This study examines the impact of education expenditure decentralization on education output in Indonesia.Education output is measured by using the channels of repetition rate,dropout rate,completion rate,net enrollment rates,and test scores at primary,pre-secondary and secondary education levels.The test scores are taken from the subjects being tested in the national examination(UN)namely natural science,mathematics,Indonesia language.Test score of the primary level is excluded from the empirical model because there is no national student evaluation at this level.The evaluation at primary level is conducted and coordinated by district school office level,while in national standard evaluation the material testing is made and coordinated by MOEC.Education production function model is used to investigate the direct and indirect effect of education expenditure decentralization on education output.In fact,Education output is affected by several factors,such as family input,school resources,and the environment outside family and school.Family characteristics that may affect student achievements include parents’income and education,family size,and peer of students.School resources that influence education output include the public expenditure in per pupil,teacher-pupil ratio,teacher education and experience,and school facility.Environment factors also have a significant impact on the student achievement,such as community participation or involvement,direct spending of local government,parents’ involvement in the education process,education council.Previous empirical results show that decentralization effect on education output is varied.This diverse result encourages to more exploration on the relationship between decentralization and education output in other countries.Formal education in Indonesia has been introduced since the era of Dutch colonialism for more than three hundred and fifty years.Education access was limited to certain people especially for the high class of social status.There was also discrimination towards woman in accessing education.During the Japan occupation,the education is directing to support Japan war in Asia Pacific.Instead of learning the knowledge and science,the students were learning physical and muscle exercises or military education.Nowadays,education in Indonesia is the 3rd large in Asia after China and India.Compared to the beginning of independence era,the net enrolment rate at primary level has been closed to 100%.This progress is satisfying and indicates that government’s effort in providing primary education is successful.However,in the pre-secondary and secondary school level,the net enrolment is far from the ideal rate.Nine years of compulsory education program have been launched in1994,which is an effort to give the people more chance to have an education until junior high school and increase enrolment of pre secondary school.Fiscal decentralization in Indonesia is emphasized more on expenditure than on revenue.The local governments are given the authority to determine their expenditure.Since the ability of each region in financing their public provision is unequal,the central government transfer fund is needed to cover the gap between the regions.The funds are divided into two categories:balance fund and special autonomy and adjustment fund.Balance fund is comprised of DAU-General Allocation Fund,DAK-Specific allocation Fund,and DBH-the shared revenue fund.The implementation of education decentralization has been running since fiscal decentralization was enacted,however,the education system and policy are still under the central government.National student evaluation(pre secondary and secondary school level),curriculum,teacher development program are still under MOEC supervision.Community participation in school learning process has been accommodated by the government through establishment education council and school committee.However,the role of two organizations has not yet been running effectively in supporting school output.The data employed in this study spans from 2007-2013,unbalance panel data of 33 provinces in Indonesia,and various regression models are used to estimate the effect.The regression results support the hypothesis that education expenditure decentralization has a positive and statistically significant impact on education output in term of dropout rate at primary and secondary education level,repetition rate at secondary level,enrolment rate at secondary level,completion rate at primary level.With regard to student achievement,the result is inconsonant with previous studies that education decentralization can improve student achievement.Student achievements are insignificant at all level for the lesson of language,mathematics,and natural science.With regard to family input,school resources,and othersvariable factor,the result is diverse.Teacher qualification is significant to repetition rate at pre secondary school,but it is insignificant at primary and secondary school level.Pupil teacher ratio is negative and significant to the completion rate at pre secondary and secondary level,but insignificant to the completion rate at primary education level.