91206 1 PILOTING PER-STUDENT FORMULA FINANCING IN MOLDOVA First Outcomes and Policy Recommendations Technical Note for the Government of Moldova This short technical note has been prepared at the request of the Ministry of Education of Moldova. It was developed under the leadership of the Scherezad Latif (TTL, Senior Education Specialist, ECSH2) by Anna Olefir (HD Operations Officer, ECSH2). Simulations of the impact of the school funding formula on school budgets and rayon education transfers in Moldova have been conducted by Rosalind Levacic (Consultant, ECSHD). The note also benefited from comments, suggestions and editing by Igor Kheyfets (Economist, ECSH2) and Indhira Vanessa Santos (Economist, ECSH4). 2013 2 Table of Contents 1. INTRODUCTION .............................................................................................................................................. 3 2. BACKGROUND ................................................................................................................................................. 5 3. KEY MILESTONES OF PER STUDENT FINANCING REFORM AND THE FUNDING FORMULA DESIGN .... 10 4. FIRST OUTCOMES ......................................................................................................................................... 13 5. SUMMARY OF CONCLUSIONS AND RECOMMENDATIONS ......................................................................... 25 3 1. INTRODUCTION 1. At present, the education sector is characterized by an oversized network of increasingly small schools and a high number of poorly paid teachers and non-teaching staff. A recent Ministry of Education sponsored feasibility study concluded that up to half of Moldova’s rural schools may need to be reorganized in the next three to five years in order to counteract the sharp population declines that have taken place over the last 20 years. It is important to underline that the restructuring of the school network is not a goal in itself. The main purpose of the reforms will be to improve quality of education through consolidation of schools and better utilization of resources — human and physical. 2. The Government of Moldova has launched a per student financing reform in the education sector as part of the broader structural reform agenda in order to bring education spending on a fiscally sustainable footing while improving the quality of the education being provided. Action Plan for Education Structural Reform Implementation developed and approved by the Government in 2011 had three objectives: (i) ensuring access of students to quality education; (ii) increasing flexibility in labor relations in education; and (iii) enabling the efficient use of financial allocations by applying per student funding nationwide. In advance of the latest, per student formula financing was piloted in selected rayons starting from January 2010. 3. It is too early to expect significant impact from the recent introduction of per student financing of schools. It takes time for the efficiency benefits of the system to be felt. The objective of this study is to review the experience so far with implementation of per student financing in Moldova’s two pilot regions/rayons? — Riscani and Causeni — launched in January 2010 before rolling out the system nationwide1. The study focuses on the following key issues:  What are the full adjustments that would be required by schools and rayons in Moldova when the transition period (that is three year period granted to the first two pilot rayons to adjust to the new financing system) is over and full rigors of the formula apply? Redistribution of financial resources within the school network will inevitably lead to some schools gaining the budget while some schools loosing that. So the simulations looked at the characteristics of the winners and losers, and what improvements might be considered?  Are there signs that the ongoing reforms created better incentives for rayons to optimize resource use? Has there been any optimization during the summers of 2010 and 2011?  What changes are happening at the school level? Are there any noticeable differences in the school principals’ attitude to their job, new financing system, school autonomy in the pilots and non-pilot rayons? 1 The evaluation of the results of the wave 2 pilots (9 additional rayons and 2 municipalities that joined the new financing scheme since January 2012 - Criuleni, Cantemir, Călăraşi, Nisporeni, Taraclia, Ştefan Vodă, Glodeni, Leova, Sîngerei, Mun.Bălţi and Mun.Chişinău) can be conducted once the data collected by IPP becomes available. The objective of this consultancy is to collect and analyze education data which will be used to measure, inter alia, enrollments, attendance, and the drop-out rates at the beginning of the school year 2012/13 as well as to get a sense of the magnitude of flows of students between schools during a summer break (when the schools are usually closed or reorganized). 4 4. The results of the evaluation of the pilots (first two-years of implementation for which the data is available) have been presented and discussed with the Ministry of Education during the Bank missions in May and September 2012. This paper is organized as follows: Section 2 explains sector context and the core of the per student financing reforms. Section 3 highlights key milestones in per student financing reform in Moldova. Section 4 reviews first outcomes of per-student financing in two pilots Riscani and Causeni. Finally, Section 5 presents the summary of key findings and policy recommendations. 5 2. BACKGROUND 6. One of the key priorities of the Government of Moldova is to improve the quality, efficiency and relevance of the education system2. However, the demographic and fiscal realities of the country have not made it easy for the Government to fulfill its mandate in education. Over the last decades, the education sector in Moldova has witnessed uneven education quality and a lack of efficiency. The education sector is characterized by an oversized network of increasingly small schools and a high number of poorly paid teachers and non-teaching staff. Of the 1,460 general education institutions in Moldova, three-quarters are located in the country’s nearly 1,000 villages, yielding an average of 1.1 schools per village. An average school in Moldova operates at 54 percent of the capacity for which it was designed, leading to wasteful expenditures in the form of heating bills and public utilities. 7. Indeed, the school network did not adjust to the new demographic realities and serves about half as many students as it did in 1991, while maintaining only around 13 percent fewer schools (figure 1). Moldova currently maintains a large network of increasingly small schools—260 students on average compared to 437 in 1991—with the median school operating at 54 percent of the capacity for which it was designed (this number is lower in rural areas). Maintaining an oversized school network is wasteful and takes away much needed resources for raising the quality of education. At present, a high proportion of the education budget goes toward the financing of personnel expenditures and maintenance costs of school buildings including expenditures on heating and utilities, crowding out much needed quality enhancing investment in capital and educational materials3. Figure 1. Trends in the number of students, teachers and schools, 1991-2011 16.0 500 13.8 Student-teacher ratio Average school size 14.0 450 14.5 3% decline since 1994 12.0 437 400 381 10.0 Student- 350 teacher ratio 10.5 8.0 Average 300 school size 48% decline since 1991 260 6.0 250 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 Source: CISSTAT, "Statistics of the Countries of the CIS", National Bureau of Statistics of Moldova. 8. The OECD Programme for International Student Assessment - PISA 2009 plus - showed that 15 year olds in Moldova are lagging behind their OECD peers by around 2National Strategy for Development “Moldova 2020”. 3All high performing systems have well-resourced schools in materials terms. On average, OECD countries spend 22 percent of current education expenditures on materials (non-staff expenditure). 6 2 years of schooling4. On the reading literacy, only 15.2 percent of Moldova’s 15 year olds scored at Level 3 or higher which is the level of proficiency frequently required in the labor market —much lower that the OECD average of 57.2 percent. Moreover, 57 percent of Moldova’s 15 year-olds remain functionally illiterate. It is important to underline that PISA is a good predictor of the economic growth (far better than the average schooling levels),5 and international evidence suggests that an increase of 50 points in PISA scores is associated with a around one percentage point higher annual growth rates (Woessman and Hanushek, 2007). 9. While improving the quality and relevance of the education system to the needs of modern economy is one of the main priorities of the Government of Moldova, the quality of education, as acknowledged by the National Strategy of Moldova for 2008-2011, ‘is negatively affected by the financing mechanism of the education system that does not create incentives for efficient use of resources’6. Indeed, the traditional input-based system of education financing in Moldova has provided few incentives to improve sector efficiency by reducing resource utilization in response to a shrinking school-age population. The existing input-based financing system ties resources to teachers not students, thus making teachers a “free resource” from the point of view of local authorities. This arrangement has resulted in some of the smallest classes and lowest student-teacher ratios in the world, in particular in Moldova’s rural areas. 10. Input-based financing systems are non-transparent leaving too much room for discretionary decision-making of local authorities. For instance, all decisions on how to allocate maintenance expenditure amongst schools within a locality are usually entirely in the hands of local authorities and are made without transparency and clear rules (compared to formula financing basis), leaving plenty of room for them to favor one school at the expense of others. The only dimension in which input-based system is excelling is by providing job security and stability among employers in the sector: despite dwindling student numbers and fewer schools, teachers maintain their jobs. At the same time, the large number of teachers (relative to students) suggests that each teacher receives a rather modest salary. 11. To improve efficiency of spending and enhance education quality, the Government has developed and embarked on a set of ambitious initiatives reflected in the “Action plan for education structural reform implementation” 7. The action plan 4 PISA tests 15 year-old competencies in reading, math and science and assesses how far students near the end of compulsory education have acquired some of the knowledge and skills that are essential for full participation in society. A difference of 40 points is roughly equivalent to what can be learnt in an additional year of schooling. 5 Since it is the quality of education - rather than mere schooling - which is contributing to economic growth. 6 For example, the results of the PISA 2009 plus assessments showed that 15 year olds in Moldova are lagging behind their OECD peers by 2 years. 7 The action plan is based on a feasibility study commissioned in 2010 by the Ministry of Education of options for rationalizing the school network to assist the MOE in developing a National Strategy of School Optimization. This study, entitled “Policy Options to Support the School Network Optimization Process in the Republic of Moldova” was finalized in January 2011 and did the following: (i) conducted a data-driven analysis of the existing school network in Moldova; (ii) identified potential schools that can be closed down or merged with others; (iii) estimated potential savings of closing down and merging schools; (iv) identified within-school optimization possibilities using survey data; and, (v) proposed a list of policy options that the MOE and Ministry of Finance (MOF) can consider to support the network optimization. 7 has three objectives: (i) equal access for all children to quality education; (ii) increasing flexibility in labor relations in education; and (iii) efficient use of financial allocations. As part of the third pillar, Moldova, similar to many other countries in the region, started reforming the financing mechanism of primary and secondary education with a view to create better incentives for local actors to use resources efficiently8 (figure 2). Figure 2. Progress Towards Per-capita Formula Financing of the Pre-University Education in ECA Source: Skills, not Just Diplomas: Managing Education for Results in Eastern Europe and Central Asia, World Bank 2012. 12. Per student financing is a key organizing principle in a set of reforms aimed to improve education quality and efficiency of the education sector. It is a common practice in most OECD countries (Figure 2). Under this mechanism, funding “follows the student” which creates a direct link between the purpose of the activity—teaching students—and the financial resources devoted to that activity. The education provider is allocated funding from the public budget for providing a specified quality of education according to the number of students receiving that education. The system should focus on outcomes and outputs rather than inputs and provide over time performance incentives. 13. Per-student financing mechanism can provide a number of advantages over the traditional input-based system in particular terms of improved efficiency, equity and transparency of budget allocations. For example, in terms of transparency, formula funding shows explicitly what allocations each local government and school receives and why (table 1). This mechanism helps to improve efficiency and lower per student costs by reducing the number of schools and classes so long as transporting students to another school costs less than teachers and buildings costs saved. It can also be used to minimize 8 The majority of ECA countries have recognized the limitations of input-based financing schemes and have come to appreciate that financing can be used as a policy instrument. At the pre-university level, a typical first step away from input-based financing is to introduce a funding formula in which the number of students enrolled in a school is the main component. In its turn, results-based education financing can be used to keep learning institutions accountable for meeting agreed learning outcomes. Source: Skills, not Just Diplomas: Managing Education for Results in Eastern Europe and Central Asia, World Bank 2012. 8 attainment disparities between different social groups so that all children in the country, regardless of family or ethnic background or location could fulfill their educational potential. Table 1. Key advantages of PSF of education compared to input-based system Input based system Per-student financing Improved efficiency due to PSF of local governments:  School budgets are determined by the type,  Funding depends on number of students: cost per quantity and prices of inputs (historic spending) student remains constant as number of students  Funding for salaries determined by number of staff falls positions and/or class size  Incentive for local government to rationalize school  As student numbers decline, student teacher ratio network by reducing number of classes and closing falls: cost per student increases some schools  Advanced systems provide performance incentives based on value added in terms of student attainment Improved efficiency due to PSF of schools  School directors have incentive to maximize  School directors limited by the school budget in the number of classes in order to maintain or increase number of staff they can employ the number of teachers  Organize classes according to number of teachers that can be paid out of the school budget Improved efficiency due to schools deciding on how to spend their budgets  Schools have no incentive to economize on utilities  Schools can spend less on utilities and use the  Local governments don’t buy the goods and money on other items e.g. learning materials, small services that schools need most repairs  Schools can more easily procure the goods and services they need Improved efficiency due to locally determined performance related pay  Salaries based on qualifications and years of  Teachers can be motivated by additional payments service: for effective teaching and greater effort  Performance makes no difference to pay Improved horizontal equity: students with similar needs funded the same  Schools which are similar in educational phase and  Schools which are similar in educational phase and numbers of students have different amounts spent numbers of students have the same amount per per student student  Similar local governments are funded different  Local governments with similar characteristics are amounts per student funded the same amount per student Improved vertical equity: students with different needs funded differently  Can be vertically equitable if students with special  Funding formulae address vertical equity by needs allocated to smaller classes or other extra allocating more money per student for specified resources e.g. learning disabilities; minority categories. language students, students in isolated rural areas  Extra weights can apply for special needs, minority language, social disadvantage, isolated rural location Improved transparency  Non-transparent: Lack of information on how  Transparent: Formula funding shows explicitly much funding per student each local government how much funding each local government and and school receives and why school receives and why  Lack of accountability: Often no requirement for  Accountability: School-based financial local officials or school directors to justify in public management: school director held accountable for how funds are allocated and spent. how school budget is spent and with what results. Very important role for School Board in holding school principal accountable. Source: based on Rosalind Levačić, Per Capita Financing of General Education and its Role in Creating Incentives for an Efficient School System: International Experience, World Bank education workshop, 2010 and 2012. 9 14. Quality improvements can be greatly facilitated under the new financing mechanism since through the consolidation of school network and better resource utilization a larger share of the budget can be made available for quality enhancing practices (improving curriculum, building strong and effective teaching force, ensuring better facilities and equipment for teachers and students and others). Per student financing can also envisage performance incentives for improved learning outcomes of students thus increasing the efficiency and accountability of local actors for the quality of education provided. Special care must, however, be taken to prevent sanctions against the schools that have larger number of low-performing students. Instead, rewards should be envisaged for improving academic achievements of students. 15. Per-student financing is a necessary but not sufficient condition for a more efficient system that provides better quality education. Strong education policies developed and led by the Ministry of Education are also important. For example, one area is the designing of long-term policies to improve quality of teaching, development and implementation of an updated program for training and remuneration of school directors and teachers (as teacher quality is one of the most important contributing factors to improving student achievement). School systems considered successful tend to “prioritize teachers’ pay over smaller classes”9. In Bulgaria, for example, the per student financing reform allowed the government to save a significant amount of resources that allowed for the considerable increase of real wages in the education sector and the reallocation of more resources for capital investment within the sector10. Another important lesson learned from a similar reform in Bulgaria shows that school consolidations may have exacerbated school dropout rates11. Therefore, a reliable EMIS is critical for monitoring the effects of the school network reforms and designing interventions to mitigate potential risks of the reform. 9 At the level of the school system, PISA shows that higher teachers’ sala ries, but not smaller class sizes, are associated with better student performance. Source: OECD (2010). What makes a school successful? Vol.IV. p.14. 10 The sweeping decentralization reform of the education system introduced in 2007 following the nationwide teachers strike aimed to promote greater school autonomy and more efficient spending. The reform produced impressive efficiency gains. The School Network Optimization was the most difficult phase of the reform with 111 small schools closed in 2007, 340 in 2008 and 44 in 2009. If the Government of Bulgaria had not implemented the reforms and consolidated the schools in 2007 and 2008, the projected total budget in 2008 would have been 4 percent higher (or more than 100 million BGN) than the observed budget. The resources thus saved were invested to increase wages in the education sector by 46 percent between 2006 and 2008. Efficiency gains allowed also allocation of more resources for capital investment within the sector. Source: A Review of the Bulgaria School Autonomy Reforms (World Bank, 2010) and http://www.worldbank.org/en/news/2010/09/13/bulgarian-education-system-more-efficient-further-focus-quality- needed. 11 Results of a rigorous impact evaluation study suggest that school closures and consequent consolidations contributed to a small but significant increase in school dropout rates in Bulgaria. A Review of the Bulgaria School Autonomy Reforms (The World Bank, 2010). 10 3. KEY MILESTONES OF PER STUDENT FINANCING REFORM AND THE FUNDING FORMULA DESIGN 16. The Government began piloting per-student financing in Causeni and Riscani in January 2010. In 2012 nine additional rayons and two municipalities joined the per- student financing with plans to expand new financing mechanism nationwide since January 2013. Table 2 presents key highlights of per student financing reform in the country. It shows that there are many changes in addition to the phased introduction of per-student financing that are actually taking place at the same time. Table 2. Key Events in Per-Student Financing Reform in Moldova October The Government adopted methodology for formula-based financing of primary and secondary schools 2009 in two pilots12. The school in pilots received autonomy to reallocate funds and determine the number of staff within the budget envelope received by the school. January Piloting launched in two rayons: Riscani and Causeni which were selected from North and South regions 2010 with very different demographic situation13. In order to give rayons time to rationalize their school networks, the pilot schemes of per student financing have proceeded with a gradual introduction of the full rigors of the formula. In the first two pilots (wave 1), schools are protected for three years from the full effects of changes to school budgets (that is budget cuts) due to the application of the formula. For that, “losing schools” had to prepare the realistic optimization plan in order to have their deficit covered for the current year and prepare for the next year budget cuts. July The Government of Moldova made adjustments to the formula as analysis of school budgets in the two 2011 pilot rayons revealed that the uniform fixed allocation funded small schools too generously. Therefore, small schools (primary schools with fewer than 41 weighted students14 and secondary schools with fewer than 91 weighted students) are provided with a smaller ‘fixed’ allocation. The pilot has been extended to additional nine rayons and two municipalities. For rayons and cities that joined PSF in 2012 and will join in 2013, protection will last for two years. January Piloting expanded to additional rayons and municipalities - Criuleni, Cantemir, Călăraşi, Nisporeni, 2012 Taraclia, Ştefan Vodă, Glodeni, Leova, Sîngerei, Mun.Bălţi and Mun.Chişinău (wave 2 pilots). Allocations are made according to the revised formula. April New legal changes created an enabling environment for nationwide efficiency reform (introduction of 2012 the per-student financing into the law on public finances; changes in the schools ownership from mayors to rayons councils aimed to support the optimization efforts at regional level; needs-based term contracting of retired staff). October The Government approved rolling out per student financing nationwide since January 2013. 2012 17. Funding formulas employed in Moldova are quite simple. Rayon-level formula that is used to calculate the state categorical transfer for education to Administrative Territorial Units (ATUs) and a school-level formula defining delegated budgets from ATUs to the schools use only the weighted number of students and the number of schools (box 12 Resolution of the Government of Moldova №612 of October 30, 2009 ‘On piloting formula -based financing of the primary and secondary schools financed from budgets of the administrative-territorial units’. 13 Rayons are quite different in terms of demographic situation. In 2010 live births in Riscani increased by 8 percent compared to 2004 (above Moldova average at 6 percent growth) while Causeni was still 7 percent below the 2004 level. 14 The formula accounted for difference in costs of educating students in primary, lower secondary and upper secondary levels of education which result from curriculum requirements. This was done through calculation of so called "weighted" students per school. Weighting coefficients that applied to actual number of students were established based on the 2007 actual data of full time teaching positions per class by levels of education and were equal to 0.75 for grades 1-4 (primary education); 1 for grades 5-9 (lower secondary education) and 1.22 - for students in grades 10-12 (upper secondary education). 11 1)15. During the first two years of implementation, the school-level funding formula employed provided the same amount per weighted student and the same amount of fixed allocation to every school regardless of size. However, analysis of school budgets in the two pilot rayons revealed that the uniform fixed allocation funded small schools too generously compared to larger schools16. As a result, since 2012 small schools (primary schools with fewer than 41 weighted students and secondary schools with fewer than 91 weighted students) are provided with a proportionally smaller ‘fixed’ allocation. Box.1. Funding formulas employed in Moldova. In October 2009 the Government adopted formulas for financing primary and general secondary schools, gymnasiums and lyceums in two pilots [categorical and additional transfers to local authorities and the allocation of delegated budgets to schools]. The rayon-level funding formula was using the weighted number of students and the number of schools in 2007. C= A*N+B*S where: • A is the cash amount or normative value per one weighted student (5021 lei in 2010); • N is the weighted number of students (based on the 2007 student enrollments and weighting coefficients 0.75 for students of primary grades; 1.0 – for students of lower secondary grades and 1.22 for students of upper secondary grades); • Parameter B is the cash normative per one school (413,854 lei in 2010); • S – number of public schools in rayon; The formula is then applied to schools (primary and secondary schools, gymnasiums and lyceums) with only minor modifications. The per-student and per-school amounts in the schools’ formula can be reduced by up to 5 percent which is retained by rayon (largely to protect schools whose formula budget is less than their regular budget during the transition period) and to give time for rayons to rationalize their school networks during three years of protection so that schools can be financially viable when funded by the formula without protection. The school funding formula was initially as follows (the same amount per weighted student was allocated to every school with additional fixed allocations given to all schools, regardless of the size): V=(A*N + B)*K + R where: • N is the number of weighted students at school i. • K is the proportion of the education transfer going to schools which cannot be lower than 95 percent 17. • R is an additional allocation funded from the rayon component of education transfer determined by a regulation approved by each rayon council. It can be used to finance home-to-school transport, dormitories or to assist schools which would otherwise have budget deficits. This formula specification was employed in Riscani and Causeni in 2010 and 2011. Analysis of school budgets in the two pilot rayons showed that the uniform fixed allocation funded small schools too generously compared to larger schools. Consequently, the new formula employed since 2012 calculates a proportionally smaller ‘fixed’ allocation for schools with fewer than N1 students. V = N*(N1*A + B)/N1*K + R where: • N is the number of weighted students at school. • A school with N1 or more students receives a fixed allocation of B regardless of the number of students. For primary schools N1 is equal to 41 students and for gymnasiums - 91 students. 15 To encourage rationalization of the school network the number of schools when calculating the transfers is based on 2007 data. 16 V.Avanesyan for the Ministry of Education of Moldova (2011). 17 In the pilot rayons in 2010 Causeni rayons retained zero and Riscani 5 percent. Similarly to input-based financing it involves bargaining and discretion, is non-transparent and unpredictable. These limitations undermine the key advantage of a funding formula as a transparent and equitable allocation mechanism. 12 18. In order to give rayons time to rationalize their school networks, the pilot schemes of per student financing have proceeded with a gradual introduction of the full rigors of the formula. In the first two pilots, schools are protected for three years from the full effects of changes to school budgets due to the application of the formula. For rayons and cities that joined PSF in 2012 and will join in 2013, protection will last for two years. Pilots of Riscani and Causeni received additional transfers in year 2010 to ensure 100 percent of their budget ceiling defined by the Ministry of Finance with an addition of 3 percent to that for the necessary adjustments during the transition. In the second year this additional transfer is reduced to 70 percent and in the third year to 30 percent, with no protection from year 4 onwards. At the same time, “losing schools” had to prepare the realistic optimization plan in order to have their deficit covered for the current year and prepare for the next year budget cuts. 19. Survey of school principals showed that nearly all schools (97 percent) have school development plans in place and principals think it is important to have one. However, such outlines of each school’s strengths and weaknesses together with a vision and action plan for making improvements should also include students’ scores on national assessment exams reported by social groups (minority students, low-income students, students with special needs etc.). These scorecards should be shared with parents, students and local authorities enabling them to assess levels of attainment and progress of students in a particular school and to compare that with other schools and with national averages (thus giving them a sense of school’s performance). For example, the Ministry of Education of Chile and the U.K. Department for Education make such school scorecards available through their web-sites18. 20. The new financing modality gave schools greater autonomy and flexibility in utilization of resources. School principals were granted authority to reallocate funds on as-needed basis and determine the number of staff within the budget envelope received by the school. This should allow schools to use resources according to their priorities – for learning materials, equipment, teachers’ training, equipment, renovations to school buildings and so on. It is important to underline that schools with greater autonomy in resource allocation tend to produce higher student learning outcomes than those with less autonomy, provided that schools are held accountable publically for that19. However, school-based management also places greater administrative burden on school staff, requiring budget management training for school principals and, occasionally, the hiring of additional accountants. 21. Also, in April 2012 new legal changes created an enabling environment for nationwide efficiency reform. This was achieved through the introduction of per-student financing into the law on public finances, the transfer of school ownership from mayors to rayon councils, and the introduction of needs-based term contracting of retired staff. Over the last decade preceding this change, the share of such teachers in Moldova has increased from around 6.6 percent to 19.6 percent. Retired teachers should be given term contracts only where they are needed to provide adequate staffing of the educational process. 18 www.simce.cl and www.education.gov.uk/performancetables 19 OECD (2010). What makes a school successful? Vol.IV. p.105 13 4. FIRST OUTCOMES 22. Per-student financing aims to improve efficiency, equity and transparency of budget allocations and we will further examine experience in two pilots along these lines and in accordance with the following logic:  What are the characteristics of the winners and losers, and what improvements might be considered before using the formula nationwide?  Simulation of the impact of the school funding formula on school budgets and rayon education transfers in Moldova  Are there signs that ongoing reforms created better incentives for rayons to optimize resource use? Has there been any optimization during the summers of 2010 and 2011?  Assessment of efficiency results  Are the noticeable differences in the attitude of school principals in the pilots and non- pilot rayons to their job, school autonomy etc?  Survey of school principals’ job satisfaction and sense of empowerment in pilots versus non-pilots 23. Efficiency results largely depend on the design of the funding formula. A simulation of the eventual impact of per student funding on school budgets 20 showed that about 40 percent of schools will experience budget losses (formula budget less than expenditure) of more than 2.5 percent, and of these 9 percent will have cuts in excess of 25 percent21. Around 48 percent of schools will experience budget gains of more than 2.5 percent (of these, 10 percent will have a budget increase of over 25 percent); 11 percent of schools will have little or no change; only 1 percent of schools will suffer a budget loss of more than 50 percent (in absolute terms). Table 3 shows the number of schools in each budget change group. 24. The new formula is more favorable to lower cost schools than was the old input-based financing system. The size of the budget gain is found to be positively correlated with the following school characteristics: average class size, student-teacher ratio, the ratio of students to non-teaching staff, building capacity utilization, average number of students per grade, and total number of students. The proportional budget gain is negatively correlated with the present level of per student spending (see Table 4 below). This indicates that the new formula is more favorable to lower cost schools than was the old input-based financing system. The relationship between proportional budget change and the factors that are the prime drivers of per student costs—class size and the ratios of students to teachers and students to non-teaching staff—is consistent with this conclusion. All three variables generally rise in line with proportional budget change (see Figure 3 below). The index of community deprivation—which measures the relative level of well- 20Rosalind Levačić, 2011-2012. 21The school budget simulation proceeded in three main steps: simulation of the school budget allocations for individual schools that would be determined by the formula; identifying the equivalent budget the schools received when funded by the previous input-based method based on schools’ actual expenditure in 2010; calculating the difference between each school’s formula budget and total expenditure and finding the ‘proportional budget change’. School expenditures for 2010 were taken from BOOST database and matched with the database used for the school network optimization study: QERM ‘Project Policy Options to Support School Network Optimization Process in the Republic of Moldova: Feasibility Study’. Chişinau, 2010. The sample that that had both complete expenditure data and non -zero total students was equal to 1234 schools. 14 being of the communities in which the schools are located—has a very small and statistically significant inverse correlation with proportional budget change, indicating that schools in more deprived areas tend to experience a positive budget change22. Table 3. Number of schools grouped according to simulated budget change under PSF Group Budget Budget Budget Budget Budget Budget classification change change change change change change Group by range of M0: M2: M5: M0: % M2: % M5: % proportional number number number schools schools schools budget of of of schools Group 1 change <-0.5 schools 13 1.1 schools 13 1.1 13 1.1 Group 2 >-0.5 <-0.25 87 7.1 98 7.9 116 9.4 Group 3 >-0.25<-0.10 178 14.4 203 16.5 246 19.9 Group 4 >-0.10<-0.05 105 8.5 108 8.8 138 11.2 Group 5 >-0.05 <- 54 4.4 80 6.5 79 6.4 0.025 Total budget losers 437 36 502 41 592 48 Group 6 >-0.025<0 76 6.2 76 6.2 74 6.0 Group 7 >0<0.025 74 6.0 66 5.3 64 5.2 Total neutral budget 150 12 142 12 138 11 change Group 8 >0.025<0.05 66 5.3 70 5.7 68 5.5 Group 9 >0.05<0.10 130 10.5 127 10.3 117 9.5 Group 10 >0.10<0.25 305 24.7 268 21.7 223 18.1 Group 11 >0.25<0.5 117 9.5 99 8.0 73 5.9 Group 12 >0.5<4 29 2.4 26 2.1 23 1.9 Total budget gainers 647 52 590 48 504 41 Note: M is equal to the proportion of the education transfer retained by rayon (M=1-K from the funding formula). Table 4. Correlations between proportional budget change and school characteristics related to cost per student or equity (new 2012 specification of the formula) Proportional Budget Number of schools Variable Change Average class size 0.34 1223 Student teacher ratio 0.50 1213 Student non-teaching staff ratio 0.20 1219 Building capacity utilization 0.08 1179 Total expenditure per student -0.52 1234 Total number of students 0.13 1234 Number of students per grade 0.14 1234 Total expenditure -0.12 1234 Proportion of students of Moldovan ethnicity 0.23 1192 Index of multiple deprivation -0.10 961 Note: all the above Pearson correlations are significant at 0.01 (2 tailed). Higher value of the multiple deprivation index shows the most prosperous communities. 22Index of deprivation of small areas a complex indicator of community development resulting from the combination of data from 7 fields: income deprivation, economic deprivation, demographic deprivation, deprivation in medical care, educational deprivation, deprivation in living condition and geographic deprivation. Rank 1 shows the most deprived community (the poorest), rank 843 – the lowest deprivation (the most prosperous community). 15 Figure 3. Proportional budget change in relation to class size, student teacher ratio and student non-teaching staff ratio 25. The funding formula promotes efficiency, benefitting larger schools over smaller schools, and should therefore incentivize rayons and cities to reorganize small, costly, and inefficient schools. It also favors rural schools over urban ones, which in the case of Moldova suggests that the formula promotes efficiency since it is easier in urban areas to reduced costs by reducing the number of small classes and schools. Overall, the uniform per student amount could give favorable efficiency incentives in the context of Moldova by encouraging rayons and cities to reorganize their school networks in order to ensure that schools can finance their costs within their respective budget allocations. 26. Nevertheless, the fixed allocation part of the formula should be reexamined. The formula rewards rayons/cities which had or still have more schools than the number required by an efficient network of schools. The number of schools in 2007 cannot be used indefinitely and if it is changed to the number of schools at a later date t his will “punish” rayons/cities which reduced the number of schools and reward those which did not. An alternative should be found that serves the same purpose of protecting rural areas that require more schools per given school-age population than urban areas. Promising indicators for this purpose are population density, average size of settlement or area per settlement. In addition, a common fixed amount could be given to all rayons which would produce more per student the smaller the rayon. Without the changes to the formula and to the current system of formation of the per-student financing budget ceiling, the fixed per- school allocations will be declining from year to year in nominal terms creating the opposition to reforms in the system, accumulating more resources for discretionary spending in hands of rayons and undermining the goal and idea of per-student based financing reform.23 27. In terms of horizontal equity, schools which are similar in educational phase and numbers of students obtained the same amount per student. In terms of vertical equity the funding formula did not include coefficients for addressing the higher cost of education of particular groups of populations (for example, children with special needs, from low social economic background or from isolated rural location, ethnic minorities etc.). At the same time, simulation showed that schools with higher proportions of non-Moldovan students are more likely to be budget losers (table 4), although this is not the case for the formula determined change in rayon transfers. The issue of closing 23declining in nominal terms proportionally to the student decline (thus the share of primary and secondary education budget allocated through per-student financing is declining). 16 minority language schools as well as Moldovan language schools is politically sensitive and can be used by various stakeholders as an argument against education reform efforts of the Government. It is important that in any of these cases the PR strategy of the Ministry was clear it is not about maintaining the buildings of the schools but the language of instruction for its students and making sure that students from optimized schools can easily and safely reach receiving schools. The strengthened EMIS system should allow the Ministry to monitor this issue at a student level. 28. The formula also does not distinguish sufficiently well between schools that should be optimized (by class consolidation or closure) without impairing access and schools that should not be optimized as this would impair access24. The formula is poor at targeting these groups of schools correctly (table 5): only 55% of schools which should be optimized would experience budget losses and 10% get budget increases of 2.5% or more. Undesirably, 60% of schools which, in order to preserve access, should not be optimized, would experience a budget loss of more than 2.5 per cent. Because the formula does not sufficiently well deliver the required funding to small schools which need to be kept open in order to provide access for rural students, it is important for the Ministry of Education to set criteria for rayons to determine and seek approval for the designation of such schools to have ‘protected’ status which makes then eligible for additional funding beyond that provided by the rayon-to-school formula. Table 5. Schools grouped by budget gain/loss and by optimization group Count or School Group 1: Group 2: Group 3: expected Group 4: Unclear: groups Optimize Do not optimize: Do not count of enough students; according to without enough students optimize: school unfavorable Total budget impairing or favorable access numbers demographics schools gain/loss access demographics impaired Count 86 187 41 111 425 Gaining Expected Count 121 139 80 85 425 Count 134 61 98 39 332 Losing Expected Count 95 108 63 66 332 Count 25 33 23 21 102 Neutral Expected Count 29 33 19 20 102 Count 245 282 162 172 859 Total Expected Count 245 282 162 172 859 Note: The expected count is the number of schools that would appear in a cell if that number were in proportion to the number in the size category and budget loss categories. For example, the proportion of gainers is 0.50 (619/1234) while the proportion of large schools is 0.128. Therefore the number of schools expected to be both large and gainers (if there is no relationship between large size and budget change) is 79 (0.50*0.128*1234). Thus the expected count is 79 while the actual count is 81. 24 The feasibility study examined the network of 970 rural primary schools, gymnasiums and lyceums and identified four groups of schools: “Clear-cut cases of institutions to close down”, “Clear-cut cases of institutions not to close down”, “Clear-cut cases of institutions not to close down because access to education would be impaired”, and “Not clear -cut cases: schools with an uncertain future”. Criteria and methodology for grouping the schools was as follows: Group 1 − Clear-cut cases of institutions to close down ( and and ); Group 2 − Clear-cut cases of institutions not to close down ( and or and ); Group 3 − Clear-cut cases of institutions not to close down because access to education would be impaired ( and and ); Group 4 − Not clear-cut cases: schools with an uncertain future ( and ). Source: Policy options to support the school network optimization process in the Republic of Moldova Feasibility Study, IPP, 2010. 17 29. International evidence with the implementation of the per-student financing scheme suggests that it takes time for the efficiency benefits of the system to be felt25. Schools were protected for three years (wave 1 pilots that is Riscani and Causeni) from the full effects of changes to school budgets due to the application of the formula provided that they submitted the school optimization plan. Figure 4 presents allocations to eligible schools26 in pilots over 2006-2011. After an increase of about 15.6 percent in nominal terms for Causeni and 24.8 percent to Riscani, the allocations grew only slightly in 2011 as the government strives to keep education expenditures growth at an average of 5% per year in nominal terms (that is below the rate of inflation) in line with the medium- term expenditure framework and Memorandum of Understanding with the IMF. Figure 4. Allocations to primary and GSE schools, gymnasiums and lyceums in pilots Allocations to pilots, MDL Share of pilots in local budgets 100,000,000 80,000,000 3.0% 60,000,000 2.0% 40,000,000 20,000,000 1.0% Causeni Causeni 0 0.0% Riscani 2,006 2,007 2,008 2,009 2,010 2,011 Riscani 2006 2007 2008 2009 2010 2011 Source: based on BOOST expenditure database. 30. There were no school closures in the wave 1 pilots so far. Closing down schools or reorganizing them represents both a risk and an opportunity. The opportunity is that rationalizing the school network will provide fiscal space to improve the quality of education in the system. The risk is that changes to schools (whether they are closed or re- organized) often involve transporting children to different schools. For academically strong children with resourceful parents, such move is unlikely to result in a greater risk of the children dropping out. But for socio-economically disadvantaged children, for ethnic minorities, and for children struggling academically in school, such move could increase the likelihood of drop out. The data currently collected by IPP under AFQERM should allow to see the magnitude of flows of students between schools during 2012 summer break and to obtain the profile of the student who ‘drops out’ from the education system in Moldova27. 31. Some initial positive efficiency results can already be seen. First of all, student- teacher ratio has finally stabilized: 10.5 in 2011 versus 10.6 in 2010 in the country; 9.5 in Riscani; 11.1 in Causeni (11.2 in 2010). Class-size has increased in Causeni and Riscani by 2.3 percent and 3.3 percent, respectively since launching the pilots, compared to 1 percent for the country as a whole in spite of the continuing demographic decline28. Overall, it is only the 2011/2012 school year in which additional eleven rayons and municipalities (that 25 Alonso et al. Reforming Education Finance in Transition Countries (World Bank, 2011). 26 Primary and general secondary schools, gymnasiums and lyceums. 27 By various characteristics - urban/rural divide, by regions, grade, gender, age, vulnerability status, minority populations, language of instruction, learning outcomes as measured by 4 th and 9th grades tests etc. 28 By 4 percent in terms of live births in Moldova, 6% in Causeni and 3 percent in Riscani over 2009-2011 period. 18 is around a third of the country) started preparations29 and joined the-per student financing (since January 2012) in which class consolidation started to exceed the pace of student decline. Still the school-size continues to decline (table 6 and figure 5). Table 6. Evolution of the class-size, 2005-2011 Changes since launch of the pilots 2005 2006 2007 2008 2009 2010 2011 2010 2011 class-size 21.1 20.5 19.9 19.4 19.2 19.0 19.3 -1% 1% Source: based on the National Bureau of Statistics data. Figure 5. Distribution of educational inputs at rayon level 14 Student-teacher ratio 12 10 8 6 4 2010 2 0 2011 Criuleni Falesti Calarasi Leova Anenii Noi Singerei Floresti Briceni Orhei Drochia Soldanesti Ialoveni Rezina Ungheni Ocnita Telenesti mun. Chisinau Causeni Riscani Cimislia Cantemir Glodeni Taraclia Nisporeni Basarabeasca ATU Gagauzia Dubasari Cahul Edinet Hincesti Soroca Straseni mun. Balti Donduseni Stefan-Voda wave 1 wave 2 pilots non-pilots pilots School-size 600 500 400 2008 300 2009 200 2010 100 0 2011 Dubasari Rezina Ungheni Causeni Criuleni Cahul Donduseni Glodeni Leova Ialoveni Orhei Straseni Riscani Cantemir Stefan-Voda Anenii Noi Basarabeasca Hincesti Telenesti Soldanesti ATU Gagauzia Ocnita mun. Balti Calarasi Nisporeni Singerei Briceni Drochia Cimislia Edinet Floresti Taraclia Soroca mun. Chisinau Falesti Moldova av. 2011 wave wave 2 pilots non-pilots 1 pilots Per student expenditures, MDL 10000 9000 8000 7000 2008 6000 5000 2009 4000 3000 2010 2000 1000 2011 0 Moldova av. Leova Riscani Cahul Causeni Falesti Rezina Donduseni Orhei mun. Balti Glodeni Taraclia Basarabeasca Ocnita Cantemir Calarasi Cimislia Ialoveni Straseni Telenesti mun. Chisinau Anenii Noi ATU Gagauzia Dubasari Hincesti Ungheni Soroca Criuleni Nisporeni Singerei Stefan-Voda Briceni Drochia Edinet Floresti Soldanesti 2011 wave 1 wave 2 pilots non-pilots pilots 29 including optimization of facilities in summer 2011 19 Figure 6. Distribution of educational inputs at rayon level (continued) Per school expenditrues 3500 2008 thousand MDL 3000 2009 2500 2000 2010 1500 2011 1000 500 Moldova av. 2011 0 Edinet Ocnita Falesti Cahul Stefan-Voda Taraclia Anenii Noi Rezina Soroca mun. Chisinau Leova Singerei Donduseni Drochia Dubasari Hincesti Ialoveni Orhei Straseni Causeni Riscani Calarasi Criuleni Glodeni mun. Balti Floresti ATU Gagauzia Nisporeni Basarabeasca Briceni Cimislia Ungheni Cantemir Telenesti Soldanesti wave 1 wave 2 pilots non-pilots pilots Source: based on BOOST expenditure database and MOE data? 32. Rayons and schools in pilots spend differently and invest more in renovations and long-term assets. Thanks to the right to reallocate funds on as-needed-basis, rayons and primary and secondary schools in Causeni and Riscani in 2010-2011 fiscal years could for the first time since 2006 afford above country’s average capital spending (figure 7). In spite of tight fiscal environment, rayons and schools in both pilots could “find fiscal space” to allocate resources for capital investments, fixed assets and capital repairs once they have obtained the mandate to manage their own budgets. Meanwhile, capital spending in the country continued to decline. Pilots could also slightly decrease the wage component since pre-pilot 2009 year. Figure 7. Capital spending, % of local budget allocations30 14% 12% 10% 8% Moldova 6% Causeni 4% Riscani 2% 0% 2006 2007 2008 2009 2010 2011 Source: based on BOOST expenditure database. 30 Budgets of primary and secondary schools, gymnasiums and lyceums. 20 Figure 8. Local budget allocations to primary and GSE schools, gymnasiums and lyceums Causeni 100% 242 Purchase of fixed assets 90% 3.9% 5.6% 2.6% 5.0% 4.1% 1.8% 14.6% 13.0% 12.3% 13.1% 12.4% 13.0% 116 Employer's share of the 80% mandatory health insurance 70% 15.7% 16.6% 241 Capital investments 15.0% 18.2% 20.5% 17.5% 60% 50% 243 Capital repair works 40% 112 Mandatory state social insurance 30% 59.2% 56.4% 61.8% 57.9% 60.8% 55.2% 20% 113 Payment for goods and services 10% 111 Remuneration of work 0% 2006 2007 2008 2009 2010 2011 Riscani 100% 242 Purchase of fixed assets 5% 4% 4% 1% 6% 90% 6% 14% 13% 14% 241 Capital investments 80% 15% 13% 13% 70% 20% 16% 116 Employer's share of the 17% 20% 18% 18% 60% mandatory health insurance 243 Capital repair works 50% 40% 112 Mandatory state social insurance 30% 58% 59% 58% 64% 58% 59% 113 Payment for goods and services 20% 10% 111 Remuneration of work 0% 2006 2007 2008 2009 2010 2011 Moldova 100% 4% 4% 4% 241 Capital investments 90% 5% 5% 6% 13% 13% 13% 80% 14% 13% 12% 116 Employer's share of the 70% 18% 20% 19% mandatory health insurance 18% 21% 23% 60% 243 Capital repair works 50% 40% 112 Mandatory state social insurance 30% 55% 59% 58% 59% 52% 52% 20% 113 Payment for goods and services 10% 0% 111 Remuneration of work 2006 2007 2008 2009 2010 2011 Source: Authors’ calculations based on the BOOST expenditure database. 21 33. Survey of school principals also showed that introduction of per student financing mechanism promoted transparency of budget allocations and created incentives for more efficient spending. Larger share of school principals from pilot rayons think that resources in their rayons are allocated transparently: 62% versus 38% in non-pilots (survey data together with the questionnaire that was developed in collaboration with the MOE and MOF are provided in the Annex 1 and 2 of this study31). School principals in pilots also became more aware of the allocations to their own school: 95% of school principals in pilot rayons compared to 83% in non-pilots know their school budget in detail (figure 9). Another important finding is that school principals in pilots also think that the new per-student financing system provides incentives for efficiency in particular as to creation of bigger classes and reduction of the number of non-teaching staff. Figure 9. Transparency of budget allocations 62% of school principals in pilots think that resources in their rayons are allocated transparently while over 61% of principals in non-pilots either don’t think so or do not know 60% 50% 40% pilot regions schools (Riscani 30% and Causeni) 20% non-pilots 10% schools 0% Strongly agree Agree Disagree Strongly disagree Don’t know Q: In my rayon, resources to the schools are allocated transparently Larger share (95%) of all school principals from pilots think that They know their school budget in detail 100% 80% pilot regions schools (Riscani 60% and Causeni) 40% non-pilots 20% schools 0% Strongly agree Agree Disagree Strongly disagree Don’t know Q: I know the budget of my school with breakdown by key expenditures Source: survey of school principals was part of social impact assessment of schools optimization reforms in Moldova, Magenta Consulting, 2011. 31Survey of school principals was conducted to examine issues of transparency of budget allocations in education system, school autonomy, empowerment of school principals, incentives for efficiency and their job satisfaction (72 school principals from pilot raions in Riscani and Causeni and 169 principals from non-pilot rayons inclusing Donduse, Anenii Noi, Drochia and Cimislia, close regionally and by school network and population density characteristics, participated in the survey). 22 34. It is important that under more flexible per student financing arrangement school principals were able to reallocate funds as needed within the budget envelope determined by the financing formula. The logic behind per-student formula financing is the following: by giving a predictable budget defined by a transparent formula, expanding school autonomy and allowing school principals and school boards to manage their own budgets, the scheme would empower them to make better informed decisions and allocate resources in a manner they judge best for their school that would produce a more efficient outcomes. The survey showed that majority of school principals in pilot rayons think that they themselves (or their school boards) have a considerable responsibility for deciding on budget allocations within the school and allocation of funds for teachers’ professional development, while in non-pilots this responsibility is viewed primarily in the mayor’s office (figure 10). Interestingly also that the majority of school directors in pilot rayons would like to have full autonomy, while in non-pilots administrators want subordination to rayon or municipal authorities (figure 11). Figure 10. Results of survey of school principals: School autonomy and sense of empowerment Q: Regarding your school, who has a considerable responsibility for this task? Deciding on budget allocations within the school 45% 40% 35% 30% pilot regions 25% schools (Riscani 20% and Causeni) 15% non-pilots schools 10% 5% 0% Principal Teachers School board Regional or Ministry of Mayor’s office local education Education authorities Q: Regarding your school, who has a considerable responsibility for this task? Allocating funds for teachers’ professional development 50% 45% 40% pilot regions 35% schools (Riscani 30% and Causeni) 25% 20% non-pilots schools 15% 10% 5% 0% Principal Teachers School board Regional or Ministry of Mayor’s office local education Education authorities 23 Figure 11. Results of survey of school principals School autonomy: Larger share of principals in pilots would like to have full autonomy for their school 70% 60% 50% 40% pilot regions schools (Riscani 30% and Causeni) 20% non-pilots 10% schools 0% Full autonomy Subordinated to Subordinate to rayon Don’t know city/town/village-hall authorities Q: What level of autonomy for your school do you think is appropriate? Source: survey of school principals was part of social impact assessment of schools optimization reforms in Moldova, Magenta Consulting, 2011. 35. It will be important to support school principals and school boards on the way ahead. This is of particular importance given that looking ahead school directors in pilot rayons are slightly less optimistic about serving in their job than their non-pilot counterparts. This finding may be the result of the transition period and the need for more information materials and training for school directors joining the new financing scheme 32. Principals to whom decision-making is decentralized under the PSF system need to be trained to obtain the necessary financial management and budget planning skills. Also, several principals complained about the procurements procedures managed centrally by local authorities even for small items like basic stationeries causing all sort of delays. It is important to identify and eliminate bottlenecks limiting the ability of local actors to efficiently manage delegated budgets. 36. Autonomy is an important but not sufficient condition for achieving good value for money. Of course, it is too early to expect for the PSF mechanism to demonstrate any improvements in students’ academic performance, but the majority of school principals both in pilots and non-pilots think that school autonomy can have a positive impact on education quality (figure 12). At the same time international evidence suggests that in countries where schools account for their results by posting achievement data publicly, schools that enjoy greater autonomy in resource allocation tend to show better student performance than those with less autonomy33. On the contrary, countries without accountability arrangements but with greater autonomy in resource allocation tend to perform worse. In other words, local actors should be held accountable for the education 32 The proposed RBF suggest the support of an updated modern system for training of school directors. 33 OECD study ‘School autonomy and accountability: Are they related to student performance?’ (2011). 24 quality provision and there should be the right combination of autonomy and accountability which tends to be associated with better student performance34. Figure 12. Results of survey of school principals Majority of school principals both in pilots and non-pilots think that school autonomy can have a positive impact on education quality 70% 60% 50% pilot regions schools 40% (Riscani and 30% Causeni) 20% non-pilots 10% schools 0% Strongly agree Agree Disagree Strongly Don’t know disagree Q: School autonomy can have a positive impact on education quality? Source: survey of school principals was part of social impact assessment of schools optimization reforms in Moldova, Magenta Consulting, 2011. 37. To improve learning achievements, it will be important to analyze Moldova’s level of school autonomy, the status of the assessment system, and the accountability of the schools in order to ensure that there are enabling conditions in place for improved learning. For that the School Autonomy and Accountability Scale of SABER tool (System Assessment and Benchmarking Education for Results) could be used which was developed by the World Bank to assist countries in the benchmarking of their system and developing the work plans that would induce maturity in autonomy, assessment and accountability. Figure 13. SABER School Autonomy and Accountability approach 34The proposed RBF suggests as the first steps to use the “school report cards” sent out to all schools with comparative performance outcomes and performance trends on regional and country level. 25 5. SUMMARY OF CONCLUSIONS AND RECOMMENDATIONS 38. It takes time for the impact from introduction of per student financing of schools to be felt since rayons and schools are protected for three years from the full effects of changes to school budgets due to the application of the formula. Still, some positive impact can already be seen: Class-size has increased in both Causeni and Riscani by 2.3% and 3.3%, respectively during the PSF implementation compared to 1% for the country as a whole. Student-teacher ratio has finally stabilized: 10.5 in 2011 versus 10.6 in 2010 in the country; 9.5 in Riscani; 11.1 in Causeni (11.2 in 2010). Budget in pilots is spent differently with more investments in renovations and long-term assets even under tight fiscal environment. 39. A recent survey of school directors’ job satisfaction and sense of empowerment also showed noticeable differences in between the pilot and non-pilot rayons.35 The study drew the following findings in each area of interest:  Transparency of budget allocations: larger share of school directors from pilot rayons know their school budget in detail (95 percent compared to 83 percent in non-pilot rayons) and think that resources in their rayons are allocated transparently (62 percent versus 38 percent);  School autonomy: majority of school directors in pilot rayons will like to have full autonomy, while in non-pilots administrators want subordination to rayon or municipal authorities;  Sense of empowerment: majority of school directors in pilot rayons think that they themselves (or their school boards) have a considerable responsibility for deciding on budget allocations within the school and allocation of funds for teachers’ professional development, while in non-pilots this responsibility is viewed primarily in the mayor’s office;  Incentives for efficiency: larger share of directors in pilot rayons compared to non-pilots think that their current system of funding provides incentives to create bigger classes and to reduce the number of non-teaching staff;  Job satisfaction: looking ahead, school directors in pilot rayons are slightly less optimistic about serving in their job than their non-pilot counterparts; and36  Education quality: The vast majority of school directors in pilot and non-pilot rayons consider that school autonomy can have a positive impact on education quality. 40. Formula employed in Moldova promotes efficiency is more favorable to lower cost schools than was the old input-based financing system benefitting larger schools over smaller ones, rural schools over urban, and incentivizing rayons and cities to reorganize small, costly, and inefficient schools. There will be winners and losers under the 35 The survey covered 241 participating school principals (72 from pilot rayons of Riscani and Causeni and 169 from non-pilot rayons of Donduseni, Anenii Noi, Drochia, Cimislia). Social Impact Assessment of School Optimization Reforms in Moldova, Magenta Consulting, March 2011. 36 This finding may be the result of the transition period and the need for more information materials and training for school directors joining the new financing scheme—therefore, the proposed Project supports an updated modern system for training of school directors. 26 new system. A readiness to change therefore needs to be generated through efficient PR campaigns. 41. At the same time, formula does not distinguish sufficiently well between schools that should be optimized without impairing access and schools that should not be optimized as this would impair access. The current solution is to permit an arbitrary amount R for non-small schools which similarly to input-based financing involves bargaining and discretion, is non-transparent and unpredictable. These limitations undermine the key advantage of a funding formula as a transparent and equitable allocation mechanism and also of failing to protect Group 3 schools which need to be financially viable in order to preserve access. In this regard, it is important for the MoE to set criteria for rayons to determine and seek approval for the designation of such schools to have ‘protected’ status which makes them eligible for additional funding beyond that provided by the rayon-to-school formula. 42. The fixed allocation part of the formula and the current system of formation of the per-student financing budget ceiling should be reexamined. The formula rewards rayons/cities which had or still have more schools than the number required by an efficient network of schools. The number of schools in 2007 cannot be used indefinitely and if it is changed to the number of schools at a later date this will “punish” rayons/cities which reduced the number of schools and reward those which did not. An alternative should be found that serves the same purpose of protecting rural areas that require more schools per given school-age population than urban areas. Promising indicators for this purpose are population density, average size of settlement or area per settlement. In addition, a common fixed amount could be given to all rayons which would produce more per student the smaller the rayon. Without the changes to the formula and to the current system of formation of the per-student financing budget ceiling, the fixed per-school allocation will be declining from year to year in nominal terms creating the opposition to reforms in the system, accumulating more resources for discretionary spending in hands of rayons and thus undermining the goal and idea of per-student based financing reform. 43. Potential inequity is also signaled by the positive correlation between budget change and the proportion of Moldovan ethnicity students on the school roll. It is important to ensure that the language of instruction for minority language students is preserved. The PR strategy of the Ministry should clearly articulate that it is not about maintaining the buildings of the schools but the language of instruction for its students. The strengthened under the proposed project EMIS system will allow the Ministry to monitor this issue at a student level. 44. Per-student financing is a necessary but not sufficient condition for a more efficient system that provides better quality education. Strong education policies developed and led by the Ministry of Education are also necessary. It is important to identify and eliminate bottlenecks limiting the ability of local actors to efficiently manage delegated budgets, as well as to hold local actors accountable for the education quality provision (review school performance and introduce performance incentives over time). 27 Linking financing to student performance will help to increase the efficiency of fiscal resources and hold local actors accountable for education quality provided. Special care must, however, be taken to prevent financing mechanism from punishing schools that have larger number of less performing students and instead reward them for improving those students’ academic achievements. 45. The emerging consensus on school performance indicates that school autonomy and accountability are key enabling factors for mobilizing individual incentives to teach and to learn. Autonomy and accountability do not generate incentives in isolation; they are interlinked with the assessment of teaching and learning at the school, and with teacher quality. The role of school assessments in school management is very important, since they allow parents and society to ask for results. To improve learning achievements, it will thus be important to analyze Moldova’s level of school autonomy, the status of the assessment system, and the accountability of the schools in order to ensure that there are enabling conditions in place for improved learning. For that the School Autonomy and Accountability Scale of SABER tool (System Assessment and Benchmarking Education for Results) could be used which is aimed to assist countries in the benchmarking of their system and developing the work plans that would induce maturity in autonomy, assessment and accountability. 46. Special attention should be paid to the development of school-level boards, mechanisms to strengthen the role of parents and the creation of the feedback systems. Parents need to be aware of the school budget and how the new system works. Amount that each school receives, calculated objectively should be over time made available to the public. The proposed project suggests as the first steps in strengthening accountability chain the “school report cards” sent out to all schools with comparative performance outcomes and performance trends on regional and country level. 47. Development of an updated modern system for training of school directors and teachers is vital. Principals to whom decision-making is decentralized under per student financing system need to be trained to obtain the necessary financial management and budget planning skills, as they together with school boards would play a pivotal role in streamlining the functioning of the general education subsector under the Government’s reform program. Building strong and effective teaching force is very important as no education system can be better than the quality if its educators. Teacher is also key on bringing about a change. There may be different reforms and policies but in the end it is a teacher that delivers it in the classroom. 28 Annex 1: Results of the survey of school principals Table 1: Regarding your school, who has a considerable responsibility for the following tasks? % (multiple choice) Regional School or Ministry N of Mayor’s Rayon Principal, Teachers, governin local of Other, respo office, s % % g board, education Educatio % ndent % % authority, n% s % Pilots 89 0 25 26 1 0 8 72 Selecting teachers for hire Non- 83 1 20 36 9 5 8 169 pilots Pilots 93 0 26 21 1 3 24 72 Firing teachers Non- 86 1 17 17 3 2 25 169 pilots Pilots 77 7 70 4 0 25 7 71 Deciding on budget allocations within the school Non- 64 3 26 4 1 70 13 169 pilots Pilots 18 28 14 17 60 1 6 72 Choosing which textbooks are used Non- 17 32 9 25 55 4 29 168 pilots Pilots 6 19 4 6 74 4 3 72 Determining course content Non- 8 16 9 10 77 0 9 167 pilots Pilots 18 11 18 7 71 4 43 72 Deciding which courses are offered Non- 12 17 18 7 80 0 48 168 pilots Pilots 59 7 48 11 3 18 0 71 Allocating funds for teachers’ professional development Non- 34 15 15 8 5 68 5 167 pilots Selecting teachers for hire 85 0 22 33 7 3 8 241 Firing teachers 88 1 20 18 2 2 24 241 Formulating the school budget 68 3 32 9 3 65 18 240 Deciding on budget 68 4 39 4 1 57 11 240 allocations within the school Choosing which textbooks are Total 18 30 10 23 57 3 22 240 used Determining course content 8 17 8 8 76 1 7 239 Deciding which courses are 14 15 18 7 77 1 47 240 offered Allocating funds for teachers’ 41 13 25 9 5 53 3 238 professional development Note: survey of school principals was part of social impact assessment of schools optimization reforms in Moldova, Magenta Consulting, 2011 29 Table 2: The current system of financing of my school provides incentives to…, % Strongly Strongly Don’t Agree, Disagree, N of Rayons agree, disagree. know, % % respondents % % % Pilots 35 58 7 0 0 72 Carefully monitor my school’s expenditures Non-pilots 5 66 25 3 2 169 Total 14 63 20 2 1 241 Pilots 7 61 31 0 1 72 Create bigger classes Non-pilots 0 38 51 6 5 169 Total 2 45 45 4 4 241 Pilots 8 50 32 8 1 72 Save on teaching costs Non-pilots 2 43 49 5 1 169 Total 4 45 44 6 1 241 Pilots 17 68 14 1 0 72 Find ways to decrease the heating bill Non-pilots 4 67 26 2 1 169 Total 7 67 22 2 1 241 Pilots 21 71 7 1 0 72 Find ways to decrease the electricity bill Non-pilots 3 73 21 2 1 169 Total 8 72 17 2 1 241 Pilots 15 71 13 1 0 72 Reduce the number of non-teaching staff Non-pilots 2 37 54 7 1 169 Total 6 47 41 5 0 241 Pilots 4 40 42 11 3 72 Approach parents to ask for financial contributions to 4 58 33 4 2 169 Non-pilots the school Total 4 53 36 6 2 241 Pilots 13 64 22 0 1 72 Approach sponsors to ask for financial contributions 8 80 11 1 0 169 Non-pilots to the school Total 9 76 15 0 0 241 Pilots 8 75 13 1 3 72 Find ways to ensure that funds are available for 4 77 18 0 1 169 Non-pilots teachers training Total 5 76 16 0 2 241 Pilots 11 79 7 1 1 72 Find ways to ensure that funds are available to 7 70 21 1 1 169 Non-pilots purchase methodological literature Total 8 73 17 1 1 241 Pilots 17 78 3 3 0 72 Find ways to ensure that funds are available to 7 69 20 2 3 169 Non-pilots purchase/upgrade computers Total 10 71 15 2 2 241 Pilots 13 83 3 1 0 72 Find ways to ensure that funds are available to other 5 71 20 1 2 169 Non-pilots expenditures important for school development Total 7 75 15 1 2 241 Note: In single choice questions the percentages may not total 100% due to rounding. 30 Table 3: How strongly do you agree or disagree with these statements, % Strongly Don’t Strongly Agree, Disagree, N of disagree, know, agree, % % % respondents % % Pilots 20 59 17 1 1 69 In my rayon the resources are allocated to Non- the schools in a transparent way pilots 7 58 30 3 1 156 Total 3 26 59 11 1 225 Pilots 11 51 26 6 6 72 As a school principal I more or less know Non- the budgets allocated to other schools pilots 2 36 31 7 23 169 Total 5 41 30 7 18 241 Pilots 71 28 1 0 0 69 I know the total sum of money allocated Non- to my institution pilots 48 44 6 2 0 156 Total 1 4 39 55 0 225 Pilots 78 17 4 0 1 72 I know the budget of my school with Non- breakdown by key expenditures pilots 54 33 10 2 1 169 Total 61 28 8 1 1 241 Pilots 49 49 3 0 0 72 An important part of my job is to present Non- new ideas to the school governing body in pilots 31 64 2 1 1 169 a convincing way Total 37 60 2 1 0 241 Table 4: What level of autonomy for your school do you think is appropriate?, % Full Subordinated to Subordinate to Rayons Don’t know, % N of respondents autonomy, % city/town/village-hall, % rayon authorities,% Pilots 63 17 17 4 72 Non-pilots 34 23 31 12 169 Total 42 21 27 10 241 Table 5: School autonomy can have a positive impact on education quality, % Strongly agree, Strongly Rayons Agree, % Disagree, % Don’t know, % N of respondents % disagree, % Pilots 32 58 7 0 3 72 Non-pilots 20 56 15 2 8 169 Total 23 56 12 2 6 241 Table 6: In your opinion who should determine the following?, %37 (multiple choice) Ministry of Principal, Teachers, School board, Regional/ local Mayor’s N of Education, Other, % % % % authority, % office, % respondents % Establishing teachers’ starting 27 8 49 24 37 7 9 156 salaries Determining teachers’ salary 25 5 37 24 49 7 17 156 increases Table 7: At your school do you have a school development plan?, % N of Rayons Yes, % No, % respondents Pilots 93 7 72 Non-pilots 99 1 169 Total 97 3 241 37 This table presents information regarding the schools where the reform did not take place 31 Table 8: In your opinion, how important is it for the school to have a school development plan?, % Very important, Somewhat Rayons Not important, % Don’t know, % N of respondents % Important, % Pilots 83 17 0 0 72 Non-pilots 73 25 1 1 169 Total 76 22 1 1 241 Table 9: How happy are you on your job today compared with 12 months ago on the scale from 1 to 5, % Neither Very Unhappy. happy, nor Very Don’t N of Rayons unhappy, Happy, % % unhappy, happy, % know, % respondents % % Pilots 3 1 19 67 10 0 72 Non-pilots 2 3 33 47 14 2 169 Total 2 2 29 53 13 1 241 Table 10: Looking ahead how optimistic are you are about serving as the school principal (on the scale from 1 to 5), % Neither Very pessimistic, Very Pessimistic, Optimistic, Don’t N of Rayons pessimistic, nor optimistic, % % know, % respondents % optimistic, % % Pilots 8 4 36 32 15 4 72 Non-pilots 7 4 21 37 30 2 169 Total 7 4 26 35 25 2 241 Table 11a: On a scale from 1 to 7, where 1 means ”completely disagree” and 7 means ”completely agree” tell to what extent do you agree with the following statements: The new system of financing can contribute to the quality of education, % N of Completely Completely Don’t Rayons 2 3 4 5 6 respond disagree agree know ents Pilots 7 1 4 14 29 28 17 0 72 Non-pilots 5 6 8 18 24 19 19 1 169 Total 6 5 7 17 26 22 18 0 241 Table 12b: Per student financing can contribute to the educational management improvement, % N of Completely Completely Don’t Rayons 2 3 4 5 6 respond disagree agree know ents Pilots 7 3 6 8 32 33 11 0 72 Non-pilots 7 6 12 17 19 17 21 1 169 Total 7 5 10 15 23 22 18 0 241 32 Table 13: Regarding your school, what would have been a better allocation of responsibilities, based on the lessons learned from the piloting new formula-based methodology?, (Riscani and Causeni only), % (multiple choice) Regional or School local Ministry of Principal, Teachers, Mayor’s N of governing education Education, Other, % % % office, % respondents board, % authority, % % Selecting teachers for hire 83 0 61 19 3 0 11 70 Firing teachers 81 1 49 13 1 1 24 70 Establishing teachers’ starting salaries 31 4 26 6 42 3 89 72 Determining teachers’ salary increases 30 4 33 7 30 4 30 69 Formulating the school budget 81 9 63 6 1 17 20 70 Deciding on budget allocations within 76 6 73 3 0 13 16 70 the school Choosing which textbooks are used 21 56 16 10 39 1 10 70 Determining course content 17 43 10 7 60 3 3 70 Deciding which courses are offered 25 29 21 21 53 0 28 68 Allocating funds for teachers’ profess. 43 7 39 19 16 13 10 69 develop. 33 Annex 2: Questionnaire QUESTIONNAIRE TO THE PRINCIPALS Name of rayon: ________________________ Name of community: __________________ Name of school: __________________________ 1. Regarding your school, who has a considerable responsibility for the following tasks? A ‘considerable responsibility’ is one where an active role is played in decision making. Please mark as many choices as appropriate in each row as follows: . Principal Teachers School Regional Ministry Mayor’s Comments governing or local of office board education Education authority a) Selecting teachers for hire      ...................... b) Firing teachers      ......................................... c) Formulating the school      budget ................ d) Deciding on budget      allocations within the school …………………………… … e) Choosing which textbooks      are used ........ f) Determining course content      …………… g) Deciding which courses are      offered ........ h) Allocating funds for teachers’      professional development ……………... 34 2. The current system of financing of my school provides incentives to: Please, mark one choice in each row as follows: . Strongly Agree Disagree Strongly Don’t Comments agree disagree know a) Carefully monitor my school’s      expenditures ………………………….. b) Create bigger classes …………………..      c) Save on teaching costs …………………      d) Find ways to decrease the heating bill …      e) Find ways to decrease the electricity bill      f) Reduce the number of non-teaching staff      g) Approach parents to ask for financial      contributions to the school ……………. h) Approach sponsors to ask for financial      contributions to the school …..………… i) Find ways to ensure that funds are      available for teachers training …………. j) Find ways to ensure that funds are      available to purchase methodological literature ………………………….. k) Find ways to ensure that funds are      available to purchase/upgrade computers l) Find ways to ensure that funds are      available to other expenditures important for school development ………………... 35 3. How strongly do you agree or disagree with these statements? Please, mark one choice in each row as follows: . Strongly Agree Disagree Strongly Don’t Comments agree disagree know a) In my rayon, resources to the schools are allocated      transparently…....……… b) As a school principal I more or less know the      budgets allocated to other schools ....... c) I know total allocations to my school ……      d) I know the budget of my school with breakdown by      key expenditures (salaries, utilities, rehabilitation etc.) ...................... e) An important part of my job is to present new ideas      to the school governing body in a convincing way Full Subordinated Subordinate Don’t Comments autonomy to city/town/ to rayon know village-hall authorities 4. What level of autonomy for     your school do you think is appropriate? Please, mark one choice as follows:  Strongly Agree Disagree Strongly Don’t Comments agree disagree know 5. School autonomy can have      a positive impact on education quality …........ Please, mark one choice as follows:  Please mark as many choices as Principal Teachers School Regional or Ministries Comments appropriate in each row as follows: governing local of  board education Education 6. In your opinion, who should authority and/or determine the following: Finance a) Establishing teachers’      starting salaries ... b) Determining teachers’      salary increases .. Yes No Comments 7. At your school do you have a school development plan?   ………………… Please, mark one choice as follows:  36 Very Somewhat Not Don’t Comments important Important important know 8. In your opinion, how important is it for the school to have a school development     plan? …………………. Please, mark one choice as follows:  1 2 3 4 5 Comment Very Very s unhappy happy 9. How happy are you on your job today compared with 12 months ago on the      scale from 1 to 5? 5 – (very happy); 1 – (very unhappy) Please, mark one choice as follows:  1 2 3 4 5 Comment Very Very s pessimistic optimisti c 10. Looking ahead how optimistic are you are about serving as the school principal (on the scale from 1 to 5)?      5 – (very optimistic); 1 – (very pessimistic) Please, mark one choice as follows:  11. Do you agree that the per student financing can contribute to the quality of education increase? Please evaluate it on a scale from 1 to 7, where 1 means “does not contribute at all” and 7 means “contributes to a great extent”. 1 (does not 2 3 4 5 6 7 (contributes to contribute) a great extent)        12. Do you agree that the per student financing can contribute to the improvement of the educational management? Please evaluate it on a scale from 1 to 7, where 1 means “does not contribute at all” and 7 means “contributes to a great extent”. 1 (does not 2 3 4 5 6 7 (contributes contribute) to a great extent)        37 For Riscani and Causeni only: Regarding your school, what would have been a better allocation of responsibilities, based on the lessons learned from the piloting new formula-based methodology? Please mark as many choices as appropriate in each row as follows: . Principal Teachers School Regional Ministry Comments governing or local of board education Education authority a) Selecting teachers for hire      ...................... b) Firing teachers      ......................................... c) Establishing teachers’ starting      salaries ... d) Determining teachers’ salary      increases ... e) Formulating the school budget      ................ f) Deciding on budget      allocations within the school g) Choosing which textbooks are      used ........ h) Determining course content      …………… i) Deciding which courses are      offered ........ j) Allocating funds for teachers’      professional development ……………...