Calendars

NBER WORKING PAPER SERIES SCHOOL RESOURCES AND EDUCATIONAL OUTCOMES IN DEVELOPING COUNTRIES: A REVIEW OF THE LITERATURE FROM 1990 TO PDF

Description
NBER WORKING PAPER SERIES SCHOOL RESOURCES AND EDUCATIONAL OUTCOMES IN DEVELOPING COUNTRIES: A REVIEW OF THE LITERATURE FROM 1990 TO 2010 Paul W. Glewwe Eric A. Hanushek Sarah D. Humpage Renato Ravina
Categories
Published
of 25
55
Categories
Published
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Similar Documents
Share
Transcript
NBER WORKING PAPER SERIES SCHOOL RESOURCES AND EDUCATIONAL OUTCOMES IN DEVELOPING COUNTRIES: A REVIEW OF THE LITERATURE FROM 1990 TO 2010 Paul W. Glewwe Eric A. Hanushek Sarah D. Humpage Renato Ravina Working Paper NATIONAL BUREAU OF ECONOMIC RESEARCH 1050 Massachusetts Avenue Cambridge, MA October 2011 This paper benefited from comments by participants at the conference on Education Policy in Developing Countries: What Do We Know, and What Should We Do to Understand What We Don t Know? University of Minnesota, February The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research. NBER working papers are circulated for discussion and comment purposes. They have not been peerreviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications by Paul W. Glewwe, Eric A. Hanushek, Sarah D. Humpage, and Renato Ravina. All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted without explicit permission provided that full credit, including notice, is given to the source. School Resources and Educational Outcomes in Developing Countries: A Review of the Literature from 1990 to 2010 Paul W. Glewwe, Eric A. Hanushek, Sarah D. Humpage, and Renato Ravina NBER Working Paper No October 2011 JEL No. H4,I25,J24,O15 ABSTRACT Developing countries spend hundreds of billions of dollars each year on schools, educational materials and teachers, but relatively little is known about how effective these expenditures are at increasing students years of completed schooling and, more importantly, the skills that they learn while in school. This paper examines studies published between 1990 and 2010, in both the education literature and the economics literature, to investigate which specific school and teacher characteristics, if any, appear to have strong positive impacts on learning and time in school. Starting with over 9,000 studies, 79 are selected as being of sufficient quality. Then an even higher bar is set in terms of econometric methods used, leaving 43 high quality studies. Finally, results are also shown separately for 13 randomized trials. The estimated impacts on time in school and learning of most school and teacher characteristics are statistically insignificant, especially when the evidence is limited to the high quality studies. The few variables that do have significant effects e.g. availability of desks, teacher knowledge of the subjects they teach, and teacher absence are not particularly surprising and thus provide little guidance for future policies and programs. Paul W. Glewwe Dept of Applied Economics, U of MN 1994 Buford Ave. St. Paul MN Eric A. Hanushek Hoover Institution Stanford University Stanford, CA and NBER Sarah D. Humpage University of Minnesota St. Paul MN Renato Ravina University of Minnesota St. Paul MN I. Introduction and Motivation Economists and other researchers have accumulated a large amount of evidence that education increases workers productivity and thus increases their incomes. 1 There are also many non-monetary benefits of education, such as improved health status and lowered crime Lochner (2011)). Finally, at the country level there is also a large amount of evidence that education increases the rate of economic growth (Hanushek and Woessmann (2008)). These analyses all highlight the value of improving a country s human capital and provide the motivation for developing countries to invest in the skills of their populations. They do not, however, indicate which types of specific investments should be pursued. Policymakers in developing countries have quite generally accepted the message of these benefits from improved human capital and have greatly increased their funding of education. As seen in Table 1, since 1980 real government expenditures on education doubled in Latin America and Sub-Saharan Africa, almost tripled in the Middle East, and increased by more than five-fold in East Asia and by almost eight-fold in South Asia. International development agencies have also called for greater resources to be devoted to education, and have increased their levels of assistance for education projects in recent years, as shown in Table 2. The most consistent focus of investment has been on increasing primary and secondary school enrollment rates, with the ultimate goal of higher levels of educational attainment. The increases in enrollment over the past three decades, particularly at the primary level, have been quite dramatic. From 1980 to 2008 primary and secondary enrollment rates have increased in all 1 The majority of this work, following the seminal studies of Jacob Mincer (1970, (1974), has focused on how school attainment relates to individual earnings, and there are now estimates of the return to schooling for a majority of countries in the world (Psacharopoulos and Patrinos (2004)). More recent work has added measures of achievement to this (e.g., Mulligan (1999), Murnane, Willett, Duhaldeborde, and Tyler (2000), and Lazear (2003)), although little of this relates to developing countries (see, however, Hanushek and Zhang (2009)). 1 regions of the developing world (Table 3), so that by 2008 gross primary enrollment rates were at or above 100 percent in all regions, and gross secondary enrollment rates were above 50 percent in all regions except Sub-Saharan Africa. 2 Similarly, Table 4 shows that primary school completion rates increased in all regions from 1991 to 2008, and were close to 100 percent in all regions except for South Asia and Sub-Saharan Africa. Much of the increased funding for education, particularly in the earlier periods, took the form of building and staffing schools in areas where no school previously existed, reflecting the simple fact that it is hard to go to school if no school exists. Moreover, there is ample evidence that enrollment increases when the distance to the nearest school decreases. When increased spending on existing schools makes them more attractive, either by reducing school fees and other direct costs of schooling or by improving the quality of the educational opportunities they provide, enrollment would be expected to increase further. 3 More recently, however, attention has begun to swing toward the quality of schools and the achievement of students and here the evidence on outcomes is decidedly more mixed. Over the past decade, it has become possible to follow changes in student performance on tests offered by the Programme for International Student Assessment (PISA). While student learning appears to be increasing in several countries, this tendency is not universal. More specifically, Table 5 presents evidence on learning among 15 year old students in 12 countries (of which 7 are in Latin America). Examining trends from 2000 to 2009, five countries show clear upward trends (Chile, Colombia, Peru, Tunisia and Turkey), while the rest show either mixed or even decreasing trends. At the aggregate level, it may simply be that expanded enrollment brings in 2 Gross enrollment rates compare numbers of school children to the size of a specific age cohort so that grade repetition, late enrollment, and the like can lead to gross enrollment rates over 100 percent. 3 Hanushek, Lavy, and Hitomi (2008) find that school dropout decisions are very responsive to the quality of the school (in terms of value-added to achievement). 2 progressively less able and less qualified students, who then pull down the average score. Yet some countries with mixed or declining trends did not show large increases in school enrollment, and were increasing real expenditures per student on education. For example, in Argentina the gross secondary school enrollment rate has been about 85 percent from 1998 to 2007, and spending per pupil was somewhat higher in than in ; yet test scores in 2007 were lower than in Similarly, Brazil s progress has been uneven at best, yet it experienced only a moderate increase in secondary school enrollment (7-13 percentage points) from 2000 to 2007, and real spending on education steadily increased over time. 4 The concern about quality becomes more significant in analyses of the impact on student learning (achievement) of demand side programs that stimulate increased enrollment. A recent survey of high quality analyses of currently popular demand side programs fee reductions, conditional cash transfers, and school nutrition programs the higher enrollment induced by these programs was not accompanied by increased achievement (Hanushek (2008)). 5 It is natural to think that bringing students into school must certainly increase their learning and achievement, but this impact may be limited to new students who were not previously in school with no effect (or even a negative effect) on current students. This discussion is related to a substantial body of literature, particularly for developed countries, that suggests that money alone is not the answer to increase student learning. Specifically, for developed countries there is substantial research indicating that overall expenditures, and common school initiatives funded by those expenditures such as lower class 4 See the World Bank s World Development Indicators. Note that Brazil s gross (net) secondary school enrollment rate increased from 99 (66) in 1999 to 106 (79) in 2005, Educational expenditures (in terms of real U.S. $ per secondary student) increased from, on average, about 1340 (350) from 1998 to 2000 to about 1510 (500) from 2004 to 2006 in Argentina (Brazil). 5 The only demand side program that increased achievement was a Kenyan scholarship program that directly related incentives to achievement (Kremer, Miguel, and Thornton (2009)). 3 sizes or more educated teachers, are not closely related to student outcomes. 6 Similar findings, although not as strong, come from the research on schools in developing countries (Fuller and Clarke (1994), Harbison and Hanushek (1992), Hanushek (1995)). In response to findings that increased educational spending has had little effect on student performance, many policymakers and researchers in both developed and developing countries have advocated changing the way that schools are run such as changing the incentives faced by teachers (and by students) and, more generally, changing the way that schools are organized. Yet it is still possible that spending that changes basic school and teacher characteristics, if properly directed, could play a role in improving students educational outcomes in developing countries. Thus it is useful to review the more recent literature on school spending and resources, extending the prior reviews that covered studies through the early 1990s. Indeed, significant numbers of new studies have appeared since More importantly, many of the newer studies employ much stronger research designs than were previously used. The appreciation of researchers for the difficulty of obtaining clear estimates of causal impacts has grown considerably over the past two decades. The sensitivity to these issues, along with more care about the underlying methodological approach, suggests that the new studies may in fact yield conclusions different from those drawn on the older research. This paper examines both the economics literature and the education literature published in the last two decades to assess the extent to which school and teacher characteristics have a causal impact on student learning and enrollment. More specifically, this paper reviews the literature that attempts to estimate the impact of school infrastructure and pedagogical materials 6 These conclusions have been controversial, and much has been written about the interpretation of the evidence. For a review of the inconsistencies of effects, see Hanushek (2003). For the range of opinions, see, for example, Burtless (1996), Mishel and Rothstein (2002), and Ehrenberg, Brewer, Gamoran, and Willms (2001). 4 (such as electricity, condition of the building, desks, blackboards and textbooks), teacher characteristics (education, training, experience, sex, subject knowledge, and ethnicity), and school organization (pupil-teacher ratio, teaching methods, decentralized management, and teacher contracts and working conditions) on student enrollment and learning. The remainder of this paper is organized as follows. The next section describes a simple interpretive framework. This is followed by a description of the parameters of this review and of how studies were selected for inclusion. Finally, we present the results of our review and draw conclusions about priorities for future research. II. Interpreting the Research on Basic Education Inputs The overarching conceptual framework employed here considers schools as factories that produce learning using various school and teacher characteristics as inputs. This is the production function approach introduced early in microeconomics courses. However, the actual application and interpretation in education differs from the simple textbook treatment. The reasoning underlying this conceptual framework is that the process by which cognitive skills are learned is determined by many different factors, and production functions are expressions, in simple terms, of this process. The relationship can be very flexible, allowing for almost any learning process. In this sense, an education production function always exists, although its existence does not guarantee that one can estimate it. In the ideal case, if one can estimate this relationship, one can use information on the costs of school characteristics, classroom materials, and even teacher characteristics to select the combination of these that is most effective in increasing enrollment and/or student performance 5 (e.g. increase in test scores per dollar spent) given a limited budget. In theory, this could also apply to pedagogical practices, which have implementation costs. A. Relationships of Interest. It is useful to step back to consider what relationships are of interest and how those relationships interact with households behavior. The theory of the firm, where analyses of production functions are generally introduced, takes the perspective of a decision maker who optimally chooses the combination of inputs for his or her firm. But this perspective ignores a key reality of education: students and parents -- both important inputs into achievement also make their own decisions in response to the school decision maker s choices. To begin, assume that the parents of the child maximize, subject to constraints, a (lifecycle) utility function. The main arguments in the utility function are consumption of goods and services (including leisure) at different points in time, and each child s years of schooling and learning. The constraints faced are the production function for learning, the impacts of years of schooling and of skills obtained on the future labor incomes of children, a life-cycle budget constraint, and perhaps some credit constraints or an agricultural production function (for which child labor is one possible input). Following Glewwe and Kremer (2006), the production function for learning (a structural relationship) can be depicted as: A = a(s, Q, C, H, I) (1) where A is skills learned (achievement), S is years of schooling, Q is a vector of school and teacher characteristics (inputs that raise school quality), C is a vector of child characteristics (including innate ability ), H is a vector of household characteristics, and I is a vector of school inputs under the control of parents, such as children s daily attendance and purchases of textbooks and other 6 school supplies. Although children acquire many different skills in school, little is lost by treating A as a single variable. Assume that all elements in the vectors C and H (which include parental tastes for schooling, parental education, and children s ability ) are exogenous. Some child characteristics that affect education outcomes (such as child health) may be endogenous; they can be treated as elements of I, all of which are endogenous. In the simplest scenario, only one school is available and parents can do nothing to change that school s characteristics. Thus all variables in Q are exogenous to the household. Parents choose S and I (subject to the above-mentioned constraints) to maximize household utility, which implies that years of schooling S and schooling inputs I can be expressed as general functions of the four vectors of exogenous variables: S = f(q, C, H, P) (2) I = g(q, C, H, P) (3) where prices related to schooling (such as tuition, other fees, and prices of textbooks and uniforms), which are also exogenous, are denoted by the vector P. Inserting (2) and (3) into (1) gives the reduced form equation for (A): A = h(q, C, H, P) (4) This reduced form equation is a causal relationship, but it is not a textbook production function because it reflects household preferences and includes prices among its arguments. 7 The more realistic assumption that households can choose from more than one school implies that Q and P are endogenous even if they are fixed for any given school. In this scenario, households maximize utility with respect to each schooling choice, and then choose the school that leads to the highest utility. Conditional on choosing that school, they choose S and I, as in the case where there is only one school from which to choose. Policymakers are primarily concerned with the impact of school and teacher characteristics (Q) and prices related to schooling (P) on years of schooling (S) and eventual academic achievement (A). For example, reducing class size can be seen as a change in one element of Q, and changing tuition fees can be seen as altering one component of P. Equations (2) and (4) show how changes in the P variables would affect S and A. In addition, equation (2) also shows how changes in school and teacher quality (Q) affect students years of schooling (S). Turning to the impact of school quality variables (Q) on student learning, there are two distinct relationships. To see this, consider a change in one element of Q, call it Q i. Equation (1) shows how changes in Q i affect A when all other explanatory variable are held constant, and thus provides the partial derivative of A with respect to Q i. In contrast, equation (4) provides the total derivative of A with respect to Q i because it allows for changes in S and I in response to the change in Q i. 7 Parents may respond to higher school quality by increasing their provision of educational inputs such as textbooks. Alternatively, if they consider higher school quality a substitute for those inputs, they may decrease those inputs. The fact that parental actions may reduce or reinforce school decisions may help to explain a portion of the prior inconsistencies in estimating the impact of school resources. Indeed, different studies could obtain different estimates of the impacts of the Q variables on 7 For an early development of this idea, see Kim (2001). 8 student learning because some studies estimate the production function, that is equation (1), while others estimate the reduced form relationship in equation (4), and it is quite possible that impacts of the Q variables will be different in these two equations. When examining the impact of school quality (Q) on academic skills (A), are the impacts in equation (1) or equation (4) most useful for policy purposes? Equation (4) is useful because it shows what will actually happen to A after a change in one or more element in Q. In contrast, equation (1) will not show this because it does not account for changes in S and I in response to changes in Q and P. Yet the impact in equation (1) is also of interest because it may better capture overall welfare effects. Intuitively, if parents respond to an increase in Q i by, for example, reducing purchases of i
Search
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x