- Sampson, Low, Marston, Searle, and Rivington. icano. Paris: Imprenta Hispano-Americana de Cosson. Hansen, W., ed. 1970. Education, Income, and Human Capi- Meyer, M., and W. Sherman. 1979. The Course of Mexican taL New York: National Bureau of Economic Research. History. New York: Oxford University Press. Hanushek, E. 1979. "Conceptual and Empirical Issues in the Meyer, J. 177. "The Effects of Schooling as an Institution." Estimation of Educational Production Functions." Journal American Journal of Sociology, vol. 83, pp. 55-77. of Human Resources, vol. 14, no. 3, pp. 351-85. Meyer, J., and M. Hannan. 1979. National Development and Haron, I. 1978. "Social Class and Educational Achievement in the World System. Chicago: University of Chicago Press. a Plural Society: Penninsular Malaysia." Dissertation Ab- Meyer, J., and others. 1979. "Public Education as Nation Build- stracts International, vol. 38, pp. 7262A-63A. ing in America: Enrollments and Bureaucratization in the Heyneman, Stephen P. 1980. The Evaluation ofHuman Capi- American States, 1870-1930." American Journal ofSociol- tal in Malawi. " World Bank Staff Working Paper no. 420. ogy, vol. 85, pp. 591-613. Washington, D.C. Mosk, S. 1950. Industrial Revolution in Mexico. Berkeley: . 1982. "Resource Availability, Equality, and Educa- University of Califomia Press. tional Opportunity among Nations." In Education and De- Myers, C. 1965. Education and National Development inMex- velopment: Issues in the Analysis and Planning of Post- ico. Princeton, N.J.: Princeton University Press. Colonial Societies, edited by Lascelles Anderson and Doug- Nash, M. 1958. Machine Age Maya: The Industrialization of a las M. Windham. Lexington, Mass.: Lexington Books. Guatemalan Community. Glencoe, Ill.: Free Press. Heyneman, Stephen P., and William A. Loxley, 1983. "The Neumann, Peter H., and Maureen A. Cunningham. 1982. Mex- Effect of Primary School Quality on Academic Achievement ico's Free Textbooks: Nationalism and the Urgency to Edu- across Twenty-nine High- and Low-Income Countries." cate. World Bank Staff Working Paper. Washington, D.C. American Joumal of Sociology, vol. 88, no. 6, pp. 1162-94. Ortiz, C. 1939. "Mexico Builds Schools: Address to the Society Heyneman, Stephen P., and others. 1984. "Textbooks in the of American Friends of the Mexican People," Mexico City, Philippines: Evaluation of the Pedagogical Impact of a August. Processed. Nationwide. Investment."EducationalEvaluationandPoli- Psacharopoulos, George. 1973. Retums to Education. San cy Analysis, vol. 6, no. 2, pp. 139-50. Francisco: Jossey-Bass. Hogan, D. 1982. "Work, Education, and Social Structure." In _ . 1983. "Educational Research at the World Bank." Work, Youth, and Schooling: Historical Perspectives on Research News, vol. 4, no. 1, pp. 3-33. Vocationalism in American Education, edited by H. Kantor Ramos, S. 1941. VeinteAnos deEducacion en Mexico. Mexico and D. Tyack. Stanford, Calif.: Stanford University Press. City: Imprenta Universitaria. Jamison, Dean T., and Lawrence J. Lau. 1982. Farmer Educa- Redfield, R. 1930. Tepoztlan: A Mexican Village. Chicago: tion and Farm Efficiency. Baltimore, Md.: Johns Hopkins University of Chicago Press. University Press. School Quality and Economic Growth in Mexico 33 Reynolds, C. 1970a. The Mexican Economy: Twentieth- Rural Education. Ph.D. dissertation, Stanford University, Century Structure and Growth. New Haven, Conn.: Yale Stanford, Calif. Processed. University Press. Tannenbaum, F. 1933. Peace by Revolution: Mexico after - . 1970b. Changing Trade Patterns and Trade Policy in 1910. New York: Columbia University Press. Mexico: Some Lessons for Developing Countries. Stanford, Thurow, L. 1974. Generating Inequality: Mechanisms of Dis- Calif.: Food Research Institute, Stanford University. tribution in the United States Economy. New York: Basic Rubinson, R., and J. Ralph. 1984. "Technical Change and the Books. Expansion of Schooling in the United States, 1890-1970." Tyack, D., and E. Hansot. 1982. Managers of Virtue: Public Paper read at the American Educational Research Associa- SchoolLeadership in America, 1820-1980. New York: Basic tion, New Orleans, La. Processed. Books. Ruiz, R. 1963.Mexico: The Challenge ofPoverty andIlliteracy. Universidad Nacional Autonoma de Mexico. 1967. Ley Orga- San Marino, Calif.: Huntington Library. nica de Instruccion Publica en el Distrito Federal, 1867- Sanchez, G. 1936. Mexico: A Revolution by Education. New 1967. Mexico City: Universidad Autonoma de Mexico. York: Viking Press. Van Young, E. 1984. "Conflict and Solidarity in Indian Village Schiefelbein, E., and J. Farrell. 1984. "Education and Occupa- hife: The Guadalajara Region in the Late Colonial Period." tional Attainment in Chile: The Effects of Educational Qual- Hispanic American Historical Review, vol. 64, no. 1, ity, Attainment, and Achievement." American Journal of pp. 55-79. Education (February), pp. 125-62. Vasconcelos, J. 1923. Boletin de la Secretaria de Educacion Schultz, Theodore W. 1971. Investment in Human Capital. Publica, Tomo I, Number 3. Mexico City. New York: Free Press. Vaughan, M. 1982. The State, Education, and Social Class in Secretaria de Instrucci6n Publica y Bellas Artes. 1905. La Mexico, 1880-1928. Dekalb, Ill.: Northern Illinois Univer- Escuela Mexicana. Vol. 1. Mexico City. sity Press. Sewell, W., and R. Hauser. 1975. Education, Occupation, and Wasserman, M. 1984. Capitalists, Caciques, and Revolution- Earnings: Achievement in the Early Career. New York: aries: The Native Elite and Foreign Enterprise in Chi- Academic Press. huahua, Mexico, 1854-1911. Chapel Hill: University of Sierra, J. 1910. Anuarios Escolares de la Secretaria de In- North Carolina Press. struccion Publica y Bellas Artes, 1910. Mexico City: Tipo- Welch, F. 1976. "The Effect of School Quality on Achievement, grafia Economia. Attainment Levels and Lifetime Earnings." Explorations in - 1969. The Political Evolution of the Mexican People. Economics Research, vol. 2, no. 4, pp. 502-36. Translated by C. Randell. Austin: University of Texas Press. Wells, D. 1887. A Study of Mexico. New York: Appleton. Sollano Ramos, A. 1961. "El Sistema Monetario Mexicano de Wilkie, J. 1970. The Mexican Revolution: Federal Expendi- 1877 a 1911." Ph.D. dissertation, Universidad Nacional tures and Social Change since 1910. Berkeley: University of Autonoma de Mexico, Escuela Nacional de Economia, Mex- California Press. ico City. Processed. Wilson, 1. 1941. A Century of Educational Thought. New Soltow, L., and E. Stevens. 1981. The Rise of Literacy and the York: Hispanic Institute of the United States. Common School in the United States: A Socioeconomic World Bank. 1980. World Development Report 1980. New Analysis to 1870. Chicago: University of Chicago Press. York: Oxford University Press. Stanley, J. 1948. La Casa del Pueblo: Mexico's Experiment in Part II implementing Investments in Quality 4 The Stages of Growth in Educational Systems C. E. Beeby 4 T his chapter is about the nature of change in educa- certain political and material reasons that prevent par- T tional systems, and the conditions that are neces- ticular projects from succeeding, only the structural sary to bring it about. My theories about school change problems intrinsic to most of the failures are considered and quality improvements are based on almost forty here. years of experience and observations in both industrial Behind every failure of this kind there lies, in my and developing countries. experience, an inability to realize that change in the Basically I have learned that change is a slow and quality of classroom practice is a matter of growth-a ragged process that does not proceed uniformly on all living, inward change in the minds of individuals, which fronts. Teachers are the frontline troops of change, and is not brought about solely by manipulating external progress depends on their own education, motivation, influences such as laws, official syllabuses, instructions, and freedom to innovate. As a result massive reforms sanctions . . . or even money. There seem to be certain that are mandated from a national department of educa- laws of growth: certain factors stimulate it, others in- tion cannot be expected to occur overnight. hibit it. When educational planners fail to recognize There is no such thing as a general definition of these laws, two kinds of miscalculations result: educational quality, since the concept of quality depends * Planners who have not had extensive experience in on one's concept of the ultimate purpose and objectives educational administration rarely understand how of education. This varies from country to country, from difficult it is for teachers to change their classroom group to group, and even among individuals. Neverthe- practice. They seldom know the type of in-service less in all developing countries there are obvious educa- training that is necessary - to help average and tional changes to be made that everybody would accept below-average teachers make even limited changes. as improvements, whatever their views on the ultimate This frequently means that the gap between an old ideal school. syllabus and a new one is too wide for many teachers For several decades now the World Bank and other to bridge. agencies have been involved in efforts to help improve both the access to, and the quality of, schools in develop- * Partly as a consequence of this and partly as a result . | . . ~~~~~~~~of failing to understand the history,of educational ing countries. Although questions of access predomi- change in older countries, planners often grossly nated inthe 1960s, the emphasis shifted to quality issues underestimate the time it will take to bring about in the 1970s. real change in classroom practice. The standards found in the best schools of the richer Success Is Illusive countries took a century to develop. Trial-and-error, pilot projects, the contribution of individual pioneers, Looking back, it is clear that many attempts to im- and the slow absorption of new ideas and attitudes prove the quality of schooling have failed. There are played a far bigger part than did overall planning. many reasons for those failures. Although there are Developing countries trying to follow the same track 37 38 Implementing Investments in Quality face a completely new problem. They want to compress Stages of Growth generations of change into a decade. This is common- place in the literature. What is not so commonly real- My thesis is that there are four stages in the growth of ized, either in theory or in practice, is that this com- a primary school system in countries of all types (Beeby pression forces processes that are by their very nature 1966); they are set out in the model below and in Table successive to occur simultaneously or nearly simulta- 4-1. Stage I (the dame school stage) can, for present neously. purposes, be ignored. The progression from the begin- If compression occurs in the operative phase of a ning of stage II (formalism) to the end of stage IV program as well as in the planning phase, it can lead to (meaning) occurs in three ways: confusion, frustration, and grievous loss of money, * The teachers become better educated and more fully effort, and good will. In the planning phase the tech- trained niques of systems analysis can help prevent some of this, but such techniques can be misleading at the opera- tedacn less regnun onmaro nds stere- tional phase, where changes in the classroom depenc4on typed and le dependention the slow and sometimes unpredictable changes in the structlon and rote memorization attitudes and skills of teachers and, to a lesser degree, of * The links between symbols and their meaning be- students. The attitude of the Bank regarding timing is come stronger and wider, more attention is given to crucial in any project concerned with improving quality the abilities and needs of individuals, and students in the classroom. Some phases of an operation may need are encouraged to solve problems and think for to be speeded up, while others may need to be slowed themselves (Beeby 1980). down and extended. Change in this direction represents improvement in Table 4-1. Stages in the Growth of a Primary School System Stage Teachers Characteristics Distribution of teachers I. Dame school III-educated, Unorganized, relatively meaningless symbols; very X untrained narrow subject content-three R's; very low standards; memorization all-important. t years I. Formalism III-educated, Highly organized; symbols with limited meaning; trained rigid syllabus; emphasis on three R's; rigid A methods-"one best way"; one textbook; exter- . nal examinations; inspection stressed; disci- pline tight and external; memorization heavily P stressed; emotional life largely ignored. Ill. Transition Better-educated, Roughly same goals as stage 11, but more effi- trained ciently achieved; more emphasis on meaning, - B but still rather "thin" and formal; syllabus and textbooks less restrictive, but teachers hesitate _' - to use greater freedom; final leaving examina- Q tion often restricts experimentation; little in classroom to cater for emotional and creative life of child. C IV. Meaning Well-educated, Meaning and understanding stressed; somewhat - well-trained wider curricula, variety of content and methods; individual differences catered for; activity R methods, problem solving, and creativity; inter- nal tests; relaxed and positive discipline; emo- tional and aesthetic life, as well as intellectual; closer relations with community; better build- ings and equipment essential. Y Stages of Growth in Educational Systens 39 the quality of classroom work; movement in the oppo- lower levels of stage III. The transition from stage II to site direction is regression. To speak of a school system stage III and the progress through most of stage III is as if all parts of it were ever at exactly the same stage relatively simple, because teachers only have to be would be an oversimplification. The last column of the helped to do better the things they have been doing table shows the probable distribution of the teachers on poorly. It is at the top level of stage III and the transition the quality scale. The vertical line XY represents the to stage IV that most of the projects fail. continuous scale of growth from the beginning of stage I At this stage teachers must accept new objectives, to the end of stage IV. The average teacher is at B on the radical changes in attitude, and much more individual quality scale just before the reforms are instituted, and responsibility. This is hard for teachers educated and the spread AC indicates that the best teachers are a full trained in very routine and authoritative systems. Cur- stage ahead of the least able. After t years of the pro- riculum changes from conventional arithmetic to the grams of reform the general standard of the teaching has new mathematics and from the factual teaching of his- improved, but there is still a wide spread of teachers on tory and geography to the more complex teaching of the quality scale PR, with the average teacher at Q. social studies have proved difficult in some countries Except under the experimental conditions established by because the change involves moving from stage III to Bloom (1976) in his research on the effects of mastery stage IV. learning, there will always be a spread in the quality of It is easier to get consensus on programs for raising teaching in any school system. (Any proposed program school systems from stage II to stage III than on pro- of reform that did not provide for this should be consid- grams that call for teachers to shift from the top of stage ered suspect.) The angle CBQ in the model is what might III to stage IV. The faults in teaching and learning at be termed the "angle of reform." Many projects, espe- stage II are so obvious that people with very different cially when supported by massive injections of financial philosophies of education are likely to agree on the aid that might intensify the sense of urgency, fail be- immediate steps to be taken. Because the move from cause they make the angle of reform too acute. stage III to stage IV means breaking with tradition and There are two hypotheses that are central to this accepting new objectives and criteria of excellence, it thesis of stages. can meet strong financial and political resistance. * There is a recognizable progression in the qualita- Schooling at stage lV is always the most expensive. tive growth of most primary school systems, and Elitist governments-South Africa is an extreme exam- one stage is a necessary prelude to the stage that ple-may think it politically dangerous to teach the follows. Except for stage I no stage can be skipped, mass of the people to think for themselves. Traditional although it is possible that, with better understand- governments may see the kind of thinking encouraged ing, the stages can be shortened. at stage IV as being disruptive of cultural, social, and ing, the stages c shorted religious values at their level of development. The role of X Passage through the stages is limited by the level of the educator is limited here, and the position of the general education ofethe teachers and by their level foreign adviser unusually delicate (Beeby 1980). and type of professional training. There may also be good educational reasons for de- Changes in teacher education are a necessary, but not laying the transition to stage IV. When the level of skill of sufficient, cause of change in the stage of a system. most teachers is too near the beginning of stage lll, it is Passage through the stages can also be limited by con- generally wiser to concentrate on helping them to do straints of finance, material facilities, and politics. Yet better the things they are already doing than to try to planners most commonly overlook the more purely pro- switch them to a radically different -concept of educa- fessional constraints. tion. At the same time if the administration of the Growth is, of course, a continuum, but in practice it is schools is flexible enough, the most able teachers can be easier to consider it as a series of stages. Although the encouraged to venture into stage IV. model in its present form applies directly only to pri- mary (elementary) schools, most of the practical conclu- sions that can be drawn from it also apply to secondary Qualitative Change in the Classroom schools. It remains to be seen whether any of the new technologies, such as mastery learning or the use of Five characteristics distinguish qualitative from computers in schools, will enable a stage to be skipped. quantitative change in classroom practice. Also it is not clear whether such new technologies are economically or professionally feasible in a poor Consensus about Change country. Many of the school systems with which the Bank will Quantitative changes are usually easy to define and be working will be at the upper levels of stage II or the measure, but there are frequently strong differences of 40 Implementing Investments in Quality opinion on what constitutes improvement in quality. strated only in mini-lessons. They have little to transmit These differences are intensified by the difficulty of de- but theory and a few simple examples. fining many qualitative objectives and of measuring Another effect of overcompressing a pre-service train- progress toward them. ing program is that in the outlying districts, classroom The best way to avoid confusion and misunderstand- teachers, school inspectors, and teacher trainers all have ing in any new program for improving work in the to be introduced to the new methods at the same time- classroom is to include representatives of all interested sometimes even in the same classes. In a hierarchical or parties in the planning process from the beginning. One authoritative school system this not only embarrasses of the dangers of rushing operations is that there is no the inspectors and teacher trainers, but it also deprives time to consult groups whose cooperation will be essen- the teachers of the help of their acknowledged leaders tial to the success of the program: teachers' organiza- and superiors at a critical juncture. tions, inspectors (supervisors), teacher training institu- If a large innovative project is to succeed with the tions, and provincial and local authorities with some average and below-average teacher, the initial period of. responsibility for schools. 4 in-service training must be followed by systematic, con- It is not simply a matter of perfunctory consultation tinuing advisory services to teachers in their own class- or passing on information. These groups must feel that rooms. This must be done not for months, but gradually they have contributed to the scheme, so share responsi- for several years. It is expensive and involves teams of bility for it, and have an obligation to explain it to their visiting subject specialists as Well as the normal inspec- members and the general public. In many countries the tors. If a country cannot afford such a follow-up service consumers of education-parents, students, and their to achieve a radical advance in teaching methods, it prospective employers-have been able to kill or distort should consider a less ambitious form of change. Skimp- educational reforms if they feel the programs run ing in follow-up services is the most common and, in the counter to their traditional idea of what education really end, the most wasteful reason that large projects fail. is. This suggestion, like some that follow, may seem Teacher training institutions should be the centers of almost too obvious to mention; yet these concepts are in-service training. Yet the quality of teacher training, often ignored, with sad results, in many places. especially for the primary school teachers, is often low. Because of this situation training schools for primary Acceptance by Teachers teachers may not be used for the initial training for a new program, and such training is then undertaken by Qualitative changes in classroom practice will occur staff from training institutions for secondary teachers, only when teachers understand them, feel secure with who probably have no experience in primary schools. In them, and accept them as their own. The most serious far too many cases there are only weak links between mistakes in planning changes in classroom practice training institutions for primary and secondary teachers occur in this area. There are several reasons why, in any as well as between primary and secondary schools. In new official program, some teachers do not sufficiently most developing countries teacher training should re- understand the new teaching methods and curricula ceive the highest priority for aid and programs of re- that they are being asked to use. form. Follow-up work in the schools should also be a The leap from the old practices to the new ones is normal function of training institutions as soon as they often too wide for the average or below-average teacher are ready to undertake it. to make in one step. The cycle of change. in a school system is often too The time allowed-often only a week-to train class- short, partly because of a rapid succession of ministers of room teachers is frequently too short when the changes education, each of whom is anxious to leave a mark and to be made are big ones. The training, moreover, is cherishes a misplaced faith in the power of a new sylla- usually too verbal and theoretical. Teachers, especially bus to change classroom practice. Minor adjustments to those with scant general education, do not learn new any new syllabus are necessary within a year or two. But practices easily from lectures and pamphlets; they need it takes at least ten years for a genuinely new syllabus, to be able to observe the new methods being effectively with a change of objectives, to become established practiced in a real classroom. If the practices to be throughout even a country of modest size. A cycle of ten learned are novel to the country and the whole program years is the minimum, but one of about fifteen is usually is unduly compressed, however, there are no classrooms better. In primary schools each teacher normally in which to observe skilled teachers in action. This teaches almost every subject and can be overwhelmed means that even the instructors who are sent from the with demands for change if the entire syllabus is altered center to introduce the innovations to the mass of the at the same time. So a continuous reform of syllabuses teachers may have no firsthand experience in the and books is preferable, with syllabuses in different sub- teaching methods and may have seen them demon- jects being altered successively. This is not so necessary Stages of Growth in Educational Systems 41 in secondary schools, in which there are specialist room. Unfortunately if the reforms have been intro- teachers. duced too fast, the principal may know no more about Teachers must be intellectually and emotionally se- them than the teacher. cure if they are to venture from the established practice and the memorized textbook of state III to stage IV, in which students are encouraged to think for themselves and ask their own questions. There are several common Teachers vary greatly in their capacity to make qual- causes for teachers to feel insecure with a new syllabus. itative changes and in the speed at which they do it. This -Deficiencies in thetache'sonmay appear to be a platitude that could be applied to any trDeficiencie occupation. But there is one thing that distinguishes the training. teaching profession from most other forms of employ- * Lack of professional libraries and of the skill in ment: apart from a relatively small number of adminis- using them. How can a teacher dare to encourage a trative and advisory posts, it has no hierarchy in which class to ask questions if no one can find answers to individuals can be slotted according to their abilities. the awkward ones? It is safer not to venture beyond Fo? any group of students the success or failure of a the textbook. reform depends on the teacher who happens to have * Necessity of teaching in a second language with been assigned to them; unless there is some form of which the teacher does not feel comfortable. team teaching, the classroom is an island. * Lack of strong and well-informed school principals, This peculiarity of the teaching profession creates a inspectors, and teacher training institutions. most complex administrative problem. How is an ad- * Lack of the moral support that comes from being a rninistrator to devise a system of control-for sylla- part of a broadly backed reform that has the support buses, textbooks, examinations, supervision, and regu- of both the ministry of education and the profes- lations-that leaves the able and imaginative teachers sional organization, as well as the understanding of reasonably free to experiment and explore, while still the immediate community. A new program in giving the less able teachers the firm framework of education does best when it rides the crest of a wave support and instruction they need? No country, rich or of social and political change. poor, has found the complete answer to this question. * Isolation of the classroom. Anything that breaks The problem is especially difficult in developing coun- that isolation down, such as team teaching, the tries in which the central control of education is tradi- open classroom, and regular professional meetings tionally authoritative. with colleagues under good leadership can increase the sense of security. Rate of Implementing Innovation But passive acceptance is not enough. Few real re- Qualitative changes in classroom practice almost forms in classroom practice are brought about by official never occur at the same rate throughout a country. This fiat. Faced with a new syllabus and with nothing to follows directly from the preceding statement, but there support it, teachers regard it as something extemal to are additional reasons why different rates of progress are themselves and show a remarkable capacity to go on found among individual teachers as well as among doing the same old things under a new name. schools and districts. If they are given freedom to ma- Teachers can be encouraged to accept new ideas and neuver, some local administrators, institutions, dis- practices as their own by involving them through their tricts, and teachers will always outstrip the bulk of the professional organizations in the devising of the pro- profession in adopting the new official practices and gram and by giving them the freedom to make the even in experimenting with innovations of their own. practices their own by modifying them to suit their The way a country regards this ragged line of advance surroundings and adapting them to their personal determines its policy for bringing about qualitative re- styles. (There are difficulties in this for the administra- form. If administrators accept this as a sign of healthy tor, as the next section will show.) growth, they will base their strategy on these differ- A teacher's faith in the new system tends to reach a ences. They will seek ways to use the natural innovators low point a few weeks after returning full of enthusiasm, to break new ground in their districts for the less able or from the initial training course. The theory that has more timid, but they must do this in such a way that the been learned and the few model lessons that have been innovators are not themselves trapped within a new given are not enough to meet the unforeseen problems system and then given neither freedom nor encourage- of a full year's teaching. At this point an understanding ment to pioneer further. principal and a visiting subject adviser are needed to This could lead to a two-phase strategy of qualitative help the teacher modify the new practices to each class- reform. It is particularly applicable to a country (such as 42 Implementing Investments in Quality Indonesia in 1970) in which there is a serious shortage of an additional reason for advocating a measure of decen- school textbooks, and in which the few existing ones are tralization that would stimulate initiative and a sense of so poor that it would be absurd to reproduce them by the commitment in individuals and institutions throughout million, no matter how urgent the need. Most of the the country. It is an expensive and time-consuming teachers are, say, at some level of stage III, and a few are technique, but it might, in the long run, save great ready to go on to stage IV. There is an immediate need expenditures on books and training that later prove for new and better books that can be produced rapidly. ineffective. Pilot projects, whether official or unofficial, More sweeping reforms in curricula, textbooks, and form an essential part of a two-phase policy of introduc- teaching methods are necessary to improve quality sig- ing innovations. nificantly, but experience has shown this to be a very lengthy process. A two-stage strategy might operate in Continuity of Change such a situation (Beeby 1979). Phase I. Make only such changes in the existing Qualitative change in education is rarely linear or continuous. If qualitative change in classroom practice curriculum as are necessary to eliminate obwioous.q.g faults and to help the mass of the teachers to do is a process of growth and not a matter of imprinting a better te t hings they are already doing and with the complete new pattern on a whole school system at once, baetter the things they are already doing and with the a project might be expected to follow some of the rules of same objectives. At the same time give freedom- organic growth: to flower, fruit, and die. This is, in fact, and a little material aid-to the able and adventur- the natural life cycle of many projects, and there is no ous teachers to encourage them to interpret thhe eontu if cye of m any ojs andvthris n curriculum more liberally and to experiment Withn on,i aetn t Sotnosinvtosi newuumethods.Produe thberaly f setof textbookst oh pedagogy have a habit of regressing toward the average this basis. It should be relatively simple to train when the bright souls who started them die or lose their teachers to use them. These books can be expected fire, although they may also eventually come to exercise tolacrstfor usoe fivem yhears,book which l ge expectimed a powerful influence on practice. Even official programs to last for somefieyear,whicwillgvetimof reform rarely retain their vigor for a long period. So new curriculum to be prepared and radically differ- determined reformers, no matter how extensive their ent books to be introduced when phase .I becomes official power, like a farmer, must plan for a succession fully operative. of crops of new ideas and practices against the day when e Phase II. This phase must overlap with phase I and some of the old ones cease to grow. But this is impossible should be started as soon as phase I takes form. It to do if all the initiative comes from the center. will follow the pattern of phase 1, except that alter- In a developing country most of the educators who ations to the curriculum will be more drastic and have studied and traveled overseas often are in the will be reflected in trial textbooks produced in some ministry of education or in institutions near the capital. economical form. These books will take longer to It is natural that new ideas should originate in the center write, try out, and modify than the books in phase I with official blessing, but, in the long run, this is not Likewise the inspectors, principals, teacher train- sufficient to maintain the liveliness of a school system. ers, and a cadre of mobile tutors and advisers will Changes of government or ministers, financial strin- have time to use the methods themselves before gency, or sheer fatigue can sap the strength of a national telling others how to use them. Since the new books campaign. If the impetus is not to be lost when the will involve changes in the very objectives of central drive flags, innovations must spring spon- teaching some subjects, the trials must be on a scale taneously from teachers or regional and local adminis- large enough to cover schools in districts that do trators. not come under the immediate supervision of the The secret of a continuing, although not necessarily original enthusiastic and skilled reformers. In addi- regular, advance lies in the symbiosis of the center and tion to the schools in which the books are systemati- the periphery of an educational system: in the interplay cally tried, some of the better and more progressive of official programs and local or individual initiatives. schools in each district should be encouraged to These two kinds of innovations are subtly interdepen- experiment with at least some of the elements in the dent, and neglect of one will weaken both. Without overt second-generation books. When the time seems fit evidence that the authorities encourage change, indi- to introduce all schools to the new curriculum, vidual teachers and institutions are less likely to strike there will be model schools in every region in which out boldly on their own. In return, while the official teachers can see the system in operation. program is being developed, the pioneers can probe its It would be difficult to apply the two-phase technique weaknesses and give new meaning to its successes. in a very large country, such as Indonesia, which has a When the program begins to lose its impetus, individual highly centralized school administration system. This is growing points will remain throughout the country to Stages of Growth in Educational Systems 43 sustain the spirit of change and suggest new lines of completely waterproof. But the transition to stage IV concerted advance as soon as conditions permit. demands much more than that, if only to store books All these characteristics must be foreseen when the and equipment and protect them from vandals and the first plans are being made to improve the quality of weather. Some poor countries must use the same build- classroom practice. ings for two or even three different schools and staffs in a day. This is tolerable at stages 11 and III but can be disruptive at stage IV. Unless there is extensive, safe External Constraints on Change storage, teachers have only the books and equipment they can carry to a lesson, and class projects have to be In order to emphasize the role of the teacher in educa- dismantled and taken away each day. The morale of tional change, the external constraints on improving the enthusiastic teachers and students declines, the class- quality of schooling have been neglected. These are well room stops being a place in which to linger, and the known to educational planners of all persuasions, but system easily drifts back to rote learning from textbook they still merit brief mention, if only to preclude the and dictated notes. possibility of teachers being blamed for all failures to 4 raise the quality of education systems. The most inspired teachers can have their efforts ruined by factors beyond Books and Equipment their control. Each step a school system takes from stage I to stage IV creates a new demand for books, writing materials, Finance and school equipment. The need for these materials increases dramatically during the transition from the Good education costs more than bad. (Beeby's Law, top of stage III to stage IV. If the links between symbols with acknowledgments to Gresham.) Admittedly an ex- and their meaning are to be strengthened, children pensive school can, for a variety of reasons, be a bad one, must have equipment and materials to reinforce the but any attempt to improve the quality of a whole educa- connections between the words and numerals they are tion system costs extra money, which is usually spent on learning and the practical things they represent. If stu- books, equipment, materials, and services and always on dents are to be encouraged to think for themselves and in-service training. to ask questions in class, both they and their teachers It remains to be seen whether some new technology in must have access to libraries. Outside the towns they are education will break this unpalatable rule and reduce likely to find one only in the school. The sciences cannot the cost of good education, but there is little sign of it as be leamed simply from a textbook, yet laboratories and yet. It is difficult to judge the effectiveness of new tech- workshops are expensive to build and maintain. nologies in conducting cheap mass education out of school, but, within the classroom, there is little chance of any combination of software and hardware replacing the teacher or making the job so simple that it can be The best of projects can fail through poor coordina- carried on by less expensive monitors. (The sad experi- tion in the ministry of education or through a break- ence of educational television in American Samoa bears down of communication with distant provinces or dis- this out.) Therefore the cost of technology must be tricts. Lack of cooperation, not to mention antagonism, added to the cost of teachers' salaries, which, especially between divisions in the ministry can be fatal to a in poor countries, is by far the biggest item of educa- project. tional expenditure. Not infrequently there is tension between a unit that The cheapest-and poorest-form of schooling oc- develops curricula and the divisions that control the curs in an institution at the bottom of stage 1, with an day-to-day work and inspection of the schools. Equally ill-educated and poorly paid teacher, a blackboard, forty as serious is the failure to include the examination or fifty children, a few tattered books, and a syllabus that division at the very beginning when a reform is being can be leamed by heart. Every national advance from planned. In the long run the people who set the external there toward the top of stage IV costs money. examinations have more influence on what goes on in the classroom than the people who write the curricula or Buildins .control the schools. Buildings In a very large and scattered school system with Quite good teaching, of the styles appropriate to highly centralized administration the lines of com- stages 11 and 111, can be done by competent teachers munication with outlying provinces and districts are under great difficulties: in buildings that are stark, often ineffectual. They are strong enough to check re- noisy, badly lit, scantily fumished (if at all), and not even gional and local growth, but not strong enough to 44 Implementing Investments in Quality stimulate, feed, and control it. Regional administrators, Conclusions district inspectors of schools, and teacher training in- stitutions (particularly for primary teachers) can be- This would be a cheerless note on which to end. Since come cut off from the mainstream of educational my first professional contacts with education in develop- thinkirrg. ing countries in 1945, however, there has been enough Primary school inspectors or supervisors are the real progress to justify mild optimism in anyone who is ministry's only continuing professional link with the not expecting miracles. classroom teacher, and yet they are often the most Many developing countries have made admirable isolated people in the whole system of control. Because efforts to provide more school places for their children. of their isolation, they tend to be conservative and bound If the progress in improving the quality of their schools to routine and are consequently ignored by the reformer has been slower and more checkered than had been who is impatient for quick change. The same is true of hoped, that is mostly because the job is far more difficult the institutions that train primary teachers. Yet often than had been naively imagined. Disappointment about the mobile teams who introduced the innovations have unreal hopes can blind one to the progress that has departed, it is the inspectors and the training inslitu- occurred. tions that must provide the follow-up program. Their In the past forty years some countries have im- work can be coordinated more effectively by a regional plemented improvements in quality that are creditable representative of the ministry, but to do that, the re- when seen against the backdrop of the four stages of gional staffer must have sufficient power. Delegation of growth. There are no schools as poor as some were back control, however, does not come easily to a government then, and the average teacher has moved up a full stage intent on establishing or maintaining national unity. on the scale. The best of the teachers already have some vision of what a school at stage IV would be like, and a Conditions of Service for Teachers handful of them run classrooms in which some of the qualities of that stage are already appearing. Some of the Any change in classroom practice necessarily involves more advanced countries, such as Malaysia, have done teachers in a great deal of additional work, especially in even better than this. Real progress in improving the the first year or two. Whether they are willing to make quality of education on a national scale is so slow in any the extra effort depends on their morale and conditions country, rich or poor, that one has to stay around a long of work. Poor housing conditions, which are common in time to see it with the naked eye. rural areas, can seriously affect morale. Low salaries will make it necessary to take on a second job, which will absorb time and energy that should go into the arduous References task of planning new courses and testing the results on their students. Nothing destroys morale more thor- Beeby, C. E. 1966. The Quality of Education in Developing oughly, however, than requests and complaints to the Countries. Cambridge, Mass.: Harvard University Press. ministry that go unanswered, salaries that arrive late, . 1979. Assessment of Indonesian Education: A Guide injustices and hard cases that are ignored, and a growing in Planning. New York: Oxford University Press. conviction that no one at the center really cares. This is . 1980. "The Thesis of Stages Fourteen Years Later." felt particularly strongly in rural areas. The brightest International Review of Education, vol. 26, no. 4, ideas and the best plans can bog down in a morass of pp. 45740. discontent over apparent trifles. Bloom, Benjamin. 1976. Human Characteristics and School Learning. New York: McGraw-Hill. 5 Lessons from Bank Experience 4 T he World Bank and its borrowers are increasingly seminar; four focused on research problems and eight Tconcerned about the quality of education. Enroll- on operational problems. Sessions that focused on re- ments are rising quickly, and the variety and sophistica- search issues included: school quality and economic tion of cognitive skills expected to be transferred efficiency, the retention of cognitive skills and literacy, through formal education are rising in parallel fashion, socioeconomic status and academic achievement, and but the capital available per student is not. The result is the distribution of school quality within countries. an increasing demand for lending for education to ame- Operational sessions included: the preparation of liorate problems of educational quality. teachers, the administration of school systems, promo- For the past five years the Bank has embarked on a tion and repetition practices, examinations and selec- series of research projects on educational quality. The tion policies, electronic media, time on task, production products have appeared in the academic and professional and distribution of textbooks, and prevocational work- literature and, in various forms, have been distributed shop subjects. Reports from the eight operational ses- and discussed with operational staff and borrowers. But sions follow. the diffusion and dissemination has not been organized or coordinated. The Bank has also broadened its experience in im- Administration of School Systems proving the quality of education through its lending program. Projects considered unusual a decade ago are Session chairman: Abdelwahed Zhiri now commonplace, such as producing and distributing Participants: Mats Hultin, Douglas Keare, textbooks, manufacturing school equipment and facili- Sherry Keith, William Loxley, James ties, and strengthening educational administration. McCabe, and Delia Pitts Likewise increased attention has been given to the prob- lems of school quality in economic and sector work. As the education sector worldwide has grown during The seminar on which this volume is based, was recent years, the problems of educational administra- designed to give both operational and research staff a tion have become more complex. Yet Bank research has chance to reflect on what has been learned. More than found that the greatest weakness of educational systems being a technical summary of the Bank's research, the in developing countries is their management system. seminar delved into policy issues, such as school man- In the industrialized countries management has im- agement, the economics of planning curriculum objec- proved, and schools provide numerous services, such as tives, and the tradeoffs between school quality and health care, libraries, compensatory education, occupa- school expansion. The seminar provided an opportunity tional counseling, and athletics. All over the world there for scholars with theoretical experience from both out- are more students with special needs, more subjects to side and inside the Bank to work alongside operations be taught, and more depth required in the teaching of staff with experience in managing investments in educa- those subjects. tional quality. There were twelve sessions in the two-day Yet the governments of many developing countries do 45 46 Implementing Investments in Quality not have the necessary resources to maintain, let alone Participants: Albert Aime, Richard Johanson, increase, the services provided by their education sys- Robert McMeekin, Abdun Noor, George tems. The rate of growth in real per capita government Psacharopoulos, Jee Peng Tan, Hans Thias, expenditures on education fell in many countries be- Jacob Van Lutsenburg Maas, and tween 1972 and 1979. The rate of growth fell in Guate- Lawrence Wolff mala by 91 percent, in Brazil by 72 percent, in Syria by 42 percent, in Zambia by 50 percent, and in Malawi by 24 percent. It is therefore necessary to find ways to The term "examinations" refers to externally set stan- make the education sector more efficient with the same dard tests for selection or diagnosis that are adminis- amount of resources. tered to students in reasonably large areas. These areas could constitute entire nations, parts of a nation, a state, Research Problems or a province. Tests designed by teachers for classes or individual schools are not included in this definition. The state of the art regarding educational manage- ment is weak. In some cases there are unclear objectives, inefficient managers, inappropriate methods of select- ing and assisting managers, and either insufficient or Examinations systems should be integral elements in inappropriate incentives. Moreover many previous stud- the highly interactive set of inputs to the quality of ies of educational administration have been too super- schooling, for several reasons. First, if properly designed ficial to provide much guidance. and properly managed, the examinations can exert Decentralization by itself is not the best way to solve powerful positive influence to improve other key vari- the problems of educational management. The "best" ables of school quality, such as curricula and supporting solution will be different for each country. The appropri- materials, teachers' classroom activity, and school ad- ate degree of centralization or decentralization varies ministration. In addition a measure of what pupils have with institutional structure, the size of a country, and learned is the best overall indicator of educational prod- the level of development of its economy. It also varies uct. Whatever else examinations may accomplish, they with the quality of the personnel that implements deci- provide information on what has been leamed, which sions at each level. In addition the political conse- can be used to improve educational planning and man- quences of administrative changes, such as teacher agement. Finally examination systems, because of their strikes in Colombia, must be better foreseen by the incentives, are among the most powerful instruments Bank. for improving school quality and management. Com- pared with the other inputs to school quality, examina- tions are relatively cheap and can have a pervasive effect. Recommendations Examination and testing systems should be evaluated The World Bank should on an equal basis with other inputs to school quality. * Analyze each administrative process in education This has several implications. In sector work the Bank should automatically examine a nation's system for step by step and level by level measuring leaming outcomes and for effectively using * Support the kind of research that could contribute the results. When the Bank finds deficiences in how to better educational management; this research information on learning outcomes is obtained or used, it should clarify issues, define objectives, and evaluate should then consider how these deficiencies might be experiences. remedied. As with other inputs to school quality, the * Make better analyses of the prerequisites to sound Bank should be ready to make the necessary invest- education management. ments. * Invest more effort in participatory management, although the area is complex Subsidiary Recommendations * Conduct comparisons of administrative systems in its normal sector work. The first priority is to make the examination as ef- fective as possible. This means orienting the systems toward measuring the basic and complex intellectual Examination and Selection Policies skills students need to understand and control their environment. It means orienting them toward widely Session chairman: Ralph Harbison agreed and practical educational objectives. Even where Presenters: Anthony Somerset and examination systems already exist, there is often much Benjamin Makau that could be improved. Lessons from Bank Experience 47 An examination system that is not somewhat auton- grades (four through six), while English-speaking and omous from direct control by govemment should be Latin American countries had the highest rate of re- assessed with special skepticism. Independence from peaters in grades one through three. political pressure and protection from educational bureaucracies is a prerequisite to the successful use of A t E examinations. ~ Analyzing the Effects of Repetition examinations. Examination systems should have a research unit: a The effect of repetition on the internal efficiency of nucleus of highly trained professionals, who analyze the school systems is normally analyzed by student flow wealth of data generated by examination systems and models. The results derived from such models are propose ways to improve both the quality and manage- affected by two types of errors. The first results from the ment of education. Often the first step toward convert- low quality of the education statistics available in many ing exams into true inputs of school quality is to graft countries: overreporting of first grade enrollments, in- this research capacity onto an examination unit whose correct distinction between new entrants and repeaters, principal function is selection and certification. and yearly variation in the coverage of data. Examinations bodies, like other inputs of school qual- the other type of error results from the assumption ity, must enjoy public acceptance. This suggests that that pupils behave homogeneously: that pupils enrolled they should have a reasonable capacity for public rela- in a given grade in a given school year have the same tions and public education so that the uses of examina- propensity to be promoted, to repeat the same grade, or tions can be fully explained and the results made avail- to drop out, regardless of their academic history. This able for public scrutiny. assumption is required because of the lack of statistics It takes a long time to develop examinations, espe- on how previous academic history affects present be- cially when starting from scratch. This is even truer havior. But working against this assumption is the fact when the prominent objectives are to improve the qual- that many developing countries have limits as to how ity and management of schools. Thus, support for ex- many times a pupil may repeat. aminations development should be planned for a decade The five most common types of limits on repetitions or more. are shown in Table 5-1, and models for analyzing the effects of these limits are developed in Fredriksen (1983). Promotion and Repetition Policies Session chairman: Aurelio Cespedes Presenter: Birger Fredriksen Table 5-1. Restrictions on the Repetition of Grades Participants: Gunnar Berlin, Hon-Chan in Ninety-Two Countries Chai, Richard Durstine, Barbara Searle, and Eric Swanson Number of countries Primary Secondary The problem of students repeating grades and the Restriction education education accompanying wastage is critical in most developing countries. A Unesco survey (1980) noted that 15 percent Promotion is automatic, that is, rep- of the students enrolled in primary education in de- grade. 10 7 veloping countries were repeaters. This figure was 12 Certain grades may be repeated but percent in Latin America, 16 percent in Africa, and 17 not others. 12 4 percent in Asia. All grades may be repeated a limited The level of repetition varies greatly among different number of times. 21 35 cultural settings. For example in the twenty French- Tnumber of titio. pemt speaking countries included in the survey the rate of during a given cycle is restricted, repetition ranged between 16 and 40 percent. Arabic- but there are no further restric- speaking countries had a similar range in repetition tions on the number of repetitions rates: Sudan had the lowest figure, 3 percent, which permitted in any given grade. 7 9 reflected automatic promotion, while Morocco had 28 The number of repetitions permitted percent, which indicated repetition practices. English- in any given grade, as well as dur- speaking countries, meanwhile, had a 10 percent repeti- ing the cycle, is restricted. 12 13 tion rate and Spanish-speaking countries had a rate of 10 Total 62 68 to 11 percent. French-speaking countries had the high- est percentage of repeaters in the upper primary school Source: Unesco (1980), Tables 7.1 and 7.2. 48 Implementing Investments in Quality Repetition and Wastage considered a remedy for inadequate academic achieve- Repetition is wasteful. If 15 percent of the students ment. repeat at least one grade, then the educational systems The arguments against repetition, and thus in favor of would need to spend seven years to produce a sixth grade automatic promotion, have been discussed in the litera- graduate. The efficiency of the rural education systems ture (Brown 1966, Jackson 1975, Goodlad 1954). of certain countries is so low that it requires twelve to * Student characteristics (age, degree of motivation, twenty student years to produce a primary school gradu- and state of nutrition and health) and supply factors ate, which is two or three times a standard primary (teachers, leaming materials, and school facilities) school cycle of six years. should be considered in school decisions concern- There are economic and educational implications of ing promotion. this economic waste caused by the repetition and pro- * In most developing countries there are no guide- motion practices in developing countries. Repetition lines or examination standards; hence, the validity limits the number of students that a given system can of tests given by classroom teachers is questionable. take in, increases the cost of education, and causes * Extemal school factors, including the family back- resources to be distributed inequitably. ground, are important variables influencing repeti- tion. Repetition can result in the fact that families Educational Effects are unable to properly feed their children. There is evidence that elementary school repetition is Structural problems may also cause repetition such associated with the students' sense of security and their as: poor quality teachers, too few textbooks and other assoiatd wih te sudens'snse f scurty ad teir relevant learning materials, inadequate guidelines for early experiences of success and failure (Intemational assessing student learning, too few national examina- Bureau of Education 1971), causes both achievement tion systems, and too few well-trained education man- and intelligence scores to deteriorate in the long term s and tor at (Goodlad 1954), and increases repetition at the second- agers and school directors. ary level (Camoy and Thias 1974). The Unesco survey shows that of the ninety-two coun- General Recommendations tries that provided information on primary education, sixty-two reported having regulations on promotion and The participants agreed with the general conclusions some kind of restrictions on grade repetition. of Haddad (1979) that although promotion seems to be There are various assumptions and arguments for and more desirable than retention or repetition, the real against the practice of repetition in developing coun- issue is to prevent failure. The main assumption is that tries. Countries that favor repetition do so based on the Bank needs to work both with borrowers that prac- three assumptions. tice repetition and with those that have automatic pro- motion. The Bank should develop the following remedial • Most pupils can learn the curriculum assigned to actions in close coordination with developing countries. each grade in the prescribed time. In Latin America, * Help borrowers develop adequate guidelines for col- for instance, Colombia, Guatemala, and Peru expect lecting, processing, and disseminating relevant sta- that first graders should leam how to read and write tistics on repetition and promotion so that the (de corrido) during the first year of school. Factors dimension of the problem can be better known by all that determine academic achievement, such as the p rned nature of the students, their socioeconomic back- parties concerned. ' * ~~~~~~~Encourage some developing countries to review grounds, and most important, the level of classroom Eirobective fopimr eductios th ex- quait to whc thyaeepsd.eentcn their objectives for primary education so that ex- sidered. pectations may be adjusted to the prevailing local sidhe best red. dy for pupils who fail to reach the conditions, curricula, school calendar, time sched- * The best remedy for pupils who fail to reach the ules, and other relevant parameters. minimum standard set for a given grade is to spend encuage oroer t documetea r t more time in that grade. ~ ~ Encourage borrowers to document teachers' atti- m Itime possinle thatughst g tudes toward slow learners, to develop in-service * It is possible, through school-designed tests and training for teachers in the use of student achieve- teachers' judgment, to single out students who have ment variables and student evaluation systems, and acquired the minimum attainment necessary to to provide special summer courses for slow learners progress to the next grade. and for students who would otherwise repeat a given Some countries, such as Singapore, Tunisia, and grade. Dominican Republic, which have promoted students * Document and research the effect of factors related automatically in the past have recently adopted policies to family background, particularly socioeconomic on repetition. The practice of repetition in these cases is conditions. Poorly nourished children, who are the Lessons from Bank Experience main candidates for grade repetition in developing have any effect on student achievement. Beyond this countries, ought to constitute a significant variable threshold there appear to be diminishing returns on the in any investigation concerning school repetition. degree to which teacher training can be expected to be Priority for examining and correcting these condi- effective. This threshold varies from country to country tions should be given to countries in which repeti- and also changes with time; what is considered to be a tion practices are based solely upon results of minimum educational requirement for primary school achievement tests. teachers will probably be too low ten years from now. * Study one or two countries that had automatic This process is a normal part of economic and social promotion in the past and have recently adopted a change and should not be treated as unusual. When policy on repetition. Among those which might be there are not enough teachers to meet enrollment de- considered are Singapore, Tunisia, and the Domin- mands, however, village adults who meet the educa- ican Republic. tional threshold as teachers should be given a chance to * Encourage the regular exchange of ideas among teach, as was done in the Pakistan and Bangladesh Pri- education staff in research and operations to facili- map' Projects. tate follow-up on suggestions and ideas, such as those presented in this section, which might be Policy Recommendations important to both the Bank and its borrowers. In-service education should be tied to training a spe- cific set of skills or a specific reform in education, such References as implementing a new curriculum or introducing new textbooks. For reasons of cost in-service training should Brown, R. I. 1966. A Survey of Wastage Problems in not be offered for its own sake or inisolation from other Elementary Education. Bangkok: Unesco Regional Office education inputs. R atin ispection sytm or for Education in Asia. education mputs Recertification, ispection systems, or Carnoy, M., and H. H. Thais. 1974. "Draft Report of Second enhanced supervision should be considered a necessary Tunisia Education Research Project." World Banko part of in-service education. Washington, D.C. Teacher salaries should be on the same scale as those Fredriksen, Birger. 1983. "Internal Efficiency of School Sys- of other civil servants, but teachers of specialized sub- tems: A Study in the Use of Pupil Flow Models for Developing jects should be paid a supplemental allowance according Countries." Ph.D. thesis, University of Lancaster. Processed. to supply and demand. For example in Jordan science Goodlad, J. 1. 1954. "Some Effects of Promotion and Non- teachers receive a rarity allowance in addition to the promotion upon the Social and Personal Adjustment of basic pay, which gives them higher salaries than history Children." Journal of Experimental Education, vol. 22, pp. teachers. 301-28. Formation of teacher resource centers where teachers Haddad, Wadi D. 1979. Educational and Economic Effect of come to share new ideas and receive some form of Promotion and Repetition Practices. World Bank Staff in-service education on a voluntary basis should be en- Working Paper no. 319. Washington, D.C.: World Bank. couraged. In Malaysia, for example, four Education Re- International Bureau of Education. 1971. Wastage in Educa- source Centers have been established and have been very tion: A World Problem. Paris: Unesco. effective. Jackson, C. B. 1975. "The Research Evidence on the Effects of Grade Retention." Review of Educational Research, vol. 45, Areas for More Research pp. 613-35. Unesco. 1980. Wastage in Primary and General Secondary The following questions are pertinent for further re- Education: A Statistical Study of Trends and Patterns search. in Repetition and Dropouts. Report no. CSR-E-37. Paris: * Which teaching styles have the greatest effect on Unesco, Office of Statistics. achievement? In what particular cultural settings are they effective? Teacher Training * Should there be merit increases for teachers in addition to seniority raises? What effect would merit Session chairman: Frank Farner increases have on teacher performance and subse- Presenter: Mae Chu Chang quently on student achievement? Participants: Victor Billeh, Shigenari * In each country what is the minimum level of basic Futagami, Himelda Martinez, Andre education required for primary teachers in light of Salmon, and Mulugeta Wodajo the demand for particular levels of complexity in curriculum content? There appears to be a minimum level of basic educa- * What is the relative effectiveness of a fully trained tion needed if training for primary school teachers is to urban teacher sent to the countryside versus an 50 Implementing Investments in Quality assistant teacher who is a local rural person with * Develop a simple set of cross-national indicators of less training? time, such as hours in the school day and days in the * Is a local rural teacher likely to be more motivated school year, that could be made available regularly than a teacher from an urban area or a different to project officers and others rural area? * Do more research on the use of time in the class- * What is the relation between age or sex of the room in various developing countries teacher and student achievement in a particular * Review the best way to introduce and sequence cultural setting? subjects within the curriculum, minimum learning * What are the criteria for determining whether the thresholds, and the problems associated with insti- salaries of teachers in a country are too high or too tuting greater selectivity in the curriculum. low as a percentage of the government budget, GNP, * Study the costs associated with manipulating school or the demographic profile? Is there another, more time. logical comparison? 4 Prevocational Subjects Time on Task Session chairman: Jacob Van Lutsenburg Maas Session chairman: Sherry Keith Presenter: William Loxley Presenter: Henry Levin Participants: Mae Chu Chang, Richard Johanson, Participants: Hon-Chan Chai, Richard Himelda Martinez, George Psacharopoulos, Durstine, Ralph Harbison, Mats Hultin, Andre Salmon, Lennart Swahn, Mulugeta and Abdun Noor Wodajo, and Abdelwahed Zhiri Henry Levin's presentation helped to identify the con- The topic of prevocational subjects has been con- ceptual problems associated with the use of time as troversial, in part because of confusion over definitions a factor affecting educational quality outcomes. He and of what may be called the "misimplementation" of stressed the intuitive appeal of time-on-task because of these ideas in Bank-financed projects. The sources of its notion that more time spent learning would lead to this confusion need to be confronted before any recom- higher achievement. His review of the existing research mendations for either research or for policy can be literature did not fully confirm this view, however. The made. knowledge to date on the concept of time suggests that total time spent in school may not be a useful indicator because a variable proportion of that time is "nonin- Conceptual Problems structional," that is, devoted to lunch recess, interrup- Prevocational subjects in secondary education have tions, and so forth and that students may or may not be been seen as part of a process of "curriculum diversifica- actually learning, even during the instructional time. tion," which would give a more "practical" emphasis to Engaged time has been studied through classroom teaching. On reflection, however, it is becoming clearer observational methods, mostly in North America. The that these key ideas or terms have actually meant dif- results of such studies suggest that students use many ferent things to different audiences in various countries techniques to appear as though they are learning during and regions. Sometimes they have meant different instruction, although they may actually not be. things within countries and even among Bank staff. There is a conceptual model for analyzing the relation One way to grasp these differences is to conceive of a of time and learning and the interaction of time with curricular spectrum rather than trying to arrive at the other critical factors that affect the learning process. "correct" definition. This spectrum is shown with some This model can be summarized as: related variables in Table 5-2. At one extreme is voca- [Student capacity] x [Student effort] tional training; at the other is general education. Arrayed in between the extremes are Models I and 11 and x [Time] x [Quality of Resources] a third model. They are arranged in order of their varying intensities for allocating specialized teaching Recommendations time and physical resources to the task of "practical teaching." Models I and 11 are borrowed directly from There is very little systematic comparative informa- the curriculum typology used in the DISCUS (Diversifica- tion regarding time and schooling in developing coun- tion of Secondary Curriculum Study). Model I provides tries. Thus the Bank should prevocational subjects in a general education curricu- Lessons from Bank Experience 51 Table 5-2. Models of Curriculum Diversification for Prevocational Subjects General "Missing Vocational education model" Model I Model ! training No subjects or Prevocational sub- Prevocational tracks are added jects are added tracks are to the curricula. to the general offered in com- Practical knowl- education curri- prehensive high edge and skills cula of high schools. are taught and schools. applied through About 50 percent general subjects, About 25 percent of class time is in science and of class time is given to pre- other core sub- given to pre- vocational sub- jects. vocational sub4 jects. Content is jects. almost "semi- vocational." This model has Example: Example: not been Tanzania Colombia adopted in Bank-financed projects. lum. Model II provides prevocational tracks, which are tendencies may have artificially pushed the curricula much more time-consuming and amount to a semi- sponsored through World Bank projects toward the vo- vocational training conducted in a comprehensive cational end of the spectrum. They may also have en- school. There should also be a third model, a so- couraged the procurement of heavier and more ex- called "missing model," which was not identified in pensive teaching equipment than was appropriate. DISCUS. It is important to evaluate the cost-effectiveness of Curriculum reform under the "missing model" would each model represented in DISCUS, which implies having not require extra subjects to be added to the school access to very thorough cost data. Student attitudes as timetable as does Model I. Instead it would require most well as cognitive and labor market performance should core subjects to be taught with a practical or applied be measured. The following questions should be ad- emphasis. This would be especially true in science but dressed in the data analysis. would also apply to the teaching of mathematics, lan- * Which model of prevocational subjects is most cost guage, and social studies. efWhich and why? The goal of giving a practical bias to the core subjects effective, and why? would be to learn application skills, in contrast to iso- * Does time spent in a prevocational track (Model II) lated bits of factual knowledge abstracted from the "real or in a prevocational subject (Model I) reduce world." It would bring relatively common technology chances for university entry? The underlying into the classroom and laboratory, yet it would not mean assumption here is that instructional time is a the procurement of heavy specialized machinery (which scarce resource, and therefore there is a tradeoff in too many instances has typified the Model II and even between university access and prevocational educa- Model I curriculum reforms). tion. In other words does prevocational education reduce academic subject achievement or standards? Or, does prevocational education tend to reduce the Misimplementation upward social mobility of students from low-income The "missing" model is missing because it has backgrounds? Since it is sometimes argued that apparently never been consciously adopted and sup- prevocational education should discourage stu- ported in a project financed by the Bank. It is not clear dents from continuing into university, the DISCUS whether this has been because of Bank-specific tenden- study should determine whether prevocational cies such as the need to find procurable items to finance graduates have a greater rate of failure in the uni- (the more expensive the item, the less time-consuming versity than graduates who had more traditional the procurement might be, and the faster the disburse- curricula. ment) or the urge to show a direct "manpower justifi- * Where should prevocational education occur-in cation" even for general secondary education. These lower secondary schools, in upper secondary 52 Implementing Investments in Quality schools, or in specialized programs outside of * Textbooks are part of a broader system of publishing school? and knowledge dissemination in all countries, and What is the best age for prevocational education? that system must be considered when designing textbook projects. A General Proposition * Textbook programs must be integrated into curric- ulum projects and other aspects of educational de- There are general conditions or criteria under which velopment. Model I and Model 1I prevocational education might be * A multi-stage approach to textbook development most successfully implemented. As a general rule of seems to be most effective. That is, it will take about thumb these models of diversification would be more a decade to develop textbooks, and at least two feasible where there is a high rate of enrollment in stages are involved. In the first stage books are secondary schools. No actual threshold enrollment rate provided promptly, even if they are not dramatically could be established, but it is now clear that where the different in content from those already in use. In the enrollment rate is low, secondary school students will second stage new textbooks with new approaches not be motivated to narrow their career horizons to any are integrated in the curricula, and other desired particular prevocational subject or track while they are changes are made. still studying. Instead they will tend to devote their * Books that related to school curricula but are not efforts to core subjects that will keep their options open necessarily textbooks should be integrated into text- once they embark seriously on their job search. book projects. The development of school libraries, The criterion of affordability is another test of of reading materials for newly literate students and whether public authorities should establish prevoca- of supplementary books should all be a part of a tional programs. It has been assumed that prevocational comprehensive textbook project. education using Models I and II would be more costly than general education. One attraction of the "missing model" is precisely its lower investment and operating The Publishing System costs. World Bank projects often have a substantial effect on a national book industry: they have the potential to contribute significantly to the development of that in- dustry or to retard its growth. Because of the sheer The Production and Distribution of Textbooks number of books published, textbooks have traditionally Session chairman: Laurence Wolff been an important part of a private sector publishing Presenter: Philip Altbach industry and can contribute to the development of eco- Participants: Albert Aime, Frank Farner, nomic health and infrastructure. Douglas KearEicwaMultinational publishers, however, have had a key Douglas Keare, Eric Swanson, role in textbook development in many developing na- and Hans Thias tions, particularly in Africa, and this role has sometimes been controversial. On the one hand multinational pub- Textbooks are a key element of many Bank education lishers can participate constructively in textbook de- programs and have a marked effect on educational qual- velopment and may even be essential in countries in ity in developing nations. This assumption is supported which a significant publishing industry does not exist. by the available literature and by the Bank's experience. On the other hand the multinationals have their own Thus the provision of textbooks should remain a key economic interests, and these may not always work to priority. the advantage of the developing countries. In consider- Although the Bank has built up considerable expertise ing any textbook project, the Bank must consider the in designing and implementing textbook programs, effect of the project not only on the educational system there is little available guidance concerning the general but also on the development and infrastructure of the guidelines on developing textbook programs. Thus each national publishing industry. project requires careful attention to the specific prob- lems and issues in the borrowing nation. Nevertheless several general points can be made concerning textbook Integrating Textbook Development development projects and the role of textbooks in de- with Other Innovations veloping societies: Textbook projects have often been integrated into * A pragmatic approach to textbook development is curricular developments and educational innovations. It required precisely because there are few general is necessary in all cases to ensure that this integration is rules that can be applied. a key part of the design and implementation of a text- Lessons from Bank Expenence 53 book project, even if this means that the ultimate goals Textbooks, Libraries, and Literacy of the project take longer than normal to achieve results. Textbooks are but one part of a broadly conceived In addition the development of a textbook must be learing and literacy program for developing countries. accompanied by the creation of appropriate teachers' Wherever appropriate, textbooks should be accompa- workbooks and especially in-service modules so that the nied by supplementary reading materials for the schools. text can be quickly integrated into the educational pro- School libraries should be encouraged, and other books gram of the school. covering the same subjects as the text should normally Higher-level staff must be trained in publishing, edit- be part of the text project. In addition attention should ing, and textbook writing. This may involve regional be given to the broader development of library resources training centers or secondment to established pub- to maintain literacy and an interest in learing in the lishers. Textbook storage facilities and minilibraries are society. also essential. Funds and appropriate expertise must be society. built into any textbook project to ensure that this inte- gration can occur. In smaller countries with limited Financing Textbooks expertise, however, it is important to limit curricular 4 and textbook innovation to that which can be realisti- Textbooks can be financed in many ways; no single cally accomplished and to adopt existingmatemeans for textbook finance is preferred aside from a ever appropriate. general commitment to provide free textbooks in the long run. There is a concern, however, about the fund- ing of textbooks once Bank assistance ends. In some instances it may be appropriate for students to purchase A Multi-Stage Approach books or for special education taxes to be imposed. But C. E. Beeby's presentation (chapter 4) and the experi- costly covers, illustrations, and other items should be ence of the Bank's textbook projects have shown that it kept to the minimum necessary to meet educational requires a rather long lead time to develop textbooks, goals. The overriding concern is to ensure that text- and this must be incorporated into project design. In books remain a priority of the educational authorities most instances a textbook project should start out with and that funds be maintained to replenish textbooks the development and production of a book that reflects once Bank-assisted texts are no longer in use. established curricular pattems. Such a book should not require significant innovations from teachers and A Research Agenda should be placed in the hands of students relatively quickly to satisfy immediate demands. The greatest Although the data show that textbooks have a positive thrust should be placed on creating innovative textbooks effect on educational quality and results, additional re- over a long period. Thus books can be adequately tested search should be done in this area. Wider research re- as part of the development process, ancillary materials suits are needed concerning the nature of textbook con- can be developed and coordinated, and curriculum au- tributions to educational quality. In particular how thorities can be involved in text development. By the many textbooks will have the maximum effect, and time these revisions are ready to be published, the first where do diminishing-returns enter? Do texts for one editions will have worn out and will need to be replaced level of the educational system have a greater effect than anyway. Thus Bank lending policies might well be ad- for others? justed for a longer-term effort, or it may be necessary for The Bank now has a wealth of practical experience in two loan programs to be involved in each text project. textbook programs. It would be very helpful to have appropriate case studies on these experiences. The de- scription and analysis of some typical projects would Participation help to provide guidelines for future involvement. Re- search on the experience of textbook programs im- The development of quality textbooks requires advice plemented by other aid agencies or by such indigenous and input from a variety of sources, so appropriate par- organizations as the National Book Trust in India or the ticipation should be built into textbook projects. The Dewan Bahasa dan Pustaka in Malaysia would also be participation of curricular experts has already been men- helpful. tioned, as has the involvement of the local publishing industry. Teacher representatives, local supervisors, and appropriate ministry officials should also be involved. Those who will design, produce, and use the books must also participate in the process of creating and producing them. 54 Implementing Investments in Quality The Use of Electronic Media sizes. Thus the goal of cost reductions has not been achieved. In fact since the production of software Session chairman: Barbara Searle has turned out to be expensive, programs using Co-chairman: Dean Jamison media are almost always more expensive than com- Participants: Victor Billeh, Gunnar Berlin, parable programs without media. Some media ex- Aurelio Cespedes, Shigenari Futagami, periments have improved achievement levels, how- Delia Pitts ever, rather than simply substituting for teachers. The question of whether the improvement in quality "Media" is a shorthand label for "the programs that is worth the cost is, of course, a value judgment that are transmitted by the media." This point is important must be made in each case. because radio and television only deliver the programs. * The experience developed in one country can be The effectiveness of a mediated program (one delivered transferred to other countries. The "Radio Math" by an electronic medium) depends far more on the lessons developed in Nicaragua have been translated quality and appropriateness of the "software" than on into Thai and are teaching Thai children as effec- the characteristics of the medium itself. Nevertht- tively as they taught Nicaraguan children. less the characteristics of the medium may limit or en- Although there has been much experimentation, few hance the effectiveness of the program in reaching its countries have adopted in-school direct teaching nation- objectives. Thus the use of media (particularly radio) wide, although Thailand is considering this with the can be examined in several instructional contexts. math lessons. Direct Teaching in Schools Direct Teaching Out of School One or two decades ago there were high hopes that Many experiments have used lessons delivered by lessons transmitted to students in schools by radio or radio or television in conjunction with a correspondence television represented a means to increase access to course. Often such courses use community residents or schools while maintaining quality at no extra cost. This underqualified teachers as monitors to lead tutoring would be accomplished because teachers would be able sessions. Several of these have led to permanently estab- to teach larger groups of students (since the media lished programs, such as the Malawi Correspondence would transmit much of the instruction) or less- College. Although the medium does not carry the full qualified teachers, usually monitors without formal burden of instruction in these cases, it paces learning qualifications, would be able to run classrooms in which and can explain difficult concepts. Such systems are students were taught by the media. During the 1970s particularly likely to become established where ex- many experiments were conducted to test this concept. aminations can be taken outside the classroom. These evaluations, which were well conceived and ade- quately funded, produced some valuable conclusions. In-Service Teacher Training * Instructional programs delivered through media can teach children. Some subjects are taught better The use of radio for in-service training of teachers has than others, and radio and television impose differ- several advantages over alternative methods, particu- ent constraints on the types of instruction that can larly for teachers in rural areas. In areas in which teach- be transmitted, but children do learn. ers are widely dispersed, it substantially reduces trans- * It is far more difficult to develop effective instruc- port costs. The programs can reach teachers in their tional materials than anyone had anticipated. It is homes and schools, and they can provide support and time-consuming and expensive because it involves training for long periods. Radio programs, in contrast to analysis of the curricula, careful specification of printed material, can model teaching behavior for larnasin ofutcomes, curricu catvefl specation, a the teachers. Finally radio programs can keep teachers up to learning outcomes, formative evaluation, and te date about new developments and help build their secu- development of instructional strategies particularly rity abourcew inform ents . Rad les- suited to the medium. In some countries in which rity as sources of information for students. Radio les- suithed tohel medium.aIn sophisticatiountisrhigh, prsons must be followed up with personal contact with the level of media sophistication is high, production supervisors and occasional face-to-face teaching ses- quality must also be high. sions, however, so that teachers and programmers alike * The (apparently reasonable) expectation that money can get feedback. could be saved on teacher salaries has not been met. In projects that have tried to use higher student- teacher ratios or underqualified teachers, the in- Buldding Support for Educational Change structors have forced the schools, through strikes Radio and television programs aimed at the popula- and other means, to raise salaries or lower class tion at large are a powerful tool for building support for Lessons from Bank Experience 55 educational change. Programs can provide parents with * The effectiveness of correspondence programs can reliable information about what schools are expected to often be improved by adding lessons delivered by do and what their children are expected to learn. This in radio or television. turn can create expectations to which schools must * Radio has not been adequately explored as a way respond. Programs can also be used to build support for to provide on-going in-service instruction to the policies and goals of the educational system. Author- teachers. ities can pave the way for important reforms by explain- * The potential for using media to build community ing the changes in advance. support for educational reforms and for educating parents about what they should expect from schools has scarcely been exploited. Observations Recommendations Media-based instructional programs have a compar- ative advantage in several special circumstances, ;he World Bank should give more emphasis through including reaching very geographically dispersed its lending program to the use of media, particularly populations (for example, the Universities of the radio, for educational purposes where analysis of the West Indies and the South Pacific); reaching very context suggests it has a comparative advantage. In large populations, where the cost of developing addition it should explore ways to increase cross- programs is small for each student (for example, the fertilization among countries so that educators and TV University of China); and teaching specialized planners can benefit from the experiences already gained subjects for which teachers are inadequately pre- in other countries by exchanging methods, ideas, and pared (particularly second or foreign languages). even programs. Bibliography of World Bank Research on the Quality of Education in Developing Countries T he word "processed" indicates works that are re- in Primary Schools." Comparative Education Review, vol. produced by mimeograph, xerography, or similar 20 (February), pp. 79-93. means; such works may not be cataloged or commonly Simmons, John, and Leigh Alexander. 1978. "The Determi- available through libraries, or may be subject to re- nants of School Achievement in Developing Countries: A stricted circulation. Review of Research." Economic Development and Cultural Change, vol. 26 (January), pp. 341-58. General Estimates of School Quality and Gains in Achievement Analyses of Specific School Qualities and Specific Countries Birdsall, Nancy. 1982. Child Schooling and the Measurement of Schooling Standards. Living Standards Measurement Arriagada, Ana-Maria. 1981. "Determinants of Sixth Grade Study (LSMS) no. 14. Washington, D.C.: World Bank. Student Achievement in Colombia." Education and Train- Hartley, Michael J., and Eric V. Swanson. 1984. "The Retention ing Department (July). Processed. of Literacy." Development Research Department. Pro- . 1981. "Determinants of Sixth Grade Student Achieve- cessed. ment in Peru." Education and Training Department (July). Heyneman, Stephen P. 1976. Comparative Education Review, Processed. vol. 20 (February), pp. 42-47. Cochrane, Susan H., and Dean T. Jamison. 1982. "Educational . 1976. "Influences on Academic Achievement: A Attainment and Achievement in Rural Thailand." In New Comparison of Results from Uganda and More Industrial- Directions for Testing and Measurement: Productivity ized Societies." Sociology of Education, vol. 49 (July), Assessment in Education, edited by A. Summers. San Fran- pp. 200-11. cisco, Calif.: Jossey-Bass. .1980. "Differences between Developed and Developing Feliciano, Gloria, and others. 1981. The Educational Use of Countries: Comment on Simmons and Alexander's 'Deter- Mass Media. World Bank Staff Working Paper no. 491. minants of School Achievement'." Economic Development Haddad, Wadi D. 1978. Educational Effects of Class Size. and Cultural Change, vol. 28 (January), pp. 403-6. World Bank Staff Working Paper no. 280. Heyneman, Stephen P., and William A. Loxley. 1982. "In- Heyneman, Stephen P. 1979.Investment in Indian Education: fluences on Academic Achievement across High and Low Uneconomic? World Bank Staff Working Paper no. 327. Income Countries: A Re-Analysis of IEA Data." Sociology of Republished in World Development, vol. 8 (1980), pp. 145- Education, vol. 55 (January), pp. 13-21. 63. - 1983. "The Effect of Primary-School Quality on . 1979. 'Why Impoverished Children Do Well in Ugan- Academic Achievement across Twenty-Nine High- and Low- dan Schools." Comparative Education, vol. 15 (June), Income Countries."American Joumal ofSociology, vol. 88 pp. 175-85. (May), pp. 1162-94. World Bank Reprint Series no. 268. . 1980. The Evaluation of Human Capital in Malawi. Simmons, John. 1976. "Retention of Cognitive Skills Acquired World Bank Staff Working Paper no. 420. 57 58 Bibliography Heyneman, Stephen P. 1983. "Education during a Period of 1982. "Mass Media, Basic Education and Agricultural Austerity:Ugandal971-1981."ComparativeEducationRe- Extension." Population and Human Resources Discussion view, vol. 27 (October), pp. 403-13. Paper no. 82-01. Population, Health, and Nutrition Depart- . Forthcoming. "Diversifying Secondary School Cur- ment, World Bank. Processed. ricula in Developing Countries: An Implementation History Perraton, Hilary, and others. 1983. Basic Education andAgri- and Some Policy Options." International Journal of cultural Extension: Costs, Effects, andAlternatives. World Education and Development. Bank Staff Working Paper no. 564. Heyneman, Stephen P., and Dean T. Jamison. "Student Learn- Perrett, Heli E. 1982. Using Communication Support in Proj- ing in Uganda: Textbook Availability and Other Factors." ects. World Bank Staff Working Paper no. 551. Comparative Education Review, vol. 24 (June), pp. 206-20. Psacharopoulos, George. 1985. "Evaluation of Curriculum Di- Heyneman, Stephen P., and others. 1978. Textbooks and versification in Colombia and Tanzania." Comparative Achievement: What We Know. World Bank Staff Working Education Review, vol. 29 (November), pp. 507-25. Paper no. 298. Available in Spanish and French; republished Psacharopoulos, George, and William Loxley. Forthcoming. in The Journal of Curriculum Studies, vol. 13 (1981), Diversified Secondary Education and Development. Balti- pp. 227-46. 4 more, Md.: Johns Hopkins University Press. . 1984. "Textbooks in the Philippines: Evaluation of the Psacharopoulos, George, and A. Zabalza. 1984. "Effect of Di- Pedagogical Impact of a Nationwide Investment." Edu- versified Schools on Employment Status and Earnings in cational Evaluation and Policy Analysis, vol. 6, no. 2, Colombia." Economics of Education Review, vol. 3, pp. pp. 139-50. 315-31. Husen, Torsten, and others. 1978. Teacher Training and Stu- Smilansky, M. 1979. Priorities in Education: Pre-School; Evi- dentAchievement in Less Developed Countries. World Bank dence and Conclusions. World Bank Staff Working Paper Staff Working Paper no. 310. no. 323. Jamison, Dean T., and E. McAnany. 1978. Radio for Education Spain, Peter, and others. 1977. Radio for Education and De- and Development. Beverly Hills, Calif.: Sage Publications. velopment: Case Studies (2 volumes). World Bank Staff Also published in Spanish. Working Paper no. 266. . 1982. "Reduced Class Size and Other Alternatives for Suppes, Patrick, and others. 1978. The Radio Mathematics Improving Schools: An Economist's View." In School Class Project: Nicaragua 1976-1977. Stanford, Calif. Stanford Size: Research and Policy, edited by C. Glass and others. University. (Several chapters by Bank staff.) Beverly Hills, Calif.: Sage Publications. Jamison, Dean T., and others. 1981. "Improving Elementary Mathematics Education in Nicaragua: An Experimental The Economic Effects of School Quality Study of the Impact of Textbooks and Radio on Achieve- ment." Journal of Educational Psychology, vol. 73, no. 4 Behrman, Jere, and Nancy Birdsall. 1983. "The Quality of (August), pp. 556-67. Schooling in Developing Countries: Quality Alone Is Mis- Lillis, K., and D. Hogan. 1983. "Dilemmas of Diversification: leading." American Economic Reviewv, vol. 73, no. 5 (De- Problems Associated with Vocational Education in Develop- cember), pp. 928-46. ing Countries." Comparative Education, vol. 19, pp. 89- Biossiere, M., and others. 1982. "Earnings: Schooling, Ability 107. and Cognitive Skills." Development Research Department. Loxley, William A. 1983. "The Impact of Primary School Qual- Processed. ity on Learning in Egypt." International Journal of Educa- Birdsall, Nancy. 1983. "The Demand for Primary Schooling in tional Development, vol. 3, no. 1, pp. 33-45. Rural Mali: Should User Fees Be Increased?" Country Policy Mahar, Dennis J. 1982. "The Teaching-Learning Role of Media Department. Processed.. Centers/Libraries." Education and Training Department. - . 1985. "Public Inputs and Child Schooling in Brazil." Processed. Journal of Development Economics, vol. 18 (May/June Metcalf, David. 1985. Economics of Vocational Training: Past 1985), pp. 67-86. Evidence and Future Evaluations. World Bank Staff Work- Heyneman, Stephen P. 1984. "Educational Investment and ing Paper no. 713. Economic Productivity: The Evidence from Malawi." Inter- Neumann, Peter H. 1980. Publishing for Schools: Textbooks national Journal of Educational Development, vol. 4 and the Less.Developed Countries. World Bank Staff Work- (February), pp. 9-15. ing Paper no. 398. Available also in Spanish. . 1984. "Research on Education in Developing Coun- Neumann, Peter H., and Maureen A. Cunningham. 1982. Mex- tries." International Journal of Educational Development, ico's Free Textbooks: Nationalism and the Urgency to Edu- vol. 4 (December), pp. 293-304. cate. World Bank Staff Working Paper no. 541. Jamison, Dean T. Forthcoming. "Child Malnutrition and Perraton, Hilary, ed. 1982. Alternative Routes to Formal School Retardation in China."JournalofDevelopmentEco- Education: Distance Teaching for School Equivalency. nomics. Baltimore: Johns Hopkins University Press. Jamison, Dean T., and Peter R. Moock. 1985. "Farmer Educa- Bibliography 59 tion and Farm Efficiency in Nepal: The Role of Schooling, . 1982. "Resource Availability, Equality, and Educa- Extension Services, and Cognitive Skills." World Develop- tional Opportunity among Nations." In Education and ment, vol. 12 (January), pp. 67-86. Development; Issues in the Planning and Analysis Moock, Peter R., and J. Leslie. Forthcoming. "Childhood Mal- of Post-Colonial Societies, edited by Lascelles Anderson nutrition and Schooling in the Terai Region of Nepal." and Douglas M. Windham. Lexington, Mass.: Lexington Journal of Development Economics. Books. Sabot, Richard. 1983. "The Quantity and Quality of Education . 1983. "Improving the Quality of Education in De- and Labor Productivity: An East African Comparison." De- veloping Countries." Finance and Development (March), velopment Research Department. Processed. vol. 20, pp. 18-21. Reprinted in Education and Develop- Sabot, Richard, and others. 1981. "Cognitive Skills: Their ment: Views from the World Bank by Aklilu Habte, George Determinants and Influence on Earnings in Two Poor Urban Psacharopoulos, and Stephen P. Heyneman. Washington, Economies." Population and Human Resources Discussion D.C.: World Bank, 1983. Paper no. 4. Population, Health, and Nutrition Department. . Forthcoming. "Diversifying Secondary School Cur- Processed. ricula in Developing Countries: An Implementation History Schiefelbein, Ernesto, and others. 1983. The Influence of and Some Policy Options." InternationalJournalofEduca- School Resources in Chile: Their Effect on Educational tion and Development. Achievement and Occupational Attainment. World Bank . Forthcoming. "The Nature of a 'Practical' Curricu- Staff Working Paper no. 530. Portion republished as "Edu- lum." Education with Production. cation and Occupational Attainment in Chile: The Effects Heyneman, Stephen P., and William A. Loxley. 1983. "The of Educational Quality, Attainment, and Achievement." Distribution of Primary School Quality within High- and American Journal of Education, vol. 92 (February 1984), Low-Income Countries." Comparative Education Review, pp. 125-62. vol. 27 (February), pp. 108-18. Leslie, J., and Jamison, Dean T. 1980. "Policy Implications of Instructional Technology: Cost and Effectiveness." In John School Quality as Policy Simmons, ed., 1980, Education Dilemma: Policy Issues for Developing Countries in the 1980s. Oxford: Pergamon Dutcher, Nadine. 1982. The Use of First and Second Languages Press. in Primary Education: Selected Case Studies. World Bank Macmillan, D. 1981. "Language Policies for African Primary Staff Working Paper no. 504. Education: Summary of the Anglophone Research Litera- Haddad, Wadi D. 1979. Educational and Economic Effects of ture." Population and Human Resources Discussion Paper Promotion and Repetition Practices. World Bank Staff no. 81-28. Population, Health, and Nutrition Department, Working Paper no. 319. World Bank. Processed. Heyneman, Stephen P. 1980. "Instruction in the Mother Somerset, A. H. 1982. "Examinations Reform: The Kenya Tongue: The Question of Logistics." Journal of Canadian Experience." Education and Training Department. Pro- and International Education, vol. 9, pp. 88-94. cessed. . 1980. "Planning the Equality of Educational Oppor- Thevenin, T. 1981. "Pedagogical Implications of Language tunity between Regions." In Regional Disparities in Educa- Policy in African Schools: A Review of the Francophone tional Development, edited by Gabriel Carron and Ta Ngoc Literature." Population and Human Resources Discussion Chau. Paris: Unesco, International Institute for Educational Paper no. 81-29. Population, Health, and Nutrition Depart- Planning. ment, World Bank. Processed. The most recent World Bank publications are described in the annual spring and fall lists. The latest edition is available free of charge from Publications Sales Unit, Department B, The World Bank, Washington, D.C. 20433, U.SA. THE WORLD BANK This volume extends the field of economics into a new area: the economic consequences of investing in the quality of education. Some specialists have long considered the possibility that -there was a causal effect between investment in education and the economic growth of develop- ing countries. For many years, however, most of the emphasis has been on growth and on the need to provide more schools and more classrooms for students. AD measurements of invest- ment in education were based on the quantityA-number of years-of education. As these schools systems have grown, however, attention has been shifting toward the qual- ity and value of the education being received. Evidence has been slowly accumulating to sug- gest that the quality of education-availability of reading material, a well-trained teacher, an adequate classroom-is as important to prospects for economic development as is the quantity of education obtained. The staff of the World Bank's education sector held a retreat to consider this evidence and to reflect on the strategies for improving the quality of education in the developing countries. This volume reports the evidence presented and the strategies considered. The papers are some of the first to address this topic. They fail into two parts: the first relates to the economic growth that might result from expenditures on the quality of education; the second concerns the best way to implement changes intended to improve the quality of education. WORLD BANK PUBLICATONS OF RELATED INTERES- sT Altemative Routes to Formal Education: Distance Teachingfor School Equivalency Hilary Perraton, editor Diversified Secondary Educadton and Development: Evidenceffrom Colombia and Taizzania George Psacharopoulos and William Loxley Educationfor 0evelopment: An Analysis of Investment Choices George Psacharopoulos and Maureen Woodhall Education Finance Simulation Model Manuel Zymelmen and Fonald K. Yee ISBN 0-8213-0702-9