CASE STUDIES OF SUCCESSFUL REFORMS TO ADDRESS THE CHALLENGES OF FINANCING EDUCATION SYSTEMS EFFECTIVELY JUNE 2023 Addressing Inefficient Distribution of Teachers Between Schools The Case of Tanzania With Malawi and the Gambia 1. Introduction Teachers are the single most important input to learning, and in many countries in Sub-Saharan Africa teachers’ emoluments account for the majority of spending on basic education (Bold et al., 2017). However, in many countries in the region teachers are poorly distributed between schools. In particular, schools in remote areas are frequently understaffed compared to those closer to towns and large villages, reflecting a reluctance among teachers to accept postings in areas with significant hardship (Mulkeen, 2010). By contrast, schools in or close to towns and larger villages, where more facilities and amenities are available, often have more teachers than required by government standards, even where the overall supply of teachers nationwide is inadequate. An estimated 28 percent of the variation in staffing between schools in the region cannot be explained by variation in the size of enrollments in schools (Majgaard and Mingat, 2012). This represents a major source of inefficiency in public education expenditure, with significant shares of finance being spent to maintain teachers in comparatively overstaffed schools where they have limited marginal impact on learning outcomes. The impacts of these inefficiencies may be exacerbated by the need to ensure a suitable range of subject expertise among the teachers at a school. These inefficiencies are often exacerbated by inefficiencies in the distribution of teachers within a school, between grades. In many regional countries, teachers are disproportionately allocated to upper primary grades, particularly male teachers, leaving lower grades with high pupil-teacher ratios (PTRs) while upper classes have more Students in Primary Seven at Zanaki Primary School in Tanzania. teachers than required. Photo © Sarah Farhat / World Bank. Approaches to rationalize the distribution of teachers between schools typically involve two strategies (Asim et al., 2019). First, countries can put in place and implement clearer rules and policies regarding the deployment of newly hired teachers to schools, and on transfers and promotions, to require teachers to be placed in schools with high PTRs. This note presents evidence of Tanzania’s use of results-based finance to improve implementation of such rules. Second, CASE STUDIES OF SUCCESSFUL REFORMS TO ADDRESS THE CHALLENGES OF FINANCING EDUCATION SYSTEMS EFFECTIVELY countries may provide an incentive or support payment to below 35.1 Inequitable deployments within schools deepened teachers in remote or hardship postings to encourage them to the problem, with the result that 40 percent of schools had accept and remain in such postings. Box 1 presents examples a PTR of more than 100 in Standard 2 in 2016,2 and class of two countries that have introduced such incentives, one sizes of more than 200 students were common in Grades 1–2 successfully (The Gambia) and one with more challenges (United Republic of Tanzania Ministry of Education, Science (Malawi). and Technology, 2020). The poor distribution of teachers reflected weaknesses in the teacher management system. Although official policy 2. What is the problem? recommended teachers to be deployed equitably between schools, in practice the emphasis of policymaking had been on the overall supply of teachers and the national PTR, Like many countries in Sub-Saharan Africa, Tanzania achieved with limited attention on distribution-related challenges. a rapid improvement in access to education following the Policy-level discussion of teacher distribution tended to introduction of free primary education in 2002, but this emphasize the lack of facilities and amenities in schools improvement in access was accompanied by stagnation in in rural areas;3 while these problems do contribute to the learning outcomes. In 2012, this stagnation became a topic reluctance of teachers to accept postings in such schools, of major public concern following a sudden decline in pass the emphasis on the long-term process of addressing these rates for the Primary School Leaving Examination (PSLE), the weaknesses distracted from the medium-term challenge high-stakes examination taken by students in the final grade of of supporting and compelling teachers to accept such primary school governing promotion to secondary school. The postings in adequate numbers. The structure of education share of students passing fell from 57 percent to just 30 percent governance in Tanzania increased the challenge, with the in 2012, leading to widespread concern about the quality of the Ministry of Education, Science, and Technology (MoEST)4 education system. Assessments suggested that students in setting standards and policies for teacher management, early grades could only read 18 words per minute in Kiswahili, the President’s Office – Regional and Local Government the main local language (Government of Tanzania, 2013). (PO-RALG)5 at national level responsible for the deployment of The crisis reflected inefficiencies in the use of Tanzania’s newly hired teachers to districts, and the districts responsible public education expenditure. Tanzania’s public expenditure for the deployment of teachers to schools, making joint on education was an average 3.7 percent of GDP between decision-making more difficult. The result was that decisions 2010–2014, in line with the average for Sub-Saharan Africa on the deployment of new hires did not prioritize equalizing (UIS, 2023). However, there were severe inefficiencies in the PTRs, leading to inequitable deployment of new hires by the utilization of this finance, particularly with regard to teachers. MoEST to districts and by districts to schools. With teacher salaries accounting for approximately 65 percent of total public expenditure on basic education, and 95 percent 1 In addition to being poorly distributed, teachers had high degrees of of education expenditure at local and regional level (World absenteeism from school and from the classroom. In 2014, 14 percent of Bank, 2019), these inefficiencies had significant implications teachers were found absent from school and, of those in school, 37 percent were not found in the classroom teaching, bringing the overall absence for overall education finance, and the large share of recurring from the classroom to 47 percent, according to the World Bank Service spending on salaries left limited space for operational costs Delivery Indicators (SDI). World Bank. 2014. Service Delivery Indicators and other recurrent expenditure. Tanzania, accessed March 4, 2023. Available at: https://documents1. worldbank.org/curated/en/415111468179674045/pdf/106090-WP- In particular, teachers were poorly distributed between P146421-PUBLIC-Tanzania2014-SDI-EducationTechReport-Final.pdf. 2 Author’s calculation using data from: schools. The national average district PTR of 46 at primary United Republic of Tanzania, Ministry of Education, Science, and 1.  level in 2014 masked wide variations between districts, with Technology. 2018. Education Sector Development Plan (2016/17– the best-staffed, Kilimanjaro, having around 32 students per 2020/21): Tanzania Mainland. Dodoma, Tanzania: United Republic of Tanzania Ministry of Education, Science and Technology; and teacher and the worst-staffed, Singida, 70 (Government of United Republic of Tanzania, President’s Office: Regional and Local 2.  Tanzania, 2014). Only 107 of the country’s 162 districts had an Government. Pre-Primary, Primary, Secondary, Adult and Non-Formal overall PTR between 35–50, identified by the government as Education Statistics in Brief 2017. Dodoma, Tanzania: United Republic of Tanzania President’s Office: Regional and Local Government. an acceptable range, with 33 having a higher PTR and 20 3 For example, the Education and Training Policy of 2014, intended to a lower PTR. PTRs also varied enormously between schools provide an overarching framework for education policy, is silent on the within a single district. As a result, only 36 percent of schools issue of the distribution of teachers between schools, but emphasizes were in the acceptable PTR range. Another 36 percent the need for investment in teacher housing in rural areas. 4 Then called the Ministry of Education and Vocational Training (MoEVT). of schools were understaffed with PTR above 50, while 5 Then called the Prime Minister’s Office – Regional and Local Government 27 percent were comparatively overstaffed with PTR (PMO-RALG). ADDRESSING INEFFICIENT DISTRIBUTION OF TEACHERS BETWEEN SCHOOLS: THE CASE OF TANZANIA The large share of understaffed schools had negative improvements in the share of districts with an average district impacts on learning: in schools with PTRs above the official PTR within the acceptable range of 35–50.6 A total of US$20 target of 40, learning outcomes were typically 15 percent million was tied to improvement in this national-level DLR over lower in Grade 4 (Asim, Chugunov, and Casley Gera, 2019). In the first five years of the program. The second rewarded addition, the comparative overstaffing in other districts and districts for improvement in the share of schools with PTRs schools had negative impacts on the efficiency of education in the acceptable range. A district would receive US$5,000 spending, with approximately 13 percent of spending on for each school whose PTR moved down into the acceptable basic education going to retain teachers in comparatively range (for example, from 55 to 45) and US$3,000 for each overstaffed primary schools where they had limited impact school whose PTR moved up into the range (for example, on students’ learning outcomes. These inefficiencies also from 25 to 35). This dual structure was intended to provide posed a challenge for Tanzania’s efforts to achieve universal the strongest incentives to districts for improving school PTRs access to education: with a primary net enrollment rate of in understaffed schools, while still providing some reward for 81 percent in 2015, expanding the education system to moving teachers away from overstaffed schools. A total of meet demand would be significantly more expensive with US$30 million was tied to achievement of the district-level an inefficient distribution of the most expensive resource. DLR over the first five years of the program. In order to maximize the impact of Tanzania’s limited education finance, achieve free education, and raise learning outcomes, there was an urgent need to improve the distribution of teachers between districts and between schools. 4. What has been the outcome on the efficiency of spending for education? 3. What has been done to tackle The national-level DLR was 100 percent achieved in the first two years of the program, reflecting a high degree the problem? of awareness within PO-RALG and MoEST of the EPforR and the incentives available. However, at district level, In partnership with development partners, the government awareness of the EPforR was much lower, and the DLR of Tanzania adopted a results-based financing approach was just 20 percent achieved in 2014/15 with only 150 to improving the distribution of teachers between schools. schools moving into the acceptable range of PTR. Following The reform formed part of Big Results Now! In Education, a series of regional workshops, which raised the awareness of later known as the Education Program for Results (EPforR), of the incentive among district-level staff, achievement of the a US$514 million project which operated from 2015–22 DLR increased to 41 percent in 2015/16 with the number of with the support of the World Bank, the United Kingdom schools in the acceptable range increasing by 7 percent. Foreign, Commonwealth and Development Office (UK FCDO), and the Swedish International Development Corporation Following the election of a new government in 2015, the Agency (Sida). EPforR provided support for a wide range of new administration introduced a Fee-Free Basic Education government reforms and interventions to improve education Policy (FFBEP) in December 2015. Building on previous outcomes, with financed tied to performance conditions reforms which had abolished formal fees for primary known as disbursement-linked results (DLRs). Two of the education, this policy abolished all formal and informal fees project’s DLRs rewarded improvements in the distribution of for basic education through lower secondary. The policy teachers between districts and between schools. Specifically, led to a large surge in enrollment throughout primary and the DLRs rewarded improvements in the share of districts and secondary school. The number of students in primary schools with PTRs in an acceptable range set at 35–50. The and secondary school surged from 9.43 million to 11.51 million acceptable range was selected as a target—as opposed to between 2015 and 2018, driven by large increases in the share simply the number of schools with a PTR below a maximum of students enrolling in primary school and progressing to lower level—to explicitly encourage the movement of teachers away and upper secondary school. While a huge step toward from comparatively overstaffed districts and schools. Tanzania’s goal of universal access to school, this surge in enrollment posed challenges for the improvement of PTRs. In Tanzania, as in many countries, decisions about the The national primary PTR rose from 44 in 2016 to 58 in 2019, placement of teachers are made at two levels: the national reflecting a lack of new teachers hired along with the increase government allocates newly hired teachers to districts, while in enrollments. Only 65 percent of districts had an acceptable districts allocate them to schools and manage transfers between schools. Therefore, the DLRs operated at these two levels: the first rewarded the national government for 6 Increased in 2019 to 35–53. CASE STUDIES OF SUCCESSFUL REFORMS TO ADDRESS THE CHALLENGES OF FINANCING EDUCATION SYSTEMS EFFECTIVELY FIGURE 1 Share of public primary schools with PTR below, within, and above the acceptable range (2015–19) 60.0% 54.2% 50.0% 48.0% 46.0% 37.8% Share of primary schools 40.0% 37.0% 37.2% 35.2% 35.6% 33.9% 34.2% 30.0% 28.3% 27.9% 20.0% 18.4% 14.8% 10.0% 11.5% 0.0% 2015 2016 2017 2018 2019 In acceptable range Above acceptable range Below acceptable range number of teachers in 2016/17, declining to a disconcerting share of districts with an acceptable PTR began to improve 45 percent by 2018/19, while the share of schools with very again, rising to 57 percent by the time the project closed in high PTRs—above 75—rose to 27 percent by 2019. 2022. However, even while the national supply of teachers failed to The reduction in the share of comparatively overstaffed keep up with demand, districts responded to the incentives schools enabled by the DLRs had significant implications by better allocating the teachers available. The result was for the efficiency of education spending. With an estimated that the share of schools with PTRs in the acceptable range US$700 million per year being spent on primary teacher continued to increase, albeit slowly, even as the national salaries alone, an estimated US$80 million per year was teacher shortage worsened. 7 In particular, districts being effectively wasted by maintaining teachers in aggressively worked to reduce the number of teachers at schools where there was a surplus—an amount reduced by comparatively overstaffed schools, with the share of schools approximately US$45 million by the improvement in teacher with a PTR below the acceptable range declining from distributions engendered by the DLRs. This efficiency gain is 27 percent to 12 percent between 2014–2019 (Figure 1). equivalent to approximately 2.2 percent of basic education However, because of the rise in the national PTR, the share expenditure.9 of schools with a PTR above the acceptable range continued to increase. By 2020, districts felt that they were unable to make significant 5. What are the lessons learned further progress in bringing schools into the acceptable range of PTR, given the constraints on overall supply. The DLR was for other countries? amended to reward the government for allocating teachers to an agreed list of understaffed schools,8 regardless of Successful action on teacher distribution may require the direct impact on PTR. Overall, 78 percent of the 4,792 action at multiple levels of government. Many Sub-Saharan teachers deployed in 2020/21 were deployed in accordance countries utilize a split-level decision-making system for with this agreed list. As the supply of teachers improved, the deployment of new teachers, with teachers allocated by the national government to districts and by districts to schools. For progress to be made, national governments need to 7 The number of schools with PTR in the acceptable range increased by ensure an adequate supply of new teachers and equitable 14 percent between 2015 and 2018 despite the increase in the national PTR, before falling in 2019 to be 5 percent above the number in 2015. 8 In addition to PTR, the agreed list included additional factors such as the need for special needs teachers at schools with significant populations 9 Cost of public primary salaries as of 2019/2020 according to World Bank of students with special needs. Task Team estimations. Exchange rate as of 1 June 2019. ADDRESSING INEFFICIENT DISTRIBUTION OF TEACHERS BETWEEN SCHOOLS: THE CASE OF TANZANIA deployment to districts, while districts need to equitably outcomes proved successful when paired with intensive deploy teachers to schools. In such a context, dual-level technical assistance for municipalities with difficulties in incentives, with some of the finance provided to districts in improving learning (Lautharte et al., 2021). response to deployment decisions, are likely to be necessary Moving teachers away from comparatively overstaffed to ensure that the entire deployment chain is incentivized. schools is possible, and still more essential, when overall Results-based financing can support improvements in teacher supply is constrained. Although understaffed schools the efficiency of spending on teachers, but are unlikely to have the more negative impacts on learning, comparatively be adequate unless backed with support and guidance. overstaffed schools contribute directly to understaffing Although the finance available to Tanzanian districts for elsewhere and have the most severe implications for improvements in the share of schools with PTR in an the efficiency of overall education expenditure. However, acceptable range were substantial, the response was initially concerted action by district-level officials can achieve weak owing to low awareness and understanding of the improvements in overstaffing even while the overall supply incentive mechanism. Awareness-raising activities and of teachers is constrained. By redeploying teachers away tools to guide officials in making deployments appropriately from schools with PTR below 35, Tanzanian districts were can help to ensure that incentives have their desired effect. able to maintain and slowly increase the share of schools The Tanzanian experience is in line with other experiments with PTR in an acceptable range even while the national PTR with results-based financing at local level, such as in Ceará, was rapidly increasing, with significant benefits for the overall Brazil, where results-based finance for improving learning efficiency of spending. BOX 1.  INCENTIVE SCHEMES FOR TEACHERS IN REMOTE SCHOOLS: LESSONS FROM THE GAMBIA AND MALAWI Once teachers have been deployed appropriately, there is a need to ensure they remain in the schools in which they have been placed. These schools are typically in remote areas with significant hardship. Rules or customs that give teachers the right to live close to medical facilities or to a spouse can be used to avoid placements in remote areas. In addition, teachers employ a wide range of formal and informal personal connections to exercise influence over placements and resist deployments to remote areas.* Provision of housing in rural schools is a common approach employed to retain teachers in hardship postings, including in Benin, Cote d’Ivoire, Ghana, and Mauritania, among others. However, it is a high-cost approach, with accommodation for three teachers typically costing more than a classroom (Bashir et al. 2018). Evidence from Ghana and Malawi suggests that offering housing is of limited value as an incentive for teachers to remain in remote postings, typically because it is of poor quality compared to what is available in nearby trading centers (Gad, 2015; Mwenda and Mgomezulu, 2018; Asim et al. 2019). Other non-monetary incentives, such as expedited promotion for those in remote postings, have been used in a number of countries, but there is a lack of reliable evidence of their impact (Evans and Acosta, 2021). A number of countries have introduced hardship or remoteness allowances to provide incentives to teachers in remote schools to remain in post and ease the hardship of these postings (ibid). However, such schemes require careful calibration and reliable data on remoteness to function effectively. The Gambia introduced a hardship allowance in 2006 for teachers in remote schools. These were defined as those more than 3km from a main road, in the regions of the country further from the capital, Banjul. The allowances used a variable formula, with additional payments for those in more remote regions; in particularly remote schools more than 9km from the main road; and for female teachers. The value of the allowance was equivalent to 30–40 percent of the typical teacher’s salary for the majority of recipients; at a maximum, for a female teacher in a school more than 9km from the main road in the most regions, the allowance was equivalent to 60 percent of a teacher’s salary. (In addition to the allowance, the hardship categorizations were also used to influence the value of school grants.) (continues on next page) CASE STUDIES OF SUCCESSFUL REFORMS TO ADDRESS THE CHALLENGES OF FINANCING EDUCATION SYSTEMS EFFECTIVELY BOX 1.  INCENTIVE SCHEMES FOR TEACHERS IN REMOTE SCHOOLS: LESSONS FROM THE GAMBIA AND MALAWI (Continued ) The allowance had an immediate impact on teachers’ preferences. Within one year, 24 percent of the existing teachers in the regions qualifying for the allowance had requested transfers to hardship posts.** Analysis conducted in 2014 confirms that the allowance had long-term impacts on staffing in the qualifying schools—increasing the share of qualified teachers by 10 percentage points and reducing pupil-qualified teacher ratios (PqTRs) by 61 percent. However, the allowance was least effective in the most remote schools, and with female teachers, despite the additional payments in these cases.*** The allowance was initially rolled out in lower basic (primary) schools in 2006 with government finance. Beginning in 2014, the financing of lower basic allowances was supported by the World Bank and Global Partnership for Education (GPE) while allowances were rolled out to upper basic (equivalent to lower secondary) schools through government finance. It is anticipated that the scheme will be fully financed by the government by 2025.**** Malawi introduced its own hardship allowance in 2010. This was intended to provide an allowance of around 30 percent of the average teacher’s salary to the 20 percent of teachers working in the most remote schools. However, the policy was weakened from the outset by the lack of a clear, data-driven definition of remoteness. District-level officials were tasked with identifying eligible schools, but while the original proposal envisioned providing the allowance to 15,000 teachers, more than double that number were identified as eligible. The scheme was known as the “rural allowance,” leading to a misperception that all schools in rural areas—more than three-quarters of Malawi’s schools—were eligible. Teachers in several districts successfully conducted legal or industrial action to gain access to the scheme. The result was that, by 2015, 87 percent of schools were eligible for the allowance, encompassing 80 percent of teachers. With this rapid expansion in the number of eligible teachers, the value of the allowance was reduced to manage costs, and by 2019 the allowance was worth only one-seventh the typical teacher’s salary.* In order to address the scheme’s shortcomings, in 2017 Malawi began work to develop a new, more data-driven definition of remoteness, based on the facilities available at a school, the distance to the nearest town or large village, and the facilities available at that location. The revised remoteness definition has been in use since 2022 for school grants, and a new hardship support scheme based on it is expected to be rolled out during 2023 with the support of the World Bank and GPE.***** Lessons learned. The divergent experiences of The Gambia and Malawi demonstrate that, to be effective, incentive and support schemes for teachers in hardship postings need to be based on reliable and data-driven definitions of remoteness. Appropriately targeted schemes, with a value of at least 30 percent of a teacher’s salary, can move the needle on staffing in moderately remote schools, but even at a higher value allowances may struggle to improve conditions in the most remote schools. Effective schemes also entail considerable expense, which may require support from development partners for an extended period of time. * Asim, S., Chimombo, J., Chugunov, D. and Casley Gera, R. 2019. “Moving teachers to Malawi’s remote communities: A data-driven approach to teacher deployment.” International Journal of Educational Development 65:26–43. ** Mulkeen, A. 2010. Teachers in Anglophone Africa: Issues in Teacher Supply, Training, and Management. Washington, DC: World Bank. *** Pugatch, T. and Schroeder, E. 2014. “Incentives for teacher relocation: Evidence from the Gambian hardship allowance.” Economics of Education Review 41:120–136. **** Through the Results for Education Achievement and Development (READ) project, followed from 2018 by the Education Sector Support Program (ESSP). ***** Through the Malawi Education Reform Program (MERP). ADDRESSING INEFFICIENT DISTRIBUTION OF TEACHERS BETWEEN SCHOOLS: THE CASE OF TANZANIA Acknowledgments Gad, B.K., 2015. “Recruitment and Retention of Public Sector Teachers in Ghana: A Discrete Choice Experiment.” Accra, Ghana: University of Ghana. This study was authored by Ravinder Casley Gera with Government of Tanzania. 2013. Early Grade Reading Assessment financial support from the World Bank Education Finance results. and Reform global team. The authors would like to thank the Government of Tanzania. 2014. “Education NKRA Lab Report.” World Bank team in Tanzania, including Kaboko Nkahiga, Available at: https://files.givewell.org/files/DWDA%202009/SDI/ Gemma Todd, and Innocent Mulindwa, for their guidance Tanzania%20-%20Big%20Results%20Now%20Presentation.pdf. and inputs; and Samer Al-Samarrai and Pedro Cerdan- Accessed March 4, 2023. Infantes for leadership and oversight. Lautharte, I., De Oliveira, V. H., Loureiro, A. 2021. “Incentives for mayors to improve learning evidence from state reforms in Ceará, Brazil.” World Bank Policy Research Working Paper 9509. Washington, DC: World Bank. References Majgaard, K., and A. Mingat. 2012. Education in Sub-Saharan Africa: A Comparative Analysis. Washington, DC: World Bank. Asim, S., Chimombo, J., Chugunov, D. and Casley Gera, R. 2019. Mulkeen, A. 2010. Teachers in Anglophone Africa: Issues in Teacher “Moving teachers to Malawi’s remote communities: A data- Supply, Training, and Management. Washington, DC: World Bank. driven approach to teacher deployment.” International Journal of Mwenda, D.B., Mgomezulu, V.Y., 2018. Impact of Monetary Incentives Educational Development 65:26–43. on Teacher Retention in and Attraction to Rural Primary Schools: Asim, S., Chugunov, D. and Casley Gera, R. 2019. “Student learning Case of the Rural Allowance in Salima District of Malawi. African outcomes in Tanzania’s primary schools: Implications for Educational Research Journal 6, 120–129. secondary school readiness.” Washington, DC: World Bank. UNESCO Institute for Statistics (UIS). 2023. Government expenditure Bashir, S., Lockheed, M.E., Ninan, E., Tan, J.-P., 2018. Facing on education, total (% of GDP) - Tanzania, Sub-Saharan Africa. Forward: Schooling for Learning in Africa. Washington, DC: World UIS Stat Bulk Data Download Service. Available at: https://data. Bank. worldbank.org/indicator/SE.XPD.TOTL.GD.ZS?locations=TZ-ZG. Bold, T., Filmer, D., Molina, E. and Svensson, J. 2017. “The Lost Accessed: March 4, 2023. Human Capital: Teacher Knowledge and Student Achievement United Republic of Tanzania, Ministry of Education, Science and in Africa.” Available at: https://riseprogramme.org/sites/default/ Technology. 2020. Pre-Primary and Primary Teacher Deployment files/inline-files/Bold.pdf. Accessed March 4, 2023. Strategy 2019–2021. Dodoma, Tanzania: United Republic of Evans, D. and Acosta, M. 2021. “How to Recruit Teachers for Tanzania Ministry of Education, Science and Technology. Hard-to-Staff Schools: A Systematic Review of Evidence UWEZO, 2013. “UWEZO Tanzania 2013: Are Our Children Learning?” from Low- and Middle-Income Countries.” Center for Global Dar es Salaam: Uwezo. Development Working Paper 595. Washington, DC: Center for World Bank. 2019. “Tanzania Public Expenditure Review 2019: Global Development. Matching Resources to Policy.” Washington, DC: World Bank. CASE STUDIES OF SUCCESSFUL REFORMS TO ADDRESS THE CHALLENGES OF FINANCING EDUCATION SYSTEMS EFFECTIVELY