Textbooks are a resource that must be renewed regularly. As such, they differ from long-lived capital investments that may be funded by long-term loans.

Textbooks have short lives. Depending on the quality of their binding and paper, if they are well used and well cared for they may last three years, or even five. Where children carry them home each day over hilly paths, in seasons of heavy rain, to homes without shelves, they may not last that long. In northwestern Pakistan, the simple saddle-stitched textbooks are not expected to last longer than one year, and those for the early grades, whose owners have not yet been trained in book care, often must be replaced part-way through the year. Textbooks must also survive domestic hazard. Younger siblings are always a threat, and an extremely unlucky student may find his illiterate father has torn a page from a textbook to roll his tobacco in or his mother has taken a page for wrapping the sugar she is selling (Ambatchew 1999). In time, moreover, and regardless of their physical condition, textbooks grow obsolete as the curriculum changes and so must be replaced.

As a result, textbooks have never fit comfortably into the large-scale projects and procurement procedures of the World Bank and other funding agencies - institutions that were designed initially to help fund long-lasting capital projects such as dams, bridges, and roads. Textbooks and other learning material are, moreover, intellectual property, the rights in which are protected by national legislation and international agreement. Open competitive bidding that is based principally on cost can be used in procuring the paper and printing for textbooks, but fails when it is applied to textbooks that are individually unique and cannot truly be compared as simple commodities. While the World Bank has never stipulated that it looks for the cheapest price, procurement has been affected by a shortage of qualified staff and a less than perfect implementation of general guidelines for procurement, including a trial set of guidelines for procurement of textbooks issued in 1994 (Sigurdsson 1999, 144-5).

State-dominated systems for the provision of textbooks have generally proved inefficient and ineffective.

In countries where state institutions wrote, produced, and distributed a single textbook for each subject/level, the systems generally have been found to be inefficient and ineffective (e.g., Askerud 1997, 57; Bgoya et al. 1997, 21). The Swedish International Development Authority (Sida) found that in the programs it supported, "by and large government publishing has not functioned well. State-aided companies are frequently hampered by inefficient bureaucracy, unawareness of real costs, or lack of motivation on the part of the staff" (Sida 1996, 15). Other observers have noted that hidden subsidies masked true costs. Civil servants - often well versed in educational theory but unskilled in printing and publishing techniques - made decisions that might be pedagogically appropriate but unnecessarily costly. Books were produced in uneconomical page sizes and in wasteful designs. Distribution systems failed because of inadequate funding for transportation from regional warehouses to the schools. Everywhere, books were frequently late in reaching the schools, if they arrived at all; too often they remained in regional and district storerooms, ill protected from water, mould, insects, and vermin.

Involving the private sector introduces a motivation for efficiency and sustainability.

Entrepreneurs who are risking their own money, or their shareholders', are more likely to be concerned about achieving economies and making use of professional skills than civil servants who have no personal stake in textbook provision beyond their own survival in office. Of course there are many civil servants who are just as dedicated and professional as any private-sector manager, but even they are likely to be overcome by the inertia of large bureaucracies and the indifference or unwillingness of their superiors to make changes.

Publishing, in contrast to bureaucracy, benefits from a personal style that marries innovation and professionalism. It is also unusually complex in the number of "product lines" - each book constituting a separate product - that must be managed simultaneously in development or revision, production, and distribution.

Because the private sector sells textbooks, either to the government or on the open market, it has funds to re-invest in the next year's sales and a powerful motivation - the need for financial turnover - to make that re-investment. Books that are available this year, therefore, are likely to be available next year if the sales potential this year is realized.

The difference between public and private sector publishing was illustrated dramatically in francophone sub-Saharan Africa, where several countries agreed on common syllabi in mathematics and, with help from French aid organizations, cooperatively developed textbooks for common use. Two approaches to publication were followed. In the primary-level series, the inter-governmental coordinating committee that developed the books acted as publisher and retained the copyright. It offered film for platemaking to countries that signed an exclusive five-year agreement with it. The countries then had to secure their own funding, by grants or loans, and arrange production of the books. The series was adopted by only three countries. In Benin, the government took three years to find funding. The state printer was given the contract for books at the first two levels but could not fulfil it for technical reasons and subcontracted it to a foreign printer at a high cost. Benin subsequently chose books for the next two levels from a French publisher. In Mali, the contracted printer delivered the books a year late. Because the financing did not cover distribution, only part of the stock was supplied to schools, mainly in the capital and its neighbourhood. The Democratic Republic of the Congo also needed three years to find funding, then chose a European printer, and faced similar problems in delivery and distribution. None of the three countries had funds to reprint books in following years.

In contrast, the committee that developed the secondary-level textbooks assigned the copyright to a single publisher, the Hachette subsidiary EDICEF, which undertook to produce and market the series in collaboration with les Nouvelle éditions ivoiriennes of Côte d'Ivoire. The secondary series was adopted in twelve countries and sold in the open market, helped by a French subsidy to lower the sale price. About 1.5 million copies were bought, either in bookshops or within the framework of a bid, at the equivalent of US$5 per copy, about half the price of a comparable Ghanaian book, also the recipient of development aid. The books continue to be available throughout the countries because they are in the copyright of a publisher with an interest in their commercial longevity (Loric 1999b, 109-12).

School books provide the foundation for a sustainable publishing industry that can meet continuing local needs while reflecting local conditions and experiences.

The importance of school books extends far beyond the classroom. A large and relatively stable continuing demand for them, year after year, can provide the financial underpinning for a more general publishing industry. With the capital generated from educational books, publishers can engage in works of fiction and poetry and in works of non-fiction - books that examine the country's history, society, economy, ecology, and culture - that will help to develop a sense of national identity and self-awareness. Most important in this context, they can develop books in all these genres written specially for children and young adults, and sell them in the marketplace to a book-hungry population. In short, publishers who have financial resources from school books can provide the books that will maintain literacy once schooling ends.

Textbooks are particularly important to publishing in countries that do not have a well-developed reading culture. In industrialized countries, textbooks and learning materials represent between 25% and 50% of all publishing income. In developing countries where textbooks are sold, they can represent up to 95% of the local publishing market (Askerud 1997, 30). In sub-Saharan Africa overall, they account for 75% of publishing revenue (Makotsi 1998, 1).

When the state dominates the provision of learning materials, the general publishing industry languishes. This was clear in many countries of Africa and Asia, where the imposition of a state monopoly devastated publishing. As one publisher (Ofei 1997, 15) explained it, "a book industry without access to the textbook market is a dying industry, because the textbook market is the side of the bread that is buttered for the publisher and bookseller."

A public monopoly may appear attractive because it cuts out the retailer and concentrates production, leading to apparently lower costs for the consumer, whether that is the government or the family. The true cost to the country is great, however. Commercial publishing lacks a financial base on which to build. Bookshops do not develop because they have no books to sell. Quality of textbooks deteriorates from lack of competition. Book industries stagnate or die. "It is not surprising," Brickhill and Priestley (1993) observed, "that book development ... is weakest where monopolies are strongest." In contrast, when textbook provision is opened to commercial competition, even partially, industries grow. Children's book production, for example, increased notably in Kenya, Nigeria, and Uganda during the 1990s, in number of titles and in copies per title printed (Faye 1998, 5).

The ownership - local or foreign - of commercial publishing houses that produce textbooks is not an issue so long as their publishing activity is truly locally based and locally oriented. Locally produced books are likely to be cheaper than imported books from Europe or North America, and far more likely to be appropriate in content, illustration, and emphasis. In contrast, the foreign imported children's books - the most common form of children's literature in the bookshops of many African countries - have been described by Chinua Achebe as "poisons wrapped in between beautiful covers" (Chakava 1997, 42). The president of Tanzania has said that it is "incumbent upon African publishers to protect African value systems and ways of life" (Mkapa 1997).

Government policies for book provision are often impeded by other government policies and practices.

The complex of ministries that impinge on books and other learning materials was described in Part III. Not infrequently the policies of other departments impede the policies of educational authorities. Most notable are high tariffs imposed by the Ministry of Trade or Finance on imported paper, other raw materials, printing equipment, and replacement parts. Such tariffs are particularly onerous in countries that have acceded to the Florence Agreement on the free movement of books and charge no duty on imported finished books but levy high tariffs on the materials for publishing books at home. Tariffs may be used to shield inefficient domestic paper mills as well as to raise revenue. An ADEA/APNET study on the intra-African trade in books has illustrated the contradictions in taxation and legislative practices that can work against the development of the book chain in countries committed to Education for All.

High interest rates set by the Central Bank also inhibit book development by commercial publishers, who may be further penalized by government subsidies to parastatal publishing and printing houses with whom they and their printers compete. The Ministry of Culture or Education may require textbooks to be produced in two, three, or even 11 local languages, a policy that is pedagogically sound but adds greatly to the investment involved in book development. The absence of intellectual property legislation, or governmental indifference to its infringements, encourage piracy of books and fragment markets that may already be marginal.

When there is no coordination among ministries in policy or regulation, the cost of providing school books generally rises. In times of fiscal austerity, that usually means that fewer books will be produced or bought, and the textbook:pupil ratio grows unacceptably high.

Books cannot be considered independently but only as part of a process involving many players and policy issues.

A school book is part of a continuum that begins with the student's needs, as defined by a curriculum institute, and ends with the student's use, inside or outside the classroom. In between there are many stages: curriculum development and authorization, identification and commissioning of authors, drafting and revising of a manuscript, preparation of illustrations, evaluation by the publisher in field tests, evaluation by the Ministry of Education, possible revision and further evaluation leading to authorization for use of the book in schools, production (typesetting, printing, binding), and finally shipping to a government warehouse and from there to individual schools or, alternatively, shipping to wholesale and retail outlets that will sell the book to families. This is not an exhaustive list of processes or players. The World Bank or another funder may be involved, with its own procedures and guidelines which may be different from the government's. Development agencies and CSOs may supply paper or expertise. Teachers will need training in the use of a new textbook, and teachers' guides must be prepared - following all the same steps - simultaneously for them. Any weak links in this book chain will, as we have seen, affect the availability of books in the classroom.

The policy issues that must be examined have been set in detail by several authors (e.g., Buchan et al. 1992, 66-7; Askerud 1997, 23-4; Bgoya et al. 1997, 104-5). They include:

· selection of subjects requiring textbooks (core subjects or all)

· use of national and local languages in instruction and textbooks

· timing of reforms and revisions · rewards for authors

· target textbook:pupil ratios (if books are provided by the state)

· technical specifications for paper, printing, binding

· procedures for evaluation and authorization of textbooks by the Ministry of Education that are - and are seen to be - objective, thorough, comprehensive, fair, timely, concerned with social content, and in the public interest

· affordable and sustainable financing

· open competitive bidding for printing and/or publishing

· practical and economical distribution

· professional development for publishing personnel and educators

· identification and provision of supplementary reading and other learning materials, including their development, publication, distribution, and financing

· support for public and school libraries and the training of teacher librarians

· positive industrial policies affecting publishing

Failures in planning and coordination will affect the provision of textbooks and learning materials adversely. External support will be meaningful only if it is rooted in detailed strategic plans, not on general policy statements (Bgoya et al 1997, 101).

Funder and national policies fail when they are directed to short-term provision of learning materials without concern for the totality of the process

Large-scale loans in the past have unquestionably reduced the pupil-to-textbook ratio in many countries from extreme scarcity to a more effective level. But book provision, as Bgoya (1997, 27) has pointed out, is not the same as book development. Textbooks have too often been viewed as commodities to be delivered in the short run to meet immediate needs. As a World Bank team leader once remarked with passion: "All I care about is getting books into the children's hands!" If the government receiving a loan for textbook provision is still unable to fund its needs when those books must be replaced, a second loan must be negotiated. The World Bank acknowledged, in its draft policy released in 1998, that "many book provision efforts over the last 25 years, some with Bank support, may have achieved their immediate objectives but have been unable to maintain the service over the longer term to sustain the educational impact that textbooks help to achieve". This has also been true of loans by the Asian Development Bank.

Newton (Sosale 1999, 158) has pointed out that is not only funders that have been responsible for short-term solutions. Ministers of Education have generally been avid for the votes that come with the promise of free books in the coming year. An insistence on international competitive bidding for all contracts above a certain threshold, while understandable from the viewpoint of the funders, has exacerbated the effects of short-term provision by channelling most of the loans for textbooks out of the recipient country. Reports to APNET indicate that in book procurement programs supported by World Bank lending, the ratio of books from local publishers to those from foreign-based publishers is 1 to 50.

A book purchase made on the basis of a one-off cost import is not a good bargain when we consider: the opportunity cost of jobs lost to the economy, since the book was not produced locally; the cost of discouraged innovative trading, which would have resulted from higher turnover and increased profitability; the cost of capacity not built; and the cost of lost spillover benefits, which in turn would have invigorated the textbook industry. As a result, these books are not cheap (Nwankwo 1999, 141).

As the same author has pointed out (Nwankwo 1996, 26), loans are negotiated between the Bank and a national government. It is the borrower who selects consultants and implements procurement. Very few governments fully understand the critical role indigenous publishers can play in national culture and, more specifically, in textbook provision, nor have they paid serious attention to the conditions that would encourage an indigenous publishing industry. The publishers have a responsibility to raise awareness of these issues if they want change.

An experienced textbook publisher, now a consultant, recommended to the Canadian International Development Agency that any project related to educational books be preceded by an in-depth analysis of the book sector; be designed to cover a minimum of ten, and preferably fifteen, years; be coordinated with other agencies; view textbook provision within an overall framework of developing local capacities in publishing, production, and distribution; give priority to financial sustainability and substantial private sector involvement; and use experts in curriculum development, publishing, production, and distribution in the project's planning, implementation, monitoring, and evaluation (Clare 1993, 86-8). The Soros Foundation's principles for support of the book sector show the same holistic approach. They include: using the commercial potential of a book wherever possible to keep subsidies to a minimum and to provide an incentive for efficiency; leaving as much initiative as possible to the publishers; avoiding monopolies; and making support part of an educational process that aims at reducing dependence on external help through carefully focused training and business planning (Pinter 1999, 46-7). These are desiderata that any funding agency should consider.

Recent policy trends - decentralization, liberalization, and multiple choice in textbooks - require a clear redefinition of the role of the state in a number of areas.

The state has the initiative and ultimate responsibility for the timely and adequate provision of textbooks in any publicly funded educational system. It decides what textbooks will be used, how they will be published, and how they will be paid for. These responsibilities remain even when some functions are performed by the private sector. The International Commission on Education for the 21 Century (1996, 160-2) defined the state's role as follows:

Education is a collective asset that cannot be left only to market forces. Thus, whatever the organization or degree of centralization or diversification of a system, the state must assume certain responsibilities to its citizens ... That role ... must not, however, be exercised as a strict monopoly. It is more a matter of channelling energies, promoting initiatives and providing the conditions in which new synergies can emerge. It is also a matter of insisting on equity and quality in education ... more specifically, the state should play a redistributive role, to the benefit of minorities and the underprivileged especially. Guaranteed educational quality moreover implies the establishment of general standards and various monitoring devices.

Ingemar Gustafsson, then head of Sida's education division, defined a further public responsibility, to ensure that textbooks reflect the culture of the student-users' own country. This in turn requires each country to have some control over and capacity to produce its own school books (Brunswic and Hajjar 1992, 37).

At a minimum, the state must be expected to prepare clear and detailed curriculum guidelines; make them available for the development of textbooks; establish an objective process of evaluation and authorization of textbooks; decide the channels to be used in funding and distributing textbooks to the schools; set minimum physical standards of production; perform the same functions with respect to other learning materials, including teachers' guides and supplementary reading; train teachers in the use and care of textbooks and learning materials; and protect intellectual property rights through appropriate legislation. The state may also be expected to support libraries and promote literacy through community-based campaigns.

Beyond this, the state may play a role in any of the permutations outlined by Searle (Table 4) and described in preceding pages. Experience suggests that no option is perfect, and what is best for one country may be unsuitable for another. The most critical choices, in financing and distribution, are discussed briefly below.

It is possible to control the cost of providing textbooks and other learning materials - but it takes professional expertise and careful planning.

Too often, inexperienced managers and editors ignore ways to save money in textbook printing that are standard practice among trained professionals. The size of a book's page can have surprising implications for its cost, especially when paper represents a high proportion of total production expenditures. The most economical page size is determined by the size of the sheet of paper used in printing, which in turn is related to the size of the printing press being used. Any mismatch between press and sheet, or between sheet size and page size, results in waste. Production costs can also be reduced by restricting the number and size of illustrations, the use of colours other than black, and the total number of pages. More generally, governments can reduce costs to state and families by limiting the number and size of textbooks required.

The state and private enterprise can work closely together to provide textbooks efficiently under a regulated monopoly.

Privatization has been successful in providing a long-term supply of textbooks in Côte d'Ivoire, where the government opted for a system of purchase by families at both primary and secondary levels and concentrated production in a very few hands. In 1991 the country's two textbook publishers, le Centre d'édition et distribution africain (CEDA) and les Nouvelles éditions ivoiriennes (NEI), both parastatal partnerships between the government and French publishers, were technically bankrupt following an economic crisis and social disorder. In order to obtain new capital they were transformed into private companies. In both, the government retained 20% of the shares and further shares were set aside for individual Ivorians to ensure majority local control. The balance of NEI shares was held by three French companies, and the balance of CEDA shares by one French and one Canadian company. The new companies were granted exclusive rights, shared equally between them, to publish primary school textbooks, teachers' guides, and work books, all written in the Ministry of Education. The two companies are also active at the secondary level, where the state issues a list of approved textbooks. The state regulates the retail prices of textbooks and in the first two years of the agreement negotiated substantial decreases; prices were frozen even after the devaluation of the CFA franc increased the cost of imported materials substantially. The price of textbooks to disadvantaged areas is subsidized. Through its shareholding and regulation, including quality control , the state acts as "school protector" while remaining committed to economic liberalism.

The results of the Ivorian policy are substantial. Textbook prices are said to be the lowest in all francophone Africa, except where prices are artificially lowered by external subsidies. The two national publishers sell more than 4 million primary schoolbooks each year for a school population of about 1.5 million students, and there is an additional market in second-hand books. It is estimated that 75% to 95% of students, depending on their level and where they live, own French and mathematics textbooks, although fewer own books in other disciplines. The system has encouraged the development of the only major printing industry in francophone sub-Saharan Africa and of a distribution network of wholesalers, bookstores, and other sales outlets. The companies have returned profits. The World Bank is supporting free distribution of the textbooks in 116 marginalized zones (Palmeri 1996; Loric 1999c; Couassi-Ble 1999; Moingeon 1999a; ADEA/UNESCO survey 1999).

It takes time to develop a competitive book sector that can sustain the production and distribution of good quality school books and learning materials.

The Ivorian policy is based on a single textbook for each subject/level. The quality of a textbook determines how well the needs of the users are met, and a good quality textbook in a single-textbook system is probably better than several competing textbooks of poor quality. Many educators and publishers argue, however, that a multi-textbook system is preferable because competition encourages publishers to innovate and improve their textbooks and teachers' guides. Gaston de Bedout, a Colombian commercial publisher, presents another reason for moving away from a single compulsory textbook in each subject, whether published by the government or bought by the government from a friendly publisher. In the final analysis, (de Bedout 1999, 41), "education through a texto unico will result in a mass of people desolately uniform, indoctrinated by the official history, philosophy, science, grammar, chemistry, biology, and even the official calligraphy."

A competitive publishing industry cannot develop overnight, however. Personnel must be trained in manuscript development, editing, design, illustration, field testing, production control, marketing, warehousing, and distribution. Authors must be identified and they too must be trained to meet the special demands of writing textbooks and teachers' guides. Training tended during the 1990s to focus on the more obvious needs of writing, editing, design, and illustration. Less attention was paid to preparing publishers for the underpinning skills of operational and financial management, marketing, and evaluation - an imbalance that is gradually being redressed.

Training takes time and practice. So does the development of learning materials. Typically, a textbook takes three years from inception to the classroom - roughly, one year or more to commission authors and for them to write a first draft; one year for field testing, revision, and evaluation by the Ministry of Education for authorization; one year to produce final books and, especially where distribution is difficult, get them to the schools. The process can be accelerated, but only if the commercial sector is sufficiently established to meet the challenge. It can be impeded by difficulties in securing capital or credit and, too often during liberalization, by bureaucratic foot-dragging.

Uganda's experience illustrates the time an industry takes to develop. Up to 1993, the National Curriculum Development Centre had a monopoly on textbook production, and the few books it produced were assigned to two transnational publishers. Other textbooks bought under a World Bank project were all imported. In 1993 the centre's monopoly was removed and selection was devolved to the schools, choosing from a centrally approved list. By the end of the decade 10 publishers were sharing the market. The change came about as a result of a general government policy of economic liberalization, lobbying by a newly formed book publishers' association, and pressure from the principal funder of a primary-level textbook provision scheme, the United States Agency for International Aid (USAID). About US$20 million was spent under the project between 1993 and 1997, and publishers responded with vigour. Sales representatives combed the countryside, visiting schools and addressing teachers' meetings - no easy job in a country with more than 8,000 government-aided primary schools, many of them in remote areas with poor roads. Local publishers built sales and skills. Some mutually beneficial joint ventures developed with foreign partners. The quality of textbooks improved. Publishers from abroad set up offices employing local people. Bookshops expanded - from five in Kampala in 1990 to 25, and others up-country. Publishers met teachers directly to offer guidance in the use of their books. The textbook:pupil ratio, which in some schools was as low as 1:20, was raised. The momentum is expected to increase as further international funding becomes available. The Ugandan book chain is developing, but its local base - even after more than half a decade of such activity - is still narrow. The number of local publishers is small, and the sales of textbooks have not yet been matched in other kinds of books. More than 90% of the primary-level textbooks are still published by British-based transnationals who have local partners that are basically only marketing agents. Most of the printing is done abroad, because local printers lack the capacity and - in a chicken-and-egg situation - cannot afford to build capacity without the lucrative textbook jobs. The government still procures textbooks centrally, consolidating orders from the various schools, and delivers them to the schools free of charge. As a result, bookstores reap no benefits, nor does the infrastructure for general publishing. Local publishers find it difficult to obtain credit because banks are reluctant to lend money to businesses with only one customer, even if it is a government ministry. After six years of rapid development, the book sector is still heavily dependent on external inputs. Without any attempt at cost recovery, the provision of primary-level textbooks remains dependent on external funding (Tumusiime 1998a, 101-2; Tumusiime 1998b, 12-13; Tumusiime 1999b, 3; Salahi 1998, 12-15; Katama 1997, 4-5).

The Ugandan experience is not atypical. Publishing is not an industry that can be jump-started. Capital is a particularly high barrier to local publishing development, and there have been several suggestions (e.g., Bgoya et al 1997, 99; Davies 1997, 73) that governments guarantee loans to publishers, as in the Dag Hammarskjøld scheme, or extend cheap credit to publishers as many do to farmers for the purchase of seeds and machinery, or obtain concessional credit for publishers through the World Bank and regional development banks. Government-guaranteed loans have proved their value in the development of the Canadian publishing industry for more than 20 years.

Countries may also impose regulations to encourage local development. Some countries, like the Philippines, have protective legislation requiring local ownership of educational publishers. Zimbabwe, which has a long history of liberalization in textbooks, passed an "import substitution" regulation in 1983 which stipulated that any textbook required in a quantity of more than 1,000 copies must be locally published. In effect, foreign publishers who wanted to tap the Zimbabwean market had to license their books to a local publisher; as a result, import expenditures were reduced and the local industry was given a powerful boost. Mozambique, recognizing that textbook shortages were too severe to be met by local development, adopted a similar policy in 1998, under which it may authorize licensing, adaptation, and co-editions of books published elsewhere, but no more than 2,000 copies of any foreign book may be imported. To meet greater demand, books must arrive under the imprint of a Mozambican publisher and/or only with special permission from the Ministry of Education (Nyambura 1998, 7-8). Ghana encourages joint ventures between local and foreign publishers but insists that Ghanaian authors and illustrators comprise 70% of any writing team and Ghanaian printers do 60% of the printing. Still, without competition at the international level, quality can suffer. A self-sustaining but complacent local industry may fossilize, as happened in Algeria (Moingeon 1999, 147).

In a transition from state monopoly to competition, due attention must be paid to preparing a commercial environment and a level playing field.

A transition from state monopoly to private competitive provision of textbooks cannot occur without a welcoming political climate and a clear, well designed and open policy - preferably one that has been developed in consultation with the book industry. A favourable environment for commercial publishing requires (Bgoya et al. 1997, 24): decentralization of funds (or at least purchasing authority) to enable schools to select the textbooks and other resources they want to use and buy; capacity building of new publishers to increase choice of textbooks and competition in quality and price; an objective and open system for approving textbooks for purchase; training in assessment, selection, and purchasing of textbooks for school administrators and teachers; adequate information systems for procurement of textbooks; and effective commercial distribution systems. The same authors argue that a multi-textbook system can be completely successful only if book purchasing is done at the lowest level, and if all the books that are approved have an equal opportunity to be purchased.

When one publisher dominates the market, the second condition cannot be met. This has happened when the state continues to produce textbooks through a commercialized parastatal that may enjoy special benefits, such as subsidies or access to subsidized government printing, in competition with small private publishers who have difficulty in obtaining credit and raw materials. This situation has occurred in many countries of Eastern and Central Europe and in some countries of Africa.

Even when the state withdraws from publishing, the field will not necessarily be level. If local commercial publishers are unprepared to deliver textbooks in sufficient quantity and quality, transnational publishers will be quick to fill demand, as in Uganda. Countries that use a metropolitan language for instruction are especially open to transnational domination. "Ultimately, it will be simply a matter of who is able to develop, publish, and place the manuscripts on the market more quickly." (Brickhill 1996b, 20).

Cooperation and co-publishing can speed the transition to sustainable, competitive provision of learning materials and reduce the risks attendant on the process.

For nascent publishing industries, the fastest route to sustainability may be through cooperation. It may be through adaptation of a textbook produced in a neighbouring country, such as the science series developed on a Kenyan model that gave Fountain Publishers of Uganda rapid access to a newly-liberalized market. Or it may be a joint venture between a local publishing house and a well-established house in an industrialized country, involving the transfer of know-how as well as capital, as happened in Jamaica, Mali, Cameroon, and Namibia (described in Part III). An impressive example of cooperation between local and non-local companies is the partnership of Heinemann Educational Publishers of the UK and New Namibia Books. When it began in 1990, the local partner consisted only of its owner, who had recently returned from England with no resources. By 1997, the collaboration had produced two core textbook courses, in primary and junior secondary science, for a country that had been almost totally dependent on South African imports. New Namibia had eight qualified employees and an extensive list of other books which it had developed with profits from textbooks (Sulley 1999, 193-5). I

ntergovernmental cooperation is also possible. In examples described above, several countries of francophone sub-Saharan Africa established common curricula and developed common textbooks in order to achieve economies of scale. The potential of intergovernmental cooperation was also recognized by a group of anglophone and lusophone policy- and decision-makers, educational planners, curriculum developers, and textbook development managers from 10 African countries at a seminar in 1991. They suggested cooperation was particularly appropriate for countries and island communities whose populations were too small to make the local production of school books economical or sustainable. Subregional clusters of such countries could form common examination councils and develop standardized curricula, opening possibilities for regional negotiations with publishers for bulk supply and bulk prices (Brunswic and Hajjar 1992, 22). Regional adoptions to achieve economies of scale in procurement were subsequently recommended for the countries of the Organization of Eastern Caribbean States (Clare 1993, Appendix E, ii). Few programs along these lines have been implemented, however. A common textbook implies a common language of instruction, normally a metropolitan language as in the francophone example, in contrast to a trend towards use of local languages. Brunswic and Hajjar have observed, moreover, that in their experience cooperation is more difficult to develop between governmental institutions than between commercial publishers.

In principle, cooperation between publishers in neighbouring countries should be practicable. Curricula differ across borders, but in many instances are similar enough for transnational companies to adapt the textbook used in one country for adoption in another. Local publishers should be able to do the same. Cooperative ventures are impeded, however, by lack of information, poor communications, ineffective marketing, national bias, customs and tariff barriers, foreign exchange restrictions, and cumbersome bureaucracies. In higher-level subjects such as science and mathematics, imported books will probably continue to play an important role. In primary-level subjects, especially those without a major socio-cultural component, co-publication and adaptations should be expected to increase. Within Africa, the value of trade in books of all kinds between countries doubled between 1992 and 1996, and the Zimbabwe International Book Fair annually reports increasing international business. At this year's fair, just over one-third of exhibitors and visitors responding to a questionnaire said they had come to negotiate, confirm, or revisit co-publishing agreements (Anonymous 1998, 5; Ling 1999).

The efficient provision of books and learning materials, like all other parts of educational planning, requires effective information systems.

The lack of reliable and up-to-date information needed for efficient textbook provision has been noted in Parts II and III. The potential for change has been demonstrated in Jordan, where an Educational Management Information System (EMIS) was installed beginning in 1990.

Up till then, Jordanian statistics had been gathered centrally on an outdated mainframe computer, in flat files not easily amenable to statistical applications, in a form principally dictated by the Educational Statistics Yearbook. The Ministry of Education and 23 regional directorates collected their own information and held it in their own locations, usually in manual records. To complicate matters further, the identification number of each school changed every year. The system could report gross enrolment and student flow rates, but the aggregated data could not be disaggregated for analysis or for comparison, for example, between districts or schools or between schools and an absolute standard. For the new system, personal computers were bought, relational database programs installed, staff trained at the central and regional levels, and EMIS operations gradually transferred to the regional offices.

At the end of five years, data entry time had been reduced from about six months to two weeks and the accuracy of data was substantially improved. The Jordanian experience demonstrates that, with proper initial technical assistance and guidance, and even modest means, a developing country can improve its educational information-gathering dramatically (Ahlawat and Billeh 1997, 274-87).

The economical provision of textbooks and learning materials requires a balance in curriculum development between the pedagogical desirability of revision and the financial desirability of stability.

Innovation in curricula is necessary to introduce new learning and new methodology. But abrupt or frequent changes can throw the textbook provision system into disarray and discourage publishers from investing in textbook development.

The provincial textbook boards in Pakistan, for example, used to complain about the almost annual small revisions that issued from the federal curriculum developers in Islamabad, forcing minor but expensive changes in the textbooks published provincially. In the Philippines, a Secretary of Education decided to re-instate science as a subject in the first two primary grades just two years after his predecessor had removed it from the curriculum; science was officially on the curriculum, but there were no textbooks to teach it and relatively few teachers trained to teach it without such help. These are minor examples compared to the changes that occurred in the post-communist countries of Central and Eastern Europe and Central and West Asia, where, as reported in Part IV, textbooks still lag behind the new curricula after almost a decade.

Commercial textbook publishers in particular seek stability in curricula. In order to keep the price of textbooks as low as possible, they are anxious to amortize the substantial costs of development (writing, editing, design, illustration, typesetting, and field testing) over three to five years of sales. If the curriculum changes significantly during that period, they cannot recoup their investment. If they have printed books in anticipation of sales under the old curriculum, they will have a warehouse full of unsaleable books. For these reasons, when the Philippines moved from state to commercial production of textbooks in 1995, the publishers requested, and were granted, a guarantee that no substantial changes would be made in the curriculum for five years.

The impact of abrupt change on commercial publishing may be seen in South Africa, where the government of Nelson Mandela made two major changes in educational structure and policy. The first was to unite the fragmented apartheid system of 18 education departments into one national department that sets policy and nine provincial departments that are responsible for operations. The second was to promulgate a new curriculum that is outcomes-based, a radical shift from the authoritarian system that existed under apartheid. The new curriculum was to be introduced at the rate of two grades per year. The schedule was stretched out subsequently, but publishers initially set out to develop, write, design, and produce textbooks for grades 1 and 7 in just 18 months - and to do so in most, if not all, of the 11 official languages. That they achieved as much as they did was remarkable, although most were forced to cut back to three to five languages besides English and Afrikaans. When the date for implementation in grade 7 was deferred, publishers were faced with development costs but no sales. There were problems with the selection and authorization of textbooks, in part because of inadequate training of evaluators, in part because of bureaucratic rigidity, and decisions were inconsistent between and within provinces. Books were often delivered after the beginning of the school year. Budgets for buying books were far less than expected. The government had promised a free and equal education for all children up to grade 9, with a book on every child's desk; in fact, however, the money for textbook acquisition dropped more than 80% between 1996 and 1998 as the bulk of the educational budget was consumed by increased teachers' salaries, and a promised increase in 1999 fell far below what was required. The consequences to what had been a well established and confident publishing industry have been devastating. Companies faced with high overhead costs for the development of new textbooks and severely reduced markets have started retrenching. An estimated 44% of the educational publishing industry has been lost in the last eighteen months - a demonstration of the fragility of an industry that depends upon a single customer (Horwitz-Gray 1999, 106-8).

The state can act to moderate inequities, which can occur whether books and learning materials are provided free of charge or under a system of cost recovery.

State purchase of textbooks for free distribution through the schools is desirable to ensure equity in education, but few of the countries surveyed have been able to meet a target textbook:pupil ratio of 1:1 or 1:2 throughout their jurisdiction or to sustain that level even with external assistance. To maintain free distribution, the state must commit funds annually to production or purchase, distribution, warehousing, and supervision. Few have been willing to do so in the requisite amounts, and budgetary shortfalls usually affect the districts already most disadvantaged. Both Jamaica and Mexico did achieve the target during the early 1990s, but at the cost of inferior books in Jamaica and ballooning cumulative expenditures in Mexico (Clare 1993, 39). Furthermore, free distribution discourages private booksellers who are the key to continuing literacy and engenders a belief that books should never be paid for. "Even politicians begin to consider books free commodities" (Muita 1999, 154).

Unless a state has the both the tax base and the political will to allocate sufficient amounts of public money for the provision of learning materials, it will continue to be dependent on external support or have to attempt some form of cost recovery. Both may be necessary.

The state can relieve itself of all or part of the financial burden of textbook provision by expecting families to pay for the textbooks their children use, just as they may be expected to pay for uniforms or work books. This approach penalizes the disadvantaged. Poor families may not be able to afford to buy textbooks, and families living in remote areas may have no access to them for lack of retail outlets stocking them. Families may have to choose which children to educate, usually at the expense of girls. They may also discriminate between subjects, buying textbooks only in those subjects that seem most likely to lead directly to future employment. This system does encourage private bookselling and instils the book-buying habit. Governments can subsidize the cost of production to reduce the cost to families, as happened in Côte d'Ivoire after the radical devaluation of the CFA franc.

Book rental schemes, in which books owned by the state or school are lent to students for the school year on payment of a fee, reduce the cost to families because the cost of the books can be amortized over three or more years. It is possible, moreover, to allocate funds to the schools so they can buy textbooks and learning materials through local booksellers, thus supporting the retail part of the book chain. The system is a compromise between state purchase and family purchase with benefits of both, but requires durable books, safe storage between school years, and educational staff trained to handle money, manage stock, and supervise loans. It works only if the fees go into a revolving fund, the fund is used only to replace worn-out or obsolete materials, the books last as long as expected, teachers and children are trained in book care and repair, few copies are lost or stolen, book prices and enrolments do not rise abruptly, revisions to the curriculum are not too frequent, and families can pay the rental fee. The system is usually cost-effective, given that the cost of a book that will last for three to five years is probably no more than 50% greater than the cost of the same book produced with poorer paper and binding that will last only one year (Buchan et al. 1991, 10).

However, book rental was found to be a less than perfect solution in a study of representative primary-level public schools, urban and rural, in Guinea. Only 32% of the books available in the schools were actually rented to students. Most of the students came from poor families, or families that were illiterate in French. The parents already had to spend money for exercise books, pencils, and other school expenses, all at the beginning of the school year. Some parents were reluctant to spend a substantial amount of money for books that might be lost or damaged by their children (Diallo 1999, 18). Problems have also arisen with late payment of fees and accountability in financial management, as reported, inter alia, in a forthcoming ADEA publication, Financing Textbooks and Training Materials in Africa. One consultant with considerable international experience suggests that revolving funds work best in small discrete communities where the disbursement and accounting pathways are relatively short, as at the school or local level or in small island nations (ADEA/UNESCO survey 1999). Once again, the government can reduce the cost to families, this time by annual grants to the revolving fund.

Externally funded initiatives can relieve inequities most effectively through demand-side financing and targeted subsidies.

Many projects in support of textbook provision have subsidized the supplier - usually the government - in ways that have either reinforced a state monopoly or distorted the market. Subsidies for production, whether in cash or in materials such as paper, tend to mask the real costs of production. Unless special care is taken, they result in prices that cannot be maintained if external assistance ends. They can help to solve short-term problems if applied with care, but may have no positive long-term results. In general, subsidies should be applied only to the one-time costs that are incurred before printing begins. Subsidization of state monopolies has usually in the past reinforced the monopoly without building sustainability. Subsidization of individual (possibly state-owned or parastatal) publishers in a competitive market has given the recipients a competitive advantage and discouraged the entry of other publishers. Subsidizing the supplier may, moreover, prove ineffective if the subsidy is inadequate. Read (cited in Askerud 1997, 72) reports that in one country where books were sold, supply-side subsidies succeeded in reducing the cost of the textbooks by half but sales did not increase; the books had simply been made cheaper for the elite who could afford them already.

Increasingly, funders are turning to demand-side subsidies as a more effective way to assist in the provision of textbooks. The subsidization becomes more effective the further down the education ladder it reaches. Direct subsidies to parents or pupils help to keep the cost to them of textbooks low and opens opportunities for competition among publishers without creating market distortions. If it is impractical to provide subsidies to families the money may be given to individual schools; and if that is impossible, selection can be placed in the hands of the schools even if the responsibility for finances and procurement is at a higher level.

Under a scheme supported by the International Development Association, for example, stipends are paid to young women in Bangladesh who are preparing for the secondary school certificate. Education among women in Bangladesh is among the lowest in the world, and access to the examination had been restricted to those who could afford examination fees, stationery, and private tutoring. The stipend includes a book allowance and also covers tuition, school fees, uniforms, shoes, umbrellas, transportation, and other costs. Stipends are paid twice a year into bank accounts in the names of the students, who thus also learn to manage money. Parents must agree that their daughters will attend school for at least 75% of the school year, a condition that recognizes that about 90% of rural girls work as household helpers. The students must obtain at least second-class marks and remain unmarried until they earn the secondary school certificate (Patrinos and Ariasingam 1997, 21-4). A similar scheme is under way in Guatemala.

Alternatively, supply-side subsidies are being targeted directly to disadvantaged populations. The development of local-language textbooks for minority groups in China and Mexico was mentioned in Part III. In both the Philippines and Bangladesh, the Asian Development Bank is funding textbook provision only in the poorest provinces, leaving it to the national government to fund free provision elsewhere in the country. Namibia introduced a policy of positive discrimination in 1998 under which it provides textbooks, posters, maps, wall charts, and encyclopedias to disadvantaged schools (ADEA/UNESCO survey 1999).

Continuing attention must be paid to issues of gender equity, minority-group needs, and local language.

Advances have been made in respect to gender equity and sensitivity to minority groups, but much remains to be done. Some ministries of education have established social content guidelines for the authors, publishers, and evaluators of textbooks. Even when efforts are made, however, stereotypes slip past. In Zambia (Brickhill et al. 1996, iii), textbook writers included many examples of role reversal and balance and female assertiveness. But in one book a man and a woman are still seen introducing themselves by saying "I am a man" and "I am a woman", the man with his hands hanging freely at the side, the woman with hands cupped in a posture of general deference.

The provision of textbooks in local languages is a matter of pedagogical as well as social concern, for programs that teach people in their own language have been proved to be more effective in both Latin America and Africa than those using the language of a former colonial power (Jung 1999, 25). Where local languages are large, as Hausa is in Nigeria, the economics of local language publication are manageable. For small language groups, the problems are more formidable because the print runs are so much smaller. There may be additional difficulties in securing typefaces for special scripts, or in agreement on orthography, or in finding writers and editors with the requisite knowledge of subject and language. There is, however, considerable scope for South-South cooperation, building on experience in the Pacific and elsewhere, and in the late 1990s there has been renewed interest in Africa in publishing in cross-border languages.

Teachers require training in the effective use of textbooks and learning materials. New textbooks and learning materials in particular are unlikely to be used unless teachers are adequately prepared for their introduction.

In the concern over the inadequate provision of textbooks and learning materials, too little attention has been paid to the use of those materials and the continuing training of teachers in their use. Generally, pre-service training should be reinforced by in-service training through various channels and by school inspectors. Some countries have tried to establish resource centres to help teachers improve the learning environment in their classrooms; others have established teacher clusters to provide peer support and in-service training. Not all such initiatives have been maintained, however. When teachers fail to use books well, they deprive themselves and their pupils of a fundamental tool for promoting reading.

In particular, pre-service training in many countries fails to give due emphasis to the preparation and use of teaching and learning aids. This failure may result from lack of hours in the program, or the inadequacy of teacher trainers in this area, or both. Creative teachers, properly trained and motivated, can overcome a lack of professionally produced materials by making their own teaching and learning aids, often with the help of the pupils or even community members. (Parents do not have to be literate to participate, as long as the meaningfulness of the work is clearly explained.) Nevertheless, few teachers in the eight countries in the UNESCO/Danida case-study survey made any effort to make teaching and learning aids, even though they seemed aware of their importance. Jamaica was the only country in which a majority of teachers made their own aids, and involved children in the activity.

Locally made aids are most likely to be appropriate for the needs and experience of the children concerned. Their value does not, however, remove government's responsibility for providing both educational materials of good quality and in-service training in the preparation and use of teaching and learning aids. Training programs in future must also prepare teachers to use many forms of educational technology and techniques, both traditional and non-traditional, including electronic media, music, dance, drama, folk theatre, newspapers, and games. Teachers will also have to become increasingly aware of intellectual property rights, in particular with respect to material found on the Internet.

In Africa, Asia, Latin America, and Central and Eastern Europe, teachers have proved reluctant to use new textbooks, especially those embodying new methods. Those accustomed to a single "official" textbook tend to stick with it even when other books are approved and available. Human nature is essentially conservative, but is more so than normally when people are underqualified and insecure.

Where competition among textbook publishers is active and the publishers experienced, the publishers undertake the training of teachers in the use of the materials they have for sale. That is part of their marketing. Where the state is the provider and in-service training is inadequate (as it frequently is), this important stage in the introduction of new materials is too often ignored, as is constructive supervision of teachers by principals and school inspectors.

In Korea, reluctance to use new textbooks surfaced when a new curriculum in the national language was introduced that changed the emphasis in teaching from repetition and detailed reading of set passages to speaking, listening, and writing. The single textbook in Korean that had been used at each level was replaced by a set of three, and the number of pages doubled. The change was made after wide consultation, research, and pilot testing, but the teachers resisted it, complaining that they now had more content to cover without an increase in the time available. Over several years, they gradually changed their opinion and recognized the value of the new books. That change coincided with a change in teaching methods towards individualized instruction. The experience demonstrated that it is easier to change textbooks than teachers' preferences and habit (Lee 1997, 99-105). In contrast, teachers in Kazakhstan were given intensive training in the methods incorporated in a new generation of textbooks and are reported to have welcomed the new materials with enthusiasm (UNESCO/Danida case studies: Kazakhstan).

With decentralization in selection and competition in provision, educators at several levels must be trained to evaluate textbooks and other learning materials.

The parallel trends towards multiple choice in textbooks and decentralization place parallel responsibilities on the state.

The first is to ensure that all the books bought with public money, or recommended by the schools for purchase by families, meet minimum standards. There must be a process of evaluation and authorization that is not only objective and thorough but can be demonstrated to have these qualities. This in turn requires a governing board with some members from outside the Ministry of Education, a set of uniform standards, a corps of trained evaluators including teachers from different parts of the country, open invitations to submit manuscripts, procedures that guarantee anonymity in submissions and confidentiality for the names and reports of the evaluators, and full and rapid reporting to publishers.

While the machinery may be put in place, uniformity in evaluation has proved more elusive. The South African criteria for selection have been described by one publisher (Horwitz-Gray 1999, 107) as often seeming arbitrary and sometimes bizarre. In the Philippines (Chua 1999), publishers complained that approvals were not only inconsistent but that evaluators tended to reject textbooks that exceeded the minimum learning competencies. The director of the government body that supervises evaluation admitted that evaluators were sometimes appointed through friendship rather than merit, and that there had not been enough time to train them.

The second responsibility of the state is to ensure that teachers and principals who have been empowered to select books can make well informed choices. At a minimum, each school must have a full list of the textbooks and other learning materials that have been authorized for use. If possible, the list should be annotated as a guide to teachers, especially those in remote areas that may receive few visits from publishers' representatives. The notes might point, for example, to textbooks specially designed for rural schools or for the needs of minority groups. Teachers and principals (who may have the final say in any selection) must be provided with suggested criteria for choosing among textbooks, and the skills of selection should become a regular part of in-service training and the training of new teachers.

Experience suggests that this is the point at which devolution of selection breaks down, especially when the transition from monopoly to choice is abrupt. Teachers who are accustomed to a single textbook throughout their career, and who may be poorly qualified in many of the subjects they teach, are ill prepared for this new responsibility. Where educational structures have traditionally been hierarchical, teachers and principals tend to be influenced in their selection by their superiors. There is no evidence that any country undergoing liberalization has adequately met the challenge.

The provision of learning materials cannot be decentralized effectively without considerable advance preparation.

Under state provision, the preparation and publishing of textbooks cannot be abruptly devolved from the central administration to regional or provincial offices without taking into account local capacities. In Ethiopia, under a new policy of decentralization, regional educational bureaus were given the mandate for developing new primary-level textbooks. Few regions had the trained staff to produce camera-ready copy or the printing capacity to produce the books. As a result, all regions except the national capital returned the production of English-language textbooks to the Institute of Curriculum Development and Research, which in turn had a British adviser do most of the writing. In other subjects, in general, one textbook was written centrally and translated into the 10 other languages of instruction. This in effect neutralized the expected benefits of local preparation and production. The books were all printed in Addis Ababa (Ambatchew 1999).

Distribution will be a continuing problem in many countries, whether it is carried out by the state or the private sector.

The physical difficulties of distribution of textbooks to the schools will not disappear. It will always be arduous and expensive to reach remote areas. In urban areas where access is easy and the market is large and concentrated, distribution is estimated to account for 25% to 30% of the cost of a textbook; in remote areas, distribution may cost twice as much as the textbook itself (Askerud 1997, 72). In many countries where the state is responsible for distribution, public funds have proved inadequate to the task. The last part of the journey, from district or regional warehouses, has usually been left to the initiative of individuals - inspectors, principals, teachers, and parents - using whatever means of transportation are available. The discrepancies in the textbook:pupil ratio between urban and rural schools throughout much of the world testify to the failures of the system. Occasionally, the army has carried books to remote areas by helicopter.

Distribution by the state has been hampered as well by human failings that should be amenable to improvement. Books have been damaged because they were inadequately wrapped or stored in insecure locations. Systems for estimating needs, accounting for inventory, and recording shipments have been poorly designed or maintained. Consolidation of orders for economy has resulted in delays when one title of several was late in delivery. Books have been shipped from central warehouses months after the school year began because there was not enough staff or money to do better. Large-scale textbook projects have not always allowed adequately for the recurrent costs of distribution. Some of these problems can be reduced by decentralizing book procurement.

Distribution through the private sector faces many of these hurdles and is further hindered by the lack of retail outlets in rural areas. Supplementary materials will be stocked only selectively, because few shopowners have the capital to hold large inventories of slow-selling goods. If there is sufficient profit to be made, however, retail outlets will spread - not necessarily bookshops, more likely shops that sell stationery and other dry goods that would, for a few months of the year, stock textbooks promising a rapid turnover and relatively stable market. The ubiquity of certain soft drinks proves that profitable goods will find even the most remote markets.

Abuses will prove difficult to eliminate under any system of provision.

The editors of a recent World Bank publication on curbing corruption observe that "corruption is a pervasive phenomenon that can be found in countries of widely varying ideology, economic conditions and social development." They continue:

Although some societies may be more vulnerable than others and may suffer more devastating effects, no country in the world today is immune from corruption's corrosive influence. Yet for all its seeming prevalence, there is no clear evidence that corruption has become more widespread today. It has been around, in one form or another, from the earliest days of social organization. What has changed is that information about corrupt practices has become more available as governments have been increasingly unable to conceal evidence of wrongdoing; the level of public tolerance for corruption has declined; and the spread of democracy seems to afford less fertile ground in which corruption can flourish. (Stapenhurst and Kpundeh 1999, 1-2)

They go on to say that malfeasance, in general, is most likely to occur where the public and private sectors meet, as they do in public procurement, contracting, and licensing - activities that relate directly to textbooks and other materials used in government-supported schools. Those who award contracts or procure goods may demand bribes, kickbacks, percentages, or other "gifts" from those seeking government business and sales. At a more petty level, underpaid and greedy civil servants demand their own bribes, and in turn must often pay their superiors for the privilege of holding public office.

This should come as no surprise, but the potential for corruption in providing learning materials may astonish even a nation accustomed to abuses of power. After the private sector took over the provision of learning materials to the public schools of the Philippines in 1998, the newspapers carried sensational stories of bribery and misuse of funds, involving all levels of government from minor clerks to congressmen and, in one instance, the office of the president (Chua 1999). Increased funding for learning materials has simply increased the amount of bribery, according to the reports. Funds were diverted from textbooks to supplementary materials that carry a higher mark-up and a higher rate of bribery, despite declining textbook:pupil ratios. A new undersecretary of education has instituted various internal controls and stringent bidding procedures that he predicts will double the number of books for the country's public schools, but he admits that no system is entirely foolproof and some abuses will occur.

Abuse of another kind occurs when a cartel of suppliers forces the government to negotiate with it for want of any other source of goods. This happened in Bangladesh, where the textbooks for government schools were produced by more than 400 small printers. The government adopted an open bidding procedure to meet funders' requirements, but ultimately acceded to the cartel because no single member of it could produce all the country's needs alone. Printers who have large and largely under-used web presses do not bid for textbooks for fear of industrial action by the small printers that would close them down.

The president of Transparency International (Pope 1999, 109-15) suggests four types of strategy to reduce corruption. The first is to work within the underlying structure of government programs and focus on improving government integrity, enforcing the existing law, designing effective penalties for bribers and bribe-takers, and developing international support mechanisms. The second is to reform the procedural basis of government to increase transparency and discourage corruption. The third is to address government programs that may give rise to corruption because they are poorly designed or unnecessary - for example, by standardizing specifications, limiting discretion of civil servants, and establishing new public-interest criteria for bidding. The fourth is to encourage self-regulation within the private sector and changes in public attitudes concerning accountability and transparency.

The involvement of local stakeholders increases innovation and accountability.

The International Commission on Education for the 21st Century (1996, 158-9) argued that one of the main aims of educational reform should be to involve stakeholders in decision-making. Decentralization, it said, would encourage innovation and participation by all. It would also allow further consideration to be given to the cultural and linguistic aspirations of minority groups and it could enhance the relevance of education through carefully worked out appropriate programs.

It has been suggested that funds generally are used more efficiently the closer decision-making is to the actual user of the material. O'Connor (1999, 116-17) argues that corruption is inevitable when there is only one customer for textbooks, as in centralized procurement by the state: "The most effective influence on the purchase decision of a sole customer is not quality of product or value for money but the size of a more or less overt 'commission' to persons of influence ... All textbooks publishers, both indigenous and international, in countries where centralized textbook approval and purchasing systems operate are compelled to adopt these practices ... they cannot do otherwise; it is a matter of marketplace survival, not business ethics." She agrees that the most positive effects of donor-provided textbooks seem to occur when the process of selection is at the lowest possible rung of the education system.

Parents' associations can act as watchdogs to ensure that money allocated to schools is spent appropriately and in response to the community's needs. They and other civil organizations are well suited to monitor, detect, and reverse the activities of public officials in their midst because they are close to and familiar with local issues (Langseth et al. 1999, 143). In Bolivia, for example, community participation and decentralization are considered essential parts of the national educational reform, not only to encourage respect for cultural diversity and multilingual education, but to help curb corruption by providing a channel for social censure (Sanchez de Lozada 1999, 73).The school community has been involved in demand-side financing of textbook provision in Armenia and Guinea, helping to decide which families or students are neediest and should receive financial assistance. A similar strategy has been proposed in Cambodia (ADEA/UNESCO survey 1999).

Decentralization is far from a perfect answer, of course. Brickhill and Priestley recognize that, especially in rural areas, books and reading are not part of the social and cultural life, which may limit the effectiveness of local surveillance of book acquisitions. Pope warns that efforts to reduce corruption by decentralizing can backfire badly unless a strong and honest team is appointed to oversee and audit public procurement. Even the strongest advocates of stakeholder participation acknowledge that civil society is a player in corruption, although it is also the major victim. But Kisubi, a former senior adviser to the Ministry of Public Works in Uganda, argues (1999, 117-19) that where genuine anti-corruption attempts have failed one ingredient has invariably been missing, and that is the involvement of civil society. Any attempt to develop an anti-corruption strategy that neglects fully to involve civil society, he says, is neglecting one of the most potentially useful tools available.

Like textbooks, other learning materials require care and planning in selection, taking into account local needs and capacities.

Everything that has been learned from experience with textbooks applies also to the other kinds of learning materials. Indeed, even more vigilance may be necessary because of the wide range of supplementary materials and the much more limited resources available to buy them.

The benefits of careful planning in the purchase of books for school libraries has been demonstrated in Bolivia (López 1999, 136-8). The country has 32 languages, of which three are major. In 1995, it began building classroom and school libraries for 11,000 urban and rural schools and for 1,800 resource centres where teachers could be trained and exchange information. Each classroom library under the scheme contains 45 titles, one-third of them produced locally. Each school library contains 10 reference books, and each resource centre, 150 titles. In all, nearly 6.5 million books were acquired through competitive bidding. Procurement followed several stages. A year before bidding was announced, the Ministry of Education made contact with more than 100 publishing houses in the Spanish-speaking world. In response to the call for bids, 120 publishing houses from 16 countries submitted proposals. About 5,000 sample books were received and reviewed by 50 readers, all classroom teachers. After six months, 798 books were selected for further review, this time by schoolchildren, parents, and teachers at book fairs in three regions. The number of selected titles was winnowed to 730, at which point their 51 publishers were asked to submit prices. Ultimately, 370 titles from 48 publishers were bought in quantities of 1,800 to 22,000 copies. About one-tenth of the titles were translations from Spanish into one of the main indigenous languages. To ensure that a reasonable number of Bolivian authors would be included, the Ministry organized a national literary contest, paid for first-time rights to the 123 winning manuscripts, and edited them for publication. The value of Bolivia's multi-stage strategy can be seen by comparing its results to similar projects in Chile, where only 16 publishers participated, and in Peru, where 24 did.

School libraries are expensive to build and maintain. They require basic storage facilities, pre-service and in-service training of teacher-librarians, and efficient systems of distribution, management, inspection, and supervision. Book box libraries have proved more cost-effective in the projects described in Part III.

National agencies have been started in some countries to direct and assist the development of school libraries by providing training, books, and advice. These school library services began in Africa with independence but, despite initial enthusiasm for the concept, they have failed to maintain their early promise (Rosenberg 1998, 13-16). The activities of those that were established in Ghana, Tanzania, and Nigeria declined; in some other countries, a service was planned but never came into being. Finances have been blamed for the failure, but indifference was also a factor. In oil-rich Nigeria of the 1980s, for example, money went to tertiary-level libraries but not to school libraries. School library services that were located in the National Library rather than the Ministry of Education proved less effective because they did not have the strong support of educational administrators. Others failed because they did not gain the support of teachers, who did not recognize the value of libraries and who were focused on formal instruction based on the textbook.

Non-traditional media are still scarce in most countries, and seem likely to remain so. Computers and other instruments that are common in the industrialized countries are beyond the budgets and the infrastructure of much of the world. The use of new educational technologies normally depends on reliable, high-quality telecommunications connections, which are seldom found in developing countries. Even where they are available, the countries may lack the necessary pedagogical and technical support to implement programs successfully (World Bank 1995, 85). Many respondents to the ADEA/UNESCO survey dismissed non-print media as irrelevant while textbooks and other basic learning materials remain scarce.

It is essential to look beyond the classroom to the development of a literate community that will support continuing education.

The lessons taught in school will dissipate if there nothing to support them outside the classroom. Continuing literacy requires a literate community, which in turn will enable people to continue learning. The availability of books and other printed material leads to social and political awareness and to continuous self-learning in practical matters such health, nutrition, agricultural practices, and small business operations. In countries in which the book trade has been destroyed by war or economic collapse, in many countries of Africa, and in remote regions everywhere, the need is great. Henry Chakava, a distinguished Kenyan publisher, has said (1997, 59) that "There is talk in the North about a 'bookless' society, meaning a post-book society, while for us in Africa 'bookless' societies are indeed pre-book societies." It is difficult with donor funding to increase access to written materials in everyday life, but funding agencies and CSOs are making the effort.

The Swedish International Development Authority (Sida) has had a broad program for support of a literate environment, ranging from the supply of Braille presses to workshops for writers of literacy materials and the development and production of supplementary readers, science kits, and radio programs. It has supported the various links in the book chain, from the collection and publication of oral literature, through writing and publishing, to promotional events such as book weeks and book fairs. It has helped to fund libraries and reading rooms in Botswana, Ethiopia, Mozambique, Nicaragua, and Tanzania, language projects in Sierra Leone, Zimbabwe, Laos, and Panama, and authors and authors' organizations such as PEN International. It has bought and distributed children's books, supported libraries, reading campaigns, and training of librarians and teachers in Gaza and the West Bank. It has also supported adult literacy and post-literacy classes, often with accompanying assistance to village reading rooms, rural newspapers, and other materials for neo-literates (Sida 1996, 22).

The Children's Book Project in Tanzania directly addressed the absence of books for children in a country where even textbooks in rural schools were scarce. To encourage the development of local writing and publishing, in 1991 it launched the buy-back scheme mentioned in Part III, guaranteeing to purchase 3,000 copies of selected new children's titles if the publisher would put at least another 2,000 copies on the open market. The 3,000 copies were distributed to rural libraries and primary schools, and their purchase was just enough to ensure the publisher a return on the initial investment; the additional copies sold would be profitable. The project was conceived by CODE and supported by Sida, Danida, the Netherlands Embassy, the British Council, the Canada Fund, and the Aga Khan Foundation. The project included training courses for writers, publishers, editors, and illustrators. Ninety titles were produced by 27 publishers, and several were reprinted. The infusion of 270,000 books into rural areas created a demand for more. There were problems in the implementation, but a review team in 1995 recommended further support and the participating donors considered it one of the most successful publishing initiatives in Africa (Lema 1997, 91-4).

In India the Rajiv Gandhi Foundation is building networks of village libraries and stocking them with books and newspapers in the local language (Ranjan 1998, 1-5). The libraries are run by village committees that are responsible for providing suitable premises for the library, identifying a librarian, and general financial management. They are expected to become financially self-sustaining through low membership fees, municipal grants, and donations. The foundation supplies initial financing, furniture, training for the librarian, and a collection of about 400 books, chosen with community assistance. Efforts are made to integrate the collection with other local rural development projects. Books are lent among libraries to obtain diversity at minimum cost.

The libraries are intended to lend books for reading at home. There is not enough accommodation in most villages for reading rooms, and there is concern that women might hesitate to use a reading room that becomes a place where men congregate.

In Egypt, Mme Suzan Mubarak, the wife of the president, initiated a national project called "Reading for All" in 1990, to encourage reading among schoolchildren during their summer vacation. Children win awards for reading in various subjects, competitions are held among schools, simplified low-cost books are published in a "Family Library", and remote villages are served by mobile libraries (UNESCO/Danida case studies: Egypt). Efforts to develop a reading culture may be doomed to failure if there is no consensus on the desirability of such a culture, particularly among policy-makers and decision-makers, an expert in East Asia has pointed out (ADEA/UNESCO survey 1999). Askerud (1997. 22-3) argues that book development projects should look beyond instructional materials and consider ways to generate a general need for and appreciation of books and printed materials, to encourage people and governments to spend money on books as a priority matter and, not least, to prevent a relapse into functional illiteracy.

Coordination and cooperation in funder/donor initiatives should be continued.

Sida is unusual in having supported basic education programs in developing countries since the early 1970s, and textbook provision for most of that time. It is also unusual in the variety of programs it supports and their geographic span across three continents. Several of its projects are co-financed: in Bangladesh and Bolivia with the World Bank; in Bolivia and Cambodia with UNICEF; in Zambia with the Finnish Ministry of Foreign Affairs. In a review of its support to educational materials programs, Sida stressed the importance in multi-agency projects of a policy framework for educational materials and in particular for textbooks:

It is essential that there is a foundation for the reconciliation of different agencies' praxis, for example regarding local and international tendering, and for institution building for educational materials production and distribution in the countries concerned. Multi-agency projects can increase the effectiveness of the support by ensuring that all the essential components in the textbook chain, from curriculum development to utilisation in the classroom, are supported (Sida 1996, 19).

Other examples of cooperation, either bilateral or through organizations like the Association for the Development of Education in Africa, have been described in Part III.

CSOs complement the work of large international funders. In Macedonia a small World Bank loan of $5 million for education had an even smaller textbook component that would pay for the printing of a few new textbooks. The Soros Foundation undertook to carry out all the pre-press work, which included helping to set up an independent evaluation board, develop criteria for evaluation, and train authors, publishers, designers, evaluators, and teachers in the new ways of providing and using new textbooks. In Georgia the foundation helped lay the basis for another World Bank project with a textbook component. While the loan was at an early stage of negotiation, the foundation undertook to fund an open competition conducted by the Ministry of Education for new textbooks and the development of nine new textbooks. In addition it provided training for authors, publishers, booksellers, designers, and evaluators and carried out a small pilot project to test the premises of rental schemes and their suitability in the Georgian context. By the time the World Bank project was to be up and running, much of the preparation needed to implement it would have been done.

Governments can support publishing development through adoption of a national book policy.

In a major study of the economics of publishing educational materials in sub-Saharan Africa, an international team found that the major obstacle to sustainable textbook publishing was the absence of clear policies, strategies, and master plans.

It is not possible to make systematic efforts towards achieving desired results if there are no logical programs created by local participants and to which they themselves are committed. At present, where plans do exist, the strategies are designed and the plans often written by outsiders, usually by or on behalf of donors. Plans are often accepted by governments as conditions for financial assistance. At its worst, this state of affairs has caused a variety of competing projects, all concentrated on publishing, but from various angles and according to targets dictated by donors (Bgoya et al. 1997, 97).

It is probably no coincidence that in Africa publishing is developing most rapidly in Namibia, South Africa, Nigeria, and Ghana. The book industries of those countries operate in tandem with important government policies, and their national publishers' associations and book development councils work closely with their Ministries of Education and Culture (Chakava 1997, 57).

We have seen the large number of players and interests involved in the move beyond the simple provision of one or more commodities - textbooks and other learning materials - to the encouragement of a literate society supported by a vital book chain. To coordinate the many-faceted activities of the public and private sectors, establish government priorities, and secure the recognition of publishing as a strategic industry, many experts recommend that governments develop a national book policy. Such a policy is

a legal instrument adopted by the national government, and binding upon all parties concerned, that recognizes the strategic importance of the publishing industry and provides a comprehensive framework - with stated objectives and specific political, economic, fiscal and legal measures - to govern all activities in the book sector and to guide the actions of all players involved ... [It] should seek to create, over a reasonable time period, the conditions necessary for the emergence or establishment of a viable national publishing industry (Newton 1999b, 13-14).

In countries with a market economy, the objectives of a national book policy include: encouraging literary creation; establishing legal protection for authors' rights; providing fiscal, credit, and administrative incentives for the publishing industry (possibly including flexible credit through state banks and temporarily reduced rates of tax on revenue from domestic publications); facilitating nation-wide distribution and free international circulation of books; establishing nation-wide library networks; introducing new methods for the teaching of reading; and training personnel in the various skills of the book sector (Garzón 1997, 22-3). Relatively few countries have formulated and publicized so comprehensive a policy. One consultant with considerable international experience has observed that the creation and imposition of a ready-made book policy is not necessarily the ideal way to create positive change. A better strategy, he recommends, is simply for government to create an environment conducive to commercial publishing and to work with the private sector towards gradual reform in the provision and procurement of learning materials (ADEA/UNESCO survey 1999).

The first step in framing a national book policy should be a detailed and comprehensive survey of the national book sector. Since 1989, some 30 such studies have been made, but not all are equally detailed and the information in them is seldom easily accessible (Askerud 1997, 103). A useful summary of nine book sector studies in Africa (Buchan et al. 1991) was prepared as background documentation for the conference on textbook provision and library development held in Manchester in 1991 under the auspices of the Donors to African Education group and hosted by the British Council and the UK Overseas Development Administration. The fact that the study is already eight years out of date draws attention to the need for the regular revision of such information.

Some countries have delegated the implementation of their national book policy to a national book development council that typically comprises representatives of ministries, publishers, authors, illustrators, book distributors and booksellers, librarians, and educational institutions. These councils have been formed in at least 15 countries of Asia, six of Latin America, and five of Africa. Their effectiveness depends in part on whether they have executive or advisory status, and in part on the status and abilities of the leadership. Opinions about the usefulness of national book development councils is mixed. In the Philippines, the national book development council and the Department of Education, Culture and Sports quickly entered into an antagonistic relationship that impeded the transition from public to private publishing of textbooks. Newton (1999b, 17) has suggested that the book sector is generally overcrowded by institutions and that the creation of new bodies, along with their additional bureaucracy, may be undesirable. Chakava (1997, 54) has gone further, arguing that programs such as national book development councils worked best in centrally controlled governments, especially in Eastern Europe, Southeast Asia, and South America, and are unsuited to a world concerned with commercialization, privatization, and sustainability.

The need for support is unlikely to disappear completely as long as countries strive for the equitable distribution of textbooks and learning materials.

A sustainable book industry will never develop in a vacuum. It must not only have the government's support and encouragement but also a favourable economic climate. The countries of Southeast Asia that surprised the world with their economic growth in the 1980s and early 1990s had invested heavily in education and training. Minowa has argued that countries that have received external aid and succeeded in developing their publishing industry have done so not because of the aid but because of internal socio-economic factors that encouraged development and because of the efforts of their own publishers. Conversely, any country that lacks the necessary socio-economic strengths will not achieve sustainability in publishing in spite of external assistance.

The smaller, poorer countries of the world are far from achieving Minowa's "take-off" point - a resultant of several factors including economic development, technological capacity, marketing infrastructure, reading environment, population, and per capita national income (Minowa 1991, 140-1; Minowa 1992, 61). One of the two local commercial publishers in Guinea has listed the constraints his fledgling book industry must overcome:

there have been very few institutional measures to facilitate publishing, nor has there been much concern shown towards encouraging writers, providing banking concessions or tax relief for imports, price regulation or a national book policy ... The book market is limited because of factors such as the high level of illiteracy (67%), the reading public's weak purchasing power, and the lack of both a reading culture and places to read. Publishers are faced with problems such as access to capital, the lack of adequately trained technical staff, high manufacturing costs, and the limited capacity of local printers. There are only fifteen or so printers in the country, modestly equipped for the most part, and among whom only three print books (Sow 1998, 11).

In eight years, the two local publishers have managed to produce fewer than 50 titles in French and two in national languages. They are unable to meet the conditions of international tender for textbooks, which continue to be won by French firms.

For the foreseeable future, countries like Guinea will require external production of textbooks and most likely some form of assistance, from external sources and/or from national governments, in the provision and distribution of learning materials. The amounts necessary may be reduced as programs of liberalization and cost recovery mature.

Ultimately, the adequate provision of textbooks and learning materials is a matter of political priorities and political will.

An equitable and adequate distribution of textbooks and other learning materials is not an impossible goal. Colclough and Lewin (1993, 9) estimated that Universal Primary Education (UPE) could be achieved with an additional investment in external aid of around US$30 billion between 1990 and 2005. Most of that money would be needed for recurrent expenses, and about half would be needed for Africa. The increase would amount to a 50% rise in aid to education, but would no more than compensate for the fall in aid since the 1970s. They recommended further that governments devote the equivalent of US$5 per pupil for learning resources - an amount they thought was realistic for a majority of developing countries and enough to have a considerable impact. More recently, Watkins (1999) estimated that UPE could be achieved within a decade in all developing regions at a cost of $7 to $8 billion annually over existing expenditures. That increment, he pointed out, is the equivalent of about four days' worth of global military spending, seven days' worth of currency speculation in international markets, less than half what parents in the United States spend on toys for their children each year, and less than half what Europeans spend each year on computer games or mineral water.

The importance of political priorities was underscored by Richard Jolly, special adviser to the administrator of the United Nations Development Programme, at the conclusion of the Mid-Decade Meeting on Education for All in 1996:

I recall my excitement in Costa Rica a few years ago when I visited a primary school with President Arias, to see in practice the pledge he had made to provide computers to half of all primary schools. He had made two conditions: that each school would have a suitable classroom, so the computers could be protected against theft; and that the class room should be open in the evening as well as during school time, so adults could use the computers. I marvelled at this practical example of bringing the frontiers of education to primary students and adults throughout Costa Rica within four years. But how can you afford it I asked? His reply - if you don't spend money on a military, you have enough in hand to afford good education for all. This is the challenge for every country - and a practical example of how saving money on the military bills can produce much of the finance required (Jolly 1996).

The International Commission on Education for the 21st Century (1996, 165) recommended that as a rule of thumb not less than 6% of GNP should be devoted to education. It felt that all need not come from government - that it is not only justifiable but desirable to raise money from private sources in order to ease the pressure on national budgets. Private funding could come from a variety of sources, including fees, although the ability of families and students to pay would vary from country to country. But, the Commission emphasized, recourse to private funding could not release states from their financial responsibilities, especially towards the most severely disadvantaged ethnic minorities or the inhabitants of remote regions. The Organization of African Unity Conference of African Ministers of Education (COMEDAF) received a report from its Finance Advisory Committee this year recommending that member states allocate a minimum of 25% of their budgets and 6% of GDP to education, that a minimum of 50% of education budgets should be allocated to primary education, and that the proportion of education budgets allocated to learning materials should be increased to 14%.