DECENTRALISATION AND EDUCATION FOR ALL IN INDONESIA

The discourse on decentralisation in developing countries is seen as a strategy in restructuring and improving economic, social and public welfare including the education sector. This idea is fundamentally in line with the goals of Education for All (EFA). The experience of Indonesia that had shifted from a strongly centralised system to a decentralised one in the early 2000s provides an interesting case. This article argues that the current decentralised system in education has seen an insignificant effect in achieving EFA. Through an extensive literature study, this article draws attention to the particular concerns of human resources, curriculum, corruption and poverty issues as contributing factors to the seemingly failing efforts in the decentralised settings, all in the light of Indonesia's historical development.


Introduction
Many countries have experienced the benefits of decentralisation on their systems of governance, including in the education sector, thereby it has been regarded as a strategy to promote Education For All (EFA). Globalisation has likewise endorsed the discourse of decentralisation by amplifying its positive effects on economy, democracy, and welfare. The experience of Indonesia in this regard provides an interesting case. Although the country has actually embarked on decentralisation since the end of the twentieth century, the radical reform has not had a significant effect on EFA. The central argument of this article is that the decentralisation has not been done and supported appropriately. In order to understand and limit the broadness of the concept, definitions and classification are necessary. In general, the discussion on decentralisation centres on the process and outcomes of transferring authority or power, responsibility or roles, from a top level of an organization to lower levels (Brown, 1990). Rondinelli, Nellis and Cheema (1983, p. 13) defined decentralisation as "the transfer of responsibility for planning, management and resource raising and allocation from the central government and its agencies to: (a) field units of central government ministries or agencies, (b) subordinate units or levels of government, (c) semiautonomous public authorities or corporations, (d) areawide, regional, or functional authorities, or (e) nongovernmental private or voluntary organizations." Regarding the degree to which responsibilities and decision-making discretion is transferred, they classified four types of decentralisation as follow. Deconcentration is understood as the handing over of certain administrative or management responsibilities to lower levels within the central government with control remains in the central. Delegation means more amount of authority being transferred to lower levels outside the regular bureaucratic. Devolution refers to creating or strengthening subnational units with authority over financial and administration. Privatization denotes the transfer of responsibility to private sector. Those forms of decentralisation can exist in one system and differ to certain extent. The next section will look at how decentralisation comes into discussion with the EFA.

Decentralisation and Education For All (EFA)
Originated in North America and Europe in the 1980s, the notion of decentralisation has since been a promising idea of transformation for many countries (Zajda, 2005;Daun, 2007). This includes Indonesia in the late 1990s, when the nation embarked on decentralisation from a highly centralised system. In education affairs, the general conceptualisation of decentralisation is the change of the role of the states in the management through a process of transferring authority and financial responsibilities (Zajda, 2005;Zajda, 2006;Bangay, 2005).
Despite the scarcity of recent literature explicitly juxtaposing the two themes of decentralisation and EFA movement in the same piece of academic text, it has been widely accepted that decentralisation is seen as a strategy of the state in restructuring mass education to meet the needs for equity, participation and diversity (Zajda, 2006), thereby in line with the goals of EFA. As the definition in the previous section implies, decentralisation has various forms in terms of how much authority is transferred. Whilst in some systems decentralisation only occurs as transferring workload and responsibility (but not power), in some others it grants major devolution of power (Zajda, 2005). The issue regarding provision of education, accordingly, revolves around who is in charge of the educational planning in terms of administration, financing, and curriculum. In this case, the level of control from the state seems to be the indicator of how decentralised a system is. Adding to this, there is no total decentralisation in politics and administration (Zajda, 2006;Hanson, 2006) and "there can only be certain degrees of centralisation or decentralisation and not one or the other" (Zajda & Gamage, 2009, p. xvii). This suggests that it is possible for decentralisation and centralisation to coexist in one system, seeking for the right balance (Caldwell, 2009), when only certain amount of responsibility is transferred.
Apart from the degree to which a nation is decentralised (or centralised), the interests towards decentralisation in education have been nurtured by its rationales. Arenas (2005) demonstrates three underlying rationales of decentralisation which are essentially rooted in politics; neoliberalism, populism, and participatory democracy. Neoliberalism encourages privatisation because it basically supports individual rights against arbitrary state government, promotes instead strong local government role as well as market forces mechanisms. In the same way, populism rejects the large state control that is unresponsive to the needs at local level, and further emphasises on the role of community or local people to accommodate their own distinct culture and knowledge. Participatory democracy advocates the significant roles of local stakeholders (parents, teachers, administrators, community members, and students) in decision-making activities. Taken together, these three ideologies defend decentralisation in term of lessening the state's control. In addition, Hanson (2006) contends that when a nation's governance transforms from autocratic to democratic, efforts to decentralise the education system come naturally.
Supporting these rationales, the reasons for decentralisation are mainly emphasised on improving quality, access and efficiency in schools, as well as increasing democratic participation at the local level (Zajda, 2005;Daun & Siminou, 2009). Often the quality improvement in education is used as the sole reason for decentralisation, and the notion implies that local people or their representatives are better than the central government in solving certain educational problem and addressing specific educational needs (Hanson, 2006;Herbst & Wojciuk, 2016). Given that the degree of decentralisation is considerably various, and that different countries have experienced different challenges in the process (Zajda & Gamage, 2009;Daun, 2007), the discussions over the outcomes can be difficult since the factors that may have affected the process are complex (Hanson, 2006). Nevertheless, even though the attempts on decentralisation have not been perfectly conducted or have brought about unexpected outcomes, the strategy must have some positive attributes that attract worldwide attention and potentially hold great promise. Moreover, case studies of some countries have served decentralisation the attention it deserves, despite the 'level of success' and the barriers that they encountered. This can also provide explanation on the reasons why it is relevant to other countries.
With regards to education quality improvement, particularly in the area of curriculum, the experience of Spain is one good example (Hanson, 2006). Hanson explains that, following decentralisation, Spain has pursued the right balance of curriculum content, where a standard or core material nationwide is combined with certain amount of regional material to accommodate different interests on the ground. He further adds that factors contributing to this outcome include the strong administrative infrastructures to manage the system at regional level. This extent of decentralisation, he argues, has brought about positive outcomes following the efforts in providing good quality education in Spain. Drawing from his past field studies on decentralisation, Hanson also exemplifies Venezuela, where the decentralisation policy led to positive outcomes, which included tackling issues of bureaucratic stagnation and corruption. Columbia had likewise experienced decentralisation that increased democratisation through devolution of power in the 1990s (ibid.). In the 1990s, in order to improve quality of education in Hong Kong, the government launched programmes that allowed transfer of responsibility to schools, making them more flexible to make decisions (Bray & Kwok-Chun, 2006). During the period, the government not only strived to improve the quality in public schools, but also empowered private sectors to build and fund schools by subsidising the operational costs (ibid.). This resulted in more schools, particularly secondary schools, to accommodate the continuation of basic education provision (ibid.). From the same perspective on seeing the positive impacts that decentralisation has brought, UNESCO and UNICEF (2013) list a number of countries, particularly in the Asia-Pacific regions, that have experienced certain improvement in education, including the Philippines. School-based management approach has been implemented by the Department of Education and some studies have recognised the success of the reform in general, although community empowerment is necessary to achieve more success (ibid.). These are some examples of how decentralisation plays a significant role as a catalyst in bridging state's policies and improving the quality and the quantity of education.
However, since decentralisation is highly political (Bjork, 2006), it is associated with problems that many countries have encountered, corruption. Joaquin (2004) argues that a decentralised system is the favourable to government's anti-corruption strategy since it is easier to detect in a smaller scale. She further contends that by centralising decision-making in a corrupt government, it will likely loosen the control over the money-flow, and lead to more corruption. On the other side of the coin, however, bringing decision and control to the lower level does not necessarily counter the problem. In fact, the possibilities of corruption are still high (ibid.).
Regarding poverty issues, according to UNESCO (2014), the good quality of education is the key to increase economic growth, which in turn reduces poverty. In the report, they argue that if all students in low-income countries gained basic reading skills when they left the school, 171 million people could escape from poverty. Many countries have actually succeeded in reducing poverty through education, for example Ethiopia had reduced poverty by half since 1995 but a third of the population remained living in poverty (ibid.).
In the context of developing countries, international agencies persistently promote decentralisation. Since funding is still a great obstacle in achieving good quality education, therefore multilateral agencies, donors and private sectors play significant roles (Rose et al., 2013). In Asia, some multilateral agencies, namely the World Bank, the European Commission, the UNICEF, and the Asian Development Bank have specifics funding instruments through Official Development Assistance (ODA) that contributes to educational development in many countries (ibid.). These agencies support the idea that decentralisation increases student enrolment as well as quality of schooling, thereby boosts economic development in the countries (ibid.). However, there is no fix formula of decentralisation that can be applied to all countries, considering the differences of historical and political development, different form of decentralisation may require different strategy to implement. This is where the significance of policy borrowing nature comes into discussion. Despite the expectedly similar future outcome of education development, policy makers should consider contextual analysis in policymaking to understand 'what works and what does not work' (Steiner-Khamsi, 2014). International agencies such as the World Bank and the Organization for Economic Cooperation and Development (OECD) have put significant contribution to education with resourceful policy recommendation (Thompson & Cook, 2014) making it no longer a matter of domestic interest only, but also international interest (Nagel, Martens & Windzio, 2010). Indeed, the successful implementation of a particular policy or reform in one country does not necessarily justify transferability to other countries.
Through an extensive study on developing countries' experiences on decentralisation, Hanson (2006, p. 24) presents a valuable overview on some key issues in order to understand how educational decentralisation works. The following are some of his propositions on effective decentralisation: • The greater the accepted vision of decentralization between the distinct centers of power (e.g., political parties, unions, bureaucrats, religious institution), the greater the chance for successful change. • Devolution rather than delegation of authority and responsibility has a greater chance for long-term success. • It is easier to initiate a decentralization initiative during times of political, economic, and social stress or turbulence, than it is during times of relative stability. • The stronger the management infrastructure at the regional levels, the greater the opportunity for success.
• It is better to transfer authority to individual regions only when they meet specific tests of readiness, rather than to all the regions at once regardless of readiness. • Decentralizing in incremental stages has a greater chance for success • Understanding the motivation behind a decentralization initiative is the key to understanding the specifics of the strategy. • The people who have been part of an organizational culture that has managed a centralized system are not very effective in managing a decentralized system (old habits and a taste for power are difficult to cast off.) • A decentralized organization should function as arts of whole rather than simply independent parts. • Once decentralization has taken place, the central ministry still must have the tools to safeguard that the regions follow national educational policy. • Educational policy on decentralization should be set through debate rather than disguised manipulations of the national budget.
Consistent to the purpose of this article, these key arguments provide important insights in assessing the case of Indonesia. Following the methodology in the next section, I will discuss how Indonesia has experienced decentralisation and the factors that have affected the EFA in the light of these propositions.

Method
The objective of this article is to identify and explain the issues and forces that play major roles in the seemingly failing reform in education in Indonesia. To meet this objective, this study makes use of available literature in the area of educational reform, particularly in the context of decentralisation. It reviews related studies, reports and researches linking the topics of education and decentralisation. The framework that it uses benefits from previous studies particularly in the context of developing countries.

Discussion
The commitment of Indonesia on achieving education for all goals has been challenged during the past fifteen years. Decentralisation, which began in the late 1990s, has hindered the EFA from achieving its targets. As I argue in the following sections, the factors that account for the failure include particular concerns on internal factors, corruption, poverty, and curriculum. The first subsection about the historical development helps to shed light on how decentralisation occurred as a radical initiative that certainly affects the education affairs. It this then followed by the thematic discussion on why decentralisation in Indonesia has had only little influence on EFA with regards to the factors.

Historical development
Decentralisation in Indonesia has to be understood in light of its colonial history and development in politics. With its historical development of democracy, the experience of decentralisation in Indonesia is distinctive (Amirrachman, Syafi'i & Welch, 2008). For more than three centuries, the Dutch colonialism was characterised as highly centralised. After the independence, "Guided Democracy" was introduced by Soekarno, the first president of Indonesia , however centralisation remained intact. Following that, during the thirty-two years regime of New Order led by Soeharto, "Pancasila Democracy" (Pancasila is Sanskrit for 'five principles', the official philosophical foundation of the Republic of Indonesia) was imposed behind the very strong centralisation system (ibid.). It was not until the 1997 economic crisis in Indonesia, that the country embarked on decentralisation, known as the 'big bang' reform during the short governance of Habibie, the third president of Indonesia (Hofman & Kaiser, 2004). Bangay (2005) sees the decentralisation in Indonesia as an endeavour to mitigate discontent in the aftermath of three decades of centralism and authoritarian governance. The underlying idea of it was about bringing power and responsibilities closer to people through local autonomy (Kristiansen & Pratikno, 2005). The strong grassroots movement insisting for decentralisation was backed up with greater pressure from international donor agencies, which had likewise pushed the government. The World Bank and the Asian Development Bank jointly supplied loans and grants, whilst the IMF offered 'post-crisis rescue package' (ibid.).

Internal factors: human resources 'legacy'
The historical development and political turbulence mentioned above explains why Indonesia was in many ways unprepared for a radical decentralisation. During the time when the decentralisation was hastily implemented, the nation was not in an ideal condition. In short, the people were not ready. Apparently there was "little experience of local democracy, unclear legislative guidelines, lack of local capacity building, on-going factional politics, resistance on the part of key officials…" (Amirrachman, Syafi'i & Welch, 2008, p. 48). The fear that Indonesia would not achieve optimal decentralisation was reasonable. This might be due to the fact that it lacked human resources at local level. The central government had to reallocate a big number of civil servants from the central level to district level, which took a lot of effort and time on manpower planning. It was something crucial at the early stage of decentralisation process but failed to meet the demand.
The long experience of strongly centralised governance was arguably the sole reason why the internal human resources never satisfied the need of the newly decentralised nation. The bureaucracy was full of people whose ideology was a by-product of centralised governance who had gained satisfaction out of top-down control. In fact, those who dominated the regime of Soeharto had remained powerful in post-regime era (Rosser & Joshi, 2013). In the education sector, the impact of those legacies apparently had embedded a distinct ethos in school leaders and teachers. For instance, many teachers and leaders were reluctant to change the norms of being loyal doers of central orders (Bjork, 2004). Their nature of being 'transmitters' of knowledge and values that the government intended to impart to the citizen had presumably lasted for generations. Thus, when decentralisation offered the opportunity to perform more autonomy on leadership and management, many failed to meet the challenge due to both the lack of experience and the unwillingness to change. Consequently, they became more dependent on the central government to acquire technical assistance that placed more responsibilities on the latter.
The criticisms, however, were reciprocal and the solutions were interdependent. The distribution of authority was vague due to the centralised control that remained strong (Rosser & Joshi, 2011). The local actors were still dependent on the direction or instruction from the central government, as they feared that they might fail to do so on their own and face the consequences. Indeed, it was a difficult transition for the educators who had insufficient leadership background; therefore the tendency was to continue following the 'safe and securing' practices instead (Bjork, 2004). In this case, strong control from the central government seemed to have hindered the eagerness of local authorities to display and nurture leadership. Taking this from a more general perspective, Indonesian government had ignored one significant dimension of decentralisation, which was the human resource management (Turner, Imbaruddin & Sutiyono, 2009).
Since the poor leadership quality corresponds to the incapability in decision-making, local government seemed to face difficulties in planning and implementing local policies. Consequently, the devolution of power appeared to result in no significant changes on the improvement of education quality at the local level. Although the system had been decentralised, the human resources that supposedly support the system was not in place. This leads to another reason why decentralisation does not necessarily improve education in Indonesia, which is corruption.

Trust issue: corruption
Another factor that hindered Indonesia from a successful transformation to decentralisation is one that related to trust issues. There is a problematic trust issue among all the stakeholders in regional education sector. These include politicians, government bureaucrats, society, school principals, teachers, and students' parents. In their study, Kristiansen and Pratikno (2005) reveal that the issue typically centres on financial management. They argue that the government often failed to convince society with accountability and transparency in educational services and financing. In fact, society has come to a point of believing that the education sector is susceptible to corruption (ibid.).
In theory, decentralisation can lead to increased transparency and accountability thereby reducing the possibilities of corruption (Joaquin, 2005), whereas in practice Indonesia has seen a quite different result. Apparently, decentralisation not only allows the transferring of responsibilities but also the transferring of possibilities of corruptions in Indonesia. The autonomy, that has actually enabled the local government to take control of the expenditure on education and to make a list of priorities to accommodate local needs, is often seen as an opportunity to take advantage for their personal interests. The more control the government has over the budget allocation at local level, the greater chances they have to misuse their authority. Therefore, the power and responsibility to manage and plan public expenditure, including for education purposes, has been abused.
Being labelled as one of the most corrupt nations in the world (Nurman et al., 2011), Indonesia has been dealing with numerous cases of corruption in social affairs including education. As the national government consistently promotes EFA with increasing budget allocation for education, the education system has become highly susceptible for corruption. According to Indonesian Corruption Watch (ICW, 2013), during the period of year 2003 to 2013, there were 296 cases of corruption in education, with the indication of total state losses of 619 billion rupiahs. This figure had included the corruption cases of school facilities and infrastructure, and also teachers' salaries. In fact, many public figures in the government who were once seen as men of integrity had been eventually proven to be involved in corruption (Prabowo, 2014).
Although an intense anti-corruption movement had begun in 2000 (Nurman et al, 2011), during the same year when decentralisation reform took place, there had been numerous cases of corruption in education revealed in the following years (ICW, 2013). In 2014, almost at the end of the EFA, corruption remained a major problem in Indonesia (Prabowo, 2014). ICW also argues that corruption within the Ministry of National Education was indeed the major factor that failed the efforts to reach education goals including reducing the drop-out rates and improving facilities and standards (Rahman, 2009). For instance, it is still widely known that corruption is evident on the central government's BOS that was intended to provide funds for schools' expenditure on materials and to help reduce poverty (Suryadarma, 2012). When corruption first took place at national level, followed by more corruption at local level, the amount of money received at school level was most likely smaller that the original budget. Consequently, particular schools had suffered from inadequate operational costs that directly influenced their overall performance (ibid.).
Other possible forms of corruption at the lowest level, i.e. at schools, had also impacted on low school enrolment rates. Rosser and Joshi (2013) found that many schools in urban and rural area had been charging the students for 'illegal' fees. These included additional fees on textbooks, uniforms, exams, excursions, and building construction and maintenance levies. Rosser and Joshi noted that 'politico-bureaucrat' (including principals and teachers) have been taking advantage of their power and positions. When the fees were set as compulsory, parents would have to pay for it fearing that their children might be disadvantaged (e.g. affecting their grade, not passing the level). Furthermore, poor parents who could not afford the fees were likely to withdraw their children from school. As a result, the provision of free basic education as intended by the government was not totally free.

Poverty
Decentralisation, as it is theorised, can ideally bring about impacts on reducing poverty and increasing equality (Hanson, 2006;UNESCO, 2013). After being decentralised for one and a half decade, Indonesia appears to experience something contrary to that. The fact that disparities in education have become evident (Amirrachman, Syafi'i & Welch, 2008) informs that decentralisation has actually not had a positive impact as intended. With regards to poverty, the number of poor people in Indonesia has been increasing every year. In 2015, the number reached 28.59 million people, which had increased from previous years (Badan Pusat Statistik, 2015). The disparities became more obvious when rural and urban areas were examined.
Indeed, poverty is a multi-dimensional problem with regards to range of aspects that it encompasses. In accordance with the worldwide perspective on the issue, Indonesia has likewise seen the importance of providing mass education in order to decrease the poverty rate, despite its national diversity and huge geographical area that it covers. In spite of the persistence of the government in fighting against poverty through basic education provision, the number of schools available in remote areas is still small. In a 2014 survey on the availability of schools in the villages, the data shows that the latest figures of total primary schools (1 st to 6 th years) and junior high schools (7 th to 9 th years) are 86.63% and 42.54% respectively (Badan Pusat Statistik, 2015). This means that more than half of the total number of villages does not have junior secondary school buildings.
Another contributing factor to the urban-rural disparities is arguably due to the nature of decentralisation itself. Decentralisation makes development partly a responsibility of the local government, thereby creating more segmented rural and urban areas unintentionally. Thus, regional autonomy may instead benefit from different paces of development in different areas due to either the lack or abundant resources available. In urban areas, relatively stronger leadership of school principals, as well as richer resources and greater support of parental and community had become a favourable condition for a good school management that boost students' performance (Amirrachman, Syafi'i & Welch, 2008), making the gap even larger compared to rural areas that lack of such resources.
At the lower level, the transfer of 'responsibilities' in decentralisation has also had an impact on parents. The basic goal to minimize government's expenditure through decentralisation could directly impact the fees charged in schools. When schools are lacking of money to support its operational costs, parents are likely to be partially responsible to ensure that their children receive a good quality standard of education. In a study of Kristiansen and Pratikno (2005), assessing the impacts of decentralisation reform, they conclude that the household expenditures on education were higher and increasing compared to the period before the reform. Arguably, the effects are most likely to be found in poor areas, where people cannot pay for any additional costs for school, resulting in withdrawing their children from schools. The additional costs may include meals and transportation fees that can be barriers for the poor in accessing basic education (Rosser & Joshi, 2013).

Curriculum
The Indonesian government holds a huge responsibility for a wide range of diversity in language, culture, and resources. The nation has a lot to manage with around 17.000 islands populated by a quarter billion people. The disparities in socio-economics and education quality are evident. The efforts to provide education and to reach as many people as possible have been positive and continuing. Increased budget for education provision throughout the archipelago through national and local government programme has been focused on building more education infrastructure. Nevertheless, having addressed the issue on quantity does not necessarily encounter the issue on quality. Such concern on the disparity is apparent not only in academia and governmental settings. The juxtaposition of education issues in access, quality and diversity against the backdrop of a struggling life in Papua is depicted through an award-winning Indonesian film "Denias", which has served as a reminder of how the socio-economic disparity affects education (Gumono, 2017).
The quality of education is sometimes overlooked when the focus of development is on providing mass education. In other words, the priority to increase the number of schools is important but the efforts are supposedly balanced with quality improvement. However, considering the complexities of interdependent factors, this is not an easy task for the government. These factors include the concerns regarding the inadequate number of teachers teaching at rural and remote areas, the language of instruction used at schools (or outside formal schools), and the literacy rates. During the last decade, the local government and the Indonesian teacher union (Persatuan Guru Republik Indonesia) through mass media have been arguing that the country needs more teachers (www.republika.co.id, 2015). The target on overall education quality improvement is far from feasible if the country is still dealing with insufficient supply of teachers. In result, the illiteracy rate is still a big concern particularly in rural areas (Badan Pusat Statistik, 2015).
Regarding the content of the national curriculum in the Indonesian education system, decentralisation reform has only had an insignificant impact. The standardised national curriculum represents the fact that the government is not promoting the so-called 'international best practices' as exemplified by international agencies. The idea is to make the curriculum as decentralised as possible to meet the local needs. However, the Ministry of National Education is unlikely to implement a total transformation on curriculum content since the current regulation still strongly emphasises on national standard and evaluation. In other words, the curriculum has not been decentralised to the extent that it is fully managed by regional authorities, let alone by schools. It is argued that the lack of appropriate teacher training, the passive attitudes toward the policy, and the scarcity of resources and funding are the factors that account for the complication and hardship in its implementation (Yeom, Acedo & Utomo, 2002). As Amirrachman, Syafi'i and Welch (2008) contend, the recommendation of the World Bank and the Asian Development Bank, both of which promote decentralised curriculum as a globally effective strategy, has not been promoted in Indonesian education system. Instead, the education system maintains the centralised curriculum, with regards to the compulsory subjects that are taught at schools and the national examinations, which can be problematic. On the one hand, it is somewhat important to provide a standardised national curriculum to ensure that no regions are left behind, in terms of students' (and teachers') performances, due to lower standards that the local government may have set (with the authority that they have). On the other hand, the standardised national curriculum may actually appear to be forcefully and ineffectively implemented at the local level without enough resources and appropriate assistance in the process.
Unlike the national curriculum issue, the unitary language of the instruction issue seems to be less problematic in its implementation despite the fact that hundreds of tribes in Indonesia use different languages or the local mother tongue. The education system in Indonesia makes Bahasa Indonesia as one compulsory language of instruction that is used in all public schools. The reason for this seems to underlie in the commitment of Indonesian government being consistent with the Youth Pledge (the literal translation of Sumpah Pemuda, a national declaration as a result of youth movement in 1928), consistently teaching and respecting the language of unity. Therefore, this is not only a matter of what works better, but it is about the commitment, which is generally accepted nationwide. However, there are cases of people who commonly use the mother tongue to educate certain primitive tribes that have intentionally isolated themselves and are not open to the outside 'modern' world including formal education (Firman & Tola, 2008). Such noble effort is quite challenging considering the difficult circumstances in the remote areas. This is perhaps another factor that deserves more attention when talking about provision of education for all, especially in Indonesia.

Conclusion and Recommendation
The debate on decentralisation versus centralisation has taken place since more than four decades ago, and the concept has been broadly discussed among policy makers. Although very often the notion is introduced as rooted in politics and economy, decentralisation has impacted social affairs including education sector. In Indonesia, the reform to decentralise began in the late 1990s. At the moment of transition in the end of the twentieth century, Indonesia was one of the most strongly centralised nations in the world (Bjork, 2004). The political and ideological legacies from the past were influential and deeply rooted in the people.
Apart from not being supported with abundant resources at the time it was implemented, decentralisation has not brought about significant impact on education, particularly in promoting Education For All. The past experiences and the 'legacies' of three decades of authoritarian government have negatively impacted the effort to implement the reform. These include the lack of human resources in the government as well as in the school level, the corruption and the trust issue in the society and the government, the huge geographical area of Indonesia, as well as the curriculum that could not cover all the need of all local people. Consequently, those interdependent factors had hindered the EFA from being successfully implemented.
Further research on the impacts the decentralisation reform has had on key sectors in education such as financial and manpower planning as well as teachers education programme is necessary. The comparison between the development of national policies on those particular sectors to the real practices in regional and school levels will indicate the extent to which Indonesian education system has progressed in the past decade.