|
Local Government and Households in Primary Education in Tanzania: Some lessons for Reform
CDR Working Paper 98.6, June 1998
Contents
Abstract The roles of the formal and informal involvement of local governments, local elites, teachers and households in decision making about primary education in selected rural districts are analysed, as are some of the outcomes of this involvement. Lessons for reforms with respect to decentralization, financing and school choice are drawn. It is concluded that local-level actors, particularly better-off households and teachers, have significant influence on primary education at the local level. Moreover, poverty rather than dissatisfaction with the quality of primary schooling may explain the low enrolment ratios in Tanzania. Finally, it is concluded that both demand- and supply-driven responses to the problems of primary education are needed and must be based on country specific analyses.
Foreword There is widespread dissatisfaction with primary education in Tanzania. Its quality is regarded as low by both educationalists and parents (Omari and Mosha, 1987; Malekela, 1995). Enrolment has stagnated for ten years, but has now begun to exceed population growth again (Tadreg, 1992; MOEC, 1995). Disparities in both quality and enrolment across districts are substantial (Sumra, 1995, table 8; Malekela, 1995, p. 70). Most other sub-Saharan African countries face similar problems (World Bank, 1988). Such dissatisfactions are as politically sensitive now as they were during colonial times (Carnoy and Samoff, 1990; Semboja and Therkildsen, 1995). In Tanzania, the government has made several attempts in recent years to respond to this (Government of Tanzania, 1984, 1992, 1995a, 1995b, 1995c), but so far the changes on the ground have been limited. Donors, who are major funders of education throughout Africa, have likewise renewed their interest in reforms of primary education, and some of them - in particular the World Bank, the European Community, Britain, Denmark and the Netherlands - are deeply involved in educational policy formulation in Tanzania (Buchert, 1997). While there is a wide agreement about the need for reform, there are conflicting views on how to go about it. In the language of current development discourse, there are, in principle, two different ways in which the serious problems of primary education can be tackled. They are based on very different assumptions, and on opposing political values regarding the proper role of the market and the state in education. One approach is to strengthen the supply-side of provision of education through the public sector because the market cannot be relied upon to provide quality education for all; the other is to strengthen the demand-side through the market, because the state has failed (Colclough and Manor, 1991; Colclough with Lewin, 1993; Colclough, 1993). In practice, elements of both approaches are used when educational policies are formulated and implemented, but their respective biases are significant. For supply-side proponents, the role of government (including local government) in primary education is important. Primary education is a national issue. Improvements in the capacity of the political-administrative system are therefore needed in order to improve primary education. Parents tend to be regarded as passive (but satisfied) recipients of services. As a consequence, they are often ignored in both research, policy formulation and actual implementation (Mosha, 1989; Tadreg, 1992). For demand-side proponents, the role of parents is a prime issue. They and their children are direct beneficiaries of education. Their influence on, and financial contributions to, the schools should therefore be strengthened. Specifically, parents should be empowered to run community schools. Local governments are regarded as marginal in the provision of primary education. Moreover, parents should be allowed an unrestricted choice of school so as to create competition for pupils in the school system (World Bank, 1995, 1996). Unfortunately, such normative views often dominate the discussions about educational policy. Moreover, present discussions of educational policy reform tend to suffer from an absence of empirically based analyses of the actual roles of local government and parents [ In many cases, it would be more correct to use the term 'family', because often decisions about and funding of primary education do not just involve the biological father and/or mother but also relatives or other adults. In certain contexts, it would also be appropriate to include children in the analyses. Unfortunately, little is known about the extent to which children influence educational decisions . ] in primary education in Tanzania. A view of the school system from below helps to highlight some of the complexities involved. [ The local-level empirical material presented here was collected between 1987 and 1993 in nine rural districts on mainland Tanzania (Mbulu, Arumeru, Babati, Pangani, Njombe, Iringa rural, Mufindi, Kibaha, and Kisarawe). Information was also collected from Arusha and Iringa municipalities, and from the rural districts of Hanang, Kiteto, Moduli and Ngorongoro. Important macro-level changes after 1993 are also included in the analyses, where these are deemed important for the arguments.] The purpose of this paper is to contribute to such analyses by focusing primarily on three issues: local-level decision making about the school system; funding of primary education; and parental school choice. They provide several important lessons for the reform of primary education. This Working Paper is a revised version of a paper prepared for workshop on "Quality and Equity Issues in Tanzania Education Policy and Practice: Insights from Research", Dar Es Salaam, December 15-16, 1993. The research was conducted in a cooperation between the Economic Research Bureau and the Centre for Development Research, Copenhagen. The research is supported by Danida's ENRECA fund.
Local-level decision making in rural districts [ The sub-district political-administrative framework in urban areas is significantly different from that established in the rural areas (Semboja and Therkildsen, 1991). Consequently, no attempt to generalize to urban areas is made. The Zanzibar situation is also quite different from that reported here (Semboja and Therkildsen, 1992).] Primary education is the most important responsibility of local governments in Tanzania. Half of all their funds is spent on this activity (although most funds are provided by the central government), and two-thirds of all local government employees are teachers. Local governmentdecision making is vested in the district council. A majority of its members are directly elected at the ward level. At the lower level, the village council has much the same functions as the district council. At the school level the school committee - in which parents are represented - is supposed to oversee the running of the school. [ There are 103 district councils on mainland Tanzania. On average, each has a population of 250,000 people, 110 primary schools, and 31,000 students. See Semboja and Therkildsen (1991) for a description of the political- administrative system at regional and district level. Analyses of the role of local governments in primary education are provided by Syrimis (1988) and Lillis (1990). ] Analyses of decision making about primary education by local governments and parents can be done in two ways. One is based on the formal division of responsibilities between the different actors as stipulated in legal and administrative documents (and, previously, party guidelines and instructions). The other approach is to focus on who actually influences decisions - an issue which also brings the importance of non-decisions to the fore. Both approaches are used in the following. The legal framework The National Education Act of 1978 states clearly who is in formal charge of education: The Minister [for Education] shall ...be responsible for the promotion of the education of the people of Tanzania ...and for securing the effective execution by local authorities, under his guidance, control and direction, ...of nationally beneficial education services (part II, section 4). This expresses the desired state of affairs: that central government is in full control of education; local authorities are agents of central government; education is for the benefit of the nation. Benefits of education for the children, parents, or the communities in which they live are not mentioned; and education is exclusively a domestic affair. In the recent 1995 revision of the Education Act, this section was left unchanged. It must therefore be regarded as a reflection of a stable and long-held normative vision of the Tanzanian political elites about the proper role of government, local institutions and parents in education. However, the quoted section of the Education Act conflicts with the powers given to local governments in the Local Government (District Authorities) Act no. 7 of 1982. [ The amendments of 1995 to the 1978 Education Act actually seek to strengthen the power of the Minister for Education over educational matters: Prior to the amendments, the minister responsible for education and local governments should agree if they wanted to modify local authority education plans; the amended law gives this power exclusively to the minister responsible for education (see amendments to part II, section 5(c) of the Education Act).] It states that -subject to national policy and plans - every local government "...shall...formulate, co-ordinate andsupervise the implementation of all plans for...social development in its area of jurisdiction" (part V, section 118). If the two acts are read together - as they should be - relations between the ministers responsible for education and local government respectively are rather ambiguous. In practice, however, the local governments (and the minister responsible) have little real influence on educational issues proper: curriculum; examinations; the relative weight given to academic and practical activities in the schools; the length of class room instruction; and so on. The national political goal to provide the same education for all has suppressed the need to adjust primary education to the often significant variations in local conditions. Even the earlier permission for local authorities to adjust school terms to the agricultural practices in the district [ Granted by circular MOE/EDTT S1/16/69 of 22 October 1968.] has been removed. Control by the Ministry of Education of primary education is, in fact, so strong that district councils [ A copy of the report is given to the District Education Officer, who could make use of the inspection reports, although not formally give the council access to them. The DEO is, however, not an employee of the council but of the central government-controlled Local Government Service Commission.] do not have access to the ministry's school inspection reports, although these contain much relevant information on the performance of teachers, schools, and pupils in each district. Such information is passed on to the ministry and left to it to act upon - which it rarely does. A second important set of relations - between communities, parents and primary schools - is also formally regulated through the establishment of school committees. Already 25 years ago it was decided that they "... will not have powers over academic matters of the school." [ Ministry circular MOE EDP/P1/124/14 of 6 May 1968.] In this 1968 circular, community was spelled with a capital C, but the practice was abandoned in later circulars. In any case, since then the committees have only had an advisory role. Their prime function is to promote the integration of the school in the life of the community and to cooperate with teachers in solving "local educational problems." In practice, this mainly means involvement in solving disciplinary problems with pupils. Similar problems with teachers, for example, are outside their authority. [ The ongoing reforms aim to strengthen the school committees substantially (see GOT, 1995a). These intentions have yet to be translated into new formal mandates for the committees however.] More specific changes have been made in the membership of school committees. In 1968, they consisted of eight parents and four teachers. Ten years later, parent representation was reduced to three (out of a typical committee of 15), while village council members occupied eight seats. The dominating position of the village council vis-a-vis the school committee is retained in theLocal Government Act of 1982 and subsequent circulars. Formally, the direct influence of parents on school affairs has therefore been reduced and that of the village council increased. [ All village council members were also members of the sole legal party Chama Cha Mapinduzi (CCM) until the introduction of multi-partyism in 1995. In practice, most members of village councils are still from CCM, because of this party's dominance in the rural areas. Councillors must be members of a party.] Finally, local governments do not have much influence on central government resource allocations to primary education. Allocation decisions are made by the central government and donors, as discussed below. Three impressions emerge from this very brief survey of the legal framework for primary education. Decision making is highly centralized with wide-ranging powers located at the ministry level. The formal influence of local governments on primary education is limited and that of parents has decreased over time. On the other hand, the legal specifications for relations among the main actors in the school system are ambiguous. Finally, the role of donors - major actors in both decision making and funding - is not specified anywhere, although they have had major influence on policy-making in education during the 1990s (Buchert, 1997, pp. 44-57). Subtle influences To conclude that local governments and parents have very little influence on primary education therefore appears justified. And this is also the standard diagnosis of much analytical work on Africa (see, for example, Wunsch and Olowu, 1990) and on Tanzania (Mwaikusa, 1985; Liviga, 1992). But more subtle ways of influencing primary education issues co-exist with the highly cen tralized decision making. The district council is just one arena among several for such decision making. For local primary education politics are dominated by various constellations of individuals from the political elite among council staff and politicians, the parties, [ Despite the introduction of multi-partyism in 1995, CCM is still very dominating in all but a few rural districts, as mentioned earlier.] central government, Parliament and, in some cases, prominent businessmen. The central issue seems to be the improvement of physical infrastructure rather than of education itself. And it is typical that the politics of education at district level often tend to be more concerned with making plans - a potent symbol of modern politics - than with finding ways of financing them. It is especially the investment (i.e. development) plans and budgets of the district council which are strongly politicized - although in a rather limited fashion. For the key recurrent issue during council budget sessions tends to be the location of development projects. Both political values and political tactics influence the outcome. Thus, to the extent that equity issues do enter the local political debate, it appears that equity between areas (i.e. wards, the constituencies of elected councillors) rather than equity between social strata (i.e. rich/poor) or institutions (i.e. schools) are of major concern. Consequently, district development plans and budgets that include many small school projects in many wards are more likely to get approval than those including fewer but larger projects. Moreover, the larger the development budget appears to be, the easier it is to satisfy various political, constituency-based demands during budget preparations, without being forced to make priorities between them with the danger of conflicts within the council that this would entail. There are two typical and direct local policy outcomes of this. One is a planned wide geographical spread of many small projects. The other is that desired expenditures are budgeted for before revenue estimates are made. Typically, therefore, the approved development budgets are much higher than the actually available funds. Consequently, most council development plans are fictitious and carry largely symbolic value. This means that many decisions with actual implications - in primary education as well as in other sectors - are made outside the council proper. Where? Generally, the acceptance by local politicians of a substantial gap between planned expenditures and actual revenues means that the influence of bureaucrats on actual resource use is enhanced. A few bureaucrats in the councils (normally the DED and the Treasurer) control the day-to-day use of very scarce funds. The DEO is normally not influential in these decisions, even when they concern education. But there are interesting variations to this trend. In one district, for example, the District Executive Director (DED) and the Member of Parliament, who was also a prominent minister, cooperated closely in mobilizing council staff and village councils to improve school buildings. They also set up competitions among schools to improve educational quality - an uncommon strategy in the 1980s, but a practice which is now widespread following the introduction of multi-partyism in 1995. In another district, the District Educational Officer (DEO) thought that the political leaders focused too much on secondary education and ignored primary education ("they grabbed the wrong end of the stick"). He lobbied for his views, primarily with the District Commissioner (a central government representative) and the District CCM Party Chairman. He also tried to influence his formal superiors in the local government (the DED and the Council chairman), although he did not regard them as very influential in educational matters. Eventually the District Party Committee decided that within three years every primary school should have an adequate number of classrooms and all teachers should have a "modern" house. The District Council later rubber-stamped this decision. As is often the case, the cost and funding implications of this decision were unclear at the time. Eventually, the standard Tanzanian default option was used: Necessary funds should be raised through special contributions from parents and schools. This failed. In May 1989, the district leaders attended a budget session of the Regional Development Committee (RDC). [ All local government budgets had to be approved by the RDC until July 1996. See Semboja and Therkildsen (1991) for a description of the budget procedure until then.] Here, they were questioned about their failure to apply for money from the President's Self-Help Fund. They reacted immediately and telephoned the DEO, asking him to prepare an application on the spot for support to primary schools (based on the ideas approved by the party committee three years earlier). This application was promptly endorsed by the RDC. The President's Self-Help Fund later approved the proposal. Subsequently, the money was sent to the District Commissioner's office. This central government organ then released the approved resources directly to the schools in question. [ Mainly corrugated iron sheets for roofing.] Actual construction at village level was organized through CCM. During this whole chain of events, the District Council, which is in formal charge of primary schools, was only marginally involved in both decision making and implementation. In a third district, the DEO proposed to the council that an Education Centre for in-service training of primary school teachers be built. The council approved the idea in 1987 without making plans for financing and implementation. The DEO estimated the cost to be some 19 mill. sh. - thirty times more than the approved primary school investment budget for the whole district (Government of Tanzania, 1987, table 5). As in the case mentioned above, schools and parents were ordered to contribute funds and bricks for the Centre. Apart from the initial spurt of activity - immediately praised by the Regional Commissioner and the RDC - no construction has taken place for some years. The DEO says that only a foreign donor can now save the project. Finally, there are many examples of prominent individuals - mostly private businessmen or parastatal executives - who donate funds to improve conditions in one or a few schools. Such initiatives are rarely aimed at the school system as such and may not even aim to benefit the recipient school. Private donations of school desks are, for example, often made with the con dition that the desks will only be used by the children of the donor(s). The borderline between the public and the private is constantly blurred. These cases illustrate the subtle and not so subtle influences exercised by local bureaucrats and local political and economic elites. They also show how the formal decision makers on primary education matters - the elected councils - tend to be by-passed. But this is not the whole picture. Non-decisions Because most district councils are re-active rather than pro-active in matters of primary education, [ Most councils are, however, very active in secondary school matters. See, for example, Ishumi (1995).] non-decisions may also have important policy implications. But evidence of outcomes of non-decisions is difficult to document. At first glance, outcomes often appear to be the result of deliberate central or local government decisions based on class, ethnicity and patronage as Oyugi (1994), for example, argues. Such considerations are thought to be particularly strong with respect to how state resources are allocated and who benefits from them; therefore, attempts are often made to explain observed inequalities in these terms. The assumption is that observed variations in public allocations to, say, primary education are the intended result of deliberate decisions reflecting the interests of the powerful public and private decision makers. But there are other forces at play as well. A closer look at how a key resource in primary education - namely teachers - is allocated to schools illustrates the point. Formally, the allocation of teachers to schools is centralized. Newly graduated teachers are allocated to regions and districts by the Directorate of Primary Education in the Ministry of Education. Allocations are based on the number of streams in each region. The Regional Education Officer, in turn, allocates these new teachers to the districts using the same stream-based formula. The DEO finally allocates the teachers to individual schools (Dom, 1995, annex 7). It is at this final stage that considerations of equity in school level teacher-student ratios can be made, although there is no national policy on this. If this system were adhered to, differences in student-teacher ratios would be reduced over time. There are, however, large variations between the best staffed districts (Liwale, Bukoba urban and Tabora urban) and the poorest staffed (Mbulu, Meatu and Bariadi) in Tanzania in terms of pupil-teacher ratio. In 1990, it ranged from 14 to 59, around an average of 35 per district. This general pattern was similar in both rural and urban districts. The variation was even greater for female teachers and the best qualified teachers (based on data in Tadreg, 1992, appendix table 5). Variation in pupil-teacher ratios across schools within districts was also substantial. In Iringa Municipal Council, for example, the average was 29, but the range was from 19 to 37 among its 26 schools. In Iringa rural district, 183 primary schools had an average pupil-teacher ratio of 42, ranging from 16 to 85. In the adjacent district of Mufindi, the average ratio was 39 among the 135 primary schools there, ranging from 17 to 72 (World Bank, 1991, p. 63). In Kibaha district, 13 percent of the teachers would have to be re-allocated to equalize the pupil-teacher ratio across schools. [ Newer data by Dom (1995) show that this problem persists.] No doubt some (unknown) part of these observed variations in the allocation of teachers is a result of deliberate decisions by central and local educational authorities, based on patronage, class and ethnic interests, as Oyugi argues. But other factors are at play as well. Changes in the pupil-teacher ratio over time provide a clue. From 1972 to 1979, inter-district disparities in this ratio rose. [ Based on data in Maro (1990) from 10 rural districts.] However, from 1984 to 1990, the ratio fell from an average of 39 to 35 in the 78 districts for which data are available. Intra-district disparities also declined. An explanation for these trends is that in the 1970s inequalities in teacher resources across districts rose because the educational authorities were unable to match the allocation of teachers to the rapidly (but unevenly spread) enrolment of children. In the 1980s, inequalities fell because especially children in districts with relatively few teachers withdrew from school. Thus, in both periods, the changing trends in pupil-teacher ratios resulted as much from changes in educational demand with which the educational authorities could not cope - as from any deliberate government policy to favour certain districts. A similar picture emerges for the allocation of teachers to schools within districts. Interviews indicate that district authorities administrate the posting of teachers to schools with a high degree of passivity. Although a DEO may order a newly recruited teacher to a remote school that badly needs staff, and many districts use peripheral schools (called "our Siberia" by deported teachers) to punish teachers that are out of favour, most DEOs regard the re-allocation of teachers within the district as tricky. It involves considerations about availability of suitable housing (which favours urban and former mission schools); the work situation for spouses (which favours the urban schools preferred by female teachers staying with their husbands, who often tend to be members of the local elites); the health situation of the individual teacher (which favours schools located near health facilities) etc. Moreover, analysis of transfers of head teachers in Arumeru and Babati districts in 1990 shows that many of them eventually succeeded in being placed in schools of their own choice. For example, the closer a teacher is to retirement, the closer (s)he is likely to work near home. In practice, therefore, the management of teachers tends to reflect the needs of the individual teacher rather than educational requirements of the schools. Decision making with respect to placement of teachers is therefore fairly individualized rather than centralized as the formal attributes of the teacher allocation system indicates.
Funding of rural primary education Officially, primary education in Tanzania is free and supposedly funded by central and local government taxes, although parents are urged to 'contribute' a small 'voluntary' fee for each child attending school. [ The Universal Primary Education (UPE) fee was sh. 200 per year in 1991 (one USD = 195 sh. in 1990). It is now sh. 1000 per year (one USD= 600 sh. in 1997). Its status is ambiguous. Payment is not compulsory and is actually paid for less than one-third of the children attending school . On the other hand, some children in some schools are denied access to school if the UPE fee is not paid.] Reality is different. There are at least five sources of funds for rural primary education. As shown in table 1, they amounted to some 5 billion sh. in 1987/88. Compared to the actual needs - and to funding levels for primary education in other African countries - these resource inputs are low (World Bank, 1988, table A-17). And when limited resources are spread across 11,000 primary schools and some 3.2 million pupils, there is - on average - very little for everyone. [ Lawson (1994, pp. 9-11) provides estimates of the magnitude of the under-funding.] This should be kept in mind in the following. Table 1 Rough estimates of resources to rural primary education in 1987/88
Sources: Government budgets (various years); World Bank (1991); Semboja and Therkildsen (1989). Notes: Government budget figures are actual but exclude (minor) funds spent by the ministry itself. Parents' contributions are based on surveys in Iringa Rural and Njombe districts. The figures represent rough orders of magnitude only. a): Also in 1994, household contributions to schooling in rural areas made up some 30% of total inflows (see footnote 19). The figures in table 1 are based on rough estimates and include donor and government funds for investments and recurrent expenditures plus direct contributions in cash and kind by communities and parents. Central government accounted for the largest share (43 percent), followed by parents (33 percent) and donors (23 percent). Regional and local governments were insignificant asfunders of primary education. They depend (almost) entirely on central government subsidies. [ Newer estimates for 1994 by the World Bank (1995, pp.38-39) show similar results: parents in rural areas pay on average 31 percent of the total expenditure per student enrolled in primary school.] Each of these sources warrants specific comments. Donor support for investment funding has been important for more than twenty years (Semboja, 1992, table 4.10; Semboja and Therkildsen, 1995). Dependency on external sources for recurrent funding started in the mid-1980s and has grown ever since. [ While foreign loans, grants and import support accounted for 14 percent of government recurrent revenues in FY 1985, they had risen to 50 percent in FY 1990 (Semboja, 1992, appendix table 2). ] Both types of donor funds are significantly underestimated in government budgets (Semboja and Therkildsen, 1989, p.94) and hence in the figures presented in the table. Central government is still, however, the largest source of funds for primary education. And contrary to widespread belief in Parliament, in university circles (e.g. Shivji, 1991, p. 21), among educationalists (e.g. Mosha, 1989, p. 63) and in some donor quarters (e.g. World Bank, 1991, p. 47), central government has put increasing emphasis on primary education in terms of funds. Out of a growing total budget (in real terms), central government allocated relatively more money to primary education in 1992 than it did in 1985 - even when population growth is accounted for (Semboja, 1992, table 4.1; Therkildsen and Semboja, 1995). [ This trend has continued to 1994 (World Bank, 1995, p. xxi).] Three major interrelated factors account for this: Increased donor funding of the central government recurrent budget as already explained; positive growth in the economy at the end of the 1980s, which made tax revenues grow as well; and the decision by Tanzanian policy makers to give priority to primary education despite resistance from the World Bank and IMF - especially during the first half of the 1980s - against "excessive" social service spending. The financial room for manoeuvre of central government with respect to primary education is therefore both enhanced and constrained by donors. Local (and regional) governments are basically conduits for central government funds. District councils' own funds are rapidly dwindling and rarely used for primary education. [ Central government allocates resources among districts, while local government in turn allocates them within their respective districts - subject to political and administrative guidelines. Allocations are based on needs, mainly the number of students. The Councils' own funds are collected through head taxes, fees, fines, crop cesses etc.] Instead, own funds are spent on mandatory salary expenses for specific staff (mainly outside social services). The UPE fees earmarked for primary schools are hardly important. Collection rates are on average below 25 percent, but seem to be positively correlated with the ability of the government to provide visible support (such as educational material) to the schools. Moreover, when UPE fundsare retained by the schools without passing through the councils, collection rates may also increase somewhat (World Bank, 1991, p. 43). [ Retention of UPE at school level was done on a pilot basis until recently, when it was formally introduced country- wide. It is, however, not uniformly complied with .] Finally, communities and parents contribute directly and substantially to primary education. In monetary terms, they paid some 495 sh. per pupil per year in 1987/88 - around one-third of total resource inflows to the sector. This share has remained fairly stable as World Bank (1995, pp.38-39) figures for 1994 indicate. On the investment side, contributions to primary schools are mostly organized through the village councils and typically consist of self-help labour for repair and ex pansion of class rooms and teachers' housing. A survey of 32 villages in Iringa and Njombe districts in 1988 showed that such activities are undertaken regularly by most villages. The "self-help spirit" had not died - contrary to conventional wisdom. The survey also showed that 60-70 percent of all village councils did make (mostly small) yearly recurrent contributions to primary schools (Semboja and Therkildsen, 1989, pp. 55-61). Surveys of Kibaha and Kisarawe districts in 1992 showed that village councils here were much less active. This indicates that there are significant regional differences in such contributions, or that there has been a general downward trend in them. On the recurrent side, households contribute directly to primary schooling both as members of their communities and as parents (community maintenance of buildings; parents purchase of exercise books, text books, supplementary school material and school uniforms; parents payment of UPE). These estimates are included in table 1. However, the table does not include contributions to school furniture, repairs of teachers houses, student labour for school and teachers, and private tuition. The direct contributions by parents and communities are therefore relatively larger than shown in the table. But certainly the evidence presented here shows that the days of "free" primary education have long gone and that parents provide a very important source of funds for primary education. The officially unrecorded household contributions to primary education are substantial and important. Their magnitude indicates the extent to which public services are de-facto privately funded. The demand-side proponents nevertheless argue that much more funds could and should be mobilized from parents (World Bank, 1995) without significant negative impact on enrolment. Direct evidence in support of that view is not yet available for Tanzania. However, it is important to note that when free primary schooling for four children from each family was introduced in Uganda last year, enrolment grew with some 30-40% according to newspaper reports. This indicates that poverty is indeed a major reason for non-enrolment of children in school.
School choice by households From a village perspective, the school system must appear very centralistic. As already shown, the formal possibilities for parental influence through school committees has decreased in the last 25 years, and the increased political representation of citizens elected to district councils that followed the re-introduction of local governments in 1984, [ Multi-partyism was introduced in 1995. It has not yet changed the political configuration of district councils significantly. The ruling party, CCM, still maintains control of all urban and rural councils. The next local government elections are in 1999.] has not strengthened the role of local governments in primary education. However, this does not mean that parents are passive recipients of state provided services. The 1980s, in particular, have witnessed many profound changes in the ways that parents influence primary education. The aggregated effects of these, mostly individual, responses can be quite substantial, as analyses of parental choice of a primary school will show. Household preferences Interviews with households indicate that a 'good' school is one that secures standard VII leavers access to government secondary school. [ Parents may also use additional criteria for accessing primary school quality - such as modest use of physical punishment of students - but access to government secondary school is by far the most important criteria.] Such access is based on the results of the standardized national examination for primary school leavers, but access is increasingly difficult to attain because the proportion of a standard VII cohort actually admitted has fallen over the years and was only 6 percent in 1990 (9 percent were admitted to private secondary schools). Nevertheless, the attraction of a place in government secondary education seems as strong as ever. There are two main reasons for the preference of government over private secondary schools: The former are widely believed to offer better quality education than most private secondary schools; and fees in private schools are on average more than twice those in government schools. The analyses of the transfer of pupils between districts and schools presented below show that certain parents are able and willing to transfer their children from one primary school to another to increase the probability of access to government secondary schools. Such transfers between primary schools can be made fairly easily, because school choice based on spatial aggregation into schools districts (which is typical in many other countries) has not been introduced in Tanzania. Transfers do require the approval of both the receiving and the sending head teacher however. If pupils are transferred across district boundaries, the DEO of the receiving district should also approve. In both cases, only one important formal reason to refusea transfer request exists: classroom overcrowding. What this means in practice is unclear, but it seems that few head teachers or DEOs refuse requests for transfer. [ Note also that bribing to gain access to a 'good' primary school is said to be widespread.] Children can therefore be transferred if families are willing and able to do so. Transfers within walking distance of the residence may not require additional resources compared to enrolment in the nearest school. Transfers beyond walking distance require special arrangements for board, lodging and adult supervision of the child. Most families are likely to depend on extended family networks for this. Grandparents, uncles, aunts, cousins, and so on, living near appropriate schools may become important resources for securing access to better education. Additional payments are required if families cannot rely on these extended networks. The feasibility of transferring children thus depends on specific circumstances. Generally, it is an option that the better-off, educated parents living in urban areas are most likely to use, but some parents in rural areas also do. As discussed in the following, the chances of access to government secondary education may increase by moving a child to: (i) a primary school within the administratively defined catchment area of a day secondary school; (ii) a district with a low examination score entrance requirement; or (iii) a 'good' primary school which 'produces' many candidates for government boarding schools. In practice, the three strategies may be combined. Each of them are discussed in turn below. Access to day school Access to government secondary day schools depends on examination results at the end of standard VII, but a Ministry of Education rule dictates that candidates should stay within 6 km radius of the day school. Consequently, children within this catchment area have a better chance of post-primary education than children living outside it. Government day schools tend to be concentrated in urban districts. This influences primary school student transfers across urban district boundaries. [ During the last decade many local governments have actively promoted the construction of day secondary schools within their jurisdiction. Often they are instrumental in such efforts although they are legally barred from running secondary schools according to the Local Government Act. To circumvent this, councils may, for example, help to establish Trust Funds to run the private secondary schools; provide staff and office facilities for these trusts; and transfer council funds to them as grants (see also Lugalla, 1989).] Arusha municipality, for example, had 2/3 of all day school places in Arusha region in 1990, but only 1/10 of the standard VII leavers. Pupils living in Arusha municipality therefore stand a much better chance of gaining access to agovernment day secondary school than pupils from the rural districts of Arusha region. This leads to many transfers from the rural to the urban district schools as illustrated in table 2. It shows the proportion of students enrolled in one grade that proceed to the next. The unique feature shown by the table is that retention rates for students in standards V to VII are very high and rising to 100 and above for the municipality, but stable or falling in the rural districts. It is a pattern that cannot be explained by general in-migration to the urban area, in which case retention rates in all grades should be affected. Disaggregated data show that the transfers are especially pronounced from Arumeru district (which completely surrounds the municipality). In the 1980s, some 10 percent of the pupils in the municipal standard VI and VII classes came from the rural districts. The trend is similar for boys and girls. Table 2 Pupil retention
Source: Regional Education Officer, Arusha Notes: a) Retention rate = Enrolment in Standard i (x 100)
b) Calculations based on information from the following years: Arusha Municipality: 1980-1990; Arumeru: 1980-85 & 1989-90; Mbulu: 1980-84 & 1986-90; Babati: 1987-91; Hanang: 1986-90; Monduli: 1980-87 & 1989-90; Kiteto: 1990-91; Ngorongoro: 1980-83 & 1985-90. c) The high retention rate for standard III-IV for rural districts is caused by the examination in standard IV which, if failed, forces pupils to repeat. This is a bigger problem in rural than in urban districts. Access to boarding secondary school Access to government secondary boarding schools also depends on the results of the final examination at the end of standard VII. But a centrally controlled quota-system aims to ensure that the same proportion of standard VII leavers in each district is admitted. It is an attempt to provide equal access for various ethnic groups to government secondary schools. However, educational performance varies across districts so that the examination score needed to qualify differs. Some parents living in districts with high entrance level requirements therefore try tocircumvent this system by transferring their children to districts with less competition. Cooksey and Ishumi (1986, app. 1) have provided some details about this phenomenon, but it is impossible to quantify it. Transfers to 'good' primary schools Transfers of pupils from a 'bad' to a 'good' [ Schools which, compared to other schools in the area, consistently sent the highest proportion of their primary school leavers to a government secondary school.] primary school is a third parental strategy for trying to gain (eventual) access to government secondary school for their child(ren). This may involve a transfer across a district border but most moves take place within districts. Analyses of standard VII examination results from selected regions show pronounced differences in quality (measured in this way) between schools within districts. This may indicate that the district - as a political and administrative entity - does not have significant influence on educational quality. It is the school -especially the head teacher - and its relation to the village(s) it serves, that are crucial in explaining differences in examination results (Temu, 1992). Hence, attending a 'good' school may increase the chance of getting access to a government secondary school. Data on transfers between 'good' and 'bad' primary schools in Arumeru and Babati districts in 1990 and 1991 provide some evidence for transfers driven by examination results. It is summarized in table 3. In the 'good' schools, chances of selection to government secondary schools is very modest, but in the 'bad' schools it is nil. This is reflected in the figures for net transfers of pupils. They are large and positive for 'good' schools and small and mostly negative for 'bad' schools. In other words, some parents seek out 'good' schools and are able to get their children enrolled in them. The relatively few 'good' schools in a district and the many 'bad' ones help to explain why the number of transfers into 'good' schools is much higher than the small number of transfers out of 'bad' schools. Table 3 Net transfers to 'good' and 'bad' schools within districts, 1990 and 1991
Source: Transfer files from 10 'good' and 10 'bad' schools in each district (excluding Gidas school, Babati; and Burka Estate school, Arumeru). Gidas (a 'good' school) was excluded because a large out-migration was caused by draught. Burka Estate, Arumeru district (a 'bad' school) was excluded because a large in-migration was caused by changes in coffee estate employment. Obviously, some transfers of pupils between schools are unrelated to the quality of schools. When families change residence, children tend to follow. Such moves may be made in search of better livelihood elsewhere or to avoid localized calamities (local draughts, significant changes in plantation employment, etc.). [ The data presented here are corrected to account for this, as indicated in the notes to the table.] The evidence presented here shows that not all families are passive recipients of state supplied primary education. Some families carefully assess the quality of different primary schools before making a choice - and they have the means to finance such choices. But the numbers involved are still limited. Some 96 percent or more of all children - except in urbanized areas - attend the nearest school. Generally, school choice is an option available only for children of the better off.
Some lessons for reform If present trends in Tanzania continue, there is a real possibility that the public primary school system will lose out. More children of richer households will seek out the 'good' private and public primary schools. Such elite schools are already a prominent feature of the school system. Mainly children of the poor will not enrol in public schools because households increasingly perceive the costs of education to exceed its benefits. This widespread dissatisfaction has helped to put the need for reform of the public primary school system on the agenda. The evidence presented here about the role of local government and households provides a number of lessons for reform. Decentralization is a key issue in the Tanzanian discussions of reform. Proponents of supply-side responses to reform - mainly found in the ministerial educational establishment - face an uphill battle. There is an obvious need to decentralize decision making so that local governments will gain a substantial and meaningful role in primary education, but the present capacity of district councils is very limited, and, as far as decision making on school issues is concerned, councils are actually quite marginal. It would therefore be a grave mistake to focus capacity building efforts on the local governments per se. Instead, the focus should be on the multiplicity of actors in the political and economic environment of public schools which influence their performance. However, proponents of demand-side inspired reforms - led especially by the World Bank - push for a much more radical decentralization. According to them the role of local governments should remain marginal, but parents and school committees at community level should be empowered to run their schools (or establish private ones) with a large degree of autonomy from central and local government. The evidence presented here clearly demonstrates the need for a much larger involvement of households in school affairs than at present. But it also shows that the relativelyfew parents already actively engaged in securing better education for their children do so as individuals. There is much less evidence of collective action by resourceful parents who seek to increase the quality of and access to the local school for all school-age children in the community. The evidence presented here also highlights the dangers of encouraging market-oriented competition for pupils between schools as a means to improve school quality, as is advocated by the most vocal demand-side proponents. As shown above, certain parents do indeed use the possibility of choosing a 'good' school for their children if local conditions permit. But since such choices are mostly available to the few families with the right connections and the required funds, competition will promote school differentiation. It may increase the choices available to the better-offs, but it will decrease them for the poor. Finally, a key reform issue concerns financing. What mix of tax financing and user payment is appropriate? Proponents of demand-side responses to reform tend to ignore this question. They assume that parents are willing and able to contribute substantially more to the schooling of their children than the official records show they do at present, if only households are given the right incentives, namely more influence on the running of schools. Thereby, they ignore the evidence presented here, which clearly shows that parents already contribute a substantial share of resources to primary education. It may well be that present levels of financial contributions indicate the upper limit that households are willing and able to pay. Proposed cost recovery measures - which must be compulsory to be financially significant - may therefore lead to increased exclusion of the poor. It follows that if education for all is taken seriously, then the key issue is not how households could be made to increase their direct financial contribution to schooling. Rather the key question is: How can general taxation be increased and government budgeting priorities changed so as to provide free primary education for all? Recent experiences from Uganda indicate that financial constraints - rather than lack of parental control of schools or their dissatisfaction with education quality - are the key impediments to significant enrolment increases. Discussions of reforms of the primary education system are often highly ideological. Proponents of demand-driven responses to reform prefer the drawbacks of market failure to state failure. Proponents of supply-driven responses tend to have the opposite preference. The evidence presented here shows that much country-specific empirical work needs to be done to assess the merits of various reform proposals.
References Buchert, L. 1997. Education policy formulation in Tanzania: Coordination between the government and international aid organizations. A report for the IWGE. The International Institute for Education Planning, Paris: UNESCO. Carnoy, M. and J. Samoff. 1990. Education and social transition in the Third World. Princeton: Princeton University Press. Colclough, C. 1993. Education and the market: which parts of the neo liberal solution are correct? Innocenti Occasional Papers, EPS 37. Spedale degli Innocenti, Florence, Italy, July 1993. Colclough, C. with K. M. Lewin. 1993. Educating all the children: Strategies for primary schooling in the Oxford: Clarendon Press. Colclough, C. and J. Manor , eds. 1991. Neo-liberalism and the development policy debate. Oxford: Clarendon Press. Cooksey, B. and A.G.M. Ishumi. 1986. A critical review of policy and practice in Tanzanian secondary education since 1967. Mimeo. June 1986. Dom, C. 1995. Tanzania: overview of public expenditures for district primary education, 1993/94-1994/95. The Local Government Support Unit, Prime Minister's Office. Draft, September, 1995. Government of Tanzania. 1984. Educational system in Tanzania towards the year 2000. Recommendations of the 1982 Presidential Commission on Education ("The Makwetta report"). Ministry of Education. Dar es Salaam: Government Printers. Government of Tanzania. 1987. Makadirio ya mapato ya halmashauri na ruzuku ya matumizi ya kawaida kwa mwaka 1987/88. Ministry of Local Government and Cooperatives, Dodoma. Government of Tanzania. 1992. The Tanzanian education system for the 21th century. Dar es Salaam: Government Printers. Government of Tanzania. 1995a. Education and training policy. Ministry of Education and Culture. Dar es Salaam: Government Printers. Government of Tanzania. 1995b. Primary education master plan: a framework. Ministry of Education and Culture. Government of Tanzania. 1995c. Tanzania: Social sector strategy. Mimeo. Dar es Salaam. Ishumi, A.G.M. 1995. Provision of secondary education in Tanzania: historical background and current trends. In Semboja and Therkildsen 1995. Lawson, A. 1994. Under-funding in the social sectors in Tanzania: origins and possible solutions. Tadreg No. 2, March 1994. Lillis, K.M. 1990. Management of decentralization of education in Tanzania: Towards the identification of training needs. International Review of Education vol. 36, no. 4: pp. 417-440. Liviga, A.J. 1992. Local government in Tanzania: partner in development or administrative agent. Local vol. 18, no. 3: pp. 208-225. Lugalla, J.P.L. 1989. NGOs, self-help projects and development in Tanzania: The Njombe district development trust. In African Development Perspectives Yearbook. Malekela, G.A. 1995. Equality and Equity in Primary Education. Papers in Education and Development no. 16: pp. 65-74. Maro, P.S. 1990. The Impact of Decentralization on Spatial Inequality and Rural Development in Tanzania. World Development vol. 18, no. 5: pp. 673-693. Mosha, H.J. 1989. Management and Administration of Education in Tanzania. A paper presented at World Bank Education Seminar, January 1989. Dar es Salaam. Ministry of Education and Culture (MOEC). 1995. Community Education Fund Pilot for primary education (mimeo). 30 October 1995. Dar es Salaam. Mwaikusa, J.T. 1985. Control of local government authorities in Tanzania. Mzumbe, July 1985. Omari, I.M. and H.J. Mosha. 1987. The quality of primary educ Nairobi: Man Graphics. Oyugi, W.O. 1994. Service provision in Kenya: Who benefits? In Semboja and Therkildsen, 1994. Semboja, J. 1992. The impact of Economic Recovery Programme (ERP) on the provision of social services. Report prepared for the World Bank, December, 1992. Semboja, J. and O. Therkildsen. 1989. Recurrent cost financing at district level in Tanzania. CDR Working Copenhagen: Centre for Development Research. Semboja, J. and O. Therkildsen. 1991. Handbook on district level administration in Tanzania. Dar es Salaam: Ministry for Regional Administration and Local Government. Semboja, J. and O. Therkildsen. 1992. National and Sub-national level administration in Zanzibar. Unpublished report prepared for the WHO/UNICEF Nutrition Programme, Dar es Salaam, 1992. Semboja, J. and O. Therkildsen. 1995. Social services under stress in East Africa. State, NGOs and . London: James Currey. Shivji, I.G. 1991. Law, Democracy and the rights struggle. UDASA Newsletter/Forum no. 13, July 1991. Sumra, S.A. 1995. Enrolment Trends in Tanzania. Papers in Education and Development no. 16: pp. 49-64. Syrimis, S. 1988. Education in Tanzania: Key issues and proposals for remedial action. Report. May 1988. Paris: UNESCO. Tadreg. 1992. Poverty focused primary education project for mainland Tanzania: A description and analysis of key data and documentation. Project 005/91-2. Tadreg, Dar es Salaam, February, 1992. Temu, E.B. 1992. Successful schools in education with production programmes in two Tanzanian primary schools: the role of leadership. Seminar paper; December 17-18, 1992. Institute of International Education, Stockholm. World Bank. 1988. Education in Sub-Saharan Africa: policies for adjustment, Revitalization, and . A World Bank Policy Study. Washington, DC: The World Bank. World Bank. 1991. United Republic of Tanzania. Teachers and the financing of education." Report No. 9863-TA. Population and Human Resources Division; Southern Africa Department; December 9, 1991. World Bank. 1995. Investing in People. The World Bank in Action. Washington, DC: World Bank. World Bank. 1996. Community Education Fund. Human resource development pilot project. Various documents, July 1996. Wunsch, J.S. and D. Olowu, eds. 1990. The failure of the centralized state: Institutions and self-governance in Africa. Boulder, CO: Westview Press. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
